4259 Prels 4/12/02 9:41 am Page vii
Chapter Title
1 2 3 4 5 6 7 8 9 10111 11 12 13 14 15 16 17 18 19 20111 21 22 23 24 25 26 27 28 29 30111 31 32 33 34 35 36 37 38 39 40111
vii
LIST OF CONTRIBUTORS Thomas Ferguson
Department of Political Science, University of Massachusetts, Boston, USA
David Greasley
Dapartment of Economic History, Universty of Edinburgh, U.K.
Jakob B. Madsen
Department of Economics and Finance, Brunel University, U.K.
John H. Munro
Department of Economics, University of Toronto, Canada
Ronnie J. Phillips
Department of Economics, Colorado State University, USA
Gary Richardson
Department of Economics, University of California, Irvine, USA
Andrew J. Seltzer
Department of Economics, Royal Holloway College, Surrey, U.K.
Fred H. Smith
Department of Economics, Davidson College, North Carolina, USA
Peter Temin
Department of Economics, Massachusetts Institute of Technology, USA
Simone A. Wegge
Department of Political Science, Economics, and Philosophy, College of Staten Island – CUNY, New York, USA vii
4259 Prels 4/12/02 9:41 am Page ix
1 2 3 4 5 6 7 8 9 10111 11 12 13 14 15 16 17 18 19 20111 21 22 23 24 25 26 27 28 29 30111 31 32 33 34 35 36 37 38 39 40111
INTRODUCTION Volume 21 of Research in Economic History is a substantial contribution in several respects. Its heft reflects the continuing increase in quality submissions to this series, which invites (although it does not require) authors to take advantage of less stringent space limitations than is typically true in a journal article. In different ways, this opportunity has been exploited by John Munro, in his encyclopedic treatment of sticky wages in Medieval England and the Low Countries, by Gary Richardson in his systematic reconsideration of Deirdre McCloskey’s explanation of scattering among medieval peasants, by Thomas Ferguson and Peter Temin in their controversial reinterpretation of the 1931 currency crisis in Germany, by Simone Wegge in her exploration of the determinants of immigration from the German principality of Hesse-Cassell, and by Fred Smith in his painstaking quantitative investigation of the history and geography of land values in the city of Cleveland. In somewhat shorter but no less substantial contributions, Ronnie J. Phillips uses the records of the San Francisco Clearinghouse to study how banking services persisted in the face of the 1906 Earthquake, and Andrew Seltzer examines personnel records from the Union bank in Australia in 1931 to determine the actual effect of a mandated 10% nominal wage cut. Finally, David Greasley and Jakob Madsen revisit econometrically the causes of the precipitous drop in consumption spending in the United States between 1930 and 1932 (this discussion is not necessarily in the order in which the articles appear in the volume). The papers offer regional diversity: two with principal focus on England, one on Germany, one on Australia, and three on the United States. There are some commonalities in themes: we have three papers on 1931, three papers that have something to do with banks, two on urban economic history, and two on wage stickiness, albeit in different countries and addressing labor markets several centuries apart. What can be said of all of these inquiries, however, is that each involves the careful consideration of quantitative and qualitative data within a well articulated theoretical framework. And in almost every case, we have original analysis of primary source material. The volume begins with Thomas Ferguson and Peter Temin’s provocative essay on the German currency crisis of 1931. Marshalling an impressive array of archival and statistical material, they challenge existing interpretations which have focused on a banking crisis and argue instead that this was a run on the ix
4259 Prels 4/12/02 9:41 am Page x
x
1 2 3 4 5 6 7 8 9 10111 11 12 13 14 15 16 17 18 19 20111 21 22 23 24 25 26 27 28 29 30111 31 32 33 34 35 36 37 38 39 40111
INTRODUCTION
currency, caused proximately by ill-advised budgetary policy and by a shortsighted foreign policy that cut the ground out from under the possibility of French assistance. In her study of nineteenth century immigration from the German principality of Hesse-Cassel, Simone Wegge links data on individuals to socioeconomic characteristics of their villages of origin. Wegge’s work is based on original archival work. She has carefully linked data on individuals from emigration lists with census data for the same individuals and village-level data. The result is an important addition to our knowledge of the determinants of nineteenth century European migration. With a similar attention to primary source material, Fred Smith develops a remarkable economic history of the city of Cleveland, exploiting assessors’ data to analyze its changing fortunes by charting the fluctuations of land values across space and time. His most striking finding is that the bulk of the substantial deterioration in land values occurring between 1950 and 1980 took place in the 1970s, and does not seem easily explicable as the consequence of the changing demographic characteristics of neighborhoods. Ronnie J. Phillips explores the response of the San Francisco banking community to the 1906 Earthquake, enriching his narrative with quotations from the minutes of the committee of the San Francisco Clearinghouse. He offers us a fascinating case study of a private financial system’s response to disaster, focusing on how bank runs were avoided through a combination of collaboration within the private sector as well as between banks and the U.S. Mint. An Appendix compares how the operation of the payments system was sustained in 1989 in the aftermath of the Loma Prieta quake. John Munro’s paper represents an extended reconsideration of M. M. Postan’s thesis about the impact of the Great Plague on wage rates. Munro insists that we systematically separate out the effect of real shocks which may have affected relative prices from monetary factors which influenced the general price level. His second major concern is to document and consider the consequences of nominal wage rigidity for real wages and living standards in the face of prolonged periods of deflation or inflation. If Munro explores the consequences of nominal wage stickiness in medieval England and the Low Countries, Andrew Seltzer examines wage flexibility half a millennium later in Australia in 1931. He uses personnel records from the Union bank to conclude that real wages were inflexible in the face of a governmentally mandated 10% wage cut, the consequence of an implicit contract between the bank and its employees. In our third paper focusing on the early 1930s, David Greasley and Jakob Madsen explore econometrically the debate about what caused the drop in
4259 Prels 4/12/02 9:41 am Page xi
Introduction
1 2 3 4 5 6 7 8 9 10111 11 12 13 14 15 16 17 18 19 20111 21 22 23 24 25 26 27 28 29 30111 31 32 33 34 35 36 37 38 39 40111
xi
consumption spending in the United States. They discount the influence of rising debt levels, the penalties for defaulting on installment contracts emphasized by Martha Olney, or changes in the supply of credit. Their conclusion instead is that income uncertainty underlay most of the spending declines between 1930 and 1932. Gary Richardson’s reconsideration of Deirdre McCloskey’s analysis of scattering in English open fields emphasizes that markets were much thicker and more fully developed than McCloskey assumed. He also finds evidence of many other mechanisms besides scattering for mitigating risk for all members of a village. Moreover, he finds little evidence of any efficiency losses from the open field system up until the eve of the Enclosures. Richardson’s analysis will be an important reference point for researchers in this area in the future, just as McCloskey’s seminal article – also published in Research in Economic History – has been for more than a quarter of a century. It’s a pleasure in this volume to publish work of scholars at all stages of their careers. We have contributions ranging from those of recently minted Ph.D.s who have excerpted and developed some of their thesis material for this journal, to those of distinguished senior scholars. Each of these articles is written with care, polish, and often passion. Academic disciplines flourish – and economic history is no exception – when scholars immerse themselves in their subjects and combine this with commitments to logic and evidence, detail, and clarity of exposition. The consequences are the fascinating papers evident here. We look forward to continuing to publish innovative, well written and carefully considered contributions to economic history, providing a niche which complements outlets such as the Journal of Economic History, Explorations in Economic History, and the Economic History Review. Potential contributors are urged to contact the editor for information on submission requirements. Alexander J. Field Series Editor
xi
4259 Ch01 4/12/02 9:42 am Page 1
1 2 3 4 5 6 7 8
MADE IN GERMANY: THE GERMAN CURRENCY CRISIS OF JULY 1931
9 10 11 12
Thomas Ferguson and Peter Temin
13 14 15 16
ABSTRACT
17 18 19 20 21 22 23 24 25
The Great Depression reached a turning point in the currency crises of 1931, and the German banking and currency crisis was a critical event whose causes are still debated. We demonstrate in this paper that the crisis was primarily domestic in origin; that it was a currency crisis rather than a banking crisis; and that the failure was more political than economic. German banks failed in 1931, but the problem was not primarily with them. Instead, the crisis was a failure of political will in a time of turmoil that induced a currency crisis.
26 27
I. INTRODUCTION
28 29 30 31 32 33 34 35
The currency crises of 1931 have emerged as pivotal events in the propagation of the Great Depression. It was during these crises that central banks and governments had to choose between domestic economic stability and maintenance of the gold standard. With the prominent exception of Great Britain, most industrial countries opted for the latter, dooming the world to the horrors of depression. To understand the Great Depression, one needs to understand the currency crises of 1931.
36 37 38 39 40
Research in Economic History, Volume 21, pages 1–53. Copyright © 2003 by Elsevier Science Ltd. All rights of reproduction in any form reserved. ISBN: 0-7623-0993-8
1
4259 Ch01 4/12/02 9:42 am Page 2
2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
THOMAS FERGUSON AND PETER TEMIN
The German crisis of July 1931 in particular has been a topic of study. Given Germany’s location, size, and subsequent ill-fated history, historians have attempted to disentangle the many factors inducing the German crisis. It was a twin crisis in the sense that there was both a currency and a banking crisis, rather like the Asian crises of 1997. The causes of this twin crisis have been found in many places by many authors – in domestic and in international events, in the currency market and the banking system, and in economic and political malfunctions. Historians have differed widely on the emphasis given to these various dimensions.1 For example, Kindleberger and Eichengreen emphasized the international aspect of the German crisis (Kindleberger, 1986; Eichengreen, 1992). James and Balderston, in contrast, stressed the domestic nature of the crisis (James, 1986; Balderston, 1994). We assert in this paper that the German crisis of 1931 was primarily domestic in origin, that it was a currency rather than a banking crisis, and that the failure was more political than economic. A currency crisis is national, affected by government policies and actions; a banking crisis is the result of private actions affected by fears for the safety of bank deposits. We argue that the German government found itself in an impossible position as the recession of the late 1920s deepened and its budget went into deficit. The Weimar government lacked the political will to deal with the impending crisis, substituting rash statements about customs unions and reparations for serious budgetary action. There is no evidence of international contagion from the Austrian crisis two months earlier; there is little evidence of structural problems among German banks or fear for their stability before the statements of the Weimar government induced fears for the value of the mark. German banks failed in 1931, but only because the Weimar government put them under unbearable strain. We document this assertion with monthly bank data and weekly Reichsbank data. We set the stage by first reviewing the literature on the German crisis of 1931 and showing that economic conditions in Weimar Germany during the first two quarters of 1931 were bad, but not getting worse. We examine banking data in that context to search for evidence of a banking crisis. We then review in detail the budgetary problems facing the German government in early 1931, showing how the government’s inability to resolve this problem amidst rising international tensions brought on the crisis. A conclusion places the German crisis into the narrative of the Great Depression as a whole.
36 37
II. EXISTING LITERATURE
38 39 40
Historians have formulated narratives of 1931 that explicitly and implicitly take positions on the cause of the crisis. The initial German response to the crisis
4259 Ch01 4/12/02 9:42 am Page 3
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
3
was to blame foreigners for withdrawing their deposits in Germany and selling marks.2 In this story, the withdrawals were triggered by changes in financial fortunes on Wall Street or by contagion from Austria. They were not occasioned by events within Germany. This story was international because it placed the movements of international “hot money” at the center of attention. It was a story about currencies rather than about banks, and it involved the German government not at all. Kindleberger may be taken as the start of the modern English-language literature on the European depression. He stated, “At the end of May 1931, Austrian financial difficulties ramified widely and led to runs on the banks of Hungary, Czechoslovakia, Romania, Poland, and Germany” (1986, p. 148). Kinndleberger’s subsequent discussion focuses on the consequent problems of German banks and makes no reference to German budgetary problems. In this context of bank fragility, the Nordwolle failed in mid-June, setting off the final stages of the crisis (Kindleberger, 1986, pp. 148–149). Kindleberger thus saw the German crisis as international in origin and mostly concentrated in banks. Eichengreen extended this internationalist tradition. He also argued that the Austrian crisis affected Germany through its effect on confidence, even though German banks were not heavily invested in Austria. Confidence in Germany collapsed because foreigners could not distinguish between Germany and Austria and because Brüning took the occasion of the Austrian default to insist on a reduction in reparations. Eichengreen did not mention the budgetary problems of the German government. Instead, he asserted that “the failure of international cooperation was the key to Germany’s defection from the gold standard system” (Eichengreen, 1992, p. 276). Eichengreen agreed with Kindleberger that the German crisis was international and carried the argument further by focusing on currency problems and failures of omission by governments outside Germany. James presented quite a different story with much more attention to internal German problems. He represented the crisis as a run on German banks that contained “structural weaknesses” and were “fundamentally unsound” (1986, pp. 294–295). The main cause of instability, in other words, was internal to Germany. The small push that toppled this unsound structure, however, was the withdrawal of foreign deposits, primarily American. International aspects were important in this story, but secondary if the crisis was the result of unsound German banking. James also argued that the Germans had brought the international crisis on themselves by tolerating capital flight in 1930–1931 and by mismanaging their debt. James only introduced the debt problem into his narrative of the crisis at the last moment, arguing that the budget crisis surfaced on June 9 when revenue figures for April and May became available, that is, 3
4259 Ch01 4/12/02 9:42 am Page 4
4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
at a late stage of the crisis (James, 1986, p. 306). In contrast to Kindleberger’s and Eichengreen’s story of international contagion, James’ story concerned German banks in a context of government financial problems. James recently articulated a view much closer to Kindleberger’s. In a book about the Deutsche Bank, he stated, “The Creditanstalt[’s] . . . failure precipitated a general banking and financial crisis in Austria that then spread to the whole of central Europe and eventually brought down German banks as well.”3 This view is consistent with his earlier emphasis on banks, but it places a much greater emphasis on international contagion than his earlier account. Balderston attempted to disentangle the banking and currency crises even as he asserted both were present. He argued that the proximate cause of the crisis was a general run on bank deposits due to conflicts over reparations and fiscal crises. The ultimate cause of the banking collapse was the inability of the German banks to replenish their reserves either from foreign loans or the Reichsbank. Balderston argued that the inability of German banks to borrow abroad was the root cause of the crisis, for only in that case did they go to the Reichsbank. Like James, Balderston focused on the problems of German banks, but like Eichengreen, he attributed the crisis to a failure of the international economy (Balderston, 1994). Balderston elsewhere presented a detailed account of the budgetary problems of the Weimar government in which he asserted that “the connection between the state’s cash crisis of June 1931 and the banking crisis which had been developing since May is too obvious to need elaboration” (1993, p. 312). But he treated the budget question and the currency crisis quite separately in his book, and he did not emphasize connections between them either there or in his subsequent paper. Balderston accordingly ends up with two rather disconnected stories of the crisis. His main story is about international aspects of German banking. But in the background runs a quite different story about the political problems of the German government. The two stories may be compatible, but they remain quite distinct. Consumers of these histories have two choices. They simply can conclude that all the above factors were important and take away a story in which all these factors – domestic and foreign, banking and governmental, economic and political – figure as separate causes. Alternatively, and preferably, they can try to disentangle this complex knot of explanations and search for a unified story. This story would not ignore any of the factors noted above, but it would state clearly where the crisis originated and specify the channels by which it spread to become a national and international tragedy. It is the purpose of this essay to provide such an account.
4259 Ch01 4/12/02 9:42 am Page 5
German Currency Crisis 1
5
III. ECONOMIC CONDITIONS IN EARLY 1931
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
The banking crisis of June and July 1931 was integral to another, even greater catastrophe – the collapse of the Weimar Republic and the Nazi seizure of power. The aura of impending doom that clings to all accounts of the period makes it easy to perceive disaster lurking everywhere, and creates a false impression of almost mythical inevitability. Accordingly, the first step in developing a unified account of the crisis is to acknowledge the old adage that history proceeds not in straight lines, but in spirals, by distinguishing as carefully as possible among the stages of the slide to the abyss. A crucial point is that at the beginning of 1931, the German economy was down, but not yet completely out. Not that the economic data, summarized in Table 1, were anything but ghastly: Industrial production had fallen to three quarters of its 1928 level. Unemployment had risen to over 30% of the unionized labor force, and the prevalence of part-time work had risen from its “normal” level of under 10% to about 20%.4 The special fund that financed the Weimar regime’s proudest achievement – the 1927 law establishing a system of unemployment compensation – teetered on the edge of bankruptcy as payouts ran far ahead of taxes paid into it. As the central government reacted by tightening eligibility, raising taxes, and reducing payments, the armies of the unemployed grew increasingly desperate. Many were recruited into the swelling ranks of paramilitary groups run by parties and groups on the right; others joined groups run by parties of the left, notably the German Communist Party (KPD). Those who exhausted their benefits or who could not qualify for any, could sometimes obtain assistance from the patchwork of local welfare programs run mostly by cities or town, but these programs were rapidly overwhelmed by the crush of applicants and financial exigency. With many cities and states (Länder) in desperate financial straits and clamoring for bailouts from the central government, community cohesion was severely strained.5 But the economic situation, while desperate, was more complicated than has often been recognized. The contraction appeared to be ending in early 1931 as economic conditions improved slightly. It is necessary to sketch briefly the political intrigues of 1930 in order to appreciate how this apparent halt to the economic decline affected the German government’s budgetary deliberations. In the Spring of 1930, only days after narrowly ratifying the Young Plan revision of the Dawes Plan, the Reichstag deadlocked over the budget. Advised by General Kurt von Schleicher, President von Hindenburg refused to allow the government headed by Chancellor Hermann Müller of the German Social Democratic Party (SPD) to make use of the President’s power to enact measures 5
11
13
14
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Source: Wirtschafts und Statistik, 1930–1931.
9
15.4 16.1 16.9 19.2 19.5 18.9 18.1 17.4 17.7
10
23.6 26 31.7 34.2 34.5 33.6 31.8 29 29.7
8
77.1 75.6 72.2 67.8 69.2 73.5 76.3 73.9 74.4
843 829 850 1085 1065 1240 972 956 1034
10,191 10,272 10,400 10,767 10,946 11,064 11,165 11,225 11,074
6
October1930 November 1930 December 1930 January 1931 February 1931 March 1931 April 1931 May 1931 June 1931
17 Savings Bank Deposits (Million RM)
5
Month
7
Bankruptcies (Monthly Total)
4
Partially Unemployed (% Trade Unionists – See Note 4)
3
Unemployment (% Trade Unionists – See Note 4)
2
Industrial Production (1928 = 100)
16
German Economic Data.
15
Table 1.
4259 Ch01 4/12/02 9:42 am Page 6
THOMAS FERGUSON AND PETER TEMIN
1
6
12
4259 Ch01 4/12/02 9:42 am Page 7
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
7
by decree under Article 48 of the Weimar Constitution. When the government resigned, the President, a former general and hero of the battle of Tannenberg, who had personally helped mint the “Stab in the back” legend after World War I, offered the Chancellorship, and the use of Article 48, to Heinrich Brüning of the Catholic Zentrum party.6 Brüning, who, in Eberhard Kolb’s words, “from start to finish regarded himself as holding office solely by virtue of the President’s authority and in order to fulfill the mission assigned to him,” immediately ran into heavy weather in the Reichstag, which could still overturn Hindenburg’s emergency decrees (Notverordungen) by majority vote.7 The major stumbling block was the same issue as before: production and trade were falling as expenditures for relief and unemployment compensation soared, unbalancing the budget and threatening further rapid increases in government debt at all levels (federal, Länder, municipalities). As shown in Table 2, Germany’s total government debt rose from 15 billion RM at the end of March 1928 to 24 billion RM in 1931.8 The federal government accounted for about half of the consolidated total, indicating that all levels of government were running deficits as the economy declined.9 While total foreign debt had risen, it did not comprise the bulk of government debt. In addition, relatively little of the official foreign debt was short-term. Most government debt was held domestically, and the volume of short-term debt rose with the total. (The sum of foreign and domestic debt falls short of the total by the amount of “old debt.” As can be verified by subtraction in Table 2, the volume of “old debt” did not change.) Pressed to get this mountain of debt under control, Brüning enacted by Notverordung a budget calling for savage spending cuts, lower taxes, cuts in the pay of government officials, and smaller transfers to the Länder and municipalities. The move outraged many on both the left and the right for different reasons. After the big business oriented German People’s Party forced the resignation of the Finance Minister and the conservative German National
30 31
Table 2.
German Government Debt (Million RM).
32 33 34 35 36 37 38 39 40
All governments Reich only Foreign debt Foreign Sh Tm Domestic debt Domestic Sh Tm
3/31/28
3/31/29
3/31/30
3/31/31
14,599 7,139 2,157
18,159 8,229 2,294 185 9,263 2,569
21,319 9,630 2,595 532 12,415 4,070
24,022 11,342 4,799 830 13,181 3,746
5,547
Source: Länderrat, 1949.
7
4259 Ch01 4/12/02 9:42 am Page 8
8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
People’s Party (DNVP) committed binary fission, a majority of the Reichstag exercised its constitutional right to reject the decree. Hindenburg and Brüning reacted with a fateful step: They dissolved the Reichstag, called new elections, and re-enacted the Notverordung with minor changes.10 A disaster for the government ensued. In the September elections, voter turnout surged and the Nazis dramatically emerged as the second largest single party in the Reichstag, while the Communists made substantial gains (Winkler, 1993, pp. 388ff.). Parties supporting the governing coalition now controlled only about a third of the Reichstag, with another third dominated by parties fundamentally opposed to the existing political system.11 Not surprisingly, investors stampeded out of Mark assets. In a single month, 800 million Reichsmarks worth of foreign capital left the country, while German securities fell sharply in foreign markets. The Reichsbank was forced to raise its rediscount rate by a full point.12 It is here that the temptation to write the history of the 1931 crisis in straight line terms has been most meretricious. Many accounts treat the crisis of June and July as a seamless projection of the late fall, 1930 panic. In some versions, the capital outflow never really abated. According to others, the 1930 crisis closed world capital markets to German enterprises and, especially, the German government, making it impossible for the government to service either its foreign or domestic debts. A recent view suggests that the overhang of Young Plan loans, which had priority for repayment in a crisis, made it impossible to float new ones.13 What actually happened is that the government responded by redoubling its resolve to continue down the path of austerity so resoundingly repudiated by the electorate. Amid a flurry of sympathetic news stories in papers from Berlin to New York, Hindenburg immediately tapped Brüning to form a new government. Solemnly promising to make fiscal reform and deficit reduction his top priority, Brüning reappointed the entire cabinet. On the basis of more Notverordungen and additional measures mandating a sinking fund for debt retirement, steep rises in workers’ contributions to the unemployment fund, higher taxes (including new levies on mineral water and individual citizens), and slashes in government spending, civil service pay, and inter-governmental transfers, the government floated a $125 million loan through a syndicate headed by Lee, Higginson.14 The spectacle of the government bowing to foreign creditors (whose connection to the austerity package was too obvious to escape acrid public comment at home) reduced the government’s popularity still more. But fears of what would happen if Brüning’s government collapsed forced the shell-shocked SPD into an agonizing reappraisal. Staring into the abyss, the party reluctantly
4259 Ch01 4/12/02 9:42 am Page 9
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
9
concluded that while it could not openly support the Chancellor, it would have to “tolerate” his government as the lesser evil. When the Social Democrats flirted with the idea of voting down some of the budget measures, the Chancellor brought them into line by threatening a Zentrum pullout of the SPD-led government in Prussia. In the murderous climate of the times, the SPD simply could not afford to lose control of the police force that controlled the largest of the German Länder.15 With this new, wider support, the Brüning government moved to reassure financial markets anxious about the domestic political scene and trial balloons the government had discreetly floated about possibilities for further concessions on reparations. Encouraged by the British government and the Bank of England (whose Governor, Montagu Norman, endorsed a plan to have the new Bank of International Settlements inject liquidity into central European financial markets), the Brüning government and the Reichsbank joined forces to strike back at former Reichsbank head Hjalmar Schacht and other critics who were campaigning for unilateral revision or repudiation of the Young Plan. Flatly ruling out action in the near future, Brüning instead preached that further reparations revisions would have to wait until Germany demonstrated to the rest of the world that the limits of her ability to pay had truly been reached by reining in the budget and reforming the fiscal system.16 Both governments and financial markets swiftly perceived that “while Dr. Brüning certainly continues to cherish hopes that, by careful internal and external preparation, he will eventually secure, by mutual consent, a further reduction of Germany’s reparation liabilities . . . [he] has become thoroughly convinced that the disadvantages which would arise from an appeal to the safeguarding clauses in the Young plan far outweigh the partial and temporary relief which Germany would obtain therefrom, and is in no mood to shortsighted pressure in this direction.”17 The government’s show of resolve rallied support in and out of Germany. In speeches that made news even in New York (the financial center with the most money of all invested in Germany), Brüning explained “like a patient and firm schoolmaster” that after cutting spending, the government would seek to consolidate its huge short term floating debt into longer term obligations (New York Times, February 6, 1931, p. 9).
35 36 37 38 39 40
In public speeches both before and after the Reichstag met on the 3rd of February the Chancellor showed courageous frankness and his increasing confidence infected the leaders of the Government parties . . .. The large majorities obtained by Dr. Brüning in the opening days of the present session of the Reichstag again impressed public opinion with his qualities of leadership, and showed that the other political parties had at last realized the necessity of standing up to the extremists.18
9
4259 Ch01 4/12/02 9:42 am Page 10
10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
International support for Brüning grew as his government made good on its tough talk. In January, the International Institute of Finance, a think tank with strong ties to financiers interested in European debt, provided timely reassurance with a widely heralded survey concluding that foreign investors need not fear that Germany might be forced to repudiate private debts in the new year. The British Foreign Office started exploring ways “His Majesty’s Government” might “give [Brüning’s government] such support and encouragement as they properly can, in order to fortify its position.” BIS officials delivered speeches calling for an increase in international lending and promising to lead the way in central Europe with the bank’s own money.19 With confidence reviving, official statistics on capital outflows and banking positions improved markedly. Studying the figures, the Economist considered that they gave “every sign of returning calm” and “the cessation of the withdrawals of money, especially foreign withdrawals.” Private bank reports mostly echoed the public media, if initially somewhat more guardedly. In a mid-January report that was certainly never intended for wider publication, Morgan et Cie., the Paris affiliate of the most influential bank in the world, reported to its New York parent that “German bankers report money beginning to return to Germany and they anticipate some weakening of rates.”20 Somewhat later the Reichsbank cabled essentially the same message to the New York Fed.21 Other favorable developments bolstered the growing confidence. First and foremost was the fact that the balance of payments was firmly in the black. Germany had been importing capital steadily since the stabilization of the Reichsmark. It had in fact borrowed considerably more than its reparations obligations, leading to what Schuker (1988) called American reparations to Germany. While the net capital inflow decreased in 1930, the decline in capital imports was offset by a fall in imports of goods and services. Relevant data are set out in Table 3. Exports reached their peak in 1929 and fell thereafter, but imports fell faster, leading to a rise in net exports. Capital imports also fell in 1929 and 1930. These two movements largely offset each other, as of course they must to make the balance of payments add up. (Other, smaller items in the balance of payments did not change as rapidly as these categories.) Net income, the resultant of all these forces, was remarkably stable. The need to pay reparations meant that Germany’s balance of payments was not in equilibrium when net imports equaled the capital inflow, but rather it was in equilibrium when the capital inflow exceeded net imports by the amount of reparations payments, which amounted to between one to two billion RM, depending on the time period one reckoned over. This represented a shift of the location of equilibrium and a burden on Germany, but it did not change the short run dynamics of the balance of payments. It was German reluctance
4259 Ch01 4/12/02 9:42 am Page 11
German Currency Crisis 1
Table 3.
11
German Balance of Payments, 1928–1931 (Million RM).
2 1928
1929
1930
1931
13,599 13,912 313 6,958 2,852 4,106 0 931 931 2,862 1,999
14,344 13,676 668 4,459 1,991 2,468 510 345 165 3,301 2,501
12,713 10,617 2,096 3,678 3,195 483 192 72 120 2,699 1,699
10,169 6,955 3,214 3,817 6,494 2,677 1,653 0 1,653 2,190 990
3 4 5 6 7 8 9 10 11 12 13 14
Exports Imports Net exports Capital inflow Capital outflow Net inflow Gold exports Gold imports Net gold exports Net income Reparations
Source: Statistisches Jahrbuch, 1933, pp. 498–499, as quoted in Schmidt, 1934.
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
to pay reparations in the longer run that made them a critical political issue.22 In early 1931, amidst the improving payments situations, the economics of the transfer issue did not seem insuperable. The consensus view was that “there will be no difficulty in making transfer” of the payments due under the Young Plan.23 Other economic statistics trickled in that fanned this optimism. Widely reported indices of German industrial production rose slightly, as can be seen in Table 1. Unemployment began to trend very slightly down. Bankruptcies – even now esteemed as one of the few clear external indicators of possible bank lending difficulties – started falling in February, rose slightly in March, and then fell sharply in April and May (Table 1). Indeed, almost every indicator that subsequent analysts have suggested as heralding banking or currency crises improved modestly in the early months of 1931. With interest rates – even long rates – falling, the German stock market rose for four months, while the bond market – including, notably, the internationally traded Young Plan bonds, surely a prime index of global sentiment toward Germany’s repayment possibilities – rallied sharply, as shown in Table 4.24 In an age that still venerated gold as the ultimate test of sustainability, Reichsbank holdings of gold and foreign exchange mounted slowly, but steadily. From 2216 million RM. at the beginning of 1931, Reichsbank gold reserves rose to 2323 million RM by the end of March. In January, Reichsbank holdings of foreign exchange dropped from 400 million to 200 million RM. As the Brüning government redeemed its budgetary promises, the outflow slowed and then reversed (see Table 8). By the third week of March – a fateful date, as 11
11
12
13
14
17
18
19
20
21
22
23
24
25
26
27
29
30
31
32
33
36
37
38
39
40 85.58 85.85 85.78 86.64 87.46 87.32 86.28
– 80.51 80.61 81.58 82.47 81.7 78.44
70.5 72.75 78.75 80 80.63 70 75
Young Bonds
1
4
28
34
35
Sources: Stocks, Overnight Money, Gold Bonds (monthly averages): Deutsches Institut für Wirtschaftsforshung, 1932, p. 9. Bank Stocks, Public Bonds (monthly averages): Reichsamt, 1932b, pp. 354–355. Young Bonds (value at end of week nearest end of month); See Note 93 on June entry): Commercial and Financial Chronicle, 1931.
5.66 5.1 5.78 5.31 5.88 5.64 7.07
9
– 110.52 111.1 112.65 112.61 106.09 100.58
8
87.3 81.8 85.6 91.1 92.4 83 75.9
7
6% Public Bonds
12
December 1930 January 1931 February 1931 March 1931 April 1931 May 1931 June 1931
10
6% Gold Bonds (Gold-pfandbriefe)
6
Month
5
Overnight Money (Tagesgeld) Percent
3
Bank Stocks (Kredit-banken) 1924–1926 = 100
2
All Stocks, 1924–1926 = 100
16
German Financial Data.
15
Table 4.
4259 Ch01 4/12/02 9:42 am Page 12
THOMAS FERGUSON AND PETER TEMIN
4259 Ch01 4/12/02 9:42 am Page 13
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
13
will shortly become evident – the rise in the Reichsbank’s total holdings of gold and foreign exchange appeared to bear out the projections of the The Economist cited earlier. Inevitably, some optimistic commentators hailed these modest upticks as evidence that the invisible hand was at last beckoning the German economy on. It is not necessary to share this touching faith – or its corollary that only subsequent political disaster aborted recovery – to acknowledge that in the early months of 1931 “authoritative banking opinion” held that “the Reich will be able to borrow sufficient funds to meet her domestic and international obligations” for the coming year. The international bankers who have had occasion recently to investigate the conditions of Germany are optimistic regarding the future. Complex problems confront the German government, but in Dr. Heinrich Brüning they see a quiet, forceful leader, capable of guiding the nation through a difficult period. His desire to permit nothing of a domestic nature to influence the calm, reasoned course of Germany’s foreign relations is accepted by the bankers as a fairly sure guarantee of a new period of stability (New York Times, Feb. 8, 1931, p. 8).
In February, through a syndicate led once again by Lee, Higginson, the government successfully returned to international money markets, borrowing RM 120 million for eighteen months to cover the deficit Brüning was pledged to reduce.25 As the Nazis and the strongly nationalistic DNVP walked out of the Reichstag in protest, giving Brüning a comfortable working majority, he momentarily acquired a stature almost larger than life. Leading international dailies echoed the views, quoted earlier, of the British ambassador, affirming that the Chancellor was “stronger than ever,” while their business pages gushed about “the particularly welcome improvement in the Berlin money market due to the continuing inflow of foreign funds, largely in the form of three month credits . . .. Bankers express the view that the ‘confidence crisis’ is ended” and that “further foreign credits will come in and correct the wide disparity between rates at home and international capital markets” (New York Times, March 23, 1931, p. 33). The revival of confidence reached its zenith in March. Highlighting his government’s success in reducing the level of short term debt, Brüning succeeded in passing his budget through normal parliamentary processes (i.e. without recourse to Notverordungen). Then he persuaded the Reichstag to recess until October. [This] probably secures the country against an internal political crisis in the next few months. The attitude of the foreign banks has correspondingly improved and money has flowed into Germany from Holland, Switzerland, France, and Belgium. Berlin banks accepted the funds at reduced rates, but in general only to maturities of three months, while they paid
13
4259 Ch01 4/12/02 9:42 am Page 14
14 1 2 3 4
THOMAS FERGUSON AND PETER TEMIN one-month credits. This policy is due to the banks’ conviction that they will soon get foreign money cheaper.26
One telegram in particular is uniquely authoritative about what “markets” were thinking about Germany at that time:
5 6 7 8 9 10 11 12
Though it has not yet reached the stage where any large foreign offerings are immediately practicable, there has already been a considerable improvement in the market position of many foreign issues and the German Government 51/2s as you will have noticed have improved from about 73 around the 1st of February to about 82 at present quotations. This reflects the market’s greater confidence in the German situation and in the prospects for political security in Europe as well as the assurances given by Dr. Luther and prominent German officials to the effect that Germany has no intention of stopping payments on her obligations to private investors, including specifically the Dawes and Young loans.27
13
A few days later, one of the authors of this rosy scenario went a step further:
14
Some years ago, I made a bet that French credit would rise to a 41/2% basis on our market. That bet I have collected. It emboldened me to make another which was that the German 51/2s would sell at par by January 1, 1935. I feel confident that if I live that I shall collect this bet also.28
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
The most dramatic development in regard to Germany’s potential creditworthiness during this period has been almost entirely forgotten. In February came sudden news that the French were reconsidering their previously hostile attitudes toward possible long term loans to Germany. Save for the Young Plan loan, French banks had not participated in any major credit to Germany since 1929. Deeply alarmed by results of the September election, they had declined any part of the earlier Lee, Higginson credit. Paris newspaper accounts at that time intimated that France, then the second largest creditor nation in the world thanks to the enormous inflow of gold that followed the Poincaré government’s undervaluation of the Franc, possessed the financial clout to bring down the German economy by calling in loans.29 In December, however, the conservative, nationalist government headed by Andre Tardieu collapsed. After a brief caretaker regime, a new government headed by Pierre Laval took over with Aristide Briand, the long time champion of Franco-German cooperation, serving as Foreign Minister. Not for the last time in French history, a Laval-led government tilted in the direction of its bigger neighbor. Briand, who had his eye on the election for the Presidency of France scheduled for May, 1931, had for some time been pushing for a change in policy to head off the rise of a still more nationalist government in Germany by easing tensions and reviving the economies of both countries. Laval was also working with important Belgian bankers and diplomats, who were anxious over the steady deterioration of Franco-German relations. In both Berlin and Paris, Lee, Higginson also actively promoted rapprochement, not least by taking
4259 Ch01 4/12/02 9:42 am Page 15
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
15
on a senior French banker with glittering policy credentials, who was close to many top French policy makers.30 Gradually a disbelieving world awoke to the fact that something remarkable was happening. First came hints that Lee, Higginson was once again considering a large loan to Germany. A few days later came the sensational announcement that at the request of the French government, the giant Banque de Paris et des Pays Bas would participate in the credit. This news “made a profound impression on the Paris bourse, where the move was widely interpreted as the initiation of a new foreign credit policy.”31 At the General Assembly of the shareholders of the Banque de France, Governor Moret “pointed to the possibility of more French lending abroad and urged the reduction of restrictions in the bond market.” In a speech in Paris on February 12, BIS President Gates McGarrah warmly endorsed Moret’s views, triggering a lively discussion of how recycling stored up French funds could help restore international financial equilibrium.32 The sudden realization that the Paris market might be far more lucrative than anyone could have imagined only a few months before momentarily threatened to disequilibrate the market for German goverment debt – but in a direction very different from that emphasized by recent German historians. Drawing mainly on official records and papers left by senior civil servants, James and, especially, Bachmann, paint decidedly pessimistic pictures of that market in this period.33 Since only a short time later demand for German government obligations did indeed collapse, annual statistics for total loans inevitably appear to validate their pessimism. But the archives of private bankers reveal a much more complex and interesting situation, brought on by widespread interest among investment bankers in German debt. Lee, Higginson, the lead bank in most of the international syndicates that brought out German government bonds at this time, was acutely aware that several rivals were after the business. It also understood that its position in Germany depended crucially on its ties to Mendelssohn & Co., the legendary Berlin private banking house. This firm was close to many prominent officials in the German government, including Reichsbank President Hans Luther and Hans Schäffer, who held the key post of State Secretary in the Ministry of Finance.34 As prospects for new German loans brightened, Mendelssohn became restive and started pressing Lee, Higginson for a better deal. It first sought preferences in northern markets where it had historically been strong. This boomeranged when a major Swedish bank made its participation in future syndicates conditional on the leadership of Lee, Higginson. In mid-February, however, with French banks flocking to take up parts of the new German credit, the 15
4259 Ch01 4/12/02 9:42 am Page 16
16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
THOMAS FERGUSON AND PETER TEMIN
German bank shocked its American partner by declaring “positively that the Paris market must be reserved for Mendelssohn and Company, Amsterdam” because of its “close connections with leading French institutions, particularly Banque de Paris et des Pays Bas.”35 At the time, Lee, Higginson was campaigning to persuade the German government to designate it as the government’s exclusive manager for future internationally syndicated loans. Still, the American bank, which had its own Paris affiliate and was trying to build up its business, flatly refused its German ally’s demands. In the end, “certain domestic factors, principally the opposition of Schacht” dashed the American house’s hopes of an exclusive franchise.36 Opposition from some key officials in the Reichsbank also blocked other concessions Lee, Higginson proposed to placate Bank Mendelssohn until the whole struggle over expected future profits from French loans to Germany became moot. But these rivalries over the German government’s debt offerings strikingly indicate just how little economic conditions in the early months of 1931 hinted at the crisis that was just around the corner. If conditions were not good in these harsh economic times, there was little anticipation that they were about to get a lot worse. Indeed, expectations appeared to argue the opposite, that conditions were on the mend. Yet in the second quarter of 1931, something went wrong. The question is whether the change was in the stability of banks or of the Reichsmark, in Germany or without, in politics or economics.
22 23
IV. THE BANKING CRISIS
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Once the crisis began, panic spread and everything fell apart. But describing the crisis does not help us discriminate between its putative causes. We must examine the process that led up to the crisis. In this section we present banking data to distinguish the effects of the budgetary crisis from two other antecedents of the crisis whose effects also have been touted in the literature: the Austrian currency crisis in May and the weakness of German banks. The tests are rather different. To see if the Austrian crisis mattered, we examine the German reaction to the Austrian crisis. Did the Austrian crisis set off a sequence of events that led to the German crisis in July? To see if banks were unstable, we examine various indicators of bank performance in the months before the German crisis to see if there were signs of distress. The source material is almost the same for the two tests. The difference is that the first set looks specifically at timing, while the second examines banking indicators in a longer-run perspective. There is an extensive literature about German banks, as noted above, but there is not a clear idea what to look for. At the risk of narrowing our focus
4259 Ch01 4/12/02 9:42 am Page 17
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
17
excessively, we make explicit the implicit model of other scholars. When we talk about a banking panic, we refer to an American-style banking panic.37 In this kind of panic, depositors become fearful for the safety of their deposits. They withdraw their deposits, starting with the most liquid since they are in a hurry to get their funds into a more secure asset – cash as opposed to bank deposits. Banks operating with fractional reserves have to draw down their reserves in order to pay out deposits. Their reserve ratio falls. They call in loans to rebuild their reserves. Being short of reserves, they exert pressure on borrowers to pay up as soon as possible. In order to pay off these loans, borrowers need to withdraw their deposits from the banks where they keep money. The bank withdrawals that may have started with a single bank or a small number of banks then spread to all banks as the ones affected initially call in loans that force other banks to pay out deposits. The banks draw down their reserves as they pay off their deposits, forcing them to call in their loans. Soon all banks are calling in loans. Everyone is scrambling for liquidity, and there is not enough liquidity to go around. There are two important steps in such a banking panic. First, scared depositors draw out their deposits. Since the spread of a banking panic is caused by the attempts of banks to rebuild reserves quickly, a marker of a panic is that people withdraw money from sight accounts. If they withdraw term deposits, they give the bank time to reconstruct its reserves without imposing as much pressure on other banks as when they need immediate help. Second, banks call in loans in order to rebuild their reserves. If they have some other way of getting reserves – say, from a central bank – they do not transmit pressure to other banks and initiate and intensify a general banking panic. We examine German banking data to see if these two conditions were present. The following tables contain relevant banking data from 1931.38 Table 5 reports deposits by bank class. The six Berlin Grossbanken are listed first. Kreditbanken are listed next. Finally, the Kreditbanken other than the Grossbanken are obtained by subtraction. Deposits are divided between those accessible within a week, labeled demand deposits (DD), and those accessible in eight days to three months, labeled time deposits (TD).39 Table 6 contains data on the largest assets of the Grossbanken.40 Table 7 reports time deposits for individual Berlin great banks.41 The first thing to note is that demand deposits did not fall at all in the crisis. If there had been a banking crisis, one would have expected these deposits to be withdrawn faster than all others. They were the easiest deposits to withdraw, and depositors should have converted them to cash or foreign exchange at once. Their stability suggests calm rather than panic. We do not mean to argue that 17
4259 Ch01 4/12/02 9:42 am Page 18
18 1
THOMAS FERGUSON AND PETER TEMIN
Table 5.
Bank Deposits by Access (Million RM) – by Bank Class.
2 3 4 5 6 7 8 9 10 11 12 13
All Grossbanken DD TD All Kreditbanken DD TD Kredit. – Gross. DD TD
28/2/31
31/3/31
30/4/31
30/5/31
30/6/31
31/7/31
3756 4627
3819 4666
3657 4801
3626 4632
3626 3519
3891 2370
4528 5578
4562 5668
4390 5850
4382 5611
4327 4432
4749 3152
772 951
743 1002
733 1049
756 979
701 913
858 782
Source: Wirtschaftsdienst, 1931, various issues.
14 15 16
Table 6.
17 18 19 20 21 22 23 24 25 26 27
Schecks, Wechsel Remboursekredite Total Short-Term Loans against securities against physical assets Deposits at other banks
Assets (Million RM) of the Grossbanken. 28/2/31
31/3/31
30/4/31
30/5/31
30/6/31
31/7/31
2497 2006 5896
2530 1894 5890
2528 1828 5834
2547 1781 5734
1914 1748 5668
1280 1599 5484
1307 3195 946
1309 3214 856
1276 3219 981
1235 3183 857
1228 3071 686
1055 3014 546
Source: Same as Table 5.
28 29 30 31
Table 7.
Time Deposits (Million RM) – Individual Grossbanken.
32 33 34 35 36 37 38
De-Di Dresdner Danat Compri Reichskredit Berliner
39 40
Source: Same as Table 5.
28/2/31
31/3/31
30/4/31
30/5/31
30/6/31
31/7/31
1778 871 977 621 181 198
1789 874 989 639 176 199
1799 931 1038 650 187 197
1760 869 967 533 215 187
1444 631 647 518 133 145
1009 490 300 363 160 109
4259 Ch01 4/12/02 9:42 am Page 19
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
19
conditions in Germany were at all calm in June 1931; rather we argue that fears of that time were not directed toward large banks. The second thing to notice is that time deposits also stayed steady, although not so long. They were steady through the end of May. In fact, they were in the range that they had been in prior years. They ranged from 4,359 million RM in May 1929 to 5,682 in August 1930. Demand deposits also were in the range of the immediate prior years: 3,434 in May 1929 and 4,289 in September 1931 (Balderston, 1994). The stability of deposits through May indicates that the failure of the Creditanstalt did not lead to a panic of German banks. The third conclusion is more complex. Time deposits fell sharply in June 1931. As shown in Table 7, all Grossbanken lost time deposits. But the fall appears to have been concentrated in these six banks; the data in Table 5 show that time deposits in other credit banks did not fall nearly as much. If this fall is a sign of panic, it suggests a peculiar sort of panic. It extended to all the great banks, but not beyond. There is no sign of the contagion from bank to bank that is the hallmark of a general banking panic. Instead we find depositors shifting from longer to shorter loans, from less liquid to more liquid assets, as if they were preparing for a possible future problem. This does not suggest panic, but rather precautionary activity by depositors on a small enough scale not to threaten any bank’s stability. This movement does not indicate a loss of faith in banks; it suggests instead that some depositors feared for the value of the mark. This conjecture is supported by the data on bank assets in Table 6, reporting the assets of the six Grossbanken. As shown in Table 5, deposits of these banks fell by 1,000 million RM in June. Assets must have fallen by the same to balance their books. The table shows the largest categories of deposits; other smaller categories did not fall enough to matter. Loans to business firms did not fall, although there was a small decline of loans against physical assets, i.e. loans against the security of the firm as a whole. Checks and bills of exchange (Wechsel) fell by 600 million RM, accounting for almost two-thirds of the fall in deposits. Deposits at other banks fell by 200 million RM, and acceptances held by the banks (reported under liabilities) rose by the same amount. There is no sign of banks calling in loans, only of a decline of short-term financial assets as the volume of trade collapsed. The fall in deposits at other banks was small. It was also offset by a fall of roughly the same amount in the deposits from other banks included among the deposits in Table 5. The Grossbanken had very small net positions in deposits to and from other banks, and their exposure did not change. Even the fall in the gross position was only one-fifth as large as the fall in time deposits. This means that the fall in time deposits was not restricted to deposits from other 19
4259 Ch01 4/12/02 9:42 am Page 20
20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
banks, and changes in interbank funds did not create hardships for the Grossbanken. The rise in acceptances is mysterious. As noted, this is not a large change, but it contrasts with the fall in bills of exchange held and the decline in time deposits. British banks had substantial short-term loans outstanding to Germany; many of these were in the form of acceptances in order to earn a higher rate of interest than on deposits.42 As a result, any bank problem from the fall in time deposits in June was a smaller one than the deposit data suggest since the rise in acceptances offset part of the fall in time deposits. Nothing in these data suggests over-expansion or other excessive activity on the part of the Grossbanken. The pattern on the asset side does not look more like a bank panic than the pattern on the debit side. Fewer bills of exchange reflect changes in the international markets, whether of trade or of currency. Trade was declining as the Depression deepened, and traders began to hold their assets outside of Germany as they feared for the Mark. None of this suggests a banking panic. There is no contraction of loans by banks under attack which weaken other banks which in turn come under attack. There was no rush to withdraw the most liquid deposits and no inter-bank pressure to recoup reserves. Comparing these data to the pattern of banking crises shows that banks were not calling in loans that forced borrowers to withdraw deposits at other banks. The Grossbanken were not acting in a banking panic. The Danat bank experienced more withdrawals in June than some other great banks, but it lost proportionately slightly less than the government-owned Reichskreditgesellschaft (RKG), as shown in Table 7. Depositors do not seem to have singled it out as dangerous in June, despite its deep involvement with the Nordwolle failure. Depositors did draw more out of the Danat bank than from RKG in May and July, but it is hard to know what they were responding to. Nordwolle was not expected to fail in May, and deposits were guaranteed in July.43 For whatever reason, time deposits at the Danat bank had fallen the most of any great bank from the end of March to the end of July. Demand deposits at the Danat bank, however, had not fallen at all – more correctly, they fell slightly and recovered in July. These data support two conclusions. First, the collapse of the Creditanstalt did not affect German banks. There was no change in either the liability or the asset side of bank balance sheets in May. Indeed, there is no sign that the Austrian crisis had any effect on German banks at all – the Reichsbank was simply stating the truth when it privately advised the New York Fed right after the collapse that “our market is scarcely influenced by the reorganization as neither the Oesterreichische Credit-Anstalt für Handel und Gewerbe has large
4259 Ch01 4/12/02 9:42 am Page 21
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
21
credit balances here, nor [were there] considerable demands of our market on Vienna. Therefore money market calm.”44 Second, there is also no evidence of longer-run problems with the Grossbanken as a group. The data in Tables 5–7 suggest nothing so much as banking stability before June. The fall in time deposits in June may have been the result of depositors preparing for a future crisis by shortening the maturity of their deposits. But even this is not a clear signal. As noted already, this net movement was the result of offsetting changes, and there is no sign of panic in the steadiness of demand deposits. Due to the articulated structure of the German banking system, tests on these data are incomplete. American banks held deposits of both firms and individuals. If ordinary people feared for their bank balances, their actions would show up in the data on demand deposits. German banks were more specialized. The Grossbanken held deposits of business firms and prosperous individuals, while other types of banks held those of ordinary people. The data in Tables 5–7 therefore show that the business community was not frightened by bank behavior, but they do not indicate if people in the street were anxious. For that, we have to turn to data for other types of banks. Data for smaller savings banks, whose business was local and accounts were almost entirely in Reichsmarks, are revealing. All through the winter and spring of 1931, small savers were not fleeing their local banks. Indeed, at the end of May, three weeks after the announcement of the failure of the Creditanstalt that allegedly triggered the banking crisis, German savings banks actually held more deposits than they did at the beginning of the month. Only in June is there any sign of withdrawals, and the losses were not large. Roughly the same is true of the so-called “cooperative” banks.45 Finally we need to examine the deposit to currency ratio to make sure that the stability of bank deposits did not mask a shift of monetary holdings into cash caused by distrust of banks. Again, the parallel with the United States is not exact. Germans economized on cash through use of the Giro system, and the data for currency in Germany do not refer to the same part of the economy as the American system.46 While conventional accounts of the banking crisis do not focus on the types of banks that held deposits of ordinary people, the deposit-currency ratio may still be informative. It fell about five percent in April, but remained constant in every other month in the first half of 1931. There was no cumulative decline in the deposit-currency ratio as there was in the United States.47 Could the “banking weakness” story be saved by abandoning all reference to the Creditanstalt, switching models, and pointing away from American style bank panics to Japanese style slow motion asset deflations? While accounts of the crisis by Balderston, James, and others all take some kind of banking panic for granted, 21
4259 Ch01 4/12/02 9:42 am Page 22
22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
THOMAS FERGUSON AND PETER TEMIN
their emphasis on long building banking weaknesses does not fundamentally involve – save perhaps in the final stages – appeals to chain bankruptcy at all. As the contemporary example of Japan reminds us, banks can also collapse and cripple an economy through the slow decay of their assets and corollary inability to lend (Goldstein, Kaminsky & Reinhart, 2000, p. 20). The Nordwolle failure did bring down the Danatbank, and other large German banks, including the giant Dresdner, also failed. Still other large banks, such as the Deutsche Bank, appear to have had narrow escapes. We therefore cannot dismiss flippantly the notion that the German banking system was structurally fragile after three or four years of depression. Perhaps the large banks took excessive risks out of the conviction that they were too big for the Reichsbank to allow them to fail.48 It is not easy to tell if this story about deteriorating assets is true in the absence of archival evidence about bank portfolios. But the available evidence is less than overwhelming. We have seen that savers were not running on the banks as the crisis approached. They evidently did not sense collapse coming. Neither do bank stocks, one of the few external indicators of the onset of banking crises recommended by contemporary analysts of financial vulnerability, appear to provide support. While no comprehensive review of the stock market evidence is known to us, existing stock indices do not indicate unusual sell offs of bank stocks relative to other sectors (see Table 4). Since the banks were widely suspected of buying their own stocks to keep up their price, this evidence is suggestive, rather than conclusive, but the general rise in the stock market through April has to have helped banks, since they held large blocs of industrial shares.49 As late as May 9, 1931, The Economist reassured readers that German banks’ “gross profits have not declined very substantially; their decline has been counterbalanced by savings in expenses and taxation.”50 Another article in the same issue favorably compared the level of capital and reserves of the German banks to those of the British (The Economist, May 9, 1931, p. 8). Bankruptcy statistics – the other obvious external indicator of asset unsoundness – were also headed in the wrong direction (Table 1). By contrast, the devastation wrought by the currency collapse is clear. Without the run on the large banks’ foreign deposits induced by the currency collapse, the Danatbank and the rest of the banking system would have been far better armored against the Nordwolle shock. The Reichsbank would have had ample means to step in quickly, guarantee the deposits, and abort the final panic.
36 37
V. THE CURRENCY CRISIS
38 39 40
Why, therefore, were depositors worried in the late Spring of 1931, if they were not anxious about the soundness of German banks? Our answer is straightforward:
4259 Ch01 4/12/02 9:42 am Page 23
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
23
They were alarmed by the latest manifestation of the Weimar Republic’s long running fiscal crisis. This now threatened to go completely out of control and lead – for reasons of domestic politics – to a moratorium or repudiation of Germany’s foreign currency debts and to default or postponement of payments even on domestic debt. The first step down the fatal path came with the decision of Hindenburg and his advisers to push Brüning towards a more aggressive and nationalistic foreign policy. Soon after taking the reins, the Chancellor ostentatiously snubbed a proposal advanced by French Foreign Minister Briand to broaden the “Locarno System” of European security arrangements.51 Tensions heightened late in 1930, when the German government denounced the Draft Disarmament Convention for reaffirming the Versailles Treaty’s restrictions on German rearmament, and putting little pressure on the other major powers to fulfill their own pledges to disarm.52 More ominously, Brüning, prodded by Hindenburg, the Reichswehr, and a vast “National Bloc” of industrialists, Junkers, and right wing political groups pushed ahead with the building of the notorious “Panzerkreuzer B,” the second in a series of pocket battleships that were designed to evade the naval limitations of the Versailles Treaty. This step made little sense in terms of either the existing military balance or the Reich’s budget situation. It also made it much harder for even sympathetic SPD leaders to assist the Chancellor, tacitly or otherwise.53 Over time, such moves eroded the basis for cooperative, internationalist strategies. Because the point can easily be misunderstood, it is worth observing that Brüning and his cabinet were not exponents of autarchy, though they were markedly more inclined to compromise on trade issues than earlier cabinets. They did not believe Germany could survive without access to international capital markets and they meant what they said about cutting the budget, consolidating public finances, and retaining access to international credit markets.54 By 1931, however, international trade was shrinking and foreign exchange was becoming scarce even for key currency countries. In both Britain and France interest group pressures to give pride of place in economic policy making to overseas territories and the empire (“imperial preference”) were growing visibly. In France, for example, various business and military groups were noisily organizing a huge exposition for the Spring of 1931 in celebration of the empire’s glories. The efflorescence of such schemes was, in no small part, a reaction to German success in stimulating exports in the midst of world depression. If brought to fruition, they threatened to effectively terminate hopes of running export surpluses. In this context, a policy of Drang nach Osten increasingly looked like either the wave of the future, or, at least, cheap insurance, to growing numbers of German firms, including several giants, such 23
4259 Ch01 4/12/02 9:42 am Page 24
24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
as I. G. Farben (which had close ties to the cabinet and political circles around the Chancellor), Siemens, and Krupp.55 While some early proposals imagined expansion into “Mitteleuropa” under the auspices of a joint Franco-German condominium, with the French contribution consisting mostly of capital, the steady shrinkage of markets turned such proposals into zero-sum games and greatly sharpened national rivalries. By 1931, an eastern-oriented strategy was increasingly emerging as an alternative, not a complement, to traditional internationalist economic policies, not least because some sort of exclusive tariff concession would be needed to ensure Germany’s primacy in these areas.56 The plans for what became the notorious proposal for a customs union between Germany and Austria did not originate with the Reichstag or other normal policy making channels. Instead, the scheme emerged from a secret planning committee convened by the aristocratic German Foreign Office, which had for years met regularly with leading German firms and enjoyed close ties to the military. While Brüning and the Foreign Minister (the project’s public champion, whose reappointment had been confirmed by President Hindenburg) knew of the proposal, the cabinet was not told until the last minute. Advised by Brüning that the timing was unfortunate but that Germany could not control this, the cabinet voted approval on March 18.57 Ever since the end of World War I, the specter of Anschluss between Germany and Austria had alternately tantalized or repelled many Europeans. While the Treaty of Versailles barred Germany from incorporating Austria, the German government’s interest in eventual amalgamation was patent, as was the impatience of much of the Right to see it brought about. Another postwar peace accord, the Treaty of St. Germain, blocked Austria from making any agreement that might threaten its sovereign existence. When news of the plan leaked, a storm of protest immediately blew up. Unimpressed by the legal justifications offered by Germany and Austria, France moved convulsively to counter the German thrust. The chilling implications immediately dawned on the press and financial markets. As the New York Times headlined in late March: “Paris loan market shut by Reich move. Union held to kill prospects for intermediate or long-term credit aid by French. Outcry in press is fatal. Bankers fear customs plan will shatter regained faith of our investors in Europe’s stability.”58 For a few weeks, hope centered on the search for a diplomatic masterstroke that might save the situation. “Financial centers are hoping, therefore, that Aristide Briand will again have recourse to one of those friendly compromises for which he is famous.”59 But Briand, who only a couple of weeks previously in the French Chamber of Deputies had declared the notion of Anschluss dead,
4259 Ch01 4/12/02 9:42 am Page 25
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
25
was seriously embarrassed. As he and the Socialists backpedaled furiously, and the right agitated, French policy shifted decisively. At the same time, the German Right was raising a storm over the opening to the French. Sudden appearances by French industrial groups offering to help bail out hard pressed German cities by buying shares in municipally owned utilities caused consternation. The government’s decision to limit such sales placed it in the position of formally discouraging precisely the capital flow that it needed to survive. On the very eve of the final decision to go ahead with the customs union, an aristocratic acquaintance of Hindenburg’s also complained directly to him about remarks by the French finance minister that seemed to imply that French loans to Germany were intended to perpetuate Germany’s position of military and strategic inferiority.60 The German government sought alternate sources of capital. With British support, Brüning’s government and the Reichsbank renewed pressure on the BIS to live up to its brave talk about leading the way to recovery in central European money markets. But now, ironically, the Germans’ earlier success in reviving international financial confidence helped undo them by driving otherwise sympathetic American bankers into coalition with the French. In late 1930, both the British government and the Governor of the Bank of England, Montagu Norman, had strongly encouraged Brüning as he firmed up his stance on the budget and reparations. After Norman submitted a memorandum to the BIS governing board, a formal proposal for a more active policy soon surfaced, championed by Sir Robert Kindersley, a prominent London private banker serving on the boards both of the Bank of England and the BIS.61 But the revival of Young Plan loans in the late winter emboldened skeptical opponents of BIS action. Their counterattack put Brüning and Luther on the spot. Political support for their stance on the budget was soft. Reichsbank President Luther was in a personal contest with Dr. Schacht for next Presidency [of the Reichsbank] and is staking his chances on a policy of fulfillment of Young Plan which Dr. Schacht opposes largely on ground that the Bank for International Settlements is not stimulating trade and capital movements[.] Dr. Luther and Brüning eager to steal his thunder by pointing to some example of our effort [to] promote capital movements as a justification of government’s change of view and abandonment of moratorium talk . . . [F]lat pronouncement at present that we shall always limit our activities to being a reserve bank specializing in 90 day bills would unloose a volcano.62
Warnings that Luther was “very despondent” and would find it difficult “to keep German papers and banking and political circles from launching radical attacks on [the] Bank,” failed to move a majority of the governing board of the BIS. Citing the dramatic rise in the price of Young Plan bonds, American 25
4259 Ch01 4/12/02 9:42 am Page 26
26 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
bankers stood on the principle of laissez faire: “we believe that Germany and other foreign borrowers will, on the whole, get as good credit as they deserve in the market from time to time without the intervention of artificial agencies.” Latin bankers suspected an “English effort to direct investment of French and American funds” while the French resisted repeated Belgian efforts to change their minds. As a result, in the late Spring, as the German financial situation grew increasingly dire, the BIS put off action on a series of proposals for more aggressive intervention in European financial markets advanced by the Belgians with support from Norman and Reichsbank President Luther.63 Closure of the international market for medium and long term loans put Brüning and his government in a box.64 At the very moment that his budgetary successes were turning heads around the financial world, the Chancellor and a handful of his closest collaborators received the dismaying news that their task was Sisyphean and that access to even short term credit would soon be at risk. At the beginning of March, Finance Minister Dietrich quietly warned Brüning that the deficit in the new budget would be much larger than anyone anticipated – perhaps as much as 430 million Reichsmarks – and that the special fund for unemployment compensation required an immediate loan or infusion of another 83 million RM. Dietrich also told the Chancellor that the seasonal flow of funds was creating a short run cash shortage for the government in the next quarter (April to June) of over 400 million RM.65 The news staggered the Chancellor and the head of the Reichsbank, who met with a small group of key ministers to discuss the situation. Issuing a “sharp warning” against any discussion outside the group, Brüning assessed the situation in dire terms. “The German people will not accept any further financial compression without decisive steps on reparations.” The Chancellor insisted, however, that no step in this direction could be taken until he had made “decisive progress” toward restructuring municipal finances and met with German ambassadors at a meeting scheduled for the end of April.66 Brüning persuaded the Reichstag to insert a special provision in the new budget allowing the government to make additional expenditure cuts if revenues continued to fall. This won him further plaudits from the financial community, which did not immediately realize why the provision had been inserted. But when the government quietly sounded out Lee, Higginson about the possibility of floating a new, large loan, the American house gasped when it heard the numbers. While the investment house and the government deliberated, the Germans took advantage of the afterglow of their budget triumph to roll over some 53.5 million RM worth of Treasury bills and sell 71.5 million RM of new T-bills.67 As Lee, Higginson prepared to approach the Banque de France with a proposal to recycle French money into German government T-bills, the
4259 Ch01 4/12/02 9:42 am Page 27
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
27
furor over the customs union exploded. Other German efforts to scout loans proved equally unavailing.68 In the meantime, the deficit projections kept growing. Desperate to avoid repeat performances, the government widened its search for possible shortfalls. By May, the Finance Minister projected a total deficit for all levels of government at 2.2 billion RM.69 But the government found it difficult to settle on a course of action. All alternatives looked disastrous. The Chancellor was convinced that a bid for complete cancellation of reparations would have to await, not only the upcoming French Presidential election and the long scheduled Disarmament Conference, but also the outcome of the 1932 American elections. He was also sure that, as bad as things were now, 1932 promised to be worse, and that the BIS would not help.70 But he also believed that if he announced yet another austerity program (only realistically possible through Notverordung) to tide the government over the rest of the current year, “the existence of the present Cabinet is in danger, if one does not seize upon (anschneide) the reparations question.” At the same time, however, Brüning believed that the fallout from a unilateral halt to reparations would be so terrible that the step was essentially unthinkable. Indeed, the protocols of their meeting suggests that the Chancellor and his ministers were not willing to surrender power to a government that would try it.71 The Foreign Minister and the Reichsbank President reinforced these forebodings. Both warned vehemently against any steps that might reawaken foreign anxieties about Germany’s commitment to paying reparations. Brüning resolved his dilemma by a kind of “doublethink.” The Chancellor finally concluded that only a truly desperate expedient offered any hope of relief:
26 27 28 29
[T]wo things are necessary: From an internal standpoint, release of the [new] Notverordung must awake in the people the impression that the revision [of reparations] has already begun; abroad, by contrast, the impression must be awakened that we are making all exertions to fulfill the plan.72
30 31 32 33 34 35 36 37 38 39 40
The Chancellor’s plan for bringing off this quixotic program rested on the combination of a diplomatic offensive with the new, draconian Notverordung. Aware that Britain had the biggest direct commercial interests in Germany and central Europe, he hoped to jolt the British into realizing that Germany could no longer afford to pay reparations by issuing a shockingly austere Notverordung on the eve of his scheduled talks with British Prime Minister McDonald at Chequers.73 But the late Spring of 1931 was not a propitious time to try to fool all of the people all of the time both in and out of Germany. Though some historians have maintained that the depths of the German budgetary crisis only emerged 27
4259 Ch01 4/12/02 9:42 am Page 28
28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
on June 9, when it became clear that tax revenues for April and May had run far below projections, in fact the government’s predicament had become apparent long before then.74 Some cities were near default or actually in arrears, forcing the Reich to step in in several instances. Since municipal loans in particular comprised a major share of the portfolios of some smaller municipal savings banks, this stirred some fears of local defaults and put additional pressures on Reich finances.75 Official publications of the republic’s monthly tax revenues and a variety of budget numbers were regularly scheduled and closely watched. Readers of the Frankfurter Zeitung, probably the leading newspaper in Germany, for example, could read virtually the same figures that economic historians now use to illustrate the size of the tax shortfall soon after they were compiled and published by the government.76 In the unlikely event that readers failed to grasp the meaning of the numbers, the paper’s increasingly alarmed articles on the growing budget shortfall could hardly have failed to alert them, or the financial community that watched anxiously as the Young Plan bonds tanked. Not surprisingly, after hitting their respective year’s highs in April (see Table 4), both the German stock and bond markets began steadily selling off, long before news of the Creditanstalt debacle arrived on May 11.77 In early May, well before news of the Creditanstalt burst upon the world, the German government attempted to float a domestic bond issue for the Reichspost. Only part of the loan could be sold, generating embarrassing publicity. This debacle plainly owed nothing to international currency pressures and everything to fears by domestic investors of heavy losses.78 By mid-May discussions between the political parties and the government over how to close the budget gap were becoming very tense. Paramilitary clashes, already running at high levels, intensified, while the national SPD leadership, not coincidentally, voiced doubts about spending cuts of the size Brüning thought was necessary. Leaders of the Prussian SPD, however, who carried enormous weight both within the party and the government, faced elections in early 1932. They spoke out in favor of making big cuts now, rather than risk having to go through the whole process again later in the year. As Brüning and the cabinet – who still lived in the shadows of 1918, the Kapp Putsch, and the myriad other disorders of early twenties German politics – geared up for the crunch, the crack of doom resounded: the Minister of the Interior warned that he could no longer guarantee the maintenance of order.79 Well aware, as some historians seem to have forgotten, that he would need votes from the right as well as the left to sustain his budget when it faced the inevitable challenge in the Reichstag, Brüning appears to have gradually shortened his time horizon, from getting all the way through to the beginning
4259 Ch01 4/12/02 9:42 am Page 29
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
29
of 1932 to just getting past the Chequers meetings.80 His language about reparations also became increasingly direct and reckless, though to the very end he did not want to admit that he was more than on the verge of repudiating them. By contrast, his Finance Minister went considerably further, in private threatening resignation if the Chancellor foreswore the right to declare a moratorium in the meetings with the British and call public attention to the need to revise reparations.81 By then the climate of both German domestic politics and European diplomacy were garishly different from the sunny optimism of early March. “Systemic risk” was rising rapidly. Just days before the German budget crisis came to a head, Aristide Briand narrowly lost his bid for the French Presidency. While factors internal to French politics surely figured in this outcome, the customs union proposal had helped torpedo Briand’s campaign. The Stahlhelm staged another in their series of gigantic rallies at which the French and their allies, the Poles, as well as, of course, “Marxists” within Germany, were roundly denounced.82 On May 19th, the launch of the pocket battleship Deutschland kicked off a new round of competitive naval construction between France, Germany, and Italy.83 On June 3, at a meeting of the celebrated Langnamverein, a trade association of heavy industrialists, including many from the Ruhr, Albert Vögler, a leading figure in the steel industry, demanded the cession of all payments of foreign “tribute” and a shift in policy toward building up the home market. General Gröner, who along with his then protégé, General Schleicher, provided critical support for the government within the Reichswehr and Hindenburg’s circle, also advised Brüning to wait no longer on reparations. This was advice that the Chancellor dared not ignore.84 For essentially domestic political reasons, then, Brüning made the fatal decision to denounce reparations and proclaim that Germany had paid all that it could. The Notverordung that Hindenburg signed on June 5 and was published, together with an accompanying manifesto, the next day stunned not only Germany but the world.85 Since at least May 25, German newspapers had been speculating that Brüning was likely to ask for some sort of relief in regard to reparations. The French government and, doubtless, others, had drawn the inevitable conclusions. Suspicions that Germany might postpone reparations transfers had been circulating for some time, fanning the run on German gold reserves and major banks that appears to have begun about May 25. On the 28th of May, Young Plan loans fell sharply all over the world, “on a rumor that the German government were contemplating the declaration of a reparations moratorium.”86 The distinction that Brüning and other German officials planned to highlight outside of Germany – that the official announcement did not formally repudiate 29
4259 Ch01 4/12/02 9:42 am Page 30
30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
reparations or invoke the Young Plan’s provisions for postponing payments – no longer mattered by the time of publication. The manifesto’s sensational references to “tributary payments”and “intolerable reparation obligations” as well as its provocative language about a people’s capacity for suffering having reached its limits, spoke in stentorian tones to investors in and out of Germany.87 Read everywhere, it was everywhere read as heralding a broader German inability to make international payments, not least because it was accompanied by official denials that Germany would soon be forced to suspend payments on both reparations and private debts.88 Investors converted Reichsmark deposits into gold, as the Reichsbank’s gold cover suddenly began to look dangerously thin. Reichsbank gold registered its first decline of the year in the week ending June 6th. This can be seen clearly from Table 8. Gold reserves did not fall in March when the customs union was announced. Neither did they fall after the Creditanstalt collapsed – indeed, both Reichsbank gold and total foreign exchange reserves rose all through the month of May, though the entries for May 30th probably reflect a final payment the government received as part of a deal struck in 1929 for a match monopoly.89 In the two weeks after June 6th, Reichsbank gold reserves declined precipitously – by 40% – before stabilizing briefly at 60% of their previous level as it became clear that the Reichstag would not overturn the Notverordung. Fears for the stability of the Reichsmark were also reflected in the premium for gold bonds over currency bonds, as shown in Table 4. Comparing the relative prices of the 6% bonds reveals that the gold premium calculated from the ratio of bond prices drifted down from 6.6% at the start of the year to 6.1% in April. It rose to 6.9% in May, and then soared to 10.0 in June, as investors became very anxious indeed about the stability of the Reichsmark. Underlying the run was the wrenching fear that the government could not make its colossal budget cuts stick. In the early days of the panic, with Brüning in England for the bilateral talks, this seemed likely. Across the political spectrum, the new budget package aroused “general horror.”90 The SPD, the parties of the Right, and the Stahlhelm (which had been afforded a share of the administration of youth programs in the budget) all vociferously denounced the cuts. Göring and other Nazis furiously attacked them. On June 11, Stresemann’s old party, the DVP, split thunderously over the cuts. Since the party was somewhat inaccurately still labeled as uniquely a party of big business, the split was widely noted, even in foreign business centers. By June 13th, the gold drain was so severe that the Reichsbank raised its discount rates two full points in an attempt to sustain the gold value of the Reichsmark. Over the next few days, the financial situation worsened,
4259 Ch01 4/12/02 9:42 am Page 31
German Currency Crisis
Table 8.
1
31
Weekly Reserves at the Reichsbank (Million RM).
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Date
Gold
Foreign Exchange
Silver
7.1.31 13.1.31 23.1.31 31.1.31 7.2.31 14.2.31 23.2.31 28.2.31 7.3.31 14.3.31 23.3.31 31.3.31 7.4.31 15.4.31 23.4.31 30.4.31 7.5.31 15.5.31 23.5.31 30.5.31 6.6.31 15.6.31 23.6.31 30.6.31 7.7.31 15.7.31 22.7.31 31.7.31
2216 2216 2244 2244 2244 2254 2266 2285 2285 2286 2286 2323 2344 2345 2348 2368 2370 2370 2370 2390 2300 1766 1411 1421 1422 1366 1353 1368
400 268 196 199 198 181 175 166 189 209 223 188 166 114 132 158 169 171 197 186 113 104 93 300 371 244 160 246
161 200 207 172 178 192 202 160 166 179 196 157 143 187 207 168 175 186 200 174 177 199 214 78 84 79 74 45
Source: Die Bank, 1931, various issues.
29 30 31 32 33 34 35 36 37 38 39 40
exacerbated by the increase in the cost of credit imposed by the Reichsbank. SPD leaders and General Gröner, the Defense Minister, claimed to see revolution around the corner. Stories – which appear to have a foundation in fact – that the military was preparing to intervene swept Berlin and other major cities. Brüning’s own day calendar (Tageszettel) for June 13 indicates that he met with General Schleicher to discuss Reichwehr plans for coping with civil unrest. Insiders were saying that the government was two weeks away from defaulting even on pensions.91 31
4259 Ch01 4/12/02 9:42 am Page 32
32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
All the tumult, along with frantic bargaining by Brüning (whose train, as he returned from Britain across Germany, was mobbed by Nazi thugs and other paramilitary groups) began to change the minds of the deputies. Pressure from various business groups and the Reichsbank intensified. Brüning and the cabinet made it clear that if the Notverordung were overturned, they would all resign. Once again, Brüning threatened to overturn the SPD-led coalition that ruled Prussia. The chancellor also offered a few, minor compromises.92 Economic historians who focus on the Creditanstalt and international factors in the banking crisis have noticed that something happened on June 16 that brought an abrupt halt to the foreign runs on German banks and the Mark. But they have typically identified it as the decision by Montagu Norman, Governor of the Bank of England, to float an emergency loan to the Austrian National Bank. No doubt this decision helped stabilize Austria, but why should a British decision to bail out Austria have stopped internal drains from local German banks? Our answer is, of course, that it did not. Rather, something else that happened the same day that stabilized both domestic and international markets in which the Mark figured: The SPD and other parties decided not to try to overturn the budget. Brüning had won, and the deputies dispersed (Winkler, 1993, p. 413). This political success brought Brüning only a few days respite. On the heels of the dispersal of the Reichstag came the first, muffled reports of the Nordwolle’s losses. The run on the Reichsbank swiftly resumed, as the shadow of death fell over the giant Danatbank, which was closely associated with the huge textile concern. The Reichbank’s action to defend the mark had increased the cost of borrowing, intensifying the tottering firm’s problems. In ordinary times, the bank could have turned to the Reichsbank for help. But the battles over the custom union, the budget, and Brüning’s ill-fated references to reparations “tribute” had left the Reichsbank in no condition to act as lender of last resort. As late as the end of May, the Reichsbank’s cover of gold and foreign exchange exceeded 60% of the value of its own notes and 47% of the money supply as a whole. The gold cover fell dramatically in June, as shown in Table 8, threatening to breach the legal limit as word of the Nordwolle’s condition began to leak out. The Reichsbank could no longer act as a central bank to its domestic constituents.93 This, of course, is what happens in countries on a gold standard. The central bank focuses on maintaining the value of the currency. It needs to buy and sell gold in response to international demands; it cannot divert its attention to care for commercial banks. The Reichsbank was a textbook example of this condition. It had maintained the value of the Reichsmark as its principal domestic constituencies and the Allies concerned about reparations payments
4259 Ch01 4/12/02 9:42 am Page 33
German Currency Crisis 1 2 3 4 5 6 7 8 9
33
desired. But it could not at the same time take care of problems within Germany. When the force of the depression caused a few large firms to default on their loans, banks clearly needed some help. The Reichsbank was unable to help and the banks failed. But the problems did not originate in bank behavior; they arose from the policy stance of the Reichsbank and from the political acts of the government that put its attention to that policy to the test.94 Grossman (1994) showed that no country that went off gold experienced a banking crisis. If Brüning had controlled trading in the Reichsmark earlier or announced the customs union later, German banks probably would not have failed.
10 11 12
VI. CONCLUSION
13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Keynes said during a discussion of the Young Plan in July 1929 that, “even if it were foreseen, the announcing of a moratorium would so damage Germany’s credit that a crisis would be precipitated by the announcement.”95 He was right. Brüning’s announcement that Germany could pay no more reparations generated a run on the mark in which banks failed and currency transactions were controlled. The ensuing run was on the Mark, not banks, because the fear was that the German governments, not the Grossbanken, would not be able to pay their bills. Could the banking crisis have been avoided? Almost certainly. Brüning abrogated his international obligations in three ways during the spring and summer of 1931. He promoted the customs union, geared up to stop reparations payments, and imposed currency controls. He did them in this order, placing great strain on banks. Had he done these three acts in any almost other order, German banks might well not have failed. Had Brüning gone off gold by imposing currency controls before one of the other announcements, German banks probably would have survived.96 Could the currency crisis have been avoided? That is a harder question. Adherence to the gold standard caused problems for countries far more stable than the Weimar Republic. It is unlikely that Brüning could have avoided going off gold in the end. Had he thought ahead, it could have been done more gracefully and with less damage to the German economy. We therefore conclude that the German crisis of 1931 was a twin crisis caused by a run on the currency triggered by domestic political actions. Banks failed, and the international economy constrained Germany, but the primary actions that led to the crisis of July 1931 were German policy initiatives, taken by Brüning in response to political pressures facing him and his supporters. The German crisis of 1931 was made in Germany. 33
4259 Ch01 4/12/02 9:42 am Page 34
34 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
THOMAS FERGUSON AND PETER TEMIN
We draw another lesson as well from this analysis: The terms of the celebrated debate over the so-called “Borchardt thesis” need to be recast. Recent work showing that political forces in favor of reflation did exist inside the Weimar regime has made an important contribution.97 But it is time to acknowledge that the truly significant political coalition that failed by only a hair’s breadth was transnational. Put simply, there were powerful political forces in both France and Germany that perceived the advantages of an internationally cooperative strategy. This cross-national coalition always faced major obstacles, but as late as the Spring of 1931, its failure was not obvious. Indeed, the cause of Franco-German cooperation, if not reconciliation, appeared to be making great strides forward. What destroyed it was the German government’s decision to pursue Drang nach Osten. It is all very well to say, as Duisberg and other apologists for the policy did, that this decision was a response to pressures for “imperial preference” among the allies. So it was. But the reckless pursuit of “Mitteleuropa” destroyed the possibility of French financing for Germany. Stabilizing Germany by this means was clearly a goal of Briand’s policy, and he came heartbreakingly close to bringing it off. What Borchardt and his followers now owe the rest of us is an answer to one simple question: Are they prepared to argue that German business groups and policy makers also had “no choice,” no “Handlungsspielraum” in this decision to swallow central Europe? That is what the denial of the possibility of an international version of reflation, if not full throated “Keynesianism” requires. As late as March, 1931, this seemed not a remote contingency, but a reasonable bet – one that even bankers were willing to take.
24 25
NOTES
26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
1. For a good inventory of the main works in both English and German, see Balderston (1994, p. 43, n. 1). 2. See, e.g. the discussion of the German reaction in James (1986, p. 291) or Born (1967); also Balderston (1994). 3. James (2001a, p.129). See also p. 16. By contrast, the emphases in James (2001b, pp. 59–63) more closely parallel his 1986 study. 4. See the discussion in Eichengreen and Hatton (1988) and Stachura (1986). Unemployment figures for Germany were published monthly in the German Statistisches Reichsamt’s widely read Wirtschaft und Statistik. The same figures also appeared in the Monthly Bulletin of Statistics published by the League of Nations, where it is explained that they are for the end of the month and cover trade unionists. These are the figures we use here. Regardless of how one might want to adjust them in the light of better information about coverage or the labor force, they are useful as an indicator of monthly trends that could actually be tracked by contemporaries. 5. The literature on the Depression’s effects on Germany is immense; besides the sources we cite for our statistics and those cited in note 1, above; see, e.g. Winkler (1993) or Feuchtwanger (1995).
4259 Ch01 4/12/02 9:42 am Page 35
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
35
6. Feuchtwanger (1995, pp. 70–71, 219–220); Winkler (1993, pp. 369ff.); Kolb (1993, p. 214). For the role played by I. G. Farben and German heavy industry in the change, see the account of the circumstances surrounding Duisberg’s letter to Finance Minister Moldenhauer in Winkler (1993, pp. 369–371). Duisberg headed the famous Reichsverband der Deutschen Industrie, the leading (but not the only) big business association in Germany. He was also a top official of the giant chemical company. The RDI had begun pressing for the fall of the government as soon as the Young Plan cleared the parliament. Plumpe (1990, pp. 529–532) offered a critique of an earlier Winkler essay on the changeover. This controversy has no bearing on our paper, but see the discussion of the dissolution of the cabinet (which does not mention Plumpe) in Winkler (1993, p. 371). 7. Kolb (1988, pp. 114ff.; the quotation appears on p. 114); his later German rendition is entirely consonant with this judgment. The discussion in Winkler (1993) is also revealing. 8. See, along with Table 2, Reichsamt (1933c, pp. 341–342). These figures underestimate the total debt, excluding, for example, another 4 billion RM borrowed by public corporations in their own names. See Reichsamt (1933c, pp. 345–346). (Here “billion” is used in the American sense of 1000 million.) 9. Estimates of total German indebtedness vary, though the differences do not add up to much for our discussion. Early German estimates were sometimes suspect, since they were prepared as part of negotiations over war debts and reparations. But the official German estimates made in the late twenties and after, appear quite good and have hardly changed at all over the years. Compare the figures in Reichsamt (1933c, pp. 341–350) with those in Länderrat des Amerikanischen Besatzungsgebiets (1949, Vol. II, p. 354) and Deutsche Bundesbank (1976, p. 313). See also the helpful discussions in Bachmann (1996, pp. 42ff.) and Thomas (1934). The 1931 Wiggin Committee’s report, sometimes referred to as the “Layton Report” after Sir Walter Layton, who played a major role in its drafting, appears to have been the source for much of the data published during this period in The Economist (which Layton edited). These two sources formed the basis for the estimates in Feinstein and Watson (1995, p. 127). Note that debts of public corporations were reckoned separately from the rest of the official debt totals, though they also borrowed through other governmental agencies; see Reichsamt (1933c, pp. 345–346). It is also important to recall that all through the Weimar period, a low intensity war raged in the courts and parliament over the possible redemption of older public debts that had been inflated away; cf., Southern (1981). Finally as emphasized by McNeil (1986), truly remarkable efforts were often undertaken to hide borrowing or evade the (usually weak) German or American efforts to restrict borrowing. 10. The literature is immense; see, Winkler (1993, pp. 380ff.); Feuchtwanger (1995, pp. 222–224). The political developments came in the context of sometimes rapidly changing strategies by major industrial and agricultural groups. See, for example, the literature discussed in Kolb (1993) or Abraham (1986). The Finance Minister, a former I. G. Farben executive, resigned on June 18. The budget package failed in mid-July. Not all spending was cut; funds for so-called Osthilfe (agricultural aid tilted heavily toward Junkers) increased. Communities were also authorized to tax beer. See Feuchtwanger, 1995, p. 224. 11. The one-third estimate does not include the 41 deputies of Hugenberg’s Deutschnationale Volkspartei, who really constituted an anti-system party of the far right. 35
4259 Ch01 4/12/02 9:42 am Page 36
36 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
The Reichstag then had 577 members. The practical effect was to increase still further Brüning’s dependence on Hindenburg. Cf. Feuchtwanger (1995, p. 233). 12. See, e.g. Ellis (1941, p.162). All analyses of the period feature a discussion of the capital outflow. 13. For the “seamless” projection of the1930 crisis, see, for example, Harris (1935) or less baldly, Ellis (1941); a variant is partly explicit and partly implicit in James (1986), though he is not entirely consistent, as noted below in the discussion of the French role. Ritschl (1998) argues that the Young Plan terminated capital flows into Germany by making private debts junior to reparations. In our view this confuses the cart with the horse. Priority became pressing once capital inflows ceased. The best disproof of the hypothesis that Germany could not borrow in the face of the Young Plan loan is the simple fact that it did so, repeatedly, sometimes in the face of warnings from authorities who went out of their way to remind potential creditors of precisely Ritschl’s point. See, for example, the Agent General for Reparation’s comments to the U.S. Secretary of State about the German loan floated by Lee, Higginson on the eve of the public offering of the Young loan in 1930 in U.S. Department of State (1945, Vol. III, pp. 96–102). Cf., also the correspondence engendered by the release of this volume in 1945. This can be found in Box 121 of the Thomas W. Lamont Collection, Baker Library, Harvard Business School. In a letter to Lamont of July 27, 1945, Russell Leffingwell of Morgan ascribed the State Department’s unwillingness to discourage the loan to intervention by Jerome Greene of Lee, Higginson. Lamont, in a draft letter of Aug. 7, 1945 to Arthur Hays Sulzberger of the New York Times, noted that Greene and the Under Secretary of State, Joseph Cotton, who played a key role in the episode, had been classmates at Harvard. That draft was not actually sent; other correspondence appears to have been substituted for it. For this paper, the important point is that major investors were well aware of the Young loan and the priority issue; as we show, they kept loaning anyway. 14. For Lee, Higginson, see, e.g. Ellis (1941, p.162) and Bachmann (1996, pp. 213ff.), though we cannot follow her judgments on several broader issues. As discussed below, the Thomas H. McKittrick Collection, now at Baker Library, Harvard Business School, contains important letters and notes concerning the Lee, Higginson deals with Germany. For the post-election maneuvers on the budget, see, again, Winkler (1993, pp. 394–398) and Feuchtwanger (1995, pp. 236–237). 15. On Prussia, the Notverordung, and the Social Democrats, see the discussion in Winkler (1993, pp. 392–402) and Feuchtwanger (1995, pp. 233–238). 16. The standard, and generally excellent, source for discussions of reparations revision is Bennett (1962). But our account, presented in more detail below, of the BIS, Norman, and German government’s stances on the “Kindersley Plan” draws on documents that were not available to him. For the British government’s efforts to influence Brüning, see Sir H. Rumbold (British Ambassador to Germany) to Mr. A. Henderson (Foreign Minister), March 4, 1931, in Woodward and Butler (1947, Vol. I, p. 573). While the letter is dated March, it discusses events that occurred in December, 1930 and January, 1931. For the campaign against the BIS, see for Schacht, e.g, New York Times, Jan. 2, 1931, p. 15; for Georg Solmssen, a Deutsche Bank director serving as President of the German Banker’s Federation, see Commercial and Financial Chronicle, Jan. 3, 1931, p. 51. In December, 1930, the Swiss government ordered the Basle head of the (German)
4259 Ch01 4/12/02 9:42 am Page 37
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
37
Allgemeine Pressebureau to leave the country after repeated French complaints that he was agitating against the BIS. See the letter of the French consul at Basle to M. de Marcilly, the French Ambassador to Switzerland, Dec. 12, 1930 in Carton B32221 in the Archives of the French Ministry of Finance at Savigny-le-Temple, France. The German journalist was said by the consul to be continuing the campaign from just across the border. 17. The quotation is again from Ambassador Rumbold’s letter to Foreign Minister Henderson, March 4, 1931, in Woodward and Butler (1947, Vol. I, pp. 576–577); and refers to the time period we discuss here. See also New York Federal Reserve Governor George Harrison’s letter of Feb. 9, 1931 to Norman Armour, which explains that “the best opinion” was that “Germany is unlikely to declare a postponement of transfers,” New York Federal Reserve Bank Archives, C261, Germany-German Gov’t 1916–1950. A contemporary account sensitive to the nuances of the reparations issue in these weeks is Dulles (1932, p. 234). The Germans believed that over time their foreign creditors would come to perceive that their own interests lay in relenting, either because they feared social chaos and the drift to the right in Germany itself or because their own industries found it impossible to compete with a major competitor whose price level was falling. See, e.g. Kolb (1993, p. 218). Brüning saw his policy of cutting the pay of German government officials as part of this strategy of general deflation. Cf. James (1986, p. 68). The wage cutting strategy was strongly supported by many industrial groups. 18. Sir H. Rumbold (British Ambassador to Germany) to Mr. A. Henderson (Foreign Minister), March 4, 1931, in Woodward and Butler (1947, Vol. I, pp. 573–574). Rumbold was generally a pessimistic observer of the German scene. 19. For the Institute of International Finance, see New York Times, Jan. 20, 1931, p. 35 and Institute of International Finance (1931). The Commercial and Financial Chronicle, Jan. 31, pp. 763–764, outlines the study’s conclusions and notes that the Institute was partially supported by the Investment Bankers Association of America. For the British government, see Mr. A. Henderson to Sir H. Rumbold, Feb. 19, 1931, in Woodward and Butler (1947, Vol. I, p. 559). For the BIS see, besides our discussion below, “Remarks of the President of the Bank for International Settlement Before the American Club in Paris,” Feb. 12, 1931. This was sent to Thomas Lamont by Leon Fraser, then of the BIS. See also N. D. Jay to Lamont, Feb. 14, 1931, which encloses a memo by Montagu Norman that had been discussed at a BIS meeting. All these are in Box 181 of the Thomas W. Lamont Collection, Baker Library, Harvard Business School. See also Dulles (1932, p. 218) and the comments below on the BIS. Her study (p. 224) is again sensitive to the improvement in the German position in this period – perhaps not surprisingly, since she was the sister of John Foster Dulles, who had been the American expert on reparations at Versailles and was then generally recognized as the leading American authority on the issue. See Ferguson (1995, p. 144). 20. The first two quotations come from The Economist of Jan. 3, 1931, pp. 17–18. The discussion there merits a longer look: “After the marked decline in deposits shown in the October bank balance sheets, in consequence of the credit crisis, the November balance sheets show every sign of returning calm . . .. The cessation of the withdrawals of money, especially foreign withdrawals, has enabled the banks to reduce their rediscounts at the Reichsbank, which had risen during October . . ..” The final quotation is from a cable from Morgan et Cie., to J. P. Morgan & Co. in New York of Jan. 12, 1931, The Pierpont Morgan Library, New York, Morgan Bank European Papers, Box 13. 37
4259 Ch01 4/12/02 9:42 am Page 38
38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
21. For the Morgan bank exchange, see the cable from Morgan et Cie., to J. P. Morgan & Co. in New York of Jan. 12, 1931, The Pierpont Morgan Library, New York, Morgan Bank European Papers, Box 13. For the Reichsbank exchange with the New York Fed, see the former’s cable to the Fed of March 11, 1931: “Offers of foreign money are said to be more numerous,” while “flowed out German capital stimulated by the rise of shares, seems returning to Germany,” New York Federal Reserve Archives, C261, Germany-Government 1916–1950. The translation is, again, clumsy but clear. 22. The reparations payments had been deliberately scaled down and made to vary over time under the Young Plan. See, e.g. Moulton and Pasvolsky, War Debts, pp. 223ff. 23. New York Times, Feb. 9, 1931; the comment occurs in a story about the “sharp advance this week” of Young Plan bonds and is clearly meant as a broader generalization. Cf. also Dulles (1932, p. 225): “The Bank [of International Settlements] was in a position to explain the creditor point of view to Berlin – the view that Germany’s balance of trade was likely to improve, and that her main difficulties were caused by internal politics.” 24. For long rates, see the tables on pp. 128–129 of Institut für Konjunkturforschung (1933). Note that quotations for the Young Bonds for the end of June were affected by the brief impact of the Hoover Moratorium. See the discussion below. Then and later, observers argued over the extent to which some of these improvements were seasonal. Completely consonant with our analysis of the key role of the budget, between January and May, industrial long rates fell considerably more than “public loans” or “communal [city] obligations.” But all fell, though we know of no conclusive data on the quantity of lending at these rates. It is a commonplace that many large German banks bought their own stocks in the wake of the 1930 crisis. See, e.g. Born (1967, p. 60). But they cannot have been buying the broader market, which was also rising. Bank stocks rose along the rest of the market in January, February, and March, with the difference in April being too small to make anything of. Born’s observation that the Berliner Handelsgesellschaft succeeded in privately placing shares in New York in January is consistent with our view of the external perceptions of the German economy then. For modern attempts to predict currency crises, see Goldstein, Kaminsky and Reinhart (2000); their discussion of potential indicators in Chapters 2–4 can be instructively compared with the available German data. 25. Authoritative contemporary accounts, such as Toynbee (1932, p. 518), reported the amount we use here; Bachmann (1996, p. 242), suggests the final amount totaled about RM 134, 400,000. The New York Times, Feb. 6, 1931, p. 11, reported a dollar figure of $32 million. The German state’s demonstrated ability to borrow made an impression on contemporaries, see, e.g. Rosenstiel (1931a, p. 18). As we discuss below, Bachmann’s determinedly pessimistic account notes the borrowing, without, in our view, absorbing its implications. Facts come through that are inconsistent with her story line: Germany borrows, bankers are competing to bring out loans, one bank is striving to become the exclusive banker for the government, etc. See Bachmann (1996, pp. 235ff.). It is only fair to note that one banking house queried by the government replied that additional loans would have to await a new reparations settlement that eliminated the possibility that in a crisis recent creditors would have to stand in line behind holders of reparations bonds This is the one case we have found that fits Ritschl’s argument, discussed above. See James (1986, p. 72); the comment came right as Lee, Higginson was floating the loan and, very likely the bank in question realized it was not going to get the business it, too, had scouted.
4259 Ch01 4/12/02 9:42 am Page 39
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
39
26. New York Times, March 30, 1931, p. 30; this article is effusive about Brüning’s success. For the short term debt reduction, see Bachmann (1996, p. 254), who recognizes that the improvement runs counter to her own pessimistic appraisal of the Reich’s credit situation. 27. J. P. Morgan, Thomas W. Lamont, and S. P. Gilbert of J. P. Morgan & Co., New York to their Paris partners at Morgan et Cie., for relay to Gates McGarrah of the BIS, March 14, 1931. The copy we draw upon is in The Pierpont Morgan Library, New York, Morgan Bank European Papers, Box 13. For more on the context, see below. Even U.S. Ambassador to Germany Frederic Sackett, a famously consistent bear on prospects for Germany – see Burke (1994, passim) – changed his tune. See Sackett to George Harrison, March 17, 1931, New York Federal Reserve Bank, C261 GermanyGerman Government 1916–1950. The ambassador relates his discussions with Luther and other prominent Germans, who are “impressed with the idea that they have got some way to raise German credit rating in the financial markets and eventually relieve themselves from the immediate pressures of short term debts.” The former Great Pessimist continues, “I have the distinct impression that things are looking better in Germany,” and that “international political moves are reassuring.” As a consequence, “a very considerable amount of German money which fled abroad is returning here, seeking the increased rates which are available.” 28. This is an excerpt from a letter in the Thomas W. Lamont Collection, Baker Library, Harvard Business School, dated March 10, 1931. Both its style and contextual evidence indicate that the author is Lamont himself. There is, however, some only partially decipherable handwriting on a page that raises a slight possibility that the author was Lamont’s Secretary, de Sanchez. But he was ill, probably out of the office, and scarcely in a position to make any wagers. The letter is in any case striking evidence of the shift in sentiment. 29. New York Times, Feb. 6, 1931, p. 11, relates details of French loan participation; on the French reluctance to loan, see, e.g. Bachmann (1996, pp. 216ff.); Bennett (1962) is an excellent general account, though he overestimates the willingness of the French to consider a loan in the Fall. Knipping (1987, pp. 198ff.), by contrast, correctly links the later French shift to the change in the French government. Note that the banks almost certainly did not believe they would be immediately able to resell all the bonds to the public, for they were still holding at least some bonds from the Fall issue; the point is that they were willing to buy them all in the context of what they perceived as an improving climate. 30. On the change of government, cf. Knipping (1987, pp. 198–199). Lee, Higginson had been targeting increased business in France since at least the fall. See, e.g. Paul Courtney to Partners, Sept. 8, 1931. This and subsequent letters in the McKittrick Collection discuss the French banker, René Charron, and the firm’s hopes for FrenchGerman detente. The bank also weighed sending some legal work to Sullivan and Cromwell in New York, in order to enlist the aid of John Foster Dulles. See the telegram to Courtney, signed “George” (Murnane), Sept. 18, 1930, Carton 6, Thomas H. McKittrick Collection, Baker Library, Harvard Business School; on Reel 11 of the microfilm of the collection. It is clear that Dulles and Murnane subsequently worked closely together, but we have no other evidence concerning this episode. See Aalders and Wiebes (1996). For Laval and the Belgians, see the discussion in Bussière (1992, pp. 397ff.), which demonstrates that Laval and the Belgian banker Francqui were communicating since at least late November, 1930. The Belgian motivation seems clear: it is a variant 39
4259 Ch01 4/12/02 9:42 am Page 40
40 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
on the answer to the old joke about what mice do when two elephants are becoming restless around them. For Lee, Higginson and Charron, see the correspondence in Carton 6 of the McKittrick Collection, Series II. Both the originals and a microfilm of this collection were consulted. Glashagen (1980, pp. 355) mentions that even the Reichswehr was then interested in exploring a possible (limited) arms accord with the French. But our discussion below relies on many sources not available to him. 31. On the French shift, see the memos from the French Ambassador to Berlin to the French Finance Minister, Jan. 15 and Jan. 29, 1931. These are now in the Archives of the Ministry of Finance, Carton B31478. See also the brief note of Feb. 10, 1931 on the discussions between Governor Moret of the Banque of France and Reichsbank President Luther in Carton B31469 in the same archive. (Unless otherwise mentioned, the Ministry’s archives were still located at Bercy, Paris when we consulted them.) The loan was sharply criticized by the right, but the cabinet voted approval. See New York Times, Feb. 8, 1931, p. 8. For the reaction of financial markets to news of the French participation, see New York Times, Feb. 6, p. 11, which includes the cited passage and mentions that other French banks were participating and Feb. 9, p. 30, which notes the immediate change in sentiment in the Berlin money market. While Briand was clearly a moving force within the French government behind the French shift, French banking opinion had also changed. See New York Times, Feb. 8, 1931, p. 8. Note that Laval formally took the reins of power at the end of January, but the new government had been in the works for some weeks. The American embassies in Paris and Berlin appear to have been surprised by the switch and somewhat in the dark about its meaning. See, the discussion in Burke (1994, p. 115) and the full texts of the key cable he mentions, Edge to Secretary of State, Jan. 29, 1931, reproduced on Reel 34 of (U.S. Department of State, 1986). It seems clear that Murnane withheld all mention of a possible French opening when he saw Sackett in January; we also think that Murnane’s comments at that time about how Lee, Higginson rather than Germany was key in selling bonds were very much comments of someone striving to bring off an exclusive arrangement with the German government. See Burke (1994, p. 114) and the full text of Sackett to Secretary of State, Jan. 21, 1931, also on Reel 34 of U.S. Department of State (1986). 32. The quotation is from Dulles (1932, p. 236), who also references McGarrah’s speech. This is the speech in the Lamont Collection, mentioned earlier. The public discussion of gold and recycling appears in, e.g. New York Times, editorial, Feb. 5, 1931, p. 22 and Feb. 9, 1931, p. 30. The discussion was much broader, however, and linked to discussions about gold distribution (e.g. The Economist, Jan. 3, 1931, p. 7). Essentially a cliché within financial circles at the time, recycling was a basic premise of the many efforts by Lee, Higginson and others made to promote rapprochement between France and Germany. Heyde (1998, pp.142–143) notes the interest of some prominent French industrialists in the developing discussions with the Germans, though his broader, more skeptical, narrative does not take sufficient account of the financial evidence we discuss here. 33. James (1986); the most detailed is Bachmann (1996), whose work is so careful and painstaking that it records a great amount of detail that is inconsistent with its central theses, as we note. Both briefly discuss the French episode, but neither pursue its implications systematically. In effect they assume that the breakdown of relations between France and Germany was inevitable, without making much of an argument.
4259 Ch01 4/12/02 9:42 am Page 41
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
41
34. Luther had long been close to the firm and owed his appointment as Schacht’s successor at the Reichsbank in part to a dramatic intervention by the bank’s top executive. See Treue (1972, pp. 64–65), which also has an excellent overall account of the firm’s place in German banking history. Bratz (1958, pp.10–11) indicates that Mendelssohn was one of the handful of private banks that could count on state help in case of trouble. For Lee, Higginson’s relation to Mendelssohn, see the papers in Box 6 of Series III of the Thomas H. McKittrick Collection, Baker Library, Harvard Business School. As one Lee, Higginson telegram to “Courtney” in Paris (unsigned but almost certainly by George Murnane, of Sept. 17, 1930) summed up: “Consider our strongest reliance is Schaeffer who is loyal to us solely through Kempner [of Mendelssohn].” The American bank’s anxieties about competitors are obvious in the correspondence referred to above; at least one rival is mentioned there. But reading Bachmann (1996, pp. 235–237). in the light of what remains of Brüning’s papers indicates several others. Her suggestion (p. 237) that the Chase Bank may have been interested, despite some noises to the contrary, is probably correct. The Chancellor’s daily calendars (Tageszettel), now in the Brüning Papers, Harvard University Archives, HUB (FP) 93.35, Box 1, show that Shepard Morgan met twice with the Chancellor around this time. He was a Chase Bank official, though the daily calendar mistakenly identifies the bank. That he may have been scouting loan possibilities seems clear; see, e.g. J. E. Crane’s memorandum to Fed Governor Harrison of Jan. 24, 1931, Federal Reserve Bank of New York Archives, C261, Germany – German Government 1916–1950. This recounts a lunch with Morgan; while Morgan was skeptical then, he was obviously closely following events and probably shared Crane’s conviction that “in view of the recent improvement in the bond market here” one could reasonably expect that “foreign loans or credits may be somewhat more popular in the near future.” It is a puzzle to us how Bachmann reconciles the fact that banks were hounding the German government to get its business with her pessimism about German chances of floating loans. 35. See the handwritten memorandum, probably by T. H. McKittrick, dated Feb. 14, 1931 in 6–14 of Carton 6 of the Thomas H. McKittrick Collection, Baker Library, Harvard Business School. Mendelssohn conducted a great deal of business from Amsterdam after World War I. See Kupferberg (1972, p. 247) for Mendelssohn specifically and Aalders and Wiebes (1996, pp. 38–40) for the broader context. As mentioned below, Lee, Higginson was then very close to Ivar Kreuger, who at the time occupied a significant position in nothern European finance; it would not be surprising if Kreuger assisted Lee, Higginson in mustering support in Scandanavia. It does appear that he supported the Belgians and others who were working to move the BIS, though any discussion takes us too far afield. 36. Handwritten memorandum, Thomas H. McKittrick Collection, Baker Library, Harvard Business School, Feb. 14, 1931. Bachmann (1996), who relies on the official archives, does not mention Schacht as a significant force in the rejection of this proposal. On p. 238, she quotes a Mendelssohn partner’s advice to a German official on dealing with Lee, Higginson, but is unaware of the real context, which only emerges from the McKittrick Collection. From the latter, it appears the negotiations about a special relationship between the bank and the German government went on rather longer than indicated in the sources Bachmann cites. 37. “American” is used here for clarity, because the best known twentieth century examples come from the U.S. In the nineteenth century, such panics were not uncommon in many European countries, including Germany. 41
4259 Ch01 4/12/02 9:42 am Page 42
42 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
38. The data come from “Zwischenbilanzen Deutscher Banken,” Wirtschaftsdienst, 16, passim. These are the same data that appear in Die Bank. They are reproduced in Institut für Konjunkturforschung (1933). For a general account of German banking in this period, see Whale (1930). 39. These figures include deposits from other banks and omit deposits of longer maturity. 40. Other, smaller assets include cash, Giro accounts, Lombards, and securities. None of them changed in any important way during the first half of 1931. 41. It should be noted that the total time deposits on 30 May 1931 from the individual banks are 100 million RM less than the total in Table 5. 42. See Forbes (2000, p. 33) and Morton (1943, pp. 30–33; 270–271). 43. The collapse of the Nordwolle was a process rather than event. We discuss some of the major dates below. But see the discussions in Born (1967) and Fiedler (1997). 44. Reichsbank to Federal Reserve Bank of New York, May 13, 1931. New York Federal Reserve Bank Archives, C261, “Germany-Reichsbank 1931–1931.” In the original translated cable, the order of the two words in brackets are reversed. The translations were often clumsy, with many showing subsequent hand corrections, and the meaning here is clear. We would be the last to deny that most official German spokespersons, including Reichsbank officials, normally took quite a different line in public and sometimes in private. But this is hardly surprising, since they could not well blame the crisis on their own actions without making everything much worse and even possibly bringing down the government. Still it is interesting that a month later, as its gold reserves disappeared in the rip tide of withdrawals and German public hand wringing about the Austrian crisis reached a crescendo, the Reichbank could still privately advise the Fed that “First reason for withdrawals probably certain political nervousness abroad with concomitant uncertainty concerning reparations.” See Luther to Harrison, June 11, 1931. New York Federal Reserve Bank Archives. C261 Policy and Procedures Germany-Reichsbank. 45. Reichsamt (1933a, p. 78) reports monthly flows in and out of savings banks. Reichsamt (1933b, p. 101) presents bimonthly data for cooperative banks. In JanuaryFebruary, there were net inflows; March-April and May-June show very small outflows of sizes that rule out talk of a run. After July, the size of the outflows increases greatly. But that only makes our point. 46. See the discussion in Whale (1930, pp. 334–339). 47. The currency data are from Reichsamt (1932, p. 326). The deposits are demand deposits for all credit banks, as discussed above. See also the more aggregrate data, which reveal the same pattern, in James (1984). 48. Various versions of the “structural weakness” view exist; the account in James (1986) certainly contains elements of such a view. Goedde (2000) presents a strong argument that German bankers over-lent in the conviction that they were “too big to fail.” 49. For bank stocks and the general market, see our discussion, above, Note 24; Born observed that banks owned industrial shares, but he had a totally different point in mind. See Born (1967, p. 59). 50. The Economist, May 9, 1931, p. 15; the article summarized “The German Banking Position.” On p. 8 of the same issue, the magazine wrote that “in the current year the marked upward movement in the share and loan market has been favorable to the banks
4259 Ch01 4/12/02 9:42 am Page 43
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
43
and their investments are worth more than at the end of last year.” For data on bank stocks and stocks in general, see (Reichsamt, 1932, pp. 354–355). 51. The celebrated Treaty of Locarno that had ushered in the famous Franco-German rapprochement of the mid-twenties only guaranteed borders in western Europe. For Brüning’s response to Briand, see, e.g. Winkler (1993, pp. 382ff.) A particularly succinct summary is Feuchtwanger (1995, pp. 239): “It had been an objective of Hindenburg and Schleicher in appointing Brüning that such a shift should occur. Their aims were fully shared by the Chancellor, though here, as in his move toward presidential government, caution was his watchword.” 52. The literature here is enormous; see, e.g. Winkler (1993, pp. 382ff.) A particularly acute contemporary overview is Wheeler-Bennett (1935, pp. 4–5). This British author was then in close touch with Brüning and many senior German military leaders. See Wheeler-Bennett (1974). 53. For the pocket battleship’s effect on the military balance, see below; for the domestic political problems, see, e.g. Winkler (1993, pp. 400ff.). The issue resonated strongly with voters on the left; by supporting the ship, the SPD opened itself to attack by the KPD. Because large navies were expensive, arms races also posed real threats to budgets in many countries. In the Spring of 1931, Lamont and other bankers were optimistic about Germany in part because of their confidence in the prospects for additional arms control agreements. See, e.g. other parts of the letter of March 10, probably by Lamont, quoted earlier or the reference to increased “prospects for political security in Europe” in the Morgan partners’ letter of March 14, 1931, above. Put simply, arms races raised country credit risks. 54. The need for foreign capital comes up several times in Cabinet deliberations in this period, recorded in the cabinet deliberations recorded in Koops (1982). 55. For the Depression and French colonial policy, see Jacques Marseille (1984, especially Chapters 6–9) and Bouvier, Girault and Thobie (1986, pp. 235ff.) On the exposition, which was a huge success and, in contrast to most such ventures, actually turned a profit, see Bouvier, Girault and Thobie (1986, p. 230). In the early thirties the French army’s Marshal Lyautey was a leader in attempting to unite business groups and the military behind policies to build up the empire. For German complaints about French “imperialism,” see, e.g. Stegmann (1978, p. 217–218); for specific German firms, see the discussion below. For the protests occasioned by German export pressures in Britain, see, among many sources, Forbes (2000, pp. 30–31 and 46–47). Forbes’ work deals mostly with the period after the British abandonment of gold, which was certainly a watershed, but the pressures were evident before the sterling devaluation. See the excellent discussion in Cain and Hopkins (1993, pp. 72–75). 56. The literature on German foreign policy and “Mitteleuropa” is too huge and contentious for this paper to attempt to review. Many studies, such as Kaiser (1980) have been dubious about the role economic factors played in these developments. But see Berghahn (1996, especially the Introduction) and Stegmann, “Mitteleuropa.” The latter’s discussion of splits within the business community, particularly over tariffs, and how France figured in the thinking of various groups bears careful reading. For German General Electric, the chemical industry, etc., see pp. 208ff. As Stegmann notes (p. 218), the famous German Mitteleuropäischen Wirtschaftstag, which agitated for German designs to the east, had been subsidized for a while by the German Foreign Office and supported by Schacht while President of the Reichsbank. It was reorganized in 1931 with financial support from Siemens, I. G. Farben, leading iron and steel firms, and a 43
4259 Ch01 4/12/02 9:42 am Page 44
44 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
major mining association. See below for additional references and discussion on the role of big business in these movements. The early German proposals definitely interested some French groups, though they imagined themselves in the leading role. For the French side, see, besides Stegmann, Knipping (1987, pp. 168–176), though, like Winkler (whose own treatment contains additional evidence on blocs within German business), we have some reservations about certain of his points. For I. G. Farben and the Brüning cabinet, see, e.g. Abraham (1986, p. 157), the sources cited earlier, but particularly the speech by Carl Duisburg discussed below, in support of the customs initiative. The vast buildup of gold reserves since the stabilization of the Franc gave the French state new abilities to project its financial power into central Europe. There is no doubt this worried German policy makers and business groups, though the French strategy of “encircling” Germany to the east was not always pursued with single-minded resolution. (See on this point, Wandycz (1988, pp. 178ff.) Still, some British policy makers who were favorably inclined toward Germany, were highly critical of the French in the run up to the summer 1931 crisis. See, e.g. Montagu Norman’s comments on French “encirclement” policy recorded in Henry Stimson’s Diary for April 8, 1931. The original of the diary is at Yale University Library; our reference is to the microfilm version, Reel 3. 57. Among the immense literature, see, e.g. Bennett (1962); Stambrook (1961); and Kaiser (1980). Stegmann, “”Mitteleuropa,” brings out the institutionalized cooperation between big business and the government, as well as the secrecy which enveloped their deliberations. For Brüning on the timing and the (false) claim that Germany could not choose it, see Stambrook (1961, p. 38). The cabinet deliberations are recorded in Koops (1982). 58. New York Times, March 26, 1931, p. 12. The remarkable speech of March 24, 1931 by Carl Duisberg of I. G. Farben defending the notion of a customs union with Austria (and other countries in southeastern Europe) placed it firmly in the context of other countries’ efforts to form tighter trade blocs and bluntly stated that reconciliation with France would come only after Germany moved east. Indeed, the speech spoke vaguely of an eventual bloc extending from Bordeaux to either Sofia or Odessa, depending on which version of the text one relies upon. See the (incomplete) text and notes reprinted in, e.g. Michalka and Niedhart (1992, pp. 245–246). Because this speech almost uncannily anticipated the subsequent direction of German policy, it has frequently been instanced as an example of the role big business played in German policy making. In recent years, however, a critical counter-current has emphasized the continuing attachment to international markets of some large firms, notably I. G. Farben and Siemens. See, e.g. Feldenkirchen (1999) for Siemens; and for I. G. Farben, Plumpe (1990, pp. 504–505) and Hayes (2001, p. 45). The latter acknowledges Stegmann’s (1978) observation that Farben helped to subsidize the Mitteleuropäischer Wirtschaftstag, but disputes its significance. For our paper, this discussion is not material, once one recognizes that the German government did attempt a customs union and that this step seriously damaged its relations with France. But Duisberg’s speech was no venture in cooperative internationalism. Hayes’ claim (p. 45) that Duisberg stressed the importance of an understanding with France misses the critical point that no one in the audience – or the French – could have: that while the Farben executive felt that an accord with France was important for a final (“endgültige”) resolution of European problems, he believed it could wait until after Germany, Austria, and southeastern Europe got moving on the “Zollunion” he advocated.
4259 Ch01 4/12/02 9:42 am Page 45
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
45
59. New York Times, March 26, 1931, p. 12. Compare a contemporary assessment in Dulles (1932, p. 239): “In fact, although March 22 has been called a turning point because of the Austro-German Customs proposal, the tragic consequences were not evident for about a month.” For the attempts at compromise, see Bennett (1962, pp. 96ff.). 60. For the discord over French attempts to buy into municipal utilities, see the ministerial discussion of March 27, 1931 in Koops (1982, Vol. II, pp. 994–996), which is quite clear about the embarrassing character of the problem. For the furor over the French Finance Minister’s remarks, see Freiherr von der Osten-Warnitz to President Hindenburg, March 2, 1931 in Opitz (1977, pp. 576–579). For the personal discussions between him and the President, see Meissner to Brüning, March 5, 2001 on p. 580; this indicates that the President directed that the letter be sent to the Chancellor, which was a characteristic way Hindenburg put pressure on him. See, e.g. Dorpalen (1964), who also notes that Hindenburg could often resist almost anything except pressures from his fellow Junkers. 61. See the discussion below on the Germans and the BIS; our dating follows the primary sources we cite; for the Kindersley Plan, see also Eichengreen and James (2001). 62. Gates W. McGarrah telegram sent through Morgan & Cie., to J. P. Morgan & Co., New York, March 18, 1931; copied to Harrison, Owen D. Young, and Jackson Reynolds. We have slightly regularized punctuation and spelling in the telegram. The telegram is in The Pierpont Morgan Library, New York, Morgan Bank European Papers, Box 13. 63. The first quotations in this paragraph come from the McGarrah’s telegram of March 18, just cited above. The American bankers response is from a telegram signed by J. P. Morgan, T. W. Lamont, and S. P. Gilbert; sent as a telegram from J. P. Morgan & Co. to Morgan et Cie., March 14, 1931; to be relayed back to McGarrah and confederates. The telegram also mentions that putting U.S. dollars in international institutions would undoubtedly arouse criticism. The telegram can be found in the The Pierpont Morgan Library, New York, Morgan Bank European Papers, Box 13. The Latin bankers’ objections are also recorded in McGarrah to J. P. Morgan & Co., March 18, 1931; also in Box 13. The Belgian efforts to move the French are related in Bussière (1992, pp. 397ff.). Details of the late Spring proposals to the BIS are in Carton B32221 of the French Ministry of Finance Archives, Savigny-le-Temple. These contain reports from the French consul at Basle. See, for example, the copy of his letter to the Minister of Foreign Affairs, April 21, 1931, which reports how the French blocked action on a proposal by Luther for BIS action in support of German export credits. A copy of his letter to the Minister of Foreign Affairs, May 20, 1931, reports on a BIS meeting in Basle in which a plan of the Belgian banker Franqui for more aggressive action by the BIS received support from Norman and Luther. The board nevertheless refused to act. See also the “Rapport du Comite de la BRI, tenu a Bruxelles, les 3 et 4 Juin, 1931” in the same Carton, in which, yet again, Melchior from Germany and Sir Charles Addis from the U.K. lined up behind Franqui, while the French opposed more vigorous intervention. 64. By itself, the customs union proposal rendered exports of long term capital impossible, but did not completely shut the market for short term capital. The nature of the German bind (before the budget blew up) was widely appreciated. As the Commercial and Financial Chronicle summed up the widely shared sentiment on the eve of the fall: “The chief difficulty of the Berlin money market is the disinclination of international bankers to make long-term issues.” See p. 3602 of the issue of May 16, 1931. 45
4259 Ch01 4/12/02 9:42 am Page 46
46 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
65. German sources disagree slightly on the exact amount of the shortfall. Winkler (1993, pp. 403ff.) discusses Dietrich’s warning to Brüning. Bachmann (1996, pp. 249–250), cites internal documents indicating numbers close, but not identical, to those. The various ministerial discussions are recorded in Koops (1982). 66. “Besprechung im Reichstag vom 6. März, 1931,” Koops (1982, Vol. II, p. 928). See also the account by a top civil servant who was one of the very few people present, Hermann Pünder (1961, pp. 92–93). 67. Again, various German sources disagree very slightly on the exact numbers. Compare Bachmann (1996, p. 249) with James (1986, p. 73). The provision for further cuts in the event of budget shortfalls was noted earlier in the New York Times, March 26, 1931, p. 2, which at the time simply treated it as one more piece of good news. For Lee, Higginson’s reaction to the confidential news, see the telegram from the New York office to Courtney in Lee, Higginson’s Paris office, March 10, 1931, Thomas H. McKittrick Collection, Baker Library, Harvard Business School, Box 6: “Any move permitting interpretation that government was again borrowing would damage it greatly.” The reason is obvious: the world was just then celebrating Brüning’s budgetary triumph, with the key votes coming in the next days. 68. See “Private and Confidential,” [L,H Paris office telegram] to Murnane, March 10, 1931, Thomas H. McKittrick Collection, Baker Library, Harvard Business School: “Have therefore thought of approaching Moret . . .. As he himself has already stated he is much interested in a financial rapprochement between the two countries and naturally the whole world particularly perhaps France is interested in doing what is possible and practical to make as easy and comfortable as possible the internal money market in Germany.” “We could give him in confidence if possible some data about German requirements over next few months and the way German treasury bills are created by Reichstag approval and so forth . . ..” 69. “Billions” is, again, meant here in the American sense. The German text speaks of “Milliarden.” See, “Ministerbesprechung vom 29. May 1931,” Koops (1982, II, p. 1131). 70. On Brüning’s awareness that the American election would affect the timing of his plans, see “Besprechung vom 7. Mai, 1931,” Koops (1982, Vol. II, p. 1054; for the BIS, see p. 1056. The role of the American elections was not a secret; see Commercial and Financial Chronicle, May 16, 1931, p. 3636. Brüning and his advisors repeatedly discussed the BIS. See, e.g. the discussions of March 6, Koops (1982, Vol. II, p. 927) or June 3 (p. 1179). 71. “Besprechung vom 7. Mai, 1931,” Koops (1982, Vol. II, p. 1054). In the protocol, what Brüning actually says is is “dann aber werde sich keine Regierung der Mitte finden, die bereit sei, dem Rechtskabinett die Verantwortung abzunehmen.” The Labor Minister, who was close to Brüning, agreed completely on the need to move on reparations. See p. 1055. 72. “Besprechung vom 7. Mai, 1931,” Koops (1982, Vol. II, pp. 1058–1059). 73. For the diplomatic issues, see especially, Bennett (1962). For the earlier stages of Brüning’s decision, see Koops (1982, Vol. II, pp. 925–928). For the connection between the British talks and the domestic crisis, see Winkler (1993, pp. 404ff.). 74. James (1986, p. 306) makes this claim. But the April tax collection numbers had been published in May and were recognized at once as a serious problem. See the discussion and references below. 75. See, e.g. The Economist, May 9, 1931, p. 1001. The influence of particular situations should not be exaggerated, however. As noted above, German savings banks
4259 Ch01 4/12/02 9:42 am Page 47
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
47
actually held more deposits at the end of May than they did at the beginning of the month. 76. They were published regularly and quickly by several publications, including Wirtschaft und Statistik and often in daily papers, such as the Frankfurter Zeitung. Major foreign papers carried also discussed the dismal numbers. See, e.g. the New York Times report from Berlin on May 15 as excerpted in the Commercial and Financial Chronicle, May 23, 1931, p. 3810: “German Tax Revenue Below Estimate – Shortage of 500,000,000 Marks Now Indicated – Expenditure To Be Cut.” In most recent literature by economic historians, the proposal for a customs union receives at most, passing attention. It is not integrated into a broader analysis, so that its pivotal significance emerges. It is worth noting, however, that some, but not all, contemporary accounts did realize the significance of the customs union proposal. Besides the New York Times story quoted above, see the discussion in Einzig (1935, p. 217), though he also touts the role of the Creditanstalt. Much sharper formulations can be found in Moulton and Pasvolsky (1932, p. 308) and especially Salter (1932). The latter, an influential figure in British financial and political circles, wrote (p. 226) that “in June, as the conditions tending to a crisis developed, men were asking each other in New York, as they had been six months before in Paris, whether there was a danger of a new war in Europe. The addition of these fears to the existing doubts as to the solvency of debtors impoverished by the depression gave a fatal shock to confidence.” 77. See, e.g. the following articles on the budget crisis in the Frankfurter Zeitung: “Wie Wird das Defizit des Reichsetats gedeckt werden,” on p. 1 of the May 6 issue; “Bemerkungen, Das Defizit,” on p. 1, May 21, 1931, along with more detailed material on p. 3. This contains a blunt assessment of the meaning of the April revenue shortfall. 78. See The Economist, May 2, 1931, p. 944; for the failure, see The Economist, May 23, 1931, p. 1102. This is vague about the date, but not the flop. The issue came out on May 4, 1931; see the data in Rosenstiel (1931b, p. 131, Table). 79. For the Prussian SPD, the street clashes, and the Minister of the Interior, see Winkler (1993, pp. 403–405). 80. Compare Brüning’s statement to his ministers on May 7 with the tone of the meeting on June 3, particularly the Finance Minister’s comments, which neither Brüning nor anyone else bothers to rebut. The latter meeting also witnesses much grasping at straws, even a mention of the possibility that Norman would finally persuade the BIS to move seriously. 81. For examples where Brüning flatly denied he planned to change policy, see, e.g. his May 7 assertion that a change in policy would be fatal, Koops (1982, Vol. II, pp. 1056–1057) or his very plain spoken June 2 comments (p. 1164). For the Finance Minister, see, e.g. pp. 1178–1179; especially N. 1 on p. 1179 for the threat to resign. For contrasting public statements by him and the Chancellor, compare Brüning in the New York Times, May 11, 1931, p. 10, with Dietrich’s remarks on the inevitability of reparations concessions, Commercial and Financial Chronicle, May 16, 1931, p. 3636. 82. For the rallies, see, e.g. Feuchtwanger (1995, p. 251). Given the Versailles Treaty’s restrictions on the size of the German army, the Army believed it had to have the cooperation of the Stahlhelm in border regions (p. 234) and, as shortly became apparent, it was also becoming far more interested in cooperation with the Nazi SA. Some prominent industrialists were members of the Stahlhelm; Hindenburg was long its nominal head, though the Stahlhelm leadership often considered him too moderate. For 47
4259 Ch01 4/12/02 9:42 am Page 48
48 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
the French election, see, e.g. Winkler (1993, p. 406). On systemic risk, see especially Salter’s remarks, quoted above. 83. For the effect on French naval planning, see Toynbee (1932, p. 71 ). For the collapse of the French-Italian naval agreement in the wake of the proposal for a customs union, see pp. 273–274. The pocket battleship created special problems because it was said either to be able to sink or outrun any ship afloat. See Wheeler-Bennett (1932, pp. 336–337). This was a major reason why Brüning’s decision, mentioned above, to press ahead with another was so controversial. It threatened a multinational round of big new arms spending. 84. For Gröner’s warning, and the urgency of heeding it, see Bennett (1962, pp. 120–121). For Voegler’s speech and the meeting, see the account in the London Times, June 4, 1931, p. 13. Other prominent industrialists, including Thyssen and Reusch, who were present, criticized Brüning for being weak and called for authoritarian methods and an anti-socialist drive. The article also discusses Hjalmar Schacht’s promotion of a moratorium before a group of industrial leaders. 85. For the dates, see Toynbee (1932). 86. A “Note du Mouvement des Fonds” of the 26th of May, 1931, bearing the title “Reparations et Dettes Interalliees,” now in the archives of the Banque de France (in a series of cartons numbered “59)” shows that the French had taken the hints in the German press and expected that Brüning would raise the reparations issue with the British. For the fall in Young Plan bonds, see Toynbee (1932, p. 68); this also mentions the Stahlhelm rally discussed earlier. 87. See, e.g. the summary in Woodward and Butler (1947, Vol. II, p. 70). 88. The London Times account of the decree also reported the official denial. See the Times, June 6, 1931, p. 11. The Commercial and Financial Chronicle, June 13, 1931, p. 4293, quoted a story by the Associated Press forecasting an imminent move on reparations by Germany, despite the denials. The Chronicle, p. 4293, also noted that the manifesto for the first time referred to reparations as “tribute.” 89. The deal was between the government and a syndicate led by Ivar Kreuger, the “Match King.” Kreuger and Lee, Higginson were so closely linked that the subsequent bankruptcy of the one spelled the ruin of the other. The $75 million dollar payment was not a surprise; but the New York Times announcement came on May 28, 1931, as the paper also reported on the previous day’s sales of German assets. See p. 43 of that issue. Simple arithmetic and a comparison with the other entries for Reichsbank total gold exchange reserves for May in our Table 8 confirms the Reichsbank’s statement to the New York Federal Reserve Bank quoted earlier on the insignificance of German bank relations with the Creditanstalt and rules out traditional stories of chain bankruptcy spreading from Austria to Germany. Note that not later than about May 25, both German newspapers and French archives agree that word of Brüning’s shift in course began leaking. As observed above, this clearly led to a first, and by later standards, rather modest, wave of selling of Mark assets in world financial markets, which does not require invoking the Creditanstalt. It is also worth noting that in a telephone conversation with Harrison of the New York Fed on June 20, 1931, Montagu Norman indicated that end of the month demands on the Reichsbank were normally heavier; the remark appears to be meant generally and not to be simply about the June situation. See New York Federal Reserve Bank Archives, File 3115.2; this is a file of memoranda by Harrison on conversations with Norman. Late Spring reverses in the American market may have made some U.S. investors nervous and led them to bring some money home. Federal Reserve behavior in this
4259 Ch01 4/12/02 9:42 am Page 49
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
49
period does suggest some anxiety and was markedly at variance with its actions during the rest of the year. Cf. the remark on this period in Epstein and Ferguson (1984). But Table 8 and our earlier discussion of banking data also rule out significant influences on the German situation from this quarter in May. 90. The characterization comes from the Berliner Tageblatt, cited in Mommsen (1996, p. 387. 91. For the possibility of running out of money, the fears of Groener and others of a revolution, the possibility of military intervention, etc., see Winkler (1993, pp. 408–413) and Feuchtwanger (1995, pp. 246–248). For Schleicher’s meeting with Brüning, see the Chancellor’s official calendar for June 13, 1931, Brüning Papers, Harvard University Archives, HUG (FP) 93.35, Box 1 of 6. A note there explains that the Reichswehr was calling in the commanders of the various military districts for consultations on procedures in case of civil unrest. 92. See, again, Winkler (1993, pp. 408–413) and Feuchtwanger (1995, pp. 246–248). 93. Temin (1989); see also Born (1967, pp. 74ff.) For the gold and foreign exchange cover, see our Table 8 and Reichsamt (1932, p. 326). Born quotes slightly different figures that do not affect any issue under discussion; see his discussion on p. 74. The actual cover requirement was somewhat complex. Reichsbank notes in circulation required a minimum cover of 40% in gold or foreign exchange. Foreign exchange could not comprise more than 25% of the total cover; this is equivalent to a 30% gold cover rule. See Madden and Nadler (1935, pp. 373ff.) which also discusses the limited role played by the handful of banks that still had powers to issue notes. The entry in our Table 4 for Young Bonds is somewhat misleading, in that it reflects the momentary revival brought about toward the end of the month by the initial offer of the Hoover Moratorium. In mid-June, the bonds fell as low as 65. Later in the summer, they collapsed, of course. By September, they languished in the thirties. 94. Balderston argued similarly that the problem with the German banks was the Reichsbank’s inability to supply reserves, but he did not connect the Reichsbank’s position in June with Weimar’s political act of March or to the German budgetary crisis. See Balderston (1993) and Balderston (1994). 95. Fleming (2000, p. 147). Fleming’s article details his notes on Keynes’ lectures in Geneva during July, 1929. 96. Note that both Brüning and Luther mention the possibility of revaluing gold in the meetings just prior to the fatal Notverordung. See Koops (1982, Vol. II, p.1058). It appears they were looking for some kind of international agreement on this, but the discussion is too abbreviated to be sure. It is perhaps of interest that Brüning’s appointment books indicate that Sir Walter Layton was the last official visitor he saw before the American ambassador arrived to discuss the text of the famous German appeal for help to Hoover and the first person the Chancellor saw afterward. 97. See Borchardt (1991); for the critics, see especially, Holtfrerich (1990) and Kim (1997).
36 37
ACKNOWLEDGMENT
38 39 40
We are grateful to seminar participants at Oxford, the London School of Economics, Harvard, and the Wharton School for helpful comments. 49
4259 Ch01 4/12/02 9:42 am Page 50
50 1
THOMAS FERGUSON AND PETER TEMIN
REFERENCES
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Aalders, G., & Wiebes, C. (1996). The Art of Cloaking Ownership. Amsterdam: University of Amsterdam Press. Abraham, D. (1986). The Collapse of the Weimar Republic (2nd ed.). New York: Holmes & Meier. Bachmann, U. (1996). Reichskasse und öffentlicher Kredit in der Weimarer Republik: 1924–1932. Frankfurt am Main: Peter Lang. Balderston, T. (1993). The Origins and Course of the German Economic Crisis: November 1923May 1932. Berlin: Halde & Spener. Balderston, T. (1994). The Banks and the Gold Standard in the German Financial Crisis of 1931. Financial History Review, 1, 43–68. Bennett, E. (1962). Germany and the Diplomacy of Financial Crisis. Cambridge, MA: Harvard University Press. Berghahn, V. (1996). Quest for Economic Empire: European Strategies of German Big Business in the Twentieth Century. Providence, Rhode Island: Berghahn Books. Borchardt, K. (1991). Perspectives on Modern German Economic History and Policy. Translated by Peter Lambert. Chicago: University of Chicago Press. Born, K. (1967). Die deutsche Bankenkrise 1931: Finanzen und Politik. Munich: R. Piper. Bouvier, J., Girault, R., & Thobie, J. (1986). L’impérialisme à la française. Paris: Editions la decouverte. Bratz, M. (1958). Der deutsche Privatbankierstand in der Nachkriegzeit: 1918–1933. Berlin: Paul Brandel. Burke, B. (1994). Ambassador Frederic Sackett and the Collapse of the Weimar Republic, 1930–1933. New York: Cambridge University Press. Bussière, É. (1992). La France, la Belgique et l’organisation économique de l’Europe: 1918–1935. Paris: Ministère de l’économie des finances et du budget: Comité pour l’histoire économique et financiere de la France. Cain, P. J., & Hopkins, A. G. (1993). British Imperialism: Crisis and Deconstruction: 1914–1990. New York: Longman. Deutsche Bundesbank (1976). Währung und Wirtschaft in Deutschland: 1876–1975. Frankfurt am Main. Deutsches Institute für Wirtschaftsforschung (1932). Wirtschaftszahlen 1925 bis 1931. Berlin: R. Hobbing. Dorpalen, A. (1964). Hindenburg and the Weimar Republic. Princeton: Princeton University Press. Dulles, E. (1932). The Bank for International Settlements at Work. New York: Macmillan. Eichengreen, B. (1992). Golden Fetters: The Gold Standard and the Great Depression. New York: Oxford University Press. Eichengreen, B., & Hatton, T. (Eds) (1988). Interwar Unemployment in International Perspective. Dordrecht: Kluwer. Eichengreen, B., & James, H. (2001). International Monetary and Financial Reform in Two Ages of Globalization. In: J. Williamson & J. Taylor (Eds), Globalization and History. Chicago: University of Chicago Press. Einzig, P. (1935). World Finance: 1914–1935. New York: Macmillan. Ellis, H. (1941). Exchange Control in Central Europe. Cambridge: Harvard University Press. Epstein, G., & Ferguson, T. (1984). Monetary Policy, Loan Liquidation, and Industrial Conflict: The Federal Reserve and the Open Market Operations of 1932. Journal of Economic History, 51, 190–200.
4259 Ch01 4/12/02 9:42 am Page 51
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
51
Feldenkirchen, W. (1999). Siemens, 1918–1945. Columbus, Ohio: Ohio State University Press. Feinstein, C., & Watson, K. (1995). Private International Capital Flows in Europe in the Inter-War Period. In: C. Feinstein (Ed.), Banking, Currency, and Finance in Europe Between the Wars (pp. 94–130). Oxford: Clarendon Press. Ferguson, T. (1995). Golden Rule: The Investment Theory of Party Competition and the Logic of Money-Driven Political Systems. Chicago: University of Chicago Press. Feuchtwanger, E. (1995). From Weimar to Hitler Germany, 1918–1933 (2nd ed.). New York: St. Martin’s. Fiedler, M. (1997). Business Scandals in the Weimar Republic. In: W. Feldenkirchen & T. Gourvish (Eds), European Yearbook of Business History (pp. 155–178). Flechtner, H.-J. (1959). Carl Duisberg: vom Chemiker zum Wirtschaftsführer. Düsseldorf: Econ Verlag. Fleming, G. (2000). Foreign Investment, Reparations, and the Proposal for an International Central Bank: Notes on the Lectures of J. M. Keynes in Geneva, July 1929. Cambridge Journal of Economics, 24, 139–151. Forbes, N. (2000). Doing Business with the Nazis: Britain’s Economic and Financial Relations with Germany: 1931–1939. London: Frank Cass. Goedde, I. (2000). The German Twin Crises of 1931. Manuscript. Goldstein, M., Kaminsky, G., & Reinhart, C. (2000). Assessing Financial Vulnerability: An Early Warning System for Emerging Markets. Washington, D.C.: Institute for International Economics. Grossman, R. (1994). The Shoe That Didn’t Drop: Explaining Banking Stability During the Great Depression. Journal of Economic History, 54, 654–682. Harris, C. (1935). Germany’s Foreign Indebtedness. London: Oxford University Press. Hayes, P. (2001). Industry and Ideology: I. G. Farben in the Nazi Era (2nd ed.). New York: Cambridge University Press. Heyde, P. (1998). Das Ende der Reparationen: Deutschland, Frankreich und der Youngplan: 1929–1932. Paderborn: F. Schoningh. Holtfrerich, C. (1990). Economic Policy Options and the End of the Weimar Republic. In: I. Kershaw (Ed.), Weimar: Why Did German Democracy Fail? New York: St. Martin’s. Institut für Konjunkturforschung (1933). Konjunktur-statistisches Handbuch. Berlin: Verlag von Reimar Hobbing. Institute of International Finance (1931). Credit Position of Germany. New York: Institute of International Finance. James, H. (1984). The Causes of the German Banking Crisis of 1931. Economic History Review, 38, 68–87. James, H. (1986). The German Slump Politics and Economics: 1924–1936. Oxford: Clarendon Press. James, H. (2001a). The Deutsche Bank and the Nazi Economic War against the Jews. Cambridge: Cambridge University Press. James, H. (2001b). The End of Globalization. Cambridge MA: Harvard University Press. Kaiser, D. (1980). Economic Diplomacy and the Origins of the Second World War. Princeton: Princeton University. Kim, H.-I. (1997). Industrie, Staat, und Wirtschaftspolitik: Die konjunkturpolitische Diskussion in der Endphase der Weimarer Republik. Berlin: Duncker & Humblot. Kindleberger, C. (1986). The World in Depression (2nd ed.). Berkeley: University of California Press. Knipping, F. (1987). Deutschland, Frankreich und das Ende der Locarno Ara, 1928–1931. Munich: Oldenbourg.
51
4259 Ch01 4/12/02 9:42 am Page 52
52 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
THOMAS FERGUSON AND PETER TEMIN
Kolb, E. (1988). The Weimar Republic, P. S. Falla (Trans.). London: Unwin Hyman. Kolb, E. (1993). Die Weimarer Republik (3rd ed.). Munich: R. Oldenbourg Verlag. Koops, T. (Ed.) (1982). Die Kabinette Brüning, Akten der Reichskanzlei (Vol. 2). Boppard am Rhein: H. Boldt Verlag. Kupferberg, H. (1972). The Mendelssohns. New York: Scribner’s. Länderrat des Amerikanischen Besatzungsgebiets (Ed.) (1949). Statistisches Handbuch von Deutschland 1928–1944 (Vol. II). Munich: Franz Ehrenwirth-Verlag. Madden, J., & Nadler, M. (1935). The International Money Markets. New York: Prentice Hall. Marseille, J. (1984). Empire colonial et capitalisme français. Paris: Albin Michel. McNeil, W. (1986). American Money and the Weimar Republic. New York: Columbia University Press. Michalka, W., & Niedhart, W. (1992). Deutsche Geschichte 1918–1933: Dokumente zur Innenund Aussenpolitik. Frankfurt am Main: Fischer Verlag. Mommsen, H. (1996). The Rise and Fall of Weimar Democracy, E. Forster & L. Jones (Trans.). Chapel Hill, N.C.: University of North Carolina Press. Morton, W. (1943). British Finance, 1930–1940. Madison, Wisconsin: University of Wisconsin Press. Moulton, H., & Pasvolsky, L. (1932). War Debts and World Prosperity. New York: Century Co. for the Brookings Institution. Opitz, R. (Ed.) (1977). Europastrategien des deutschen Kapitals: 1900–1945. Köln: Pahl-Rugenstein Verlag. Plumpe, G. (1990). Die I. G. Farbenindustrie AG: Wirtschaft, Technik und Politik: 1904–1945. Berlin: Duncker & Humblot. Puender, H. (1961). Politik in der Reichskanzlei. Stuttgart: Deutsche-Verlags Anstalt. Reichsamt, Statistisches (1932). Statistisches Jahrbuch für das deutsche Reich. Berlin: R. Hobbing Reichsamt, Statistisches (1933a). Die deutschen Sparkassen in den Jahren 1931 und 1932. Vierteljahrshefte zur Statistik des deutschen Reichs, 42(2), 76–100. Reichsamt, Statistisches (1933b). Die gewerblichen Genossenschaftsbanken in den Jahren: 1931/1932. Vierteljahrshefte zur Statistik des deutschen Reichs, 42(2), 101–105. Reichsamt, Statistisches (1933c). Deutsche Wirtschaftskunde: Ein Abriss der deutschen Reichsstatistik (2nd ed.). Berlin: R. Hobbing. Ritschl, A. (1998). Reparations Transfers, the Borchardt Hypothesis, and the Great Depression in Germany. European Review of Economic History, 2, 49–72. Rosenstiel, F. (1931a). Die Kreditmärkte. Wirtschaftskurve, 10 (April), 17–21. Rosenstiel, F. (1931b). Die Kreditmärkte. Wirtschaftskurve, 10 (July), 125–131. Salter, A., Sir (1932). Recovery the Second Effort. London: G. Bell and Sons. Schmidt, C. (1934). German Business Cycles, 1924–1933. New York: National Bureau of Economic Research. Schuker, S. (1988). American ‘Reparations’ to Germany, 1919–1933: Implications for the Third World Debt Crisis. Princeton: Princeton University Department of Economics: Princeton Studies in International Finance. Southern, D. (1981). The Impact of the Inflation: Inflation, the Courts, and Revaluation. In: R. Bessel & E. J. Feuchtwanger (Eds), Social Change and Political Development in Weimar Germany. London: Croom Helm. Stachura, P. (Ed.) (1986). Unemployment and the Great Depression in Weimar Germany. New York: St. Martin’s. Stambrook, F. (1961). The German-Austrian Customs Union Project of 1931: A Study of German Methods and Aims. Journal of Central European History, 21, 15–44.
4259 Ch01 4/12/02 9:42 am Page 53
German Currency Crisis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
53
Stegmann, D. (1978). Mitteleuropa: 1925–1934. In: D. Stegmann, B.-J. Wendt & P.-C. Witt (Eds), Industrielle Gesellschaft und politisches System: Festschrift für Fritz Fischer (pp. 203–221). Bonn: Verlag Neue Gesellschaft. Temin, P. (1989). Lessons from the Great Depression. Cambridge MA: MIT Press. Thomas, B. (1934). Germany. In: H. Dalton, B. Thomas, J. Reedman, J. Hughes & W. Leaning (Eds), Unbalanced Budgets. London: George Routledge. Toynbee, A. (Ed.) (1932). Survey of International Affairs, 1931. London: Oxford University Press. Treue, W. (1972). Das Bankhaus Mendelssohn als Beispiel einer Privatbank im 19. und 20. Jahrhundert. In: Mendelssohn-Gesellschaft (Ed.), Mendelssohn Studien (pp. 29–80). Berlin: Duncker & Humblot. U.S. Department of State (1945). Papers Relating to the Foreign Relations of the United States 1930 (Vol. III). Washington, D.C.: U.S. Government Printing Office. U.S. Department of State (1986). Microfilm. Confidential U.S. State Department Central Files: Germany, Internal Affairs, 1930–1941. University Publications of America. Wandycz, P. (1988). The Twilight of French Eastern Alliances: 1926–1936. Princeton: Princeton University Press. Whale, P. (1930). Joint Stock Banking in Germany. London: Frank Cass. Reprint, Augustus Kelley. Wheeler-Bennett, J. (1932). Disarmament and Security since Locarno: 1925–1931. New York: Macmillan. Wheeler-Bennett, J. (1935). The Pipe Dream of Peace. New York: William Morrow. Wheeler-Bennett, J. (1974). Knaves, Fools, and Heroes. London: Macmillan. Winkler, H. A. (1993). Weimar 1918–1933: Die Geschichte der ersten deutschen Demokratie. Munich: Verlag C. H. Beck. Woodward, E., & Butler, R. (Eds) (1947). Documents on British Foreign Policy: 1919–1939, Second Series, Vol. 2. London: His Majesty’s Stationery Office.
22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
53
4259 Ch02 4/12/02 9:43 am Page 55
1 2 3 4 5 6 7 8 9 10
THE HOUSEHOLD BALANCE SHEET, CREDIT, AND UNCERTAINTY AT THE ONSET OF THE GREAT DEPRESSION IN THE USA
11 12 13 14
David Greasley and Jakob B. Madsen
15 16 17 18
ABSTRACT
19 20 21 22 23 24 25 26 27
The influence of the household balance sheet, the supply of credit, and uncertainty on consumer spending during the early years of the Great Depression in the USA are assessed within a unified life-cycle consumption function framework. Income uncertainty played the dominant role in the spending declines of 1930 and 1932. The depletion of households’ financial assets contributed modestly to the consumption falls in 1931–1932. Indebtedness, the severe penalties surrounding installment debt default, and the supply of credit had little effect on the consumer spending slump.
28 29
I. INTRODUCTION
30 31 32 33 34 35
A key strand in the historiography of the Great Depression highlights the role played by American consumers (Eichengreen, 1992, pp. 215–216; Fearon, 1987, pp. 151–153). Their real spending fell precipitously, by 19.3% in the period 1929–1933 (Lebergott, 1996, pp. 148–159). Considerable effort has been directed towards understanding this consumption fall since Temin (1976, pp. 62–83)
36 37 38 39 40
Research in Economic History, Volume 21, pages 55–77. Copyright © 2003 by Elsevier Science Ltd. All rights of reproduction in any form reserved. ISBN: 0-7623-0993-8
55
4259 Ch02 4/12/02 9:43 am Page 56
56 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
DAVID GREASLEY AND JAKOB B. MADSEN
emphasized its central role in the Great Depression, and failed to explain its causes. Subsequent historiography has debated whether uncertainty surrounding households’ incomes, consumers’ financial liquidity, credit supplies, or the severity of debt default penalties were causes of the collapse, but, typically, the effects are evaluated individually. This article assesses possible explanations of the consumption decline by considering the competing hypotheses within unified models. The approach estimates life-cycle-type consumption functions for various categories of spending, including durables, semi-durables, perishables, and services. Incorporating liquidity and uncertainty measures within the same consumption function allows their respective contributions to the decline of consumption during the period 1930–1933 to be gauged. Romer (1990), Olney (1999) and Mishkin (1978) provide different interpretations of the consumption collapse during the early years of the Great Depression. Mishkin’s (1976) liquidity hypothesis, potentially, has roles for both the risks associated with illiquidity, and income uncertainty, in explaining the falls in durable purchases. Mishkin’s (1978) model postulates that households’ risks of distress selling of their durable possessions are related to their holdings of financial assets, their debts, but also to their expected income and its variance. He shows that consumers are less likely to purchase durables the higher their debts, the lower their financial asset holdings, and the greater the uncertainty surrounding their expected income. Elsewhere, and rather differently, Mishkin (1990) contends that financial crises, by promoting information asymmetries and adverse selection, played an important role in prolonging the contraction through 1932.1 His measure, based on the spread of interest rates paid by high and low quality borrowers, indicates that lending risks of financial institutions rose sharply in 1931–1932, implying that financial disintermediation may have accentuated the collapse in spending. While much of this discussion centers on the implications for investment, Mishkin (1990, p. 23) notes the relevance of his argument for Romer’s (1990) interpretation of the consumer durables spending. Bernanke (1983a) also highlights the possible importance of financial disintermediation for explaining the great contraction, and utilizes alternative indicators, including the deposits of suspended banks. Romer (1990) takes a different view, and her influential discussion of income uncertainty drew on Bernanke’s (1983b) concept of irreversibility to explain why consumers reduced their durable purchases in 1930. The theory of irreversible purchases considers the value of waiting relative to the immediate acquisition of durables, and shows how heightened uncertainty may lead to the deferring of purchases of durable consumption goods. For Romer, the rise in share price volatility in the year following the Wall Street Crash indicates the
4259 Ch02 4/12/02 9:43 am Page 57
Household Balance Sheet in U.S. Depression 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
57
income uncertainty that led to the decline in consumer durable spending in 1930. Both Romer’s approach invoking the theory of irreversible purchases, and Mishkin’s liquidity model, postulate that income uncertainty reduced durable goods spending. Their positions are separated by Mishkin’s empirical finding that liquidity effects explain most of the durable spending fall during the first year of the Great Depression, and his denial that there was a sharp rise in uncertainty in 1930. He explains around 69% of reduced consumer durable and residential housing spending in 1930 by liquidity balance-sheet effects. Romer does not deny completely that liquidity effects reduced consumption in 1930. Her model though does not allow explicitly for liquidity effects, and shows that heightened income uncertainty explains fully the fall in durable spending in 1930. For Mishkin, uncertainty played a more important role in later years, following the banking and gold standard crises of 1931. Ferderer and Zalewski (1994) agree that the financial crises heightened uncertainty. They show that interest rate volatility was unimportant in 1930 but rose sharply in 1932–1933 as the doubts surrounding the USA’s adherence to the gold standard strengthened. Whether uncertainty played an important role in reducing consumption in 1932 deserves further attention. This year witnessed the largest consumption falls of the Great Depression, and Greasley, Madsen and Oxley (2001) show that both life cycle and permanent income consumption functions, for both durable and non-durable spending, have large negative residuals in 1932.2 New research by Lebergott (1996) shifts the attention from durables by emphasizing the breadth of the post-1929 decline across most categories of consumer spending. Olney (1999) extends the liquidity model to non-durables by postulating that consumers might reduce their propensity to consume, and thus their non-durable consumption, to avoid default on installment debt. Noting the unprecedented rise of installment debt to 1929, she argues that severe debt default penalties led consumers to curtail all spending in 1930 to reduce the likelihood that their durable goods would be repossessed. Ostensibly, Olney’s results explain most of Temin’s autonomous fall in 1930 consumption, although her empirical model excludes uncertainty. Brown (1997) reinforces her view that households were concerned greatly by their debts in 1930, postulating that there was a sharp rise in the voluntary early repayment of debt in that year. A fear of default, though, might not be the only reason why consumers reduce their non-durable consumption. Hahm and Steigerwald’s (1999) theoretical model and empirical findings show that consumers adjust their precautionary savings and non-durable consumption in response to changes in income uncertainty. A variety of explanations have been put forward to explain the consumption slump at the onset of the Great Depression. Here we estimate the effects of 57
4259 Ch02 4/12/02 9:43 am Page 58
58 1 2 3 4 5
DAVID GREASLEY AND JAKOB B. MADSEN
heightened uncertainty and deteriorating household balance sheets within a unified life-cycle consumption function framework using Lebergott’s new data for consumers’ expenditures. Romer’s share price volatility measure of income uncertainty is used, and new measures of financial assets and liabilities are estimated and deployed.
6 7 8
II. UNIFIED MODELS OF CONSUMPTION
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Most studies of the consumption decline at the onset of the Great Depression, including those of Temin (1976), Mayer (1978) and Olney (1999), are based on the life-cycle hypothesis (LCH) of Ando and Modigliani (1963). The model used here adapts the life-cycle hypothesis to consider explicitly the ratio of consumption to income within a modern time series framework. While the aggregate consumption-income ratio tends to be stationary in the long run, counter-cyclical variations in the ratio are an implication of the LCH, because a cyclical decline in income is expected to be temporary, and the ratio of consumption and income will consequently increase in recessions. The important issue is to gauge whether heightened uncertainty, illiquidity, or credit constraints at the onset of the Great Depression disrupted the consumption-income ratio, to result in a catastrophic fall of consumption that was largely independent of changes in income and wealth. Much of the historiography of the consumption collapse investigates particular components of total consumption. The ratios of durable and various categories of non-durable consumption to income may vary in the long run, and the model adopted allows for this possibility. The model for categories of non-durable consumption is formulated as follows:
28 29
ln Ct = 0 + 1 ln Yt + 2 ln FASt + 3 ln LIAt + 4Vt
30 31
+ 5 (ln Ct – 1 – ln Yt – 1) + 6 (ln Ct – 1 – ln At – 1) +
(1)
32 33 34 35 36 37 38 39 40
where C is real per capita consumption, Y is per capita real disposable income, FAS and LIA are households’ per capita financial assets and liabilities, V is uncertainty, A is real per capita net value of household assets and is a disturbance term. This model allows for the dynamic adjustment of consumption to changes in income and wealth, with the error correction terms (ln Ct1ln Yt1), and (ln Ct1ln At1) certifying that consumption gravitates towards its equilibrium relationship with income and net wealth in the long run. A decrease in income, which is anticipated to be temporary, may lead to a temporary rise
4259 Ch02 4/12/02 9:43 am Page 59
Household Balance Sheet in U.S. Depression 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
59
in the C/Y ratio; however, the error-correction term certifies that a constant C/Y is maintained in the long run. For consumption of durables and semi-durables, their prices relative to that of total consumption are added to Eq. (1) and the real interest rate is added to the equation for durables. Hamburger (1967) derived the desired stock of consumer durables as a function of income, wealth, relative prices, and the real rate of interest. Using a simple stock adjustment mechanism he shows durables’ consumption to depend on these variables, and the stock of durables in the previous period. Thus Eq. (1), extended to durables and semi-durables, includes relative prices, and the real interest rate. In addition, in the case of durables, bank suspensions are included to investigate the possibility that financial disintermediation influenced this category of spending.3 In sum, the formulation of the LCH for non-durables specified as Eq. (1), investigates, how, if at all, influences other than income collapse contributed to the consumption decline at the onset of the Great Depression. Including uncertainty and liquidity variables in a life-cycle model provides a natural route to investigating the competing hypotheses of the consumption downturn at the onset of the Great Depression. For the balance sheet variables, A, the real per capita net value of household assets, is decomposed into financial assets and liabilities for the estimation of the empirical models since Mishkin (1978, p. 927) argues that a rise in debt has more powerful effects on consumption than a fall in financial assets. In the case of durables Romer’s uncertainty hypothesis considers whether income uncertainty reduced spending by increasing the value of waiting relative to spending immediately on irreversible purchases. However, the uncertainty hypothesis can be easily extended to non-durables. Hahm and Steigerwald show that an increase in income uncertainty lowers the consumption of non-durables under the assumption that consumers suffer from absolute risk aversion. Thus, a rise in precautionary savings in circumstances of heightened income uncertainty may lead to falls in all types on consumer spending.4 Rather differently, Mishkin (1990) postulates that financial institutions’ uncertainty surrounding their risks led to adverse selection and reduced lending, including to consumers for the purchase of durables. He uses the interest rate spread between a Treasury and a Moody’s Baa corporate bond to indicate information asymmetries between borrowers and lenders. However, the interest rate spread measures the uncertainty associated with fixed investment, default risk, and other factors, and is, therefore, not directly related to the adverse selection and financial disintermediation.5 Accordingly, Bernanke’s (1983a) measure of credit barriers, the deposits of suspended banks, is used in the following empirical sections. 59
4259 Ch02 4/12/02 9:43 am Page 60
60 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
DAVID GREASLEY AND JAKOB B. MADSEN
Olney’s extension of the liquidity hypothesis to non-durables postulates that installment debt had greater effects on their consumption before 1932 because default penalties were severe during these years. In the case of non-durables the simple uncertainty hypothesis maintains that consumers increase their precautionary savings by reducing non-durable consumption in circumstances of heightened income uncertainty. Olney, alternatively, argues that indebtedness has time varying effects depending on the severity of default penalties on installment credit. When penalties were severe, Olney argues consumers will reduce all spending, including that on non-durables, to avoid uncompensated repossession of their durable assets. Default penalties gradually became more severe through the 1930s, but Olney’s results show it was 1938 before installment debt positively influenced consumption at conventional levels of statistical significance. Her findings also show a coefficient for installment debt for the years to 1932 that is statistically significantly different to that for post-1938, although individually the coefficient for the pre-1932 period is not significantly negative. Data for installment debt are only available for years from 1919, which limits the utility of the cointegration-error correction model to test Olney’s avoiding default hypothesis. A variant of Olney’s own life-cycle model, which incorporates installment debt in different periods, and dummy variables to capture variation in the severity of debt default penalties, is adopted, and augmented as Eq. (2) to gauge the robustness of her model to the uncertainty hypothesis:
24 25 26 27 28
ln Ct = 0 + 1 ln Yt + 2 ln At + 3Vt + 4 ln ID2132, t1 + 5 ln ID3840, t1 + 1t
(2)
29 30 31 32 33 34 35 36 37 38 39 40
where ID21–32 is per capita real installment debt over the period from 1921 to 1932, and zero elsewhere, and ID38–40 is per capita real installment debt over the period from 1938 to 1940, and zero elsewhere. This formulation reflects that the severity of default penalties was ambiguous in the years 1933–1937.6 Note that the equation is estimated over the period from 1921 to 1940.7 This variant of the Olney model predicts that 4 < 0 and 5 > 0, and thus importantly that consumption falls with increased installment debt in the previous year, when the price of default is high. The data needed to estimate the augmented LCH-type consumption functions are described and discussed in the next section and the results are reported and discussed in Section IV.
4259 Ch02 4/12/02 9:43 am Page 61
Household Balance Sheet in U.S. Depression 1
61
III. THE DATA
2
Consumer Expenditures
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
The expenditure series used in the consumption function estimation are taken from Lebergott (1996). Lebergott’s data cover over one hundred items, and are consistent with and benchmarked to the U.S. Bureau of Economic Analysis (BEA) data in 1929. Most importantly, and in contrast to Shaw’s (1947) estimates used by Romer, these data cover both commodities and services. Components of Lebergott’s consumption data are allocated here between durables, semi-durables, perishables and services. To illustrate, elements of these data are presented in Table 1 for the period 1928–1933. While durables spending shows the sharpest falls, of around 50%, during the contraction, non-durable (perishables and services) spending accounts for 60% of the fall in total expenditures in the period 1929–1933. Although Mishkin and Romer focus on explaining the decline in durables spending, a satisfactory understanding of the consumption decline during the early years of the Great Depression also needs to account for shifts in non-durables expenditures. Services, according to Lebergott’s estimates, accounted for 51% of consumer spending in 1929. In addition to providing estimates of services’ spending,
21
Table 1.
22 23 24 25 26 27 28 29
Consumer Expenditures ($ billions, 1987 prices).
Year
Perishables
Services
Semi-durables
Durables
Total
1928 1929 1930 1931 1932 1933
164.4 168.3 164.8 164.4 154.1 151.1
265.7 280.2 265.2 254.1 233.6 231.1
59.9 61.7 56.1 53.6 46.9 42.7
42.4 44.6 34.8 30.0 22.5 22.6
532.3 554.8 520.8 501.9 456.9 447.6
30 31 32 33 34 35 36 37 38 39 40
Source: Lebergott (1996, Tables A2–A4). Numbers in parentheses refer to column numbers in Lebergott. Perishables. Total food, alcohol and tobacco (1), wood, gas and coal (34). Services: personal care (15), housing (18), domestic service (36), other household operation (37), medical care, except ophthalmic (38), personal business (44), auto repair (54), gas and oil for motor vehicles (55), auto installation and tolls (56), recreation, except books and maps, durable toys and wheel goods, and music, radio, and TV (59), education (67), electricity (31), gas (32), water (33), telephone and telex (35), religion and welfare (71), for traveling by U.S. residents (75), local transport (57), and intercity transport (58). Semi-durables. Clothing (7), semi-durable house furnish (28), clean and polish (29), stationary (30), and tires and accessories (53). Durables. Furniture and mattresses (24), kitchen appliances (25), china (26), other durable furniture (27), motor vehicles and wagons (52), and the recreation and medical components excluded from services (40), (60), (63), and (64).
61
4259 Ch02 4/12/02 9:43 am Page 62
62 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
DAVID GREASLEY AND JAKOB B. MADSEN
Lebergott adjusts Shaw’s commodity data to allow for changing distributive margins, and provides new estimates for food expenditures. His commodity spending estimates do not adjust Shaw’s output data for fluctuations in inventories. The utility of Lebergott’s data depends on the robustness of his estimates for 33 categories of services. Some of these data are direct estimates, including those for education and foreign travel, but other elements, including domestic and medical services, are based on employment and wages. In similar fashion to Shaw’s construction of commodity output data from occasional census returns, Lebergott interpolates certain categories of services’ spending, including legal services and intercity transportation. The advantage of his data lies principally in their breadth of coverage (they include services, the share of which in total spending grew from 41 to 51% between 1900–1929), and in the consistent benchmarking they provide to post-1929 BEA estimates. The disaggregated consumer spending data show a variety of relationships with disposable income during the early years of the Great Depression. The ratio of durables consumption to income, which had more than doubled in the period 1900–1929, fell around 30% between 1929–1933. Semi-durables consumption in relation to income also fell, by around 4% between 1929–1933, whereas the ratios of perishables and service consumption to income rose respectively 24% and 14%. Romer argues that in periods of heightened income uncertainty consumers switch their spending away from irreversible commitments to buy durables towards spending on perishables. Total consumer spending as a proportion of disposable income, however, rose around 11% over the years 1929–1933. Nevertheless, the decline in consumption may have led to a fall in income since contra-cyclical variations in the consumption-income ratio are an implication of the LCH. For consumers’ income uncertainty or deterioration in their balance sheets to have played a major role in the consumption decline at the onset of the Great Depression, the effects need to have extended beyond durable goods to perishables and services. Heightened uncertainty or balance sheet effects may have played important roles in reducing consumption, even for those categories where the ratio of spending to income rises over the period 1929–1933. Some categories of consumption will show counter-cyclical variation in relation to disposable income, when income decline is expected to be transitory. The central issue to gauge is whether or not the rise in perishable, service, and, indeed, total spending relative to income would have been substantially higher in the absence of heightened income uncertainty or deterioration in balance sheets. The model outlined in the previous section, which articulates both long run relationships between categories of consumption and income, and shorter-term variations in consumption, including those
4259 Ch02 4/12/02 9:43 am Page 63
Household Balance Sheet in U.S. Depression 1 2
63
in response to changes in households’ balance sheets and income uncertainty, permits investigation of the central issue.
3 4
Financial Assets and Liabilities.
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Ando and Brown’s annual series for net wealth, which is based upon Goldsmith’s (1955) benchmarks, provides the starting point for measuring households’ financial assets and liabilities.8 Their aggregate data do not distinguish financial assets, non-financial assets, and liabilities. To disaggregate, a Mishkin-type measure is constructed where financial assets are calculated as a residual by subtracting tangible assets from and adding liabilities to net wealth.9 The measures used here for tangible assets and liabilities do not rest on interpolation, and the resultant financial assets data should better represent annual movements than Mishkin’s data, which are interpolated between 1929, 1933, and 1939 benchmarks. Our approach provides a measure of financial assets corresponding to items included by Goldsmith in the national balance sheet, including life insurance and pensions. In estimating financial assets, liabilities are defined as the sum of consumer mortgage and non-mortgage debt, and tangible assets as the sum of residential and consumer durable wealth. Data for non-mortgage consumer debt are available on an annual basis from 1900, and mortgage debt from 1916.10 The mortgage data are back-projected for earlier years using estimates of the housing stock’s value. Occasional pre-1916 data are available for mortgage debt, and these show a near constant relationship with the value of the housing stock. The ratio of mortgages to the value of residential and non-residential structures is in 1900, 15.98%, in 1912, 15.61%, and in 1922, 14.70% (Goldsmith, Lipsey & Mendelson, 1963).11 For tangible assets, annual data for market values of residential wealth are available, and a series for consumer durable wealth is constructed by applying the perpetual inventory method to the real consumption of durables using a 22% depreciation rate.12 Real consumer wealth is set equal to the average for durable expenditures in the first six years that data are available (1885–1890), multiplied by 1/0.22, which is the average life of a durable at the given depreciation rate. Table 2 compares the growth in our estimates of financial assets, and liabilities, deflated by consumer prices, with Mishkin’s estimates. The disparities highlight the possible weaknesses of using interpolation during a period of macroeconomic instability. The discrepancies are most pronounced for liabilities. Mishkin’s interpolated figures show a near 20% increase in 1930, while our estimate shows an increase of 0.5%.13 Mishkin’s data showing a sharp rise in real liabilities in 1930 appear implausible. Olney reports nominal installment debt fell 12.4% in 63
4259 Ch02 4/12/02 9:43 am Page 64
64 1
DAVID GREASLEY AND JAKOB B. MADSEN
Table 2.
2
Changes in Households’ Real Financial Assets and Liabilities (percent).
3 Financial Assets (Mishkin)
4 5 6 7 8 9 10 11 12
1930 1931 1932 1933
Financial Assets
3.9
1.7
5.6
8.1
7.8
8.7
2.2
6.1
Liabilities (Mishkin) 19.6 1.8 1.9 6.3
Liabilities 0.5 2.9 4.1 6.7
Notes: Mishkin (1978, p. 920) is the source of his data. Consumer prices are used as the deflator. Our measure of financial assets reported here incorporates Lebergott’s (1996) consumer durables data.
13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1930, and consumer prices fell only around 3% in that year. Nominal mortgage debt rose around 2.5% in 1930, which explains why our estimate of real debt rises slightly in 1930. Part of the discrepancy may arise from Mishkin’s use of beginning year values for debt. The empirical models estimated here do include previous year debt, and thus allow for the possibility that sharp rises in debt in 1929 diminished consumption in 1930. Mishkin’s interpolated data for financial assets, although less at variance with our estimates, show values which fall continuously between 1929 and 1932. Our measure shows a rise in the real value of financial assets in 1930. Part of the difference arises from direct measures for tangible assets used in our constructions. Lebergott’s estimates of consumer durables estimates were not available to Mishkin. The direct estimates of the market value of residential wealth show a fall in nominal terms of around 25% between 1929–1933. These data are constructed by adding increments each year of net residential capital formation to an initial year value. This approach should indicate more reliably annual movements than interpolation between benchmarks, although the differences are not great.
31 32
Uncertainty
33 34 35 36 37 38 39 40
Different approaches have been used for measuring consumers’ income uncertainty. Most indicators are based on the volatility of prices (including interest rates and share prices), output, or exchange rates. Wilson, Sylla and Jones (1990) discuss alternative measures of share price volatility. That used by Romer, and deployed here, as a proxy for consumers’ income uncertainty is the standard deviation of monthly share prices, averaged over 12 months to September, and illustrated in Fig. 1.
4259 Ch02 4/12/02 9:43 am Page 65
Household Balance Sheet in U.S. Depression 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
65
Most debate surrounding the use of share price volatility to measure preWorld War Two income uncertainty concerns the low numbers, around 2% of Americans in 1929, holding shares. Romer, though, argues that the stock market can be considered an imperfect predictor of the real economy, citing the views of forecasters in the financial press of 1929. Flacco and Parker’s (1992) econometric estimates of income variance for the 1920s, which they use as a measure of income uncertainty, support her position.14 They conclude that the unprecedented negative shocks to the stock market dominate their statistical measure of income variance in late 1929. Further, they show that the decline of one measure of consumption, department store sales, at the onset of the Great Depression, can be explained substantially by their income variance estimate of uncertainty. Flacco and Parker replicated their analysis of consumption using share price volatility as an alternative to their statistical estimate of income variance, and obtained similar results. They conclude that the stock market volatility measure seems to be a good proxy for income uncertainty during the 1920s. For later years there is supporting statistical evidence that consumers generally are influenced by stock market events. The findings of Otoo (1999), and Madsen and McAleer (2000) for the post-1945 period show that consumer confidence is strongly related to share prices. Most importantly, Otoo found that share price movements influenced consumer confidence among owners and non-owners of shares equally. For the 1920s there are good reasons to believe that the uncertainty spawned by share price volatility affected not only business forecasters and holders of stock but consumers more generally. The public’s attention to the stock market grew strongly during the bull market in the 1920s. This was an inevitable consequence of increased coverage by the press and by the periodical literature. Shiller (2000), for example, shows that the percentage of articles in the Reader’s Guide to the Periodical Literature dealing with the stock market quadrupled during the 1920s. Galbraith (1955) goes further, arguing that by the summer of 1929, only the most aggressive of eccentrics maintained a detachment from the stock market. Consumers, generally, were probably aware that events in the stock market during 1929 were unprecedented. For the years to 1914, Wilson, Sylla and Jones (1990) argue that banking panics and the stock market crashes that accompanied them occurred in the USA every ten to twenty years, and they maintain that these events, though unsettling, were not especially surprising. In contrast, they contend that the stock market crash of 1929, in the era of the Federal Reserve, was different for three reasons. First, the extent of the share price decline in October and November 1929 was unprecedented relative to any other two month decline since 1834. Second the stock market 65
4259 Ch02 4/12/02 9:43 am Page 66
66
DAVID GREASLEY AND JAKOB B. MADSEN
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Fig. 1.
16 17
Real Share Price Volatility.
Source: Romer (1990).
18 19 20 21 22 23 24 25 26 27 28 29
crash was unrelated to a banking panic. And third, stock price volatility persisted, and rose to new heights over the following three years. The stock market crash captured the imagination of the public generally, and its unprecedented characteristics meant its consequences were uncertain, for households and businesses. Figure 1 illustrates the measure of share price volatility used to indicate income uncertainty in the results section. This measure of income uncertainty shows a sharp rise to 0.009 in the year to September 1930, compared to its average pre-1929 value of 0.001, illustrated as the horizontal line in Fig. 1. The stock volatility index falls to 0.006 in the year to September 1931, before peaking at 0.026 in the year to September 1932.
30 31
IV. EMPIRICAL MODELS
32 33
Estimation Results.
34 35 36 37 38 39 40
The section reports results from the life-cycle type consumption functions formulated within a modern time series framework. All variables were first tested for order of integration using augmented Dickey-Fuller tests and the results are reported in Table 3. The table shows that all variables are integrated of order 1 at the 10% level, except share price volatility and the real interest rate, which are stationary variables, and disposable income. The tests for disposable income give
4259 Ch02 4/12/02 9:43 am Page 67
Household Balance Sheet in U.S. Depression 1
Table 3.
67
Augmented Dickey-Fuller Tests for Unit Root.
2 logs, no trend
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Disposable income Financial assets Liabilities Relative price of durables Net wealth Share price variability Installment debt Deposits of suspended banks Real interest rate Consumption, total Consumption, durables Consumption, perishables Consumption, semi-durables Consumption, services
logs, trend
log first difference, no trend
3.22
3.24
3.49
1.63
2.03
6.31
2.23
2.30
3.86
1.14
1.98
2.63
1.57
1.80
6.38
3.54
3.81
+ 0.04 2.29 4.49 2.53 1.21 1.71 2.56 1.39
1.38
2.86
2.13
6.51
4.61
–
–
3.29
2.72
1.33
2.44
1.73
3.66
2.53
4.25
2.51
2.96
Notes: The critical values at the 10% level are: 2.57 in logs with no trend, 3.13 in logs and time-trend included, and 2.57 in log first differences and with no time-trend included. The number of lags in the augmented Dickey-Fuller tests are based on the Akaike Information Criteria. Estimation period: 1902–1939. Sources: See Section III, and for deposits of suspended banks, Historical Statistics, TablesX741–67. The real interest rate is the nominal short-term interest rate minus the expected inflation rate based on prices of durables (see footnote 30 for data sources). The expected inflation is generated as a forecast of an ARMA (2,2) model. The short interest rate is Prime bankers’ acceptances, 90 days, Historical Statistics, Table X449.
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
ambiguous results, and in the estimations below it is treated as an integrated variable. The implication of the stationarity tests is that the models can be formulated in error-correction form, such as Eq. (1). Equation (1) is expanded with one-period lags of all first-difference variables lagged one period and estimated using the general-to-specific estimation strategy where insignificant variables are sequentially deleted. A 10% significance level is used. The unrestricted and restricted estimation results for durables spending are shown in Table 4. The results for other categories, and for total spending shown in Table 5 simply are for the restricted model following the sequential deletion of the statistically insignificant variables. The diagnostic tests do not give evidence against the model specifications. The statistical insignificance of the error correction term implies, for each category of spending, that shifts in consumption were not influenced by the deviation of the ratio of consumption to income from its long-run equilibrium. 67
4259 Ch02 4/12/02 9:43 am Page 68
68
DAVID GREASLEY AND JAKOB B. MADSEN
Table 4.
1 2
Unrestricted and Restricted Parameter Estimates of Durable Consumption.
3 4
ln (Y)t
–0.94 (0.48)
5
ln (Y)t1
–0.46 (0.54)
6
ln (FAS)t
0.28 (1.27)
7
ln (FAS)t1
8
ln (LIA)t
9
ln (LIA)t1
10 11 12 13 14 15 16 17 18
0.51 (2.69) 0.52 (1.20)
Vt
–9.70 (2.02)
Vt1
–8.23 (1.14)
ln (P du /Pt) t
–1.08 (2.36)
/Pt1) ln (P du t1
–0.07 (0.15)
RRt
0.79 (2.01)
RRt1
0.15 (0.40)
ln (SUSt)
0.01 (0.24)
ln (SUSt1)
0.45 (3.18)
–0.29 (1.13) –8.30 (2.39) –1.90 (5.75)
–0.04 (1.58)
ln (Ct1 /Yt1)
0.07 (0.24)
ln (Ct1 /At1)
–0.07 (0.23) –0.12 (0.23)
22
Constant ¯2 R
23
DW
24
B-P
25
RESET
0.58
0.34
26
ARCH
0.23
0.12
19 20 21
27 28 29 30 31 32 33 34 35 36 37 38 39 40
0.65 1.82 16.4
–0.01 (0.40) 0.62 2.20 3.79
Notes: RR is the real interest rate and SUS is deposits in suspended banks. t-statistics are in ¯ 2 = adjusted multiple correlation coefficient. DW = Durbin-Watson statistics for firstparentheses. R order serial correlation. B-P = Breusch-Pagan test for heteroscedasticity, where the squared residual is regressed on original regressors, and is distributed as 2(k) under the null hypothesis of no heteroscedasticity, where k is the number of regressors excluding the constant term. RESET = RESET test for functional form, adding the predicted value squared to the regression, and is distributed as F(1, nk + 2) under the null hypothesis of correct functional form, where k is the number of regressors in the original regression excluding the constant term and n is the number of observations (39). One period lags of the following variables measured in log first differences are used as instruments for ln Yt: per capita real disposable income, manufacturing wages, consumer prices, and M2 deflated by consumer prices. Estimation period is 1902 to 1939. Consumer prices and M2 are from Historical Statistics, Tables E135 (consumer prices) and X 415 (M2). Disposable income is from Lebergott (1996). Manufacturing wages are from Williamson (1995) and for after 1919 from Historical Statistics, Table D802.
4259 Ch02 4/12/02 9:43 am Page 69
Household Balance Sheet in U.S. Depression 1
Table 5.
2 3 5 6 7 8 9 10 11 12 13
Parameter Estimates of Semi- and Non-Durables Consumption Functions. Semi-durables
4 ln Ct1 ln Yt Vt Vt1 ln FASt Constant ¯2 R DW B-P RESET ARCH
69
Services
Perishables
Total
0.43 (3.14) 5.29
(2.64) 6.77 (3.23) 0.26 (2.02) 0.00 (0.49) 0.32 2.04 6.99 0.50 0.73
3.41
(3.16)
0.11 (1.77) 0.00 (0.94) 0.29 1.88 3.11 0.91 0.11
1.35 (4.49) (1.92)
2.64
0.11 (2.08) (1.86) 0.45 1.88 1.86 2.36 0.45
0.01
0.69 (2.67) (3.49) 2.93 (2.42) 4.19
0.00 (0.26) 0.46 1.84 2.10 2.18 0.03
14 15
Notes: as for Table 4.
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
This suggests that a long-run relationship between consumption, income and wealth is not present in the sample period considered here.15 Since the variables are not cointegrated, lnY is not exogenous, and is instrumented, as reported in the notes to Table 4. The estimated coefficient of the uncertainty variable is significant in all the estimates and therefore provides robust support for the uncertainty hypothesis.16 After controlling for the possible effects of shifts in consumption towards equilibrium levels determined by income and wealth, as well as for shorterterm fluctuations in income, consumers’ income uncertainty is shown to significantly affect all categories of their spending. The influence of uncertainty appears strongest in the case of durables consumption, with the size of the estimated coefficient for the effects of the change in current year uncertainty over twice that for total consumption. The findings provide only partial support for the liquidity hypothesis. Changes in households’ holdings of financial assets are shown to statistically significantly affect changes in all categories of consumption other than services, but as is shown below, the quantitative importance of this influence appears modest. Indeed, Mishkin’s own model applies strictly to durables spending, and lays most emphasis on the impact of debt on the downturn on consumer durables spending at the onset of the Great Depression. Our empirical models do not support the hypothesis that changes in households’ liabilities influenced durables consumption. Further, for no categories of spending do the results show households’ debts to significantly influence consumption. The simple liquidity 69
4259 Ch02 4/12/02 9:43 am Page 70
70 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
DAVID GREASLEY AND JAKOB B. MADSEN
hypothesis does not appear robust to the inclusion of income uncertainty within a unified consumption model. Olney, though, refines the liquidity model to take account of variations in the severity of default penalties on installment debt. The results from estimating Eq. (2) are shown in Table 6.17 The estimates show that adding a measure for income uncertainty to an Olney-type model undermines further the liquidity hypothesis. The debt default variant of the liquidity model postulates negative coefficients for lagged installment debt prior to 1932 when penalties were severe, and positive ones from 1938, when they were not. The results do not show statistically significant negative coefficients for pre-1932 installment debt, or significant negative coefficients for 1938–1940, for any category of spending. Indeed, for total spending, higher previous year installment debt before 1932 appears associated with higher spending in the current year. In contrast, the uncertainty hypothesis receives further support from the results in Table 6, which incorporate the effects of changes in income and wealth on consumption, and show for all categories of consumption that heightened income uncertainty led to diminished spending. Olney’s interpretation that a diminished supply of credit was not central to the collapse of consumer durables spending in 1930–1932 receives support from the results in Table 4. The indicator, the deposits of suspended banks, does
21 22
Table 6.
Consumption, Income Uncertainty and Installment Debt.
23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
ln Yt ln (P du /Pt) t ln At Vt Vt1 ID2132, t1 ID3840, t1 Constant R¯ 2 DW B-P RESET ARCH
Durables
Semi-durables
4.55 (3.86) –0.44 (2.98)
0.66 (2.44)
–14.3 (5.41) 0.22 (0.95) –5.28 (2.81) –0.07 (2.42) 0.54 1.60 39.70 19.30 0.01
0.61 –4.47 –7.30 –0.05 0.06 0.01 0.53 1.73 30.7 20.87 5.10
(3.00) (2.15) (3.89) (0.36) (0.49) (0.65)
Services
Perishables
0.92 (3.36)
1.21 (4.59)
–3.66 (3.83) 0.20 (2.55) 0.00 (0.03) 0.00 0.99) 0.39 1.82 39.85 17.02 0.03
0.28 –2.67 –2.45 –0.03 0.03 –0.01 0.63 1.58 37.13 24.67 1.69
(3.19) (3.30) (2.51) (0.58) (0.41) (0.63)
Total
–4.62 (4.42) 0.18 (1.93) –0.01 (0.20) –0.01 (1.36) 0.49 1.51 39.86 16.38 0.03
Notes: See notes to Table 4. Estimation period 1921 to 1940. The t-values are based on White’s heteroscedasticity consistent covariance matrix.
4259 Ch02 4/12/02 9:43 am Page 71
Household Balance Sheet in U.S. Depression 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
71
not exhibit a statistically significant relationship with durables spending. Furthermore, the banks did not supply much installment credit before the great contraction. Their share of the total lending was only around five percent in 1929, and thus this statistical finding appears unsurprising. Moreover, the severity of default penalties (missing a single payment could lead to uncompensated repossession of the durable good by the finance company) diminished the relevance of adverse selection. Olney (1991) notes that finance companies, in stark contrast to the banks, were largely unaffected by bankruptcies.18 That neither debt nor credit restrictions contributed significantly to the collapse of consumer spending at the onset of the Great Depression strengthen greatly the likely central role played by uncertainty surrounding incomes. Finally, the results for durable consumption in Table 4 indicate that the real interest rate does not have the expected negative effect, where the real interest rate is measured as the nominal rate on a short-term government bond minus an ARMA(2,2) forecast of the rate of inflation in prices of durables. The estimated coefficients of the real interest rate were positive. Sensitivity analysis using other measures of the real interest rate did not change this result.19 The estimated coefficients were statistically insignificant and positive.
19 20 21
Simulation Results and Implications
22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
The respective effects of the uncertainty and balance sheet variables on consumption at the onset of the Great Depression are gauged by simulating their estimated effects, utilizing the results in Tables 4 and 5, and focusing especially on 1930 and 1932, the years of largest spending declines.20 For total consumption in 1930, the estimated effect of heightened income uncertainty as measured by share price volatility reduced per capita spending by 3.23%, compared to the actual fall of 7.47%. Since real financial asset depletion did not occur until 1931–1932 (see Table 2), the deterioration on households’ balance sheets had its largest estimated effect in 1932, reducing total consumption per capita in that year by 0.90%. In contrast, income uncertainty is estimated to have reduced total consumption per capita in 1932 by 7.47%, compared to the actual fall of 10.03% in that year. Neither the depletion of households’ financial asset nor income uncertainty had substantial effects in 1931. Overall spending per capita fell 4.47% in that year, with the effects of reduced income estimated to have diminished spending by 4.09%. Income uncertainty was the dominant reason for the falls in total consumption in 1930 and 1932. A comparison with the experience of 1921 helps inform what distinguished the Great Depression. In 1921 when total consumption per 71
4259 Ch02 4/12/02 9:43 am Page 72
72 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
DAVID GREASLEY AND JAKOB B. MADSEN
capita fell by 6.59%, income uncertainty was unimportant. The simulation results show that a depletion of financial assets and the fall in income respectively reduced spending per capita by 1.95 and 3.70% in 1921, while income uncertainty is estimated to have reduced consumption by 0.21%. On these results, what set the onset of Great Depression apart was the rise in consumers’ income uncertainty, which, according to the estimates, explain 43% of the fall in overall consumption per capita in 1930. While uncertainty, rather than deterioration in households’ balance sheets, appears central to understanding the total consumption collapse at the onset of the Great Depression, the continuing fall in consumer spending through 1931 resulted largely from the fall in income. However, a further heightening of income uncertainty in 1932 lies at the heart of the Great Depression’s most substantial falls in consumption, which occurred in that year. The results for both categories of non-durables mirror those for total spending. Thus, income uncertainty is estimated to have reduced services spending per capita by 2.39 and 6.74% in 1930 and 1932 respectively, while the comparable figures for perishables are 1.85 and 5.22%. The depletion of financial assets is estimated to have reduced spending on perishables modestly, by less than 1% in 1931 and 1932, and to have had no effect on services spending. For durables consumption, the position is complicated by the estimated effects of higher real prices for this category of consumption, especially during 1931–1933. In 1930 and 1932 income uncertainty is estimated to have reduced durables spending per capita by 5.84 and 16.45% respectively. Higher real prices for durables are estimated to have reduced their consumption per capita by 1.09% in 1930, and the fall in income reduced durables consumption in the same year by 6.39%. Actual durable spending per capita fell by 15.56% in 1931, and the fall in households’ financial assets in the previous year and rising durables prices are estimated to have respectively reduced spending in 1931 by 3.7 and 9.14%. In contrast, the further heightening of income uncertainty in 1932 is estimated to have reduced durables spending per capita by 16.45% in that year, compared to the actual fall of 29.48%. The results for semi-durables are somewhat distinct from those for other categories of spending. Semi-durables consumption per capita fell steadily throughout the years 1930–1933. For example, the fall of 9.80% in 1933 was almost as severe as the 10.82% fall in this category of spending in 1930. The estimates show that heightened income uncertainty explain 40% of the fall in semi-durable spending per capita in 1930, and between 60–66% of the falls between 1931–1933. The depletion of households’ financial assets contributed to the spending falls for this category of consumption in 1931–1932, but not in 1930 or 1933.
4259 Ch02 4/12/02 9:43 am Page 73
Household Balance Sheet in U.S. Depression 1 2 3 4 5 6 7 8 9 10 11 12
73
Overall, the results in Tables 4 and 5 point clearly to uncertainty rather than illiquidity depressing consumer spending at the onset of the contraction. For all categories of consumers’ expenditures, income uncertainty as measured by share price volatility is shown to reduce spending. Since share price volatility rose most in 1930 and 1932, this measure of uncertainty had greatest effect in these years, although substantial lagged effects, and thus impact in 1931 and 1933, were found in the case of semi-durables. Depletion of households’ financial assets did contribute to the spending falls of 1931–1932, but except in the case of semi-durables, the effect was modest. For durables, the rise in the relative price of this category of spending had a more deleterious effect on consumption, especially between 1931–1933, than the deterioration of households’ balance sheets.
13 14
V. CONCLUSION
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Previous attempts to explain the falls in consumption during the early years of the Great Depression focus separately on the household balance-sheet, especially on the role of debt, the supply of credit, and on uncertainty, and pay most attention to 1930 and to durable spending. This article considers the various hypotheses within a unified consumption model, and gives greater attention to non-durable spending, and to the events of 1931–1933. Our key findings show that income uncertainty precipitated and prolonged the consumption declines at the onset of the Great Depression. Reduced holdings by households of financial assets contributed to the spending declines of 1931–1932, but the magnitude of the adverse balance sheet effect was modest for total spending, and does not contribute to explaining the spending falls of 1930. Nor does excessive indebtedness, or reduced credit supplies, from the evidence of our estimates, explain the consumption downturn at the onset of the Great Depression. Mishkin and Olney have highlighted the role of debt, but neither version of the hypothesis appears robust. The rise in the real value of households’ liabilities, broadly defined, was lower in 1930 than Mishkin contends, and did not significantly influence durables consumption. Nor do the results here support Olney’s view that the unprecedented rise in installment debt, and the severe default penalties prior to 1932, explain most of the consumption collapse in 1930. No difference was found here between the estimated impact of debt on consumption in years to 1932, when default penalties were severe, and after 1938, when they were not, after incorporating the effects of uncertainty. Heightened income uncertainty was chiefly responsible for the collapse in all categories of consumer spending in 1930 and 1932. Uncertainty was less important in 1931. The pace of the consumption decline ameliorated in 1931, 73
4259 Ch02 4/12/02 9:43 am Page 74
74 1 2 3 4 5 6 7 8 9 10 11 12
DAVID GREASLEY AND JAKOB B. MADSEN
and was driven principally by income falls and increases in the real price of durables. The prolongation of the contraction through 1932, though, was associated with a further rise in consumers’ income uncertainty. Our interpretation postulates that consumers increased their precautionary savings by reducing spending at times of uncertainty surrounding their future income. Additionally, for durables, higher prices and the irreversibility of purchases once made, reinforced the effects of a rise in precautionary savings. The force of income uncertainty was in 1930 and 1932 sufficient to account for most of the collapse in total spending, irrespective of the levels of households’ debt. Uncertainty was reduced greatly in 1933. By 1935 consumers’ certainty had recovered to the extent that non-mortgage debt had risen to a greater proportion of average incomes than in 1929, and households’ spending grew strongly.
13 14
NOTES
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
1. Although Mishkin (1990) does not focus explicitly on consumption its content has relevance for the debates surrounding durables spending. 2. In the case of the life cycle model the unexplained falls in total consumption are 9% and 7%, respectively, in 1932 and 1930, while the comparable figures for non durables are 6% and 4%, respectively. 3. Bernanke (1983a, pp. 269–270) utilizes monthly data and makes adjustments for the bank holiday of March 1933. Our annual data do not include a similar adjustment since the deposits of banks which were re-opened by June 1933 are not included in the suspensions data. 4. If the amount of consumption the consumers are willing to give up to avoid a given amount of uncertainty, is an increasing function of wealth, then consumption is a declining function of income uncertainty. That is, the marginal utility of consumption is a convex function of consumption, which implies that the marginal utility of consumption is a declining function of consumption but at a declining rate. The convexity of the marginal utility implies that marginal utility of consumption is higher for the average consumption than the sum of the marginal utilities for various values of consumption. More formally income uncertainty lowers consumption if the third derivative of a von Neumann-Morgenstein utility function is positive. 5. This is shown by Black and Scholes (1973). 6. Courts in several states deemed default penalties unduly harsh during these years, but the legal landscape nationwide was not changed until 1938. See Olney (1999, p. 327). 7. Error correction terms analogous to those in Eq. (1) were incorporated in initial estimates but were found statistically insignificant. In light of the modest sample size, and that Olney’s own model does not include error correction terms, they are not shown in the specification of Eq. (2). 8. Ando and Brown’s data are reported in Modigliani (1966, pp. 215–216). 9. Details of Mishkin’s procedure are in Mishkin (1978, appendix). 10. Non-mortgage consumer debt is from Olney (1991) for years to 1919, and thereafter from Historical statistics, Table X551. For years from 1916 mortgage debt is from Historical statistics, Tables X393–409.
4259 Ch02 4/12/02 9:43 am Page 75
Household Balance Sheet in U.S. Depression 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
75
11. The non-residential structures included here are those defined as part of household wealth, and include, for example, garages. 12. For residential wealth see, Grebler, Blank and Winnick (1956) Consumer durable spending is from Lebergott (1996) for 1900–1941, and Kendrick (1961) for 1885–1899. The consumer durable deflator is from Balke and Gordon (1986) for 1919–1941, for the last quarter of each year, and Historical statistics, Tables P 318–374 for 1889–1918. 13. Mishkin has only four data points for the interwar years, 1922, 1929, 1933, and 1939. He uses a combination of linear and constant growth approaches to interpolation. These methods are unlikely to represent accurately annual movements in households’ assets and liabilities. 14. In the absence of monthly data for disposable income they use industrial production as a proxy. The correspondence between output variance and share price volatility adds credence to the view that stock market volatility can be considered an imperfect predictor of the real economy. 15. The absence of a long-run relationship is likely to reflect the small sample and macroeconomic instability. To test for the possibility that the cointegration tests are sensitive to the choice of estimator, the tests of Johansen were undertaken using three lags of all variables. The null hypothesis of no cointegration vectors could not be rejected at the 10% level. 16. Since the consumption of housing services, which formed around 20% of services spending, cannot easily be varied, the regression for services was also run excluding housing services, but the results of Table 4 were not materially affected. 17. The diagnostic tests indicate heteroscedasticity and functional form problems for all consumer categories and in some cases also the possibility of first order serial correlation, thus suggesting that the Olney-type model is mis-specified. Although White’s heteroscedasticity consistent covariance matrix is used, the parameter estimates may be biased. 18. She reports evidence that only five banks suffered defaults on loans to finance companies in the period 1930–1935. 19. The estimates were based on different measures of the interest rate and different measures of inflation expectations. The estimates included both the short interest rate, as in the estimates above, and the interest rate on a long-term government bond. Inflation expectations were measured as the actual inflation rate, a one-period lag of the inflation rate and an AR(1) forecast of the inflation rate which includes a lagged dependent variable, lagged growth in M2 and the lagged short-term interest rate. Inflation rates were based on prices of durable goods. 20. The simulations are dynamic in the case of services to allow previous year uncertainty via its effect on previous year consumption to influence current year consumption. The results in Table 5 show shifts in lagged consumption are statistically significant only in the case of services spending.
35 36
ACKNOWLEDGMENTS
37 38 39 40
Comments and suggestions by a referee, and by Les Oxley, are gratefully acknowledged. The paper was written while the second author was employed at the University of Western Australia. 75
4259 Ch02 4/12/02 9:43 am Page 76
76 1
DAVID GREASLEY AND JAKOB B. MADSEN
REFERENCES
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Ando, A., & Modigliani, F. (1962). The Life Cycle Hypothesis of Savings: Aggregate Implications and Tests. American Economic Review, 53, 53–84. Balke, N. S., & Gordon, R. J. (1980). Appendix B. Historical Data. In: R. J. Gordon (Ed.), The American Business Cycle: Continuity and Change (pp. 781–850). Chicago: University of Chicago Press. Bernanke, B. S. (1983a). Irreversibility, Uncertainty and Cyclical Investment. Quarterly Journal of Economics, 98, 85–106. Bernanke, B. S. (1983b). Non-Monetary Effects of Financial Crises and the Propagation of the Great Depression. American Economic Review, 73, 257–276. Bernanke, B. S. (1985). Adjustment Costs, Durables, and Aggregate Consumption. Journal of Monetary Economics, 15, 41–68. Black, F., & Scholes, M. (1973). The Price of Options and Corporate Liabilities. Journal of Political Economy, 81, 637–654. Brown, C. (1997). Consumer Credit and the Propensity to Consume: Evidence From 1930. Journal of Post Keynesian Economics, 19, 617–638. Cowles, A., & Associates (1939). Common-stock indexes. Bloomington: Principa Press. Eichengreen, B. (1992). The Origins and Nature of the Great Slump Revisited. Economic History Review (2nd series XLV), 213–239. Engle, R. F. (1982). Autoregressive Regressive Conditional Heteroscedasticity With Estimates of the Variances of United Kingdom Inflations. Econometrica, 50, 987–1007. Ferderer, P., & Zalewski, D. (1994). Uncertainty As a Propagating Factor in the Great Depression. Journal of Economic History, 54, 825–849. Flacco, P., & Parker, R. (1992). Income Uncertainty and the Onset of the Great Depression. Economic Inquiry, 30, 154–171. Friedman, M., & Schwartz, A. J. (1963). A Monetary History of the United States, 1867–1960. Princeton: Princeton University Press. Galbraith, J. K. (1955). The Great Crash. London: Penguin. Goldsmith, R. W. (1955). A Study of Savings in the United States. Princeton: Princeton University Press. Goldsmith, R. W., Lipsey, R. E., & Mendelson, M. (1963). Studies in National Balance Sheet of the United States. Princeton: Princeton University Press. Grebler, L., Blank, D. M., & Winnick, L. (1956). Capital Formation in Residential Real Estate, Trends and Prospects. Princeton: Princeton University Press. Greasley, D., Madsen, J. B., & Oxley, L. (2001). Income Uncertainty and Consumer Spending During the Great Depression. Explorations in Economic History, 38, 225–251. Fearon, P. (1987). War Prosperity and Depression: 1917–1945. Oxford: Philip Allan. Hahm, J.-H., & Steigerwald, D. (1999). Consumption Adjustment Under Time-Varying Income Uncertainty. Review of Economic and Statistics, 81, 32–40. Hall, R. E. (1978). Stochastic Implications of the Life Cycle-Permanent Income Hypothesis: Theory and Evidence. Journal of Political Economy, 86, 971–987. Hamburger, M. (1967). Interest Rates and the Demand for Consumer Durable Goods. American Economic Review, 57, 1131–1153. International Monetary Fund (various years). International Financial Statistics. Washington. Kendrick, J. W. (1961). Productivity Trends in the United States. Princeton: Princeton University Press.
4259 Ch02 4/12/02 9:43 am Page 77
Household Balance Sheet in U.S. Depression 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
77
League of Nations (various years). Monthly Bulletin of Statistics. Geneva. Lebergott, S. (1996). Consumer Expenditures: New Measures and Old Motives. Princeton: Princeton University Press. Leisner, T. (1989). One Hundred Years of Economic Statistics. Oxford: Economist Publications Ltd. Madsen, J. B., & McAleer, M. (2000). Direct Tests of the Permanent Income Hypothesis Under Uncertainty, Inflationary Expectations and Liquidity Constraints. Journal of Macroeconomics, 22, 229–252. Mankiw, N. G. (1985). Consumer Durables and the Real Interest Rate. Review of Economic and Statistics, 62, 353–362. Mayer, T. (1978). Consumption in the Great Depression. Journal of Political Economy, 86, 139–145. Mishkin, F. S. (1976). Illiquidity, Consumer Durables Expenditure, and Monetary Policy. American Economic Review, 66, 642–654. Mishkin, F. S. (1978). The Household Balance Sheet and the Great Depression. Journal of Economic History, 38, 918–937. Mishkin, F. S. (1990). Asymmetric Information and Financial Crises: A Historical Perspective. National Bureau of Economic Research Working Paper No. 3400, Cambridge, MA. Modigliani, F. (1966). The Life Cycle Hypothesis of Saving, the Demand for Wealth and the Supply of Capital. Social Research, 33, 160–217. Olney, M. L. (1991). Buy Now, Pay Later. Chapel Hill: University of North Carolina Press. Olney, M. L. (1999). Avoiding Default: The Role of Credit in the Consumption Collapse of 1930. Quarterly Journal of Economics, 114, 319–335. Otoo, M. W. (1999). Consumer Confidence and the Stock Market. Working Paper, Board of the Governors of the Federal Reserve System. Romer, C. (1990). The Great Crash and the Onset of the Great Depression. Quarterly Journal of Economics, 105, 597–624. Shaw, W. (1947). Value of Commodity Product Since 1869. New York: NBER. Shiller, R. J. (2000). Irrational Exuberance. Princeton: Princeton University Press. Temin, P. (1976). Did Monetary Forces Course the Great Depression? New York: Norton. United States Bureau of Census (1975). Historical Statistics of the United States: From Colonial Times to the Present. Washington: Government Printing Office. Warren, G. F., & Pearson, F. A. (1932). Wholesale Prices in the United States for 135 Years, 1797–1932. New York: Cornell University Agricultural Experiment Station Memoir 142. Wilson, J. W., Sylla, R. E., & Jones, C. P. (1990). Financial Market Panics and Volatility in the Long Run, 1830–1988. In: E. White (Ed.), Crashes and Panics: The Lessons of History. Homewood: Solomon Brothers Center for the Study of Financial Institutions. Williamson, J. G. (1995). The Evolution of Global Labor Markets Since 1830: Background Evidence and Hypothesis. Explorations in Economic History, 32, 141–196.
33 34 35 36 37 38 39 40
77
4259 Ch03 4/12/02 9:44 am Page 79
1 2 3 4 5 6 7 8 9 10
COPING WITH FINANCIAL CATASTROPHE: THE SAN FRANCISCO CLEARINGHOUSE DURING THE EARTHQUAKE OF 1906
11 12 13 14
Ronnie J. Phillips
15 16 17 18
ABSTRACT
19 20 21 22 23 24 25 26 27 28
The assumption today is that the Federal Reserve stands behind the financial system in case of a catastrophic shock. There has been little research on how the payments system functioned during economic catastrophes prior to the establishment of the Federal Reserve System. This paper examines the 1906 San Francisco earthquake when a private sector response was required after disaster occurred. The research question addressed is how well the private sector responded when there was a large external shock to the payments system such as an earthquake. The San Francisco Clearinghouse is examined as a case study.
29 30
I. INTRODUCTION
31 32 33 34 35
Today the bank regulatory structure that includes deposit insurance through the FDIC and a lender of last resort function through the Federal Reserve provides for the ultimate safety and stability of the financial system. In the event of a natural or financial “shock” to the system, it is presumed that these agencies
36 37 38 39 40
Research in Economic History, Volume 21, pages 79–104. Copyright © 2003 by Elsevier Science Ltd. All rights of reproduction in any form reserved. ISBN: 0-7623-0993-8
79
4259 Ch03 4/12/02 9:44 am Page 80
80 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
RONNIE J. PHILLIPS
will enable the financial system to continue to function. Thus, while many people were concerned about the Y2K problem, they were reassured that the Federal Reserve was printing up a few extra billion in Federal Reserve Notes. But prior to the establishment of this regulatory structure, what happened to the financial system when there were shocks such as earthquakes and panics? How did the private sector respond? Did the system sink into chaos, or were normal operations resumed quickly? Though there have been studies of panics under the national banking system, mostly for Chicago and New York, the experience of the other clearinghouses in crisis times has not been recently examined (Dwyer & Gilbert, 1989; Tallman & Moen, 1995; Roberds, 1995). The purpose of this paper is to supplement knowledge of the actions of a clearinghouse during shocks by examining the experience of the San Francisco Clearinghouse during the earthquake of 1906. Coordinated actions by clearinghouses in the U.S. in the nineteenth century represented a private sector solution to the problems inherent in fractional reserve banking. They were attempts to mitigate the effects of weaknesses of the National Banking System and establish a safe and stable banking system. Clearinghouses are an example of Alfred Chandler’s “Visible Hand” whereby the market mechanism is supplanted by management by American businessmen, in this case bankers, in coordinating the activities of the economy (Chandler, 1977, p. 1). The Federal Reserve System, established in 1913, was itself based on clearinghouse experience (Timberlake, 1993). The next two sections of the paper cover the early history of California banking and the role of clearinghouses in the nineteenth and early twentieth centuries. This is followed by a detailed review of the actions of the San Francisco clearinghouse during the earthquake of 1906. The paper concludes with the lessons of history for the operation of the payments system.
28 29 30
II. CALIFORNIA BANKING AND THE SAN FRANCISCO CLEARINGHOUSE
31 32
Early Banking in California
33 34 35 36 37 38 39 40
The California constitution prohibited the state from “granting any charter for banking purposes” (Article IV, Section 34). The banks that operated were private institutions that survived on reputation (Chandler, 1986, p. 3). During the early Gold Rush days, the bankers provided two fundamental services: paying miners in gold coin for the gold dust, and selling the drafts drawn on Eastern banks as a means of sending money through the mail (Chandler, 1986, p. 3; Wright, 1910, p. 15). The essential requirement for taking up banking at the time was
4259 Ch03 4/12/02 9:44 am Page 81
San Francisco Clearinghouse 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
81
apparently to be the owner of a safe in a well-protected building (Wright, 1910, p. 15). Express companies (including Wells Fargo) also collected the gold at the point of production in exchange for local or Eastern bank drafts (Wright, 1910, pp. 15–16). Between 1849–1856, private mints supplied the gold coins until the U.S. Mint in San Francisco was in full operation. Gold coin was the primary medium of exchange in California, and paper money was viewed with suspicion and seldom accepted. Even after the establishment of the National Banking System in 1864, there were initially no national banks established in California. On July 1, 1864, the Bank of California was organized as a joint stock commercial bank (the second of its kind in the U.S.) and reorganized as a national bank shortly thereafter. At the same time, the private bank of Donohoe, Kelly, and Co. was formed (Wright, 1910, pp. 19–20). In 1870, the U.S. Congress authorized the establishment of National Gold Banks that issued currency redeemable in gold. The First National Gold Bank of San Francisco was organized in November 1870 and began issuing banknotes in March 1871 (Wright, 1910, p. 51). Though the notes were issued at face value, brokers across the street from the bank were discounting the notes at an average of 90 cents on the dollar in 1871 (Wright, 1910, p. 88). It was not until January 1, 1879, when Congress resumed convertibility of greenbacks into gold, that paper money circulated at par in California. The first incorporated savings bank in California was the Savings and Loan Society of San Francisco, organized in July 1857 (Wright, 1910, p. 28). The first president of the bank was E. W. Burr, who at the time was the mayor of San Francisco; he retained his position with the bank until 1880. Other savings banks that came into existence around this time were the Hibernia Savings and Loan Society, French Savings and Loan Society, the California Building Loan and Savings Society, and the San Francisco Savings Union (Wright, 1910, p. 32). By 1906, there were 14 savings banks in San Francisco (Wright, 1910, p. 38). Though savings banks had been providing annual financial statements for some time, chartered banks were not required to prepare statements until July 1876. The private banks generally ignored the law (Wright, 1910, p. 75). The Board of Bank Commissioners was authorized by the legislation to begin regular bank examinations and they did so starting in the fall of 1878 (Wright, 1910, p. 108). Of the first five banks examined (all savings banks in San Francisco), four were closed shortly after the examinations. These were followed by voluntary liquidations of at least a half dozen interior savings banks (Wright, 1910, pp. 108–109). There was a net loss of eight savings banks and a net gain of two commercial banks in the first two years of the operation of the new law (Wright, 1910, p. 135). In 1887, legislation required private banks to submit 81
4259 Ch03 4/12/02 9:44 am Page 82
82 1 2 3
RONNIE J. PHILLIPS
semi-annual reports on the same dates as the incorporated banks, but it was not until 1905 that the private banks were placed under the jurisdiction of the Board of Commissioners (Wright, 1910, p. 136).
4 5
The Establishment of the San Francisco Clearinghouse
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
The establishment of the San Francisco Clearinghouse (SFCH) was in part a response to demands by the public for more disclosure of bank conditions in the wake of bank failures in 1875. The SFCH was established on February 4, 1876 with fifteen banks as members. Payments at the SFCH were made in gold coin unless checks were specifically payable in silver or currency (Chandler, 1986, p. 6). On June 1, 1883, the Clearinghouse began using U.S. Treasury Gold Certificates for settlements, though there was a temporary return to gold coin after the Panic of 1893 and certificates were not used again until August 19, 1899. By 1902, 70% of the settlements were in Treasury Gold Certificates (Chandler, 1986, p. 7). During the Panic of 1893, the clearinghouse permitted banks to settle balances with clearinghouse loan certificates. Table 1 indicates the growth of banking in San Francisco over the period 1858–1908. The impact of the Panic of 1907 and the national economic downturn is presumed to account for the decline in total deposits in 1908. Table 2 lists the founding members of the SFCH and the type of charter. At the time, three types of charters were possible: state, national, or foreign. In addition, banks could operate without a charter as private institutions. The growth in payments activity of the SFCH is indicated in Table 3. The decline in 1893 can be attributable to the panic in that year, and the decline in 1908 to the Panic of 1907. Table 1.
29
Deposits of Chartered San Francisco Banks, Selected Years, 1858–1908 (thousands of dollars).
30 31 32 33 34 35 36 37 38 39 40
Year 1858 1868 1878 1888 1898 1906 1907 1908
Savings Banks $20 $17,166 $60,631 $63,154 $105,192 $169,538 $160,965 $147,095
Source: Wright, 1910, pp. 69–71.
State chartered commercial banks ––– ––– $39,806 $25,526 $28,659 $80,875 $101,902 $70,225
Nationally chartered banks
Total
––– ––– $1,331 $3,048 $14,916 $36,954 $52,135 $31,785
$20 $17,166 $101,768.00 $91,728.00 $148,767.00 $287,367.00 $315,002.00 $249,105.00
4259 Ch03 4/12/02 9:44 am Page 83
San Francisco Clearinghouse
Table 2.
1 2
83
Founding Members of the San Francisco Clearinghouse, March 1876.
3 4 5 6 7 8 9 10 11 12 13 14 15
Bank
Type of Charter
The Anglo-Californian Bank, Ltd. The Bank of British Columbia The Bank of British North America The Bank of San Francisco B. Davidson & Co. Ir. Belloc Donohoe, Kelly & Co. Hickox & Spear London and San Francisco Bank, Ltd. The Merchants’ Exchange Bank Sather & Co. Swiss-American Bank Wells Fargo & Co.
Foreign (1873) Foreign (1840) Foreign State Private Private Private (1864) Private Foreign State Private Foreign Private (1852)
16 17
Source: Records of the San Francisco Clearinghouse.
18 19
Table 3.
20 21 22 23 24 25 26 27 28 29 30 31
Annual Clearings of the San Francisco Clearinghouse, 1876–1908, Selected Years.
Year
Clearings
1876 1878 1888 1893 1898 1906 1907 1908
$476,123,238 $715,329,320 $836,735,954 $699,285,778 $813,153,024 $1,998,400,779 $2,133,883,626 $1,757,141,850
Source: Wright 1910, pg. 123.
32 33 34 35 36 37 38 39 40
In 1905, the California legislature passed a provision in the banking laws establishing minimum levels of capital for banks graded by the population of the community. This provision was ruled unconstitutional and was repealed in 1907 (Wright, 1910, p. 141). Instead, a provision was passed that required that the capital should be at least 10% of deposits, with an upper limit of paid-up capital of $1,000,000, and no bank could be organized with less than $25,000 capital (Wright, 1910, p. 141). 83
4259 Ch03 4/12/02 9:44 am Page 84
84 1
RONNIE J. PHILLIPS
III. THE ROLE OF CLEARINGHOUSES
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Though government involvement in the payments system has been regarded as the norm, the theory and practice of self-regulated payments systems has been a topic of recent research (Gorton, 1989; Gorton & Mullineaux, 1987; Selgin & White, 1994). In their study of the Suffolk banking system, Charles Calomiris and Charles Kahn (1996) present the arguments in favor of private sector payments systems vs. government mandated systems. The potential benefits, they note, are as a means to provide stability during periods of financial stress and as a means to clear payments during normal times (Calomiris & Kahn, 1996, p. 772; McAndrews & Roberds, 1995). Because cooperative systems involve coinsurance, they reduce the likelihood and costs of financial crises. In addition, interbank coordination reduces transactions and information costs during normal times. The cooperative arrangements reduce the costs of using bank liabilities as a means of payment and increase the demand for banknotes. Bankers, because of the information specialization that arises with financial intermediation, are able to determine the optimal allocation of investment funds at their disposal. However, an incentive problem arises between the banker and the depositors. As Calomiris and Kahn (1991) have shown, a demandable debt contract with a “first come, first served” rule can provide a solution to this problem (Calomiris & Kahn, 1996, p. 773). A cooperative arrangement whereby one bank places deposits in another bank helps the standardization of products, services, and practices. It also provides an enhanced mechanism for its liability holders to run the bank. Further, private regulation reduces the costs of monitoring one another, and membership in a cooperative arrangement provides a credible signal to uniformed depositors about a bank’s quality (Calomiris & Kahn, 1996, p. 773). Though government intervention could provide these advantages, with private sector solutions the costs are paid by the private sector. Because of incentive problems of government regulators, it may also be the case that private arrangements may be able to act more quickly in preventing fraud and other problems that could threaten the entire arrangement (Calomiris & Kahn, 1996, p. 774). Clearinghouses were private sector solutions that arose in the nineteenth century and served three important functions: (1) settlement of payments; (2) regulatory body; and (3) lender of last resort (U.S. National Monetary Commission, 1910a). These three functions incorporated a mix of competitive and anti-competitive actions, and, consequently were encumbered with mixed costs and benefits. The first and foremost role of a clearinghouse was the multilateral netting of payments. As stated in Article II of the Constitution of the San Francisco Clearinghouse:
4259 Ch03 4/12/02 9:44 am Page 85
San Francisco Clearinghouse 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
85
The objects of this Association shall be the effecting at one place of the daily exchanges between the several associated Banks, and the payment at the same place of the balances resulting from such exchanges.
The clearinghouse was responsible only for those balances actually paid into the hands of the clearinghouse manager. In the event of losses of balances while in the hands of the manager, for which he could not be held responsible, these were borne and paid by the members of the association in an equal proportion. Clearly, to the extent that the settlement of payments is the primary function of the clearinghouse, then the agreement to meet at a central location, and at a certain time, enables all banks to minimize their costs of settlement. The need for personnel to make individual settlements with each bank is reduced. As long as there is never a question of final settlement, then there is no risk to the clearinghouse. There is an incentive for the clearinghouse association to allow only those institutions that are able to make settlements good with regularity. In order to reduce costs for individual settlements further, the clearinghouse would require that each member bank place on deposit with the clearinghouse a certain amount of the final means of settlement. In the 19th century this was primarily gold coin, but could also be silver coin, various types of government currency, and possibly national banknotes. A clearinghouse serving to reduce the transactions costs of payments of individual banks also reduces the overall social costs of operating the payments system. If this was the sole function of clearinghouses, then they would have undeniably served a function that was consistent with a cooperative private sector solution to the costs of operating the payments system in a competitive market economy. To this extent, the clearinghouses were beneficial to the overall economy. However, clearinghouses did not limit their activities to merely reducing the social costs of the payments system. Rather, clearinghouses were a means to regulate competition in the industry since clearinghouse membership in some sense represented a “seal of approval” for banking firms from a body that represented the most prominent banks in the city. Clearinghouses went further in many instances by imposing minimum fees, interest rate ceilings, reserve requirements, examinations, and so on, which were clearly attempts to restrain competition. In this respect, the actions of clearinghouses may not have been consistent with the promotion of a competitive market economy. After all, in a market economy, others besides banks would be better at assessing the viability and performance of a member bank. The final function of a clearinghouse was as a lender of last resort during a financial panic. It is this function, together with the regulatory function, that Gorton and Mullineaux have in mind when they refer to the clearinghouse’s 85
4259 Ch03 4/12/02 9:44 am Page 86
86 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
RONNIE J. PHILLIPS
role as the “joint production of confidence.” (Gorton & Mullineaux, 1987). The fundamental question is why firms that are in competition with each other would voluntarily enter into the joint production of confidence. For Gorton and Mullineaux, the answer lies in the problem of the distribution of asymmetric information in the banking industry. The problem lay in the mix of demand deposits and banknotes that came to predominate in the late nineteenth century. Gorton and Mullineaux argue that the capacity of the market to monitor and control banks eroded as demand deposits came to supplant bank notes in the nineteenth century (Gorton & Mullineaux, 1987, p. 458). Hierarchical organization and control replaced the market mechanism as Coase (1937), Williamson (1975), Chandler (1977) and Stiglitz (1985) have argued. In periods of financial panic, the clearinghouses began to issue clearinghouse loan certificates (CHLC). Because each member bank had made a contribution in the final means of payment (gold for the most part), the clearinghouse could effectively pool these reserves and loan them to members experiencing problems. This action brings attention to the fundamental contradiction in the clearinghouse arrangement. Banks that were not experiencing problems had no incentive to loan to their competitors to keep them in business. Indeed, the problems of competitors should have been welcomed since it would have meant that depositors would shift to the banks viewed as better operated and managed. However, there is the problem of contagion – a run on one bank leading to runs on other banks. This would arise if depositors did not have complete information about the solvency of other banks. Hence, solvent banks might have had an incentive to help out banks in trouble if they feared a contagion. The fear of contagion provided the rationale for the clearinghouse to lend the pooled reserves to a member in trouble and it did this through the issuing of clearinghouse loan certificates. The bank in trouble pledged collateral (securities of some sort) and the clearinghouse advanced a discounted amount of clearinghouse loan certificates. The bank paid out the CHLC to depositors who wished to withdraw. The holder of the CHLC could then, if he/she wished, present the CHLC for payment in gold to the clearinghouse. Alternatively, the CHLC could function like any other paper currency. However, the issuing of CHLC to be used as currency was not legal until the passage of the AldrichVreeland Act in 1908. Numerous studies have been made of the issuing of CHLC and in general the conclusion is that the system worked sometimes and at other times it did not. The role of the Treasury was often important in the panics (Timberlake, 1993; Dwyer & Gilbert, 1989). The Federal Reserve Act provided the established clearinghouses with competition from the Fed – which was able to subsidize check clearing, unlike the existing clearinghouses (Spahr, 1926;
4259 Ch03 4/12/02 9:44 am Page 87
San Francisco Clearinghouse 1 2 3 4 5 6 7 8 9
87
Stevens, 1996; Stevens, 1998; Gilbert, 2000). In summary, the development of clearinghouses represented an economic organization formed by economic agents (bankers) in order to counter the impact of unrestricted competition. As Gorton notes with regard to clearinghouses: That such an economic entity should have endogenously arisen in the banking industry suggests important links between the characteristics of the product and institutional and contractual forms of economic organization. While much work remains to be done on these links, the existence of the clearinghouse suggests that private agents can creatively respond to market failure (Gorton, 1985, p. 283).
10 11 12
IV. THE EARTHQUAKE OF 1906
13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
At 5:12am on Wednesday, April 18, 1906, an earthquake with a magnitude of 8.3–8.6 on the Richter scale struck San Francisco. The earthquake and fires that burned until Saturday, the 21st, destroyed 500 city blocks, 25,000 buildings, left 250,000 homeless, killed 450–700, with total damages in 1906 dollars at between $350 and $500 million. The destruction included the financial district of San Francisco and most bank buildings (Thomas & Witts, 1971). By early in the morning, San Francisco Mayor Eugene Schmitz had issued a proclamation in which he announced that federal troops, policemen, and special officers had the authority to kill any persons found looting or involved in other criminal activity. All gas and electricity was cut off indefinitely (to avoid spreading fires), all citizens were to remain in their homes, and the Mayor appointed a Citizen’s Relief Committee of Fifty (Muscatine, 1975, p. 433). Much has been written on events of the earthquake, some of which is pertinent to the impact on banking and the payments system. The earthquake damage itself was extensive, but the fact that water mains were broken and fires had been started from busted gas mains, meant that those who survived the earthquake faced the prospect of losing their homes and businesses to fires. In addition, some fires were intentionally set and some buildings dynamited in order to help contain the fires. Thus, the earthquake did not level the banks for the most part, but the subsequent fires destroyed, or heavily damaged, the buildings of virtually all of the banks in the city. The fact that the initial earthquake, though creating panic and chaos in the city, did not destroy all the banks (or even very many), meant that one could have hours or days in which to make plans. One banker who was neither a member of the Mayor’s Committee of Fifty nor the SFCH, but who did respond quickly was A. P. Giannini. 87
4259 Ch03 4/12/02 9:44 am Page 88
88 1
RONNIE J. PHILLIPS
A. P. Giannini and the Bank of Italy
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
When his father-in-law passed away in 1902, A. P. Giannini, who had been a successful entrepreneur and businessman, assumed his position as a director of a bank, the Columbus Savings and Loan Society, a small bank catering almost exclusively to Italian-Americans in the community of North Beach in San Francisco (Nash, 1992, pp. 22–23). After disagreements with other members of the Board of Directors, Giannini established his own bank, the Bank of Italy, in 1904. When the earthquake struck, the Bank of Italy was a small but growing and successful bank. Giannini was not in San Francisco when the earthquake struck, but managed to arrive shortly thereafter. He realized that, even though the initial earthquake had not destroyed his bank building, it was quite likely that it would be consumed by fire within a short period of time. When Giannini arrived at the bank around noon, he found that his Assistant Cashier and clerk were standing at their desks. Both men had arrived at the bank early in the morning and, finding the building intact, decided to open for business as usual. They went to the Crocker Woolworth National Bank where they collected three heavy canvas bags which held over $80,000 in gold and silver coins – every cent of hard cash held by the Bank of Italy (Thomas & Witts, 1971, p. 118). They then went back to the bank and opened for business, though by 11 AM they were apparently questioning the wisdom of doing so. When Giannini arrived he reportedly told them “only an insane man would open a bank on a day like this. You should have left the money at Crocker. At least it would have been safe” (Thomas & Witts, 1971, p. 120). By 5 p.m. the day of the earthquake, Giannini and his staff loaded the bank’s available cash and $80,000 in gold, along with bank records, onto a wagon and proceeded to make their way to Giannini’s home in San Mateo. Before departing, he parked his wagons at his brother-in-law’s home near Francisco and Jones streets at the far end of the North Beach Area (Nash, 1992, p. 32). Around 8 PM, Giannini began his trip to San Mateo, and arrived there – amid great panic and chaos – in the early morning hours of April 19. If Giannini could get his cash out, what did the other bankers do with the time between the earthquake and fire? At the Crocker Bank on the morning of the 18th, customers brought ledgers, records, and money for storage in the vaults. However, William Crocker, the bank’s president, had already hired a boat to take all the bank records out into the center of the Bay – “and just sit there” (Thomas & Witts, 1971, p. 119). At the Anglo-California Bank, tellers carted negotiable bonds in a wheelbarrow down Market Street to the Ferry building for shipment across the Bay to Oakland. In three hours they shifted a
4259 Ch03 4/12/02 9:44 am Page 89
San Francisco Clearinghouse 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
89
million dollars worth of unregistered bonds (Thomas & Witts, 1971, p. 119). At the London, Paris, and American Bank, sacks of U.S. registered mail were placed in the vault along with deposits of jewel boxes and family deeds (Thomas & Witts, 1971, p. 119). On Friday the 20th, Giannini returned to San Francisco where he learned that the Bank of Italy building had indeed been burned to the ground. The bank’s safe had become a lump of molten iron. On Saturday, April 21st, Giannini attended a meeting of businessmen near the waterfront. In response to the view that the banks would not be able to open until November, he responded: Gentlemen, to follow the course you are suggesting will be a vital mistake . . .. We’ve got to fight our way out of this spot. If you keep your banks closed until November . . . there will be no city or people left to serve. Today is the time they need you. The time for doing business is right now. Tomorrow morning I am putting a desk on Washington Street wharf with a Bank of Italy sign over it. Any man who wants to rebuild San Francisco can come there and get as much cash as he needs to do it (quoted in Nash, 1992, p. 33).
On Sunday, the 22nd, just four days after the earthquake, Giannini opened for business, bringing $10,000 in cash from his San Mateo hoard. He took out an ad in the San Francisco Chronicle and announced that he was open for business (Nash, 1992, p. 34). Giannini actively began to make loans, and his actions after the earthquake made him something of a local hero, and of course were great publicity for his bank. The actions of the SFCH can be contrasted with those of an individual banker who was not bound by the agreements reached by the clearinghouse members. Indeed, Giannini was viewed as an outsider, “someone to be watched, possibly, but never to be admitted to that intimate circle dominated by the American National Bank and Wells Fargo” (Thomas & Witts, 1971, p. 43). Giannini’s bold decision to reopen so quickly boosted his image in the community and aided the growth of his bank in later years. The earthquake also provided the rationale for branch banking, an area where Giannini was very aggressive. Some bankers removed their gold and cash while others left everything in the vaults. Removing the gold and cash was an enormous risk both from the threat of robbery and the danger that it would provoke a further banking panic. Giannini was bold and lucky and reaped the benefits in subsequent years as his bank grew in assets and reputation.
36 37
The Saving of the U.S. Mint
38 39 40
On the morning of the 18th, rumors spread that an armed gang was massing to storm the U.S. Mint, which contained $300 million in gold coin and bullion. 89
4259 Ch03 4/12/02 9:44 am Page 90
90 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
RONNIE J. PHILLIPS
This was indeed a rumor, but in any event a company of troops from the Sixth Infantry took up defensive positions on the roof that gave them an excellent vantage point for firing upon intruders (Thomas & Witts, 1971, p. 127). The Mint in San Francisco was established in April 1854, and the building at Fifth and Mission was a monolith of granite and sandstone blocks and suffered only minor damage in the earthquake. However, as the fire approached, the troops closed and bolted the heavy iron shutters across the ground-floor windows. The tarred roof was ripped away with iron bars and pick shovels, then buckets of blue vitriol were hoisted up from the refinery and mopped over the exposed roof beams. Though the Mint had its own water supply, the quake had broken the pump; but it was repaired so that it could be hand-operated. The Mint was saved during the fire due to its construction, the distance of the façade from the street, and an artesian well on the premises, and employees who helped fight the fire (Wright, 1910, pp. 164–165). The Sub-Treasury building on Commercial street, which had at one time been occupied by the Mint, was destroyed in the 1906 earthquake and fire. The contents of the vault were not damaged, however (Wright, 1910, p. 169).
18 19
The Response of the San Francisco Clearinghouse(1)
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
There was a brief meeting of the SFCH at 9 AM on the day of the earthquake where it was decided that the banks should remain closed. By the evening of that day, virtually all of the bank buildings had been destroyed. This meeting was followed by two brief meetings on April 20 and 21, for which minutes do not exist. The first formal meeting of the San Francisco clearinghouse banks was called for Monday, April 23, meeting at the home of Mrs. Elenor Martin, on the corner of Broadway and Buchanan. The group passed a resolution asking that the governor call a special session of the legislature and declare legal holidays on a day to day basis until the legislature enacted legislation. A special meeting of the SFCH was held at noon in the same location. The sense of the meeting was that the banking problems of the populace be resolved as quickly as possible in temporary quarters either in a large building on a public square or at private residences of a bank officer. At a meeting on April 24th, the SFCH created an Emergency Finance Committee (EFC). The task of the EFC was to prepare a “plan for meeting the pressing needs of Depositors through limited Payments on account, prior to opening our vaults.” The plan consisted of the use of the U.S. Mint’s gold coin to make payments to the depositors of San Francisco banks. Specifically,
38 39 40
(1) Arrangements to be made with Superintendent Leach of the U.S. Branch Mint, that he will handle these payments on behalf of the banks.
4259 Ch03 4/12/02 9:44 am Page 91
San Francisco Clearinghouse 1 2 3 4 5 6 7 8 9 10 11 12 13 14
91
(2) Each bank to supply Superintendent Leach with coin to such extent as it may desire for the purpose. (3) Each Bank to have quarters where it will meet its customers, and, at its discretion, mark their checks – “Good at the U.S. Mint,” or words to that effect. (4) Superintendent Leach, through a clerical force to be furnished by the Clearinghouse, to honor these checks to the extent of the balance of each of the several banks. (5) San Francisco non-member banks may join the plan by depositing money for the purpose, or may specially act through members. (6) The amounts to be paid as above by the Banks should be absolutely in the discretion of each bank, provided that no payment of more than Five Hundred (500) Dollars should be made on account of any customer until further action by the SFCH.
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
The report was unanimously adopted and a committee of nine was appointed to prepare a plan to enable all banks to open and to provide the necessary funds to construct the new San Francisco. By April 25, most of the banks had established temporary offices in residences in the Western Addition of the city. Thus branch banking came to California. At the meeting on Wednesday, April 25, just a week after the earthquake, the Vice President of National City Bank in New York assured San Francisco bankers that National City “intends to extend to you its resources and influence.” The Committee on Protection recommended that each bank make arrangements for the patrolling of its own premises and no joint action of the SFCH Banks would be necessary. Though cooperating in the reopening of the banks, there remained an element of competition since it was known that some banks were in better shape physically and financially than others. The votes for cooperation were not unanimous, but evidently no individual clearinghouse bank broke the agreements in order to gain a competitive advantage. At the April 26 meeting, representative of interior banks in California assured the San Francisco banks that their banknotes “would be treated liberally until the vaults could be opened.” On April 27 Gov. Pardee notified the SFCH that he didn’t want to be hasty about convening the legislature, noting “God save us from hasty and vicious legislation.” At this meeting, it was revealed that the vaults of the following banks were deemed entirely safe:
37 38 39 40
American National Bank; * Bank of California; * Crocker Woolworth National Bank; 91
4259 Ch03 4/12/02 9:44 am Page 92
92 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
RONNIE J. PHILLIPS
Canadian Bank of Commerce; Central Trust Co; * Donohue-Kelly Banking Co; French-American Bank; German Savings Bank; Mercantile Trust Co; Mutual Savings Bank; National Bank of the Pacific; San Francisco Savings Union; Savings and Loan Society; Security Savings Bank; Union Trust Co.; * Wells Fargo-Nevada National Bank. * SFCH members Of course, the claim that the contents of the vaults were ok was relayed confidently to the public, but there was no way to know for sure if the vaults could be opened, or if, once opened, the contents would be all right. Only later was it determined that no bank had lost items stored in the vaults. If the doors to the vaults were too hot to open, it was also obvious that the contents inside would at least be a little warm. In the case of a fire in Baltimore in 1904, the contents of many safes and vaults ignited spontaneously when opened before their interiors cooled. It was also the case that brick vaults would be better able to survive the fire than metal safes (Douty, 1977, p. 169). At the April 28th meeting, a plan was presented to open up under the head of “SF Clearinghouse Bank” at the U.S. Branch Mint, between the hours of 11am and 3pm until further notice, and make small advances to various clients through the medium of a form of promissory note, which was approved by the bank. It was understood that the total amount advanced to each client would not exceed 500 dollars. This plan met with the approval of the members present. A report was presented on the country banks. The conclusions and recommendations were:
33 34 35 36 37 38 39 40
(1) open country banks ASAP, many already open; (2) banks are protected by bank holiday declared by the governor; (3) the banks in each locality should cooperate. “Any tendency for individual banks to reopen, taking advantages of their relative strength of position to gain prestige in their communities would be destructive to public confidence and would finally react on such banks themselves. In SF the situation is well under control, because the Banks
4259 Ch03 4/12/02 9:44 am Page 93
San Francisco Clearinghouse 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
93
are taking action in concert. Besides their general measures for resumption, they are not arranging a plan to make limited cash advances to meet the personal needs of depositors. Some such course has been followed in Oakland, Sacramento and elsewhere.” The advances were to be limited so as to be used for necessities – $20, $50 and $100. (emphasis added). (4) We should all remember that we are not facing a money or credit panic; but a problem forced upon us by this unlooked for catastrophe; that business conditions are sound; the prospects assuring, and that there is no good reason why any solvent enterprise should suffer; except for the short, unavoidable delay required for readjustment; (emphasis added). (5) We therefore recommend that the country banks in each locality discourage independent action, and work in strict cooperation with each other; while their ultimate aim is to reopen as soon as possible they should proceed cautiously and with due regards to their available cash resources, and the attitude and demands of their customers; meanwhile, they may take full advantage of the legal holidays proclaimed by the Governor, although it may be wise to meet the pressing personal needs of depositors by small advances; that they should keep in mind the essential soundness of business conditions and discountenance apprehension, in exaggeration of the emergency, in themselves no less than in their customers (emphasis added). The Clearinghouse was clearly worried that the competitive and independent spirit might lead some banks to open early. In squashing the competitive impulse, the Clearinghouse invoked the fear of bank contagion. This is despite the fact that they also noted that the present catastrophe was not a financial panic – implying that a panic situation would be worse! In the face of disaster, the response was clearly that competition was bad and cooperation was good. This was despite the fact that the Bank of Italy was open for business – though admittedly, with no more than $80,000 in gold coin. At the April 30th meeting of the SFCH, a telegram of support from the Chicago Clearinghouse sent on the 18th was read. The Committee on Vaults did not inspect any more vaults since the banks were guarding their own. The Insurance Committee was meeting with the insurance companies. The SFCH was to open at the Mint on May 1st. At the May 1st meeting, plans were presented for new business with depositors: (1) Deposits: Clients deposit money and checks with Banks Money, i.e. coin and currency to be immediately available. Checks on San Francisco to be entered for collection unless marked by drawee banks “SFCH Bank Please pay” Checks on outside points to be available when collected. 93
4259 Ch03 4/12/02 9:44 am Page 94
94 1 2 3 4 5 6
RONNIE J. PHILLIPS
(2) Checks: Checks against funds available as above to be presented to respective bank and marked “SF CH Bank please pay” Checks against these funds when presented by other SF banks to be paid by check on “SFCH Bank” Later if necessary a method of clearing these checks and settling balances may be devised. No bank to use this system to evade, directly or indirectly, the agreement to withhold payment on balances existing before the fire.
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
Thus, a depositor would go to the temporary office of his bank, give a promissory note to the bank, which the bank then endorses for payment, and the depositor then goes to the Mint to receive payment. The SFCH began operating through the Mint on May 1st and continued until the banks reopened three weeks later. Temporary offices for all banks had been established and on May 2nd the plan to open the banks was ready. On May 3rd there were reports from the commercial banks’ on their readiness to reopen and on May 7th for Savings Banks. Table 4 gives the banks dates when they could be expected to open with business as usual. Several SFCH members had already begun some banking services. By May 2, Crocker-Woolworth National Bank, the Central Trust Company, and the Mercantile Trust Company had opened tented offices near the camps of refugees near Lafayette Square (American Banker, May 3, 1906). The city government also paid April salaries on the 2nd with a deposit of $2,000,000 in cash at the Western National Bank. At the May 4 meeting there was concern about various plans that had emerged from some quarters regarding the rebuilding of the city, some of which the Executive Committee of the SFCH believed were “ill-advised.” The Report of the Executive Committee recognized that there was a need for financial aid for the destitute, and that assistance “from a distance will be necessary.” However, with regard to financing the rebuilding of the city, the Report stated that with $150,000,000 from the insurance companies, the banks in a strong solvent position, a bountiful harvest promised in the state, and the general underlying soundness of business, help from the outside would not be essential. The bankers were thus optimistic on the possibilities for financing the rebuilding of the city:
33 34 35 36 37 38 39 40
In other words, since the insurance indemnity will largely replace the destroyed buildings, as well as stocks of merchandise, and our banks will be able to meet general commercial requirements, individuals or institutions who may come to find it necessary or expedient to apply to outside sources for loans, should do so, as occasion may arise, on the usual basis of good security. We cannot rebuild in a day. We shall shortly have more money than can be immediately used; so it seems premature to assume that our resources will prove inadequate, and especially that the occasion demands the introduction of untried methods of finance which may be found to be illegal or otherwise impracticable (emphasis added).
4259 Ch03 4/12/02 9:44 am Page 95
San Francisco Clearinghouse 1
Table 4.
95
Dates Banks Expected to Open.
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Canadian Bank of Commerce International Banking Corporation Crocker Woolworth National Bank Bank of British North America Security Savings Mutual Savings Bank San Francisco National Bank First National Bank Donohue-Kelly Banking Co National Bank of the Pacific French Savings Bank American National Bank Bank of California Central Trust Company Anglo-California Bank London Paris and American Bank Ltd Savings and Loan Society San Francisco Savings Union German Savings and Loan Society Italian American Bank Wells Fargo-Nevada National Bank
May 3rd May 3rd May 13th May 13th May 14th May 14th May 17th May 17th May 17th May 17th May 17th May 18th May 24th May 24th May 24th May 24th June 1st June 1st June 1st uncertain uncertain
20 21
Source: Records of the San Francisco Clearinghouse.
22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
The Report further stated that a “newly awakened spirit of co-operation” had emerged among the citizens after the fire and there was the hope that “A realization of interests in common has been developed, which we can all wish should continue and increase.” This report was from a group of bankers who were concerned that funding from the outside would undermine the profitable lending opportunities that would come to them from rebuilding the city.2 On May 7th, the safe-deposit vaults of the Union Trust Company of San Francisco, the Crocker-Woolworth National Bank and the Mercantile Trust Company of San Francisco formally reopened for business. During the period from the 7th to the 19th, business was done through special accounts of these banks. Also on the 7th, the first SFCH clearings occurred at the Mint location. These balances were from the special accounts that arose from the Mint-SFCH arrangement, the regular checks from April 17 had yet to be cleared. The dollar value of the special accounts cleared between May 7 and 21 was $3.7 million. At the May 7 meeting there was continued discussion of a date to reopen the banks and on May 8 there was discussion to set the date of reopening, though some were not ready, and it was resolved that the date not be later than 95
4259 Ch03 4/12/02 9:44 am Page 96
96 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
RONNIE J. PHILLIPS
June 1. At the May 11 meeting, it was passed that beginning May 12 there should be weekly reports made to the clearinghouse on bank deposits. At the May 16 meeting the Committee recommended May 23 as opening day for banks. Balances were to be cleared among clearinghouse banks on May 19. Checks that were deposited on April 17 were to be cleared on Sat May 19 at 10 AM. Balances between April 17 and May 21 to be cleared on May 21 at 10 AM. Thus, it took just under five weeks to reopen the banks for full business. By May 23, all but four of the SFCH banks were open in their old offices, which had been temporarily refitted. The four that moved were: San Francisco National Bank (two blocks away), Wells Fargo-Nevada (which moved to the Union Trust Building three blocks away), the International Banking Corporation (opening in the Western Addition at Fillmore and Sutter), and the Italian-American Bank (a block and a half from its original location). The savings banks as well reopened in their own offices, except for the San Francisco Savings Union whose building and office were totally destroyed.3 By May 19, the SFCH had relocated to 464 California Street. On June 6, the SFCH received a letter from Senator W. C. Ralston chairman of the Banking Committee regarding Senate Bill No. 33. The SFCH Association resolved that the SFCH President notify Senator Ralston to the effect that “it is the opinion of this Association that no emergency exists which warrants the consideration of such a measure at a special session of the Legislature, and that such amendment should be considered in the regular course of ordinary legislation.” The SFCH saw no need for further special legislation. However, problems were beginning to arise with the financial plan envisioned by the bankers when, by July 12, some insurance companies were refusing to pay. The Policy Holders League was subsequently formed. Ultimately, the insurance companies paid $229 million, with several insurance companies paying off in full, some failing and only paying off partly, and several German firms totally refusing to pay (Thomas & Witts, 1971, p. 271; Bronson, 1959, p. 111).
32 33
The Associated Savings Banks and the Earthquake of 1906
34 35 36 37 38 39 40
Savings banks, though important in the financial system, served a different function from commercial banks. Savings banks held long term assets (home mortgages) and short-term liabilities (passbook savings accounts). The Associated Savings Banks of San Francisco came into existence July 13, 1906, nearly three months after the earthquake, though they operated as an unofficial association during the catastrophe.4 There was on April 23, 1906, a meeting of
4259 Ch03 4/12/02 9:44 am Page 97
San Francisco Clearinghouse 1 2
97
the various savings banks at the home of Mrs. Elenor Martin, Broadway and Laguna Streets. On April 24th meeting the items for discussion were:
3 4 5 6 7 8 9 10 11
(1) (2) (3) (4)
when to open; amount to pay out; form of the notice to be posted in front of each bank for info; advisability of making arrangements to rent all Safe Deposit boxes in the City to prevent the public from withdrawing their money from the banks and hoarding same. (5) that no report of our meeting be given to the daily papers unless same was in writing and signed by the proper parties.
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
The sign on banks to read: THE VAULTS OF THIS BANK ARE IN PERFECT CONDITION AND DATE OF OPENING WILL BE ANNOUNCED LATER BY THE BOARD OF DIRECTORS. They agreed to request that the commercial banks not open until all savings banks could open. There was discussion of the rumor that the commercial banks would reopen the first of the week. Clearly, then the primary concern of the savings banks was that the commercial banks would reopen before the savings banks and this would lead to an erosion of confidence in the savings banks and a contagious run. It is interesting that the savings banks discussed buying up all of the remaining safe deposit boxes in the area in order to discourage runs on the savings banks. Of course the liabilities of the savings banks were not redeemable on demand. They also had the protection of the Bank Holiday declared by the governor.
28 29
The Role of Correspondent Banks
30 31 32 33 34 35 36 37 38 39 40
Though the saving of the U.S. Mint and the subsequent opening of the damaged banks’ vaults ultimately provided the liquidity needed during the disaster, correspondent banks also played a role. Under normal conditions, San Francisco would experience a net outflow of shipments of cash to the rest of the United States, especially New York. Such shipments would indicate that San Francisco had a balance of payments deficit with New York. The normal range of this deficit would be somewhere between $200,000 and $300,000. Though the figures are reported for Pacific States and Eastern States, the former included only San Francisco and Oakland, and the latter is dominated by New York. Table 5 below gives an indication of the massive inflow of cash to San Francisco 97
4259 Ch03 4/12/02 9:44 am Page 98
98 1
RONNIE J. PHILLIPS
Table 5.
2
Cash Shipments to the Pacific States and to and from the Eastern Region, 1906, Monthly (thousands of dollars).
3 4 5 6
Date
Cash Shipments to Pacific States from the Eastern Region
Cash Shipments to the Eastern Region from the Pacific States
0 0 0 1650 2615 0 0 0 0 0 0 1250
194 347 334 35 150 1790 936 323 37 194 163 179
Pacific States Net Inflow
7 8 9 10 11 12 13 14 15 16 17 18 19
January 1906 February 1906 March 1906 April 1906 May 1906 June 1906 July 1906 August 1906 September 1906 October 1906 November 1906 December 1906
194 347 334
1615 2465 1790 936 323 37 194 163 1071
Source: Kemmerer (1910), Table IX, pp. 72–73.
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
in the months of April and May 1906 – $4.2 million. Subsequently, in June and July there were large outflows. The large inflow to the Pacific States in December 1906 is not related to the earthquake (see Kemmerer, 1910, pp. 121–123). Changes in Banking Regulation and Legislation after the Earthquake The SFCH did not play a large regulatory role prior to 1908. This means that its primary function was for payments settlement. As noted above, reducing the costs of payments settlement is entirely consistent with a competitive economy. In July 1908, the SFCH amended its constitution to institute regular examinations of non-member banks wishing to clear through a member bank. The expense of the examinations was charged to the non-member bank. Non-member banks were also required to submit statements of condition, and to be . . . subject to all rules and regulations in matters of common interest, or for the fostering of sound and conservative methods of banking, as have been or may from time to time be adopted by this Association, . . .
The state legislature responded with increased regulation and supervision, along with branch banking, after the earthquake. Branch banking in California was a
4259 Ch03 4/12/02 9:44 am Page 99
San Francisco Clearinghouse 1 2 3 4 5 6
99
direct result of the 1906 earthquake. In San Francisco, branching continued until the close of 1909 when all of the branches were abolished and there was a return to unit banking (Wright, 1910, p. 142). Giannini was a strong proponent of branch banking. The 1909 legislation permitted branching upon the approval of the state superintendent of banking. Giannini took advantage of this opportunity, while others did not (Nash, 1992, pp. 41–42).
7 8 9
V. CONCLUSION
10 11 12
Alton Gilbert (1996), in writing about the lessons from the financial history of the U.S., concluded that:
13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
(1) Banks were vulnerable to runs by their depositors. Runs by depositors on some banks tended to undermine the confidence of the depositors of other banks. (2) Banks did not rely on market mechanisms to allocate reserves during financial crises. Instead, they worked together through their clearinghouses in attempting to deal with such crises. Each bank benefited if all banks in its community could meet the demands of their depositors for cash, since depositor runs that closed one bank were likely to induce depositors to run on other banks (Dwyer & Gilbert, 1989; Roberds, 1995). (3) Clearinghouses that included as members all of the providers of payment services in their communities were more effective in dealing with banking panics than the clearinghouses that excluded substantial numbers of payments providers (Tallman & Moen, 1995). (4) Speed of actions by clearinghouses in dealing with panics was essential to success. Government sanction made such actions more effective (Roberds, 1995). (5) Conflicts of interest among banks as competitors may have limited the effectiveness of clearinghouses in dealing with financial crises. In particular, conflicts of interest may have limited the willingness of the members of the clearinghouses to commit their resources to ensure that their competitors were able to meet the demands of their customers for cash. Conflicts of interest may have limited the speed of response by clearinghouses when banks faced contagious runs by depositors (Goodhart, 1985).
37 38 39 40
My examination of the SFCH during the earthquake of 1906 leads me to agree with four of Gilbert’s conclusions: (1) There was a danger of bank runs in 1906. (2) The banks acted cooperatively and sought to dampen competition. 99
4259 Ch03 4/12/02 9:44 am Page 100
100 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
RONNIE J. PHILLIPS
(3) The wide membership in the SFCH enabled them to better deal with the crises. (4) Action was taken quickly and the help of the federal government through the U.S. Mint was crucial to the success of the plan. (5) However, the actions of SFCH members do not seem to have limited the effectiveness of the clearinghouse, and neither did the actions of non-member banks such as the Bank of Italy. What would have been the likely result if the SFCH had not existed and instead the competitive spirit would have been unbridled? This would essentially mean that every banker would have been free to do what A. P. Giannini did. Would this have been possible or desirable? Giannini owned a small bank, even though $80,000 in gold was a considerable sum. The Bank of Italy would not have had the cash if its clerks had not unwisely withdrawn it from Crocker. It would have been impossible for other bankers to do the same thing, i.e. move the gold to a new location. It would have also been very risky and indeed it is likely that the information that the gold was being moved would create a contagious bank run on all banks. The depositors would have been quite rational to run on the bank under such circumstances: get your gold before the banker moves it out of town. Another obvious factor is the risk involved in moving that much gold, even if you had the time and resources to do so. The rumors of armed men attacking the Mint might have discouraged a few bankers from doing what Giannini did. Hence, it seems reasonable to conclude that had all bankers followed Giannini’s example, a contagious run and total financial collapse would have occurred after the earthquake. Of course the crucial element is the fact that the U.S. Mint was saved. The bankers thus knew that it was highly likely that arrangements could be made to make the Mint gold available. Hence, there was no need to move their gold that was probably much safer in the strong brick vaults. If the Mint had not been saved, and if there had not been access to the Mint’s gold, this again would have dramatically changed the situation. Under those circumstances, what would have been required would have been an even greater shipment of cash from other parts of the country. Though the private sector has the incentive to prepare for disaster, there is little doubt that today people would expect the Federal Reserve or some agency of the U.S. Government to intervene in the event of a catastrophe. The private sector might handle the problem very well, but it is likely that both business and consumers feel more comfortable with the ultimate backup role of the government in disasters. This conclusion should not be upsetting to those who are ardent advocates of private enterprise, but rather reflects the fact the institutional and legal structure in the U.S. is strong enough to generate a positive externality of confidence in the private enterprise system.
4259 Ch03 4/12/02 9:44 am Page 101
San Francisco Clearinghouse 1
101
NOTES
2 3 4 5 6 7 8 9 10 11 12 13 14
1. All discussions that follow about the actions of the SFCH are taken from the notebook that contains the minutes of the clearinghouse meetings. The notebook is in the possession of the Western Payments Alliance in San Francisco, California. 2. Though it is not the purpose here to get into much discussion of the “economics of disaster,” the lesson that comes through time and time again is that a disaster is always an economic stimulus in the aggregate. See Douty, 1977. 3. This information is from a letter written to the American Banker on June 7, 1906 by John D. McKee, cashier of the Mercantile Trust Company of San Francisco. 4. The members of the Association on June 29, 1906 were: Hibernia Savings and Loan Society, San Francisco Savings Union, Savings and Loan Society, Humboldt Savings Bank, Security Savings Bank, French Savings Bank, Mechanics Savings Bank, and Scandinavian American Savings Bank. The minutes discussed used in this section were the “unofficial” association minutes, though the notebook continues with the “official” minutes.
15 16
ACKNOWLEDGMENTS
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Earlier versions of this paper were presented in a seminar in the Economics Department of the Office of the Comptroller of the Currency and at the annual meeting of the Economic History Association, Baltimore, MD, October 8–10, 1999. David Wheelock, Kerry Odell, Lawrence H. White and an anonymous referee provided valuable comments on a previous draft. I would also like to thank Gerard F. Milano and the staff at Western Payments Alliance in San Francisco, California for providing access to, and assistance with, the archival records of the San Francisco Clearinghouse and for their hospitality during my visit.
REFERENCES Bronson, W. (1959). The Earth Shook, The Sky Burned. Garden City, NJ: Doubleday and Company. Calomiris, C. W., & Gorton, G. (1991). The Origins of Banking Panics: Models, Facts, and Bank Regulation. In: R. G. Hubbard (Ed.), Financial Markets and Financial Crises (pp. 109–173). A National Bureau of Economic Research Project Report. Chicago and London: University of Chicago Press. Calomiris, C., & Kahn, C. (1996). The Efficiency of Self-Regulated Payments Systems: Learning from the Suffolk System. Journal of Money, Credit and Banking, 28(4, Part 2), 766–797. Chandler, A. (1977). The Visible Hand: The Managerial Revolution in American Business. Cambridge: Harvard University Press. Chandler, R. J. (1986). San Francisco Clearinghouse Certificates: Last of California’s Private Money. Reno, Nevada: McDonald Publishing. Douty, C. M. (1977). The Economics of Localized Disasters: The 1906 San Francisco Earthquake. New York: Arno Press. Dwyer, G. P., Jr., & Gilbert, R. A. (1989). Bank Runs And Private Remedies. Federal Reserve Bank of St. Louis Review, (May), 43–61.
101
4259 Ch03 4/12/02 9:44 am Page 102
102 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
RONNIE J. PHILLIPS
Gilbert, R. A. (1996). Financial Regulation in the Information Age. Remarks prepared for the Cato Institute’s 14th Annual Monetary Conference (June 25). Gilbert, R. A. (2000). The Advent of the Federal Reserve and the Efficiency of the Payments System: The Collection of Checks, 1915–1930. Explorations in Economic History, 37, 121–148. Goodhart, C. (1988). The Evolution of Central Banks. Cambridge: MIT Press. Gorton, G. (1985). Clearinghouses and the Origin of Central Banking in the United States. Journal of Economic History, 45, 277–283. Gorton, G. (1988). Banking Panics and Business Cycles. Oxford Economic Papers, N. S.; 40, 751–781. Gorton, G. (1989). Self-Regulating Banking Coalitions. Finance Department, The Wharton School, University of Pennsylvania, Working paper. Gorton, G. (1994). Bank Regulation When ‘Banks’ and ‘Banking’ Are Not the Same. Oxford Review of Economic Policy, 10, 106–119. Gorton, G., & Mullineaux, D. J. (1987). The Joint Production of Confidence, Endogenous Regulation and Nineteenth-Century Commercial-Bank Clearinghouses. Journal of Money, Credit and Banking, 19(November), 457–468. Kaufman, G. G. (1994). Bank contagion: A Review of the Theory and Evidence. Journal of Financial Services Research, 8, 123–150. Kemmerer, E. W. (1910). Seasonal Variations in the Relative Demand for Money and Capital in the United States: A Statistical Study. Washington, D.C.: Government Printing Office. McAndrews, J. J., & Roberds, W. (1995). Banks, Payments, and Coordination. Journal of Financial Intermediation, 4, 305–327. Muscatine, D. (1975). Old San Francisco: The Biography of a City from Early Days to the Earthquake. New York: G. P. Putnam’s Sons. Nash, G. P. (1992). A. P. Giannini and the Bank of America. Norman: University of Oklahoma Press. Plehn, C. C. (1909). The San Francisco Clearing House Certificates of 1907–1908. Publications of the Academy of Pacific Coast History, 1. Berkeley: University of California. Roberds, W. (1995). Financial Crises and the Payments System: Lessons from the National Banking Era. Federal Reserve Bank of Atlanta Economic Review, 80(Sept.–Oct.), 15–31. San Francisco Examiner (1989). October 19. Selgin, G., & White, L. H. (1994). How Could the Invisible Hand Handle Money? Journal of Economic Literature, 32, 1718–1749. Spahr, W. (1926). The Clearing and Collection of Checks. New York: Bankers Publishing Company. Stevens, E. (1996). The founders’ intentions: sources of the payments services franchise of the Federal Reserve Banks. Financial Services Research Group (Federal Reserve Bank Of Cleveland) Financial Services working paper No. 96-03. Stevens, E. (1998). Non-Par Banking: Competition and Monopoly in Markets for Payments Services. Discussion Draft, Financial Services Research Group, Federal Reserve Bank of Cleveland. Tallman, E. W., & Moen, J. R. (1995). Private Sector Responses to the Panic of 1907: a Comparison of New York and Chicago. Federal Reserve Bank Of Atlanta. Economic Review, (March), 1–9. Thomas, G., & Witts, M. M. (1971). The San Francisco Earthquake. New York: Stein and Day. Timberlake, R. H. (1993). Monetary Policy in the United States: An Intellectual and Institutional History. Chicago: The University of Chicago Press. U.S. National Monetary Commission (1910a). Volume 6: Clearinghouses. Washington, D.C.: Government Printing Office.
4259 Ch03 4/12/02 9:44 am Page 103
San Francisco Clearinghouse 1 2 3
103
U.S. National Monetary Commission (1910b). Volume 6: The Use of Credit Instruments in Payments in the United States. Washington, D.C.: Government Printing Office. Wright, B. C. (1910). Banking in California. San Francisco: H. S. Crocker Company.
4 5 6 7 8
APPENDIX: THE PAYMENTS SYSTEM AND THE BAY AREA EARTHQUAKE OF 1989
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Bankers Clearing House (BCH) was created in1978 with the merger of the San Francisco and Los Angeles Clearing Houses. In 2001, Western Payment Alliance (WesPay) and BCH completed a merger and took the Western Payments Alliance name. Wespay is a non-profit association of financial depository institutions that provides a variety of payments-related services to its more than 1,000 member financial institutions throughout the Western U.S. and Pacific regions. On October 17, 1989 at 5:04 PM, a 7.0 on the Richter scale quake hit the Bay Area. The tremor collapsed a section of the San Francisco-Oakland Bay Bridge. There were 63 deaths and nearly 14,000 injuries. Property losses were estimated at 1,118 homes destroyed; 23,408 homes damaged; 366 businesses destroyed; 3,530 businesses damaged. The total lost was estimated at $5.9 billion. Though the Federal Reserve Bank of San Francisco’s computers were knocked out, they were up and running by 8:15 PM. The Fed processed about one-third the normal amount of checks on the 18th. Backup systems in Los Angeles and in Virginia were utilized. Bank of America, First Interstate, Security Pacific, and Wells Fargo closed their San Francisco branches on the 18th due to power outages, bad phone connections, structural damages, and the Mayor’s pleas to stay home (San Francisco Examiner, 10/19/89; p. A-13). Wells Fargo utilized its communications center in Los Angeles. According to the San Francisco Examiner, “the main concern among San Francisco residents was getting cash. Long lines at automatic teller machines around town were common. Not all of the machines were open.” Wells Fargo had 45 ATMs out of service, Bank of America 100 out, First Interstate 20; and Security Pacific 35 out of 215. Eighty of 433 Bank of America branches were closed in the Bay Area. At BCH, all exchanges were cancelled for the day of the earthquake. The building housing the BCH, which was in San Francisco at that time, suffered little damage. The staff of BCH relocated to Sanwa Bank’s San Francisco operations center by 9 PM on the 17th and communicated with members by 103
4259 Ch03 4/12/02 9:44 am Page 104
104 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
RONNIE J. PHILLIPS
radio. The Executive Director, Gerard Milano, was enroute by plane to San Francisco, but was diverted to Las Vegas. Milano served as a communications link from his hotel room for the Operations Advisory Committee (OAC) members and staff. A conference call of OAC members and the Federal Reserve Bank of San Francisco (FRBSF) was arranged at 10:30 AM. The 8 AM October 18 exchange was cancelled and the OAC members agreed to a 2 PM exchange and settlement, which would also include a final settlement through the FRBSF. A schedule was established for exchanging checks at noon and midnight through Monday, October 23. The Air Transport schedule for moving the checks was left intact. BCH personnel were not permitted access to BCH offices on the 18th to run settlement because the building had not yet been inspected. Bank of America personnel worked with BCH staff to improvise a manual settlement procedure and settlement was computed and delivered to the FRBSF on the 18th. The next day, the BCH office resumed settlement on its computers. On Monday, October 23, it was agreed to return to the regular exchange and settlement schedule. Thus, daily settlements were achieved during the emergency and five days after the earthquake, operations were back to normal. In personal conversation, Gerard Milano stated that as President and CEO of BCH (now Wespay), he noted a strong incentive for him to assure that operations were returned to normal as soon as possible, since he was under a mandate in the association’s regulations to restore normal operations within five days or report to the Board of Directors why this could not be accomplished.
4259 Ch04 4/12/02 9:45 am Page 105
1 2 3 4 5 6 7 8 9 10
CAN INCOMES POLICIES REDUCE REAL WAGES? MICRO-EVIDENCE FROM THE 1931 AUSTRALIAN AWARD WAGE CUT
11 12 13 14
Andrew J. Seltzer
15 16 17 18
ABSTRACT
19 20 21 22 23 24 25 26 27 28
Industry-specific minimum wages in Australia have long been set by government tribunals. Although creating microeconomic inefficiency, the system may facilitate incomes policies, such as the 10% wage cut in 1931. This paper uses personnel records from the Union Bank of Australia to examine the effectiveness of the 1931 policy. It is shown that the bank responded to the cut by increasing the frequency of payments over the minimum rates, and that between 1924–1934 tenure-adjusted real wages were essentially constant. Finally, it is shown that the bank maintained a policy of real wage shielding as part of its internal labor market.
29 30
I. INTRODUCTION
31 32 33 34 35
Australian wages have been set by several state and federal tribunals since the 1907 Harvester Decision that firms had to pay “fair and reasonable” wages in order to qualify for tariff protection. This was institutionalized as the federal “Award System” in which tribunals set industry-specific and occupation-specific
36 37 38 39 40
Research in Economic History, Volume 21, pages 105–133. Copyright © 2003 by Elsevier Science Ltd. All rights of reproduction in any form reserved. ISBN: 0-7623-0993-8
105
4259 Ch04 4/12/02 9:45 am Page 106
106 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
ANDREW J. SELTZER
award rates covering a wide range of employees across Australia. An award set a binding wage floor; however, individual firms were free to pay their workers “over-awards”. Since its establishment, most federal awards have prescribed the “basic wage”, a standard minimum rate for unskilled workers, and perhaps an additional margin for skill. Beginning in 1922 the basic wage was indexed to inflation, although margins for skill were never indexed and were subject to frequent renegotiation. It has long been alleged that the system creates microeconomic inefficiencies because all workers covered under a given award receive the same award rate, regardless of ability or local cost of living (Gregory et al., 1988b; Norris & Wooden, 1996). Other authors have argued that this is offset by increased macroeconomic efficiency.1 They argue that a centralized system results in greater wage flexibility in times of recession or deflation and that it offers the possibility of trades of wage restraint for employment security or other working conditions. One of the most heavily debated episodes in the history of the system is whether during the 1930s it played a role in the recovery or prolonged the depression. Because of its dependence on primary exports, the Australian economy suffered severely from the collapse of world trade that began in 1929. In order to promote recovery, the Commonwealth Court of Conciliation and Arbitration (CCCA) cut the real value of the basic wage by 10% on February 1, 1931. Contemporaries generally, but not universally, believed that there was a flow-on from the cut in the basic wage to overall wage rates.2 However, recent scholarship has renewed the debate on this question. In their influential work on the Great Depression in Australia, Gregory, Ho and McDermott (1988a) concluded that the attempt to implement a 10% cut was unsuccessful. They showed that the mean all-industry real wage and non-farm real wage did not fall in response to the 10% cut to the basic wage. If anything, real wages were higher during the early 1930s than they had been in the 1920s. Moreover, they concluded that wages behaved in a similar manner in Melbourne, Sydney, and Australia as a whole, thus the failure of real wages to fall by 10% cannot be explained by differences between state and federal awards. More recently, Forster (1990) argued that the federal courts were able successfully to implement the 10% real wage cut to workers on the basic wage. However, this was not fully reflected in the all-industry and non-farm wage series because there were lags in the flow-on effects of the cut; the cut did not apply to margins above the basic rate, and federal awards only covered a fraction of the workforce. He concluded, “Nothing in the court proceedings, in the view of the judges, employers, and employees, lent support to any extreme conclusion regarding the ineffectiveness of [the Court’s] decision” (Forster, 1990, p. 26).
4259 Ch04 4/12/02 9:45 am Page 107
1931 Australian Wage Cut 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
107
Perhaps one reason why this debate has yet to be settled is that neither Gregory et al. (1988a) nor Forster (1990) examined the behavior of individual wages. Gregory et al. (1988a) drew their evidence primarily from aggregated wage series. Forster (1990) supplemented these broad statistical aggregates with evidence from state wage boards, employers, employees, and the judges on the federal courts. The conclusions that can be drawn from aggregated statistical series are limited because these data may be heavily influenced by changes in labor force composition. For example, it is intuitively plausible that firms only retained their best workers during the Great Depression, thus the apparent failure of real wages to drop by 10% may simply reflect an increase in average ability of those still employed.3 Furthermore, it is not possible to precisely control for award rates using these series, thus, they cannot be used to directly estimate changes in the pattern of over- and under-award payments.4 This paper avoids the problems associated with interpreting short-term movements in aggregated series by examining micro-data from the personnel and payroll records of one firm, the Union Bank of Australia (UBA). The approach has several advantages, but one inevitable disadvantage. Micro-data do not suffer from composition effects because the same individuals appear in the sample prior and subsequent to the wage cut. Furthermore, any changes in composition, most notably tenure, can be identified and controlled for using statistical techniques. The payroll records precisely identify the actual wages paid to each individual and allow inference of award rates and off-award payments for each employee at any given point in time. In addition, the data contained in personnel and payroll records is likely to be more accurate than evidence presented before the Commonwealth Court or state wage boards. The accuracy of personnel and payroll data was essential for the bank to efficiently allocate its employees. By contrast, individuals testifying before the courts or wage boards had an incentive to manipulate information to their advantage. The Courts themselves recognized that this was potentially a problem, stating with respect to one of the wage provisions in the 1924 Banking Award, “No evidence was given by disinterested persons or experts” (19CAR, p. 295). The cost of such precision is, of course, a loss of generality. The results of this study are necessarily idiosyncratic to the UBA and the extent to which the results apply to other firms can only be determined by further study. The first Banking Award was in 1921, but covered only New South Wales. The point of departure for this study is the 1924 Commonwealth Banking Award, which created a pay scale for the first 18 years of employment for workers with no prior experience. Table 1 shows the 1924 and 1931 Banking Award pay scales. The Banking Award was never linked to the basic wage, and, unlike the basic wage, it was not linked to the price level. Thus the real 107
4259 Ch04 4/12/02 9:45 am Page 108
108
ANDREW J. SELTZER
Table 1.
1
Pay Scale Under the 1924 and 1931 Banking Awards.
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Tenure 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
1924 Scale
1931 Scale
75.00 95.00 115.00 140.00 160.00 220.00 232.00 247.00 262.00 282.00 302.00 322.00 337.00 352.00 367.00 382.00 392.00 402.00
67.50 85.50 103.50 126.00 144.00 198.00 208.80 222.30 235.80 253.80 271.80 289.80 303.30 316.80 330.30 343.80 352.80 361.80
19 20
Source: 19CAR, p. 288.
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
value of the award rates declined steadily after 1924. Retail prices declined by about 7.5% from 1930 to 1931, thus had each worker been paid exactly the award rate, real wages would have declined by only 2.5% for the year. However, the earlier decline in the real value of the scale wage due to inflation meant that by 1931 the real value of the award had dropped by almost 7% from 1924 levels, a magnitude not dissimilar to the 10% drop in the basic wage. The Banking Award stipulated that receiving pay according to the scale was conditional on “good conduct, diligence and efficiency”(16CAR, p. 58; 19CAR, p. 298). This was to be certified by the superintendent, general manager, or chief manager of the bank concerned, thus there existed rules that provided for under-award payments at the discretion of the individual banks. This paper proceeds as follows. The second section examines the organization of the Union Bank of Australia and its personnel and payroll data. The third section analyzes the wage data to determine whether the 10% nominal cut in Banking Award rates affected the actual wages that the UBA paid its employees. It is shown that the 10% cut was applied far from universally. The bank increased its use of over-award payments and decreased its use of award and under-award payments to partially offset the reduction prescribed in the Banking
4259 Ch04 4/12/02 9:45 am Page 109
1931 Australian Wage Cut 1 2 3 4 5 6 7 8 9 10
109
Award. After controlling for years of tenure, nominal wages for a sample of 1917–1927 entrants were down by only about 5% and real wages actually increased by about 2.5% in 1931. The average tenure-adjusted real wage varied little between 1924 and 1934, and regression analysis does not show any relationship between real wages and the real value of award rates. The fourth section examines the reasons for the failure of wages to decline. It is argued that wages at the UBA behaved differently from wages in the manufacturing sector and that the UBA maintained constant tenure-adjusted real wages over the business cycle as part of an implicit contract with its workers. The fifth section concludes.
11 12
II. DATA
13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
By contemporary standards the UBA was fairly small; however, during our sample period, it was one of Australia’s largest service-industry firms. From the late 19th century it maintained branches in every Australia colony, New Zealand, and London. In 1900 it was Australia’s second largest bank with 11.0% of the total assets held by the nation’s trading banks and in 1928 it was the sixth largest with 6.2% (Butlin et al., 1971, p. 133). The UBA employed a total of 539 workers in its 84 Australian branches in 1900 and 1,011 workers in 161 Australian branches in 1930.5 The primary data employed in this study are the personnel and payroll records of employees who fit three criteria: first, initial hiring between 1917 and 1927; second fewer than two years work experience prior to joining the UBA; and third, still being employed at the bank in 1932. The requirement of under two years outside experience was imposed to ensure that sample employees were covered by the Banking Award. The requirement that individuals had to be present in 1932 was necessitated by the UBA’s recordkeeping procedures.6 A total of 359 employees met these conditions. A sample record from an employee who entered in 1922 is shown in Fig. 1. The 10% award cut is specifically noted on the record. For the purposes of this paper, the most important information contained in the records is the wage data, which is available for each worker’s entire career. Following Baker, Gibbs and Holmstrom (1994a, b), the unit of analysis used in this paper is the annual wage, which I have recorded at a single point in time, October 1 of each year.7 Real wages were computed using the retail price deflator PC31 from Vamplew (1987) with 1924 set as the base year. In addition to the 1917–1927 entry sample, I utilise the personnel records of 170 employees who entered between 1881 and 1891, had no experience prior to entry, and were still present in 1896.8 109
4259 Ch04 4/12/02 9:45 am Page 110
110
ANDREW J. SELTZER
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
Fig. 1.
A Sample Personnel Record.
30 31 32 33 34 35 36 37 38 39 40
Because the personnel records provide the date of entry for each employee, it is possible to determine the exact tenure at the time of each wage observation. This information can also be used to infer the appropriate rate from the Banking Award scale in Table 1. The Banking Award stated that “automatic increases shall date from the first pay day falling after the completed year of service” (16CAR, p. 59; 19CAR, p. 301). However, the majority of employees who began their employment between October 1 and December 31 received at least the award rate for a completed year of service. Thus, for the purpose of this analysis I have assumed that all employees who began in the same year were
4259 Ch04 4/12/02 9:45 am Page 111
1931 Australian Wage Cut 1 2
111
covered by the same award rate of pay. Off-award payments are calculated as the difference between actual pay and award rates.
3 4 5 6
III. WAS THE WAGE CUT SUCCESSFUL? EMPIRICAL EVIDENCE
7 8
Evidence From Entry Cohorts
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
As a first pass at examining the effects of the 1931 wage cut, Fig. 2 shows the average nominal wage of each of the 11 entry cohorts and the sample crosssection of wages between 1917 and 1934.9 The individual cohorts’ profiles show a slowing down of nominal wage growth beginning in 1931. Members of the more senior cohorts earned the same on average in 1933 as they had in 1930. In addition, the average wage of each cohort reaching a given tenure in 1931 was lower than that of the previous cohort in 1930 or two years previous cohort in 1929. While this suggests that the award cut did have an influence on wages, it does not address whether the CCCA was able to achieve its desired cut in real wages. Figure 3 shows the real wage profiles of the 11 entry cohorts along with the cross-sectional average for the period 1917 to 1934. It is evident from casual inspection of the individual series that 1931 was an unremarkable year in terms of real wage growth. For most of the individual cohorts and for the sample as a whole, real wages continued to grow at rates similar to those exhibited in previous years. If anything, there appears to have been a slight acceleration in real wage growth in 1931. The mean real wage increase in 1931 was larger than the increase in 1930 for 10 out of 11 cohorts, the increase in 1929 for nine of 11 cohorts, and the increase in 1928 for 8 of 11 cohorts. Table 2 compares the wages of successive cohorts in 1930 and 1931. The comparison is made between successive cohorts in order to eliminate the effect of increasing tenure on wages; e.g. a 1926 entrant would have reached their fifth year of tenure in 1930, whereas a 1927 entrant would have reached their fifth year in 1931. As is evident from columns 4 and 5, the later cohorts earned less in nominal terms in 1931 than the earlier cohorts earned in 1930. However, in each case the difference was less than 10% and in 7 cases the difference was statistically significant. On average nominal wages in 1931 were only 5.25% less than in 1930, implying that the UBA implemented policies that offset almost half of the award wage reduction.10 Columns 7 and 8 compare real wages in 1930 and 1931. In each case, with the exception of the 1917 and 1918 entrants, the later cohort earned more in 1931 than the earlier cohort earned in 1930. On average this difference was approximately 2.65%. 111
4259 Ch04 4/12/02 9:45 am Page 112
112
ANDREW J. SELTZER
1 2 3 4 5 6 7
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Sources: Z/87/1, Z/87/2, and Z/87/3.
11
Fig. 2.
9 10
Average Annual Nominal Wages by Entry Cohort, 1917–1934.
8
4259 Ch04 4/12/02 9:45 am Page 113
1931 Australian Wage Cut
113
1 2 3 4 5 6 7 8
Average Annual Real Wages by Entry Cohort, 1917–1934.
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Fig. 3.
29 30 31 32 33 34 35 36 37 38 39 40
113
Source: Z/87/1; Z/87/2; Z/87/3; and Vamplew (1987), PC31.
9 10
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
31
32
33
39
40
30
34
35
36
37
38 181.9 217.3 232.9 249.8 267.2 287.2 308.2 323.1 342.3 361.6
Sources: Z/87/1; Z/87/2; Z/87/3; and Vamplew (1987), PC31.
171.4 212.8 221.7 236.0 254.3 272.3 289.4 305.5 323.6 333.7
173.73 207.56 222.46 238.55 255.24 274.31 294.42 308.58 326.91 345.38
7.72
5.46#
5.45
6.10#
5.19##
4.83##
5.52##
4.81##
2.07##
5.77
177.31 220.22 229.46 244.21 263.13 281.76 299.49 316.03 334.80 345.23
Real 31
2.06 6.10** 3.15* 2.37 3.09** 2.72* 1.72 2.41 2.42 0.04
% Diff.
114
* = Significantly different from zero at the 5% level (real wages only). ** = Significantly different from zero at the 1% level (real wages only).
# = Significantly different from 10 at the 5% level (nominal wages only). ## = Significantly different from 10 at the 1% level (nominal wages only).
20 35 29 37 45 49 56 19 8 31
13
35 29 37 45 49 56 19 8 31 30
9
5 6 7 8 9 10 11 12 13 14
10
Real 30
7
% Diff.
6
Nominal 31
5
Nominal 30
3
Obs. 31
2
Obs. 30
1
Tenure
12
Nominal and Real Wages, 1930 and 1931.
11
Table 2.
4259 Ch04 4/12/02 9:45 am Page 114
ANDREW J. SELTZER
4
8
4259 Ch04 4/12/02 9:45 am Page 115
1931 Australian Wage Cut 1
115
Changes in Use of Over-Award and Under-Award Payments
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
The fact that the difference in nominal wages between successive cohorts was well less than 10% suggests that the bank increased its use of over-awards or reduced its use of under-awards to partially offset the effect of the cut in the award rate. Examining the behavior of individual wages provides a more direct evidence of this hypothesis. Table 3 shows some summary statistics on the extent and magnitude of over-award and under-award payments in 1930 and 1931. In 1930 the majority of workers (57.23% of sample employees) received exactly the award rate. Just over a quarter of workers were paid more than the award rate and 16.22% were paid less than the award rate. In 1931, after the 10% award reduction, the incidence of both award and under-award payments dropped dramatically. Only 9.42% of sample employees received less than the award rate, and the average under-award payment dropped by nearly a half, from £17.80 to £9.63. By contrast the incidence of over-award payments increased more than two and half fold to 69%, although the average over-award payment dropped slightly from £32.78 to £26.30.
18
Long-Term Changes in Wages and Off-Award Payments
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Tables 2 and 3 only show the short-term changes between 1930 and 1931, and do not offer any insight to the behavior of wages over the longer period. It is possible to analyze the behavior of wages over the first 11 years of the Banking Award by creating several fixed weight series using the 1924 scale as the weights. Such series have been created as follows: (1) (2) (3) (4) (5)
NOMAWARDt, i = (AWARDt, i /AWARD24, i )*100 NOMWAGEt, i = (WAGEt, i /AWARD24, i )*100 REALAWARDt, i = ((AWARDt, i /PRICEt )/AWARD24, i )*100 REALWAGEt, i = ((WAGEt, i /PRICEt )/AWARD24, i )*100 OFFAWARDt, i = ((WAGEt, iAWARDt, i )/AWARDt, i )*100
AWARDt, i denotes the award rate for an individual, i, at time, t, given their tenure at time t. AWARD24, i denotes what would have been their award rate under the 1924 scale, given their tenure at time t. WAGEt, i denotes the actual nominal wage rate received by individual i at time t. Finally, PRICEt is the retail price deflator PC31 (divided by 100) from Vamplew (1987). By linking current wages or award rates to the 1924 scale, the fixed weight series provides a simple means of controlling for tenure. An individual receives a value of 100 if their current wage (or real wage or award wage) is exactly the rate prescribed under the 1924 Banking Award scale, given their current tenure. 115
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
39
40
34
35
36
37
38
Sources: Z/87/1; Z/87/2; Z/87/3; and 19CAR, p. 288.
126.00 144.00 198.00 208.80 222.30 235.80 253.80 271.80 289.80 303.30 21.58
25.00 22.86 24.14 29.73 17.78 24.49 25.00 26.32 12.50 0.00
57.23
69.00
70.00 74.29 65.52 59.46 71.11 67.35 62.50 63.16 75.00 90.32
26.55
42.86 20.69 18.92 17.78 22.45 28.57 26.32 25.00 29.03 36.67
9.42
5.00 2.86 10.34 10.81 11.11 8.16 12.50 10.53 12.50 9.68
16.22
2.86 10.34 18.92 20.00 14.29 25.00 21.05 25.00 16.13 10.00
26.30
39.36 20.98 21.25 25.50 26.79 28.98 29.89 27.82 27.62 19.59
32.78
51.33 1.00 23.00 33.00 34.36 34.50 38.00 25.50 26.00 31.27
9.63
4.00 27.00 9.47 13.85 5.40 12.55 8.40 18.00 3.30 2.80
17.80
5.00 28.00 18.29 15.56 17.43 18.57 18.00 21.00 14.00 18.67
116
Total
14
5 6 7 8 9 10 11 12 13 14
13
1931
12
Total
9
54.29 68.97 62.16 62.22 63.27 46.43 52.63 50.00 54.84 53.33
8
140.00 160.00 220.00 232.00 247.00 262.00 282.00 302.00 322.00 337.00
7
5 6 7 8 9 10 11 12 13 14
6
1930
5
Average Average % Over Award % Under Award Over-Award Under-Award
4
% On Award
3
Award Rate
2
Tenure
1
Year
11
Over-Awards and Under-Awards, 1930 and 1931.
10
Table 3.
4259 Ch04 4/12/02 9:45 am Page 116
ANDREW J. SELTZER
4259 Ch04 4/12/02 9:45 am Page 117
1931 Australian Wage Cut 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
117
Figure 4 shows the mean values of NOMAWARD, NOMWAGE, REALAWARD, REALWAGE, and OFFAWARD between 1924 and 1934. The most striking feature of figure 4 is that REALWAGE varied little over the period. The standard deviation of the annual average of REALWAGE is only 1.27, whereas the standard deviation of the annual averages of NOMAWARD, REALAWARD, NOMWAGE, and OFFAWARD are 5.05, 2.64, 6.20, and 2.62 respectively. Figure 4 implies that the real wage profile was effectively constant over the period. Prior to the depression most employees earned exactly the award rate; whereas after the award cut, over-awards sharply increased. However, as the price level continued to decline, reaching below 90% of its 1924 level in 1933, the UBA switched to a policy of more frequent underaward payments. The causes of individual variation in REALWAGE can be analyzed using regression analysis. The data for the regression are pooled individual observations between 1924 and 1934. If the Banking Award was binding, a REALWAGE should be determined by REALAWARD, their age at entry, and their ability.11 Age at entry should have a positive effect on individual wages because the Banking Award contained provisions relating to the wages paid to workers appointed “late in life” (beyond age 16 or 17), or, alternatively, because
20
Table 4.
21
The Determinates of Real Wages.
22 (1)
23 24 25 26 27 28 29
Independent Variable REALWAGE REALAWARD 0.0727 (1.188) Age at Entry
(2) REALWAGE 0.0172 (0.241) 1.817* (9.445)
(3)
REAL- LN(REALWAGE WAGE) 0.0729 (1.322)
LN(REALAWARD)
0.0518 (0.916)
30 31 32 33 34
Constant
92.109* (15.396) Model Specification OLS
67.227* (8.768) OLS
Adjusted R2 Sample Size
37
0.025* 3772
38 39 40
0.000 3772
90.455* (15.127) Fixed Effects 0.194* 3772
(6)
LN(REALWAGE)
LN(REALWAGE)
0.027
0.0511 (1.017)
4.356* (16.814) OLS 0.000 3772
0.021* 3772
Notes: Absolute value of the t-statistic in parentheses. * = Significant at the one-percent level. Sources: Z/87/1; Z/87/2; Z/87/3; and Vamplew (1987), PC31.
117
(5)
(0.414) 0.263* (8.614) 3.977* (12.696) OLS
LN(Age at Entry)
35 36
(4)
4.346* (18.775) Fixed Effects 0.215* 3772
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
33
34
35
36
37
38
39
40
Sources: Z/87/1; Z/87/2; Z/87/3; Vamplew (1987), PC31; and 19CAR, p. 288.
32
Average Values of Four Fixed Weight Series, 1924–1934.
118
Fig. 4.
4259 Ch04 4/12/02 9:45 am Page 118
ANDREW J. SELTZER
1
2
3
4
5
6
7
8
9
4259 Ch04 4/12/02 9:45 am Page 119
1931 Australian Wage Cut 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
119
the UBA chose to pay higher wages to employees with outside experience or more education. Ability cannot be directly observed. However, in this sort of panel data, innate ability can be controlled for indirectly by using a fixed effects model with individual dummies included as independent variables.12 The regression results are shown in Table 4. The coefficient on REALAWARD is fairly small (a one percentage point increase in REAL AWARD causes approximately a 0.08 percentage point increase REALWAGE) and insignificant at standard levels in each specification. In other words variation in real wages between 1924 and 1934 cannot be explained by variation in the real award rate. The results for the other variables are as expected. The coefficient on age at entry is positive and strongly significant. The fixed effects specifications have considerably more explanatory power than the ordinary least squares specifications and a number of the individual dummies were statistically significant, indicating that individual ability was an important determinate of wage. These results consistently support the Gregory et al. conclusion that the award cut was not successfully implemented. Had the UBA perfectly followed the award scale throughout, the average value of REALWAGE would have been 6.875% lower in 1931 than in 1924. In reality the difference was only 0.583%, a figure consistent with Gregory et al.’s claim that real wages were essentially given exogenously over the period. The evidence suggests that the UBA maintained a policy of constant real wages and allowed off-awards to adjust to offset changes in the real value of the award rate. This is further evidenced by the fact that the correlation between the annual average of offaward and the annual average of REALAWARD is a massive 0.898. Similarly, pooling the individual observations of OFFAWARD and REALAWARD between 1924 and 1934 yields a correlation coefficient of 0.245.
28 29
Comparisons to 1894–1895
30 31 32 33 34 35 36 37 38 39 40
Prior to the passage of the Banking Award, wages were determined by market forces and were set by the individual banks. Between 1889 and 1897 the UBA maintained the following wage schedule: year 1 – £40, year 2 – £60, year 3 – £80, year 4 – £100, year 5 – £120, year 6 – £135, year 7 – £150.13 The Bank had no explicit rules governing wages after 7 years of tenure; however, Seltzer and Merrett (2000) argue that the very low variance of individual wages of employees with the same tenure suggests the existence of an implicit set of rules. As was later to be the case under the Banking Award, moving up the wage scale was not an entitlement. The UBA’s rules specifically stated that “[m]erit will be in all cases the first claim to consideration” for salary increases 119
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
39
40 101.34 114.88 130.76 136.84 146.10 157.80 179.00 177.63 176.07 199.14
–9.44 –14.57 –12.05 –14.87 –15.71 –21.88 –9.37 –6.69 –19.18 –10.50
203.34 244.35 270.15 292.10 314.96 367.05 358.87 345.89 395.86 404.30
191.14 216.68 246.62 258.08 275.56 297.63 337.60 335.01 332.09 375.60
–6.00 –11.32** –8.71** –11.65** –12.51** –18.91** –5.92 –3.14 –16.11* –7.10
% diff
120
31
32
33
34
35
36
37
38
# = Significantly different from –10 at the 5% level (nominal wages only). ## = Significantly different from –10 at the 1% level (nominal wages only). * = Significantly different from zero at the 5% level (real wages only). ** = Significantly different from zero at the 1% level (real wages only). Sources: U/271/1; U/271/2; U/271/3; and Vamplew (1987), PC31.
111.90 134.47 148.68 160.75 173.33 202.00 197.50 190.36 217.86 222.50
11
25 21 19 34 20 15 10 8 14 14
9
21 19 34 20 15 10 8 14 14 10
14 Real 95
1
5 6 7 8 9 10 11 12 13 14
10 Real 94
8
% diff
6
Nominal 95
5
Nominal 94
4
Obs. 95
3
Obs. 94
2
Tenure
13
Nominal and Real Wages, 1894 and 1895.
12
Table 5.
4259 Ch04 4/12/02 9:45 am Page 120
ANDREW J. SELTZER
7
4259 Ch04 4/12/02 9:45 am Page 121
1931 Australian Wage Cut 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
121
(U/195/1, p. 5). However, most junior employees received the prescribed rate for their first 7 years (Seltzer & Merrett, 2000). In 1895, after five years of rapid deflation, the bank decided to implement an across-the-board 10% wage reduction.14 The 1881–1991 entrants’ personnel records can be used to replicate the analysis of the effects of the 1931 reduction. Table 5 replicates the analysis shown in Table 2. Columns 4–6 of Table 5 show the change in nominal wages, adjusted for years of tenure. It can be seen that the wage reduction was fully implemented; the cohort reaching a given level of tenure in 1895 earned on average 13.77% less than the previous cohort in 1894. For each of the 10 cases the null hypothesis that the later cohort’s earnings in 1895 were 10% less than the earlier cohort’s earnings in 1894 cannot be rejected. Columns 7–9 show the average real wage of successive pairs of entry cohorts in 1894 and 1895. Although there was some deflation during the year, it only partially countered the effects of the wage cut. The cohort reaching a given level of tenure in 1895 earned on average 10.50% less in real terms in 1895 than the previous cohort in 1894.
17 18 19
IV. SPECULATIONS: WHY WAS THE 1931 WAGE CUT UNSUCCESSFUL?
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
The fact that the UBA maintained a policy of constant real wages between 1924 and 1934 is somewhat surprising and counterintuitive. The Great Depression severely effected output and labor markets in Australia; the peak to trough decline in real output was 22.1% and the increase in unemployment was 15.55 percentage points.15 This paper examines two explanations for the stability of real wages in the face of such a large output shock. First, it is possible that the UBA was a price taker in a competitive labor market with sticky real wages. Gregory et al. (1988a, b) argue real wage stickiness characterized Australian labor markets in the 1930s, and Dighe (1997) argues that a similar phenomenon occurred in American labor markets. A second explanation that is examined in this paper is that the UBA did not face a competitive labor market, but rather insulated its workers from external shocks as part of an internal labor market. As a test of whether the UBA acted as a price taker in the labor market, I have compared the time paths of REALAWARD to a fixed weight index of real annual earnings in manufacturing.16 Figure 5 plots REALAWARD and REALMAN, real annual earnings in manufacturing indexed to 1924 levels, between 1924 and 1934.17 If the UBA were a price taker in the broader labor market, one would expect the two series to move together. While both series show evidence of real wage stickiness and high wages during the 1930s, there are substantial differences in the time paths of the two series. In particular, the 121
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
36
37
38
39
40
Sources: Z/87/1; Z/87/2; Z/87/3; and Vamplew (1987), PC31 and LAB201.
35
Fixed-Weight UBA Earnings and Average Hourly Earnings in Manufacturing.
122
Fig. 5.
4259 Ch04 4/12/02 9:45 am Page 122
ANDREW J. SELTZER
1
2
3
4
5
6
4259 Ch04 4/12/02 9:45 am Page 123
1931 Australian Wage Cut 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
123
UBA wages remained roughly constant in 1925 and 1931 when there were large changes in REALMAN. As a further test I have disaggregated the two series by states and plotted them over the same period in Figure 6.18 As with the aggregated series, there are considerable differences in the state-level movements of REALAWARD and REALMAN. The variance across states is greater for REALWAGE in the 1920s, but is greater for REALMAN in the 1930s. Furthermore the rankings of the states for the two series differ considerably over time. For example, Tasmania is the lowest wage state in the manufacturing series throughout the period, but is about average in the UBA series. This divergence between REALAWARD and REALMAN suggests that the UBA was not acting as a price taker in the broader labor market. In order to test the hypothesis of wage shielding as part of an internal labor market, it is necessary to examine the nature of the UBA’s personnel practices prior to the 1930s. The UBA operated an internal labor market from at least as early as the 1880s.19 This internal labor market insulated employment and wages of employees from external shocks in several ways, including the following. The frequency of terminations changed little with increases in unemployment. Nominal wage cuts were used extremely infrequently; outside of 1895 and 1931–1933 fewer than 1% of nominal increments were negative.20 Real wages were positively related to the current unemployment rate and the minimum unemployment rate experienced to that point of a worker’s career.21 Nominal wages in the late 19th century were governed by the scale in operation when a worker first entered the UBA, even during the severe deflation of the 1890s after the Bank reduced the scale for new entrants. Figures 3 and 4 and Table 4 suggest an additional form of wage shielding. Individuals could expect a fairly constant annual rate of real wage growth, assuming that their work was of a suitable standard. Furthermore, they could form very accurate expectations of their own future earnings by examining the current earnings of earlier entry cohorts. The observed pattern of similar rates of real wage growth in 1931 as in the years immediately prior and subsequent for the 11 entry cohorts is consistent with this sort of policy. Workers would find such a policy highly desirable because it would make their future real income more predictable. This explanation is also consistent with the UBA’s 10% wage cut in 1895. The sharpest deflation in Australian history occurred between 1890 and 1894 and this resulted in extremely high real wage growth for UBA employees. Thus, the 1895 cut may have served the purpose of bringing real wage growth back down to the normal rate. Whether the Bank could offer such a contract would depend on the nature of the employment relationship. A firm can only afford to maintain abovemarket wages during economic downturns if it can recoup any short-term losses 123
4259 Ch04 4/12/02 9:45 am Page 124
124
ANDREW J. SELTZER
1 2 3 4 5 6 7 8 9 10 11 12
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Sources: Z/87/1; Z/87/2; Z/87/3; and Vamplew (1987), PC31 and LAB195-LAB200.
15
Fig. 6.
14
Fixed Weight Indexes by State.
13
4259 Ch04 4/12/02 9:46 am Page 125
1931 Australian Wage Cut
125
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Continued.
17 18 19 20
Fig. 6.
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
125
4259 Ch04 4/12/02 9:46 am Page 126
126 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
ANDREW J. SELTZER
by underpaying its workers during boom periods. However, underpayment during a boom would normally lead to high worker turnover, which would have been unacceptable for a bank because of high search and training costs in the industry.22 Labor economists have identified several impediments to employee mobility; two of which seem particularly relevant to the industry: firm specific human capital and inability of outsiders to observe a worker’s marginal product. The existence of these impediments to labor mobility also make it highly unlikely that the UBA was acting as a price taker towards employees in this sample, who would have had between 3.5 and 14.5 years tenure at the time of the Award cut. The UBA’s records provide considerable evidence of specific human capital. A 1926 Sub-Inspector’s Report described the tasks performed by each employee at the Melbourne branch (U/93/20, pp. 17–26). Most of the tasks performed by clerical staff involved preparation and checking of various bank records. Younger employees were required to have mastered a number of different record-keeping procedures before they could move up the hierarchy. These tasks were likely to have been specific to the industry, rather than the individual banks. These tasks were learned on the job over a period of several years, and as a consequence promotion at the UBA was a slow process. The mean time until first promotion for 1888–1900 entrants was almost nine years for workers with no prior experience.23 At higher levels of the hierarchy even more specific human capital was required. Tellers had to know their customers in order to speed up transactions and prevent fraud. Accountants had to know their staff members very well in order to optimally assign employees to tasks and ensure that the tasks were performed in accordance with the bank’s rules. The most specific skills, however, belonged to the branch manager who was expected to be familiar the Bank’s lending rules, local business conditions and property values, and the financial position of the bank’s customers (U/195/1, p. 1). Although some of these skills were undoubtedly general to the banking industry, many were highly specific to the UBA or even to individual branches. An alternative reason that workers might be unable to command better wages in the external labor market is that it might have been difficult for the external labor market to make inferences about individual worker’s productivity.24 Two of the UBA’s personnel practices, slow promotion and wage compression, might have resulted in workers of a similar age but different innate abilities being observationally identical to outsiders. The evidence on turnover is consistent with the existence of impediments to mobility. Quit and total turnover rates were low by historical and contemporary standards, with annual averages of 3.7 and 6.3% respectively, between 1887 and 1899.25 Although quits increased during boom periods (the correlation
4259 Ch04 4/12/02 9:46 am Page 127
1931 Australian Wage Cut 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
127
between GDP growth and quits is 0.68), the annual rate never exceeded 6.5 even though real GDP growth averaged nearly 9% per year between 1896 and 1898. Moreover, the quit rate decreased dramatically after about five years tenure, the period of tenure covered by the sample used in this paper. Whereas the quit rate for employees with five years tenure was approximately 6%, for employees with 14 years tenure it was barely 2%. Furthermore, many quits appear to have been driven by issues of lifestyle rather than salary. In a number of cases the reason for quitting is given on the personnel record. The majority of these quits were for reasons such as leaving the country, taking up farming, going into the family business, going to university, or joining the clergy. There is considerable evidence that it was uncommon for banks to attempt to poach each other’s workers.26 Finally, employees were more likely to resign if they were promoted in the previous year, which is consistent with signals from the Bank’s personnel practices providing information to the external labor market.27 Although it is beyond the scope of this paper to search for direct evidence of implicit contracts that smoothed real wages over the business cycle at other Australian firms, it should be noted that by the 1930s a large proportion of the Australian workforce were governed by internal labor markets. At least half of all Australian workers were employed in firms that were sufficiently large to support internal labor markets.28 Large firms existed in a number of industries including: mining, metallurgy, food processing and distribution, textiles, brewing, confectionery, soap, rubber, automobiles, oil, packaging, retailing, sugar, timber, paper, glass, electrical goods, publishing, banking, utilities, public transportation, and communication (Merrett, 1997). There is considerable anecdotal evidence that several of these industries had extensively internalized their labor markets by this time. All public servants, teachers, and railway workers had published wage scales from the 1890s onwards (Seltzer & Merrett, 2000, p. 578). Sammartino (2002) showed that the Victorian Railways maintained internal labor markets from the 1860s and finds some evidence that wages were shielded from external shocks. Wright (1995) noted that many large firms were moving toward replacing managerial discretion with more formal personnel policies. Dunningham (1937) claimed that in the 1930s employment in most non-rural industries was permanent and that employment in mining was becoming less intermittent (Dunningham, 1937, pp. 12, 23). Permanent hiring was sufficiently prevalent that by 1920 the awards set higher wage standards for workers who were hired on a nonweekly basis (Anderson, 1929). The administrative requirements of the Award System almost certainly necessitated the sort of book-keeping practices that are also necessary to maintain internal labor markets.29 Whether these firms also shielded their workers from the external labor market by 127
4259 Ch04 4/12/02 9:46 am Page 128
128 1 2
ANDREW J. SELTZER
increasing their use of over-award payments in 1931 is an important topic for further research.
3 4 5
V. CONCLUSIONS
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
This paper uses the personnel records of one firm, the Union Bank of Australia, to examine the impact of the 1931 Australian incomes policy that attempted to cut wages by 10%. Recent scholars generally agree that economy-wide real wages did not fall by 10%, the level desired by policy makers. However, there has been an extended academic debate about why there was wage stickiness. Forster (1990) argues that the federal courts were able to successfully implement a 10% real wage cut for most workers under federal jurisdiction. However, federal coverage was incomplete and state awards and margins over the basic rate did not drop by a similar amount. Other scholars such as Gregory, Ho and Dermott (1988a, b) have argued that the real wages of workers covered by the federal awards failed to drop 10%, and in fact behaved much the same way as workers not covered by these awards. Gregory et al. (1988a) conclude that real wages during the Great Depression were given largely exogenously. Previous quantitative research has been limited to the examination of aggregate wage series, which are subject to composition changes and can not precisely control for the award rates. This paper uses microdata from one firm, the Union Bank of Australia, to more precisely analyze the effect of the 1931 award wage cut. The Banking Award only specified a nominal wage cut of 10%, while prices fell by nearly 8% for the year. The evidence presented in this paper indicates that even the 10% nominal cut was not reflected in the wages that the UBA paid its workers. After controlling for years of tenure, nominal wages declined by just over 5%. Real wages for the year increased by about 2.65, one of the largest annual increases between 1924 and 1934. The bank increased its use of over-award payments to more than offset the award reduction. In the following years the bank cut nominal wages further and increased its use of under-award payments as the price level continued to drop. However, over the period 1924–1934 real wages remained roughly constant, varying only within a 3% band. Regression analysis shows no relationship between the real value award rate and real wages paid to individual workers over the period. Finally, this paper argues that the UBA cut wages by well less than the 10% specified in the award because it sought to maintain constant real wages over time as part of its implicit contract with its workers.
4259 Ch04 4/12/02 9:46 am Page 129
1931 Australian Wage Cut 1
129
NOTES
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
1. Gregory et al. (1988b) summarizes the views of several 1930s authors. Also see Borland (1997) and Kenyon (1997) in a recent policy forum on Australian unemployment. 2. See Gregory et al. (1988a, b) and Forster (1990) for a summary of contemporary views on the issue. 3. Dighe (1997) examines the changes in labor force composition and their implications for real wages in the United States during the 1930s. 4. Schedvin (1970) argues that under-award payments were widespread during the Great Depression. However, he does not present any statistical evidence on legal underaward payments or violations of the law to support this claim. The data presented by Gregory et al. (1988a, b) suggest the opposite, firms increasingly paid over-awards in order to offset the effect of the award wage cut. 5. U/218, U/219, U/220, U/221, U/222. These figures exclude employees at the Head Office and the inspection staff. 6. The ANZ Group Archive contains several boxes of personnel records of individuals who were present in 1932 (Z/87/7, Z/87/8, Z/87/9). A small proportion of the records are missing (there is only one individual whose surname begins with K or L), however, this is unlikely to systematically bias the results. The fact that individuals who survived to 1932 may be systematically better employees than the cohort of entrants between 1917 and 1927 is of little concern for the purposes of this study because the CCCA intended to reduce each individual’s wages with the 1931 cut. 7. October 1 was used in preference to December 31 because fewer employees were on short-term assignments replacing workers taking annual leave. This is likely to reduce the number of short-term wage adjustments in the data. 8. The records of all employees present in 1887 and all entrants between 1888 and April 1900 have been preserved. (U/271/1, U/271/2, U/271/3). Employees who were not present in 1896 were dropped from the sample in order to ensure comparability with the 1917–1927 entry sample. 9. The simplest way to determine the effects of the Banking Award would be to examine the cross-sectional distribution of wages prior and subsequent to 1931. However, cross-sectional data will suffer from a composition bias. The nature of the sampling process results in artificially increasing average tenure throughout the sample period and thus in an overestimate of wage growth. The average tenure for the annual cross-sections were as follows: 1917 – 0.43, 1918 – 0.98, 1919 – 1.79, 1920 – 2.25, 1921 – 2.46, 1922 – 2.46, 1923 – 2.95, 1924 – 3.42, 1925 – 4.05, 1926 – 4.57, 1927 – 5.25, 1928 – 6.24, 1929 – 7.24, 1930 – 8.24, 1931 – 9.24, 1932 – 10.24, 1933 – 11.24, 1934 – 12.24. 10. The figure of 5.25% was computed using a weighted average of the percentage decline for each tenure group. The size of the cohorts differs for each year of entry, thus slightly different results will be obtained depending on whether the 1930 or 1931 number of observations is used as the weight. However, this difference proved to be only 2/100 of a percentage point. The results reported are an average of the two alternative weightings. 11. The Banking Award also prescribed a minimum rate for workers required to live away from their usual home. It is not possible to control for this effect in the regression analysis because the records do not identify workers’ usual homes. 129
4259 Ch04 4/12/02 9:46 am Page 130
130 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
ANDREW J. SELTZER
12. It is not possible to control for age and ability simultaneously because the fixed effects model only allows the inclusion of independent variables that vary over time. 13. U/87/9. Prior to 1900 number of employees came to the UBA with previous banking experience. These employees typically were paid well more than proscribed under the internal wage scale. Seltzer and Merrett (2000) and Seltzer and Simons (2001). 14. Between 1890 and 1895 prices dropped by 31.3%. This contrasts markedly with the period between 1929 and 1931 when prices dropped by only 7.7%. Vamplew (1987), PC31. Even at the trough of the business cycle in 1933, prices were only down 15.4% from their 1929 peak. 15. Vamplew (1987), series ANA64 and PC31 (output and retail prices) and LAB97 (unemployment). 16. Ideally one would like to compare earnings at the UBA to earnings at other whitecollar employers; however, I am unaware of any such series for the 1930s. The use of manufacturing wages in this analysis leaves open the possibility that the UBA was a price taker in a clerical labor market that had very different properties than the manufacturing labor market. 17. The index for real earnings in manufacturing is given by 100 (AHRt/AHR24) / (PC31/100), where AHR is average nominal hourly earnings in manufacturing from Vamplew (1987), series LAB 201. 18. The state-level indexes were constructed the same way as REALMAN using Vamplew (1987), series LAB195 – LAB200. 19. See Seltzer and Merrett (2000) and Seltzer and Simons (2001) for a discussion of internal labor markets at the UBA. 20. However, real wages were much more flexible from year to year. Among 1887–1900 entrants, approximately 26.6% of annual real wage increments were negative over the course of their careers. Among the sample used in this paper 9.65% of real wage increments between 1924 and 1934 were negative. Most of these negative real increments were received by mid- or late-career employees remaining in the same position and receiving the same nominal wage between two years. 21. Beaudry and DiNardo (1991) offer a theoretical explanation for an expected negative relationship between wages (W) with the current unemployment rate (UC) and the minimum unemployment rate up to that point in a worker’s career (UM). They argue that a spot market would be characterized by a negative relationship between W and UC. Firms would be forced to pay market wages, which would be influenced by workers’ outside options. They also consider an impure contract model where firms could make credible commitments but workers could not because of costless mobility between jobs. This would be characterized by a negative relationship between W and UM. Firms would have to match workers outside options during booms, but any wage increments they received at these times would remain during subsequent downturns. 22. See Seltzer and Merrett (2000) and Seltzer (2000) on search and training costs in the banking industry. 23. This figure excludes the 45% of employees who were never promoted beyond the level of clerk. 24. See Itoh (1994) on the information extraction problem in the context of Japanese firms. 25. All turnover figures are computed using U/271/1, U/271/2, and U/271/3. 26. Only 11 of the full sample of 1767 UBA entrants between 1850–1900 and 1917–1927 have noted that the individual left to join another Australian bank. This was
4259 Ch04 4/12/02 9:46 am Page 131
1931 Australian Wage Cut 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
131
true throughout the industry. It was noted in the 1924 Banking Award that “it has not been the practice of the banks to give appointments to applicants beyond the age of 16 or 17 years.” 19CAR, p. 281. This preference for younger workers precluded poaching from other banks. 27. Seltzer and Simons (2001) show that all else equal, promotion in the previous year increased the probability of resignation by about 35%. 28. Wright (1995) notes that 6% of Australian firms employed half the workforce. 29. Jacoby (1985) argues that the administrative requirements of the New Deal ware an important pre-condition for the internalization of labor markets in the United States.
ACKNOWLEDGMENTS I have benefited from suggestions made by Tim Leunig; Robert Margo; Ian McDonald; an anonymous referee and the participants at seminars at ClaremontMcKenna College, Royal Holloway College, the 1998 Economic History Society Conference, the 1999 Western Economic Association Meetings, and the 1999 Economic History Association Meetings. I gratefully acknowledge skillful data collection and encoding by Brendon Avallone, David Choa, Jeremy Feiglin, Sarah Martin, Geoffrey Sinclair, and Ming Yu; and the help of the staff at the ANZ Archive (Trevor Hart, Peggy Kennedy and Tony Miller). This research has been funded by grants from the Department of Economic History and Faculty of Economics and Commerce at the University of Melbourne and the Australian Research Council. Remaining errors are entirely mine.
23 24
REFERENCES
25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Anderson, G. (1929). Fixation of Wages in Australia. Melbourne: MacMillan & Co. Ltd. The Australia New Zealand Banking Group (1932–1970). Male Staff Record Cards. Melbourne: ANZ Group Archive,, VIC, Z/87/7, Z/87/8, Z/87/9. Baker, G., Gibbs, M., & Holmstrom, B. (1994a). The Internal Economics of the Firm: Evidence from Personnel Data. Quarterly Journal of Economics, 109, 881–919. Baker, G., Gibbs, M., & Holmstrom, B. (1994b). The Wage Policy of a Firm. Quarterly Journal of Economics, 109, 921–955. Beaudry, P., & DiNardo, J. (1991). The Effect of Implicit Contracts on The Movement of Wages over The Business Cycle: Evidence from Micro Data. Journal of Political Economy, 99, 665–688. Borland, J. (1997). Unemployment in Australia – Prospects and Policies: An Overview. Australian Economic Review, 30, 391–404. Butlin, S. J., Hall, A. R., & White, R. C. (1971). Australian Banking and Monetary Statistics, 1817–1945. Sydney: Reserve Bank of Australia, Occasional Paper No. 4A. Commonwealth Arbitration Record 16 (1922). The Bank Officials Association and the Bank of Australasia and Others. Commonwealth Arbitration Record 19 (1924). The Bank Officials Association and the Bank of Australasia and Others.
131
4259 Ch04 4/12/02 9:46 am Page 132
132 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
ANDREW J. SELTZER
Dighe, R. S. (1997). Wage rigidity in the Great Depression: Truth? Consequences? Research in Economic History, 17, 85–134. Dunningham, J. M. (1937). Report upon employment and unemployment in the State of New South Wales. Government Printer, Sydney. Forster, C. (1990). Wages and Wage Policy: Australia in the Depression, 1929–1934. Australian Economic History Review, 30, 23–42. Gregory, R. G., Ho, V., & McDermott, L. (1988a). Sharing the Burden: The Australian Labour Market during the 1930s. In: R. G. Gregory & N. G. Butlin (Eds), Recovery from the Depression: Australia and the World Economy in the 1930s (pp. 217–245). Cambridge: Cambridge University Press. Gregory, R. G., Ho, V., & McDermott, L. (1988b). The Australian and U.S. Labour Market in the 1930s. In: B. Eichengreen & T. Hatton (Eds), Interwar Unemployment in International Perspective (pp. 397–430). Kluwer Academic Publishers. Itoh, H. (1994). Japanese Human Resource Management from the Viewpoint of Incentive Theory. In: M. Aoki & R. Dore (Eds), The Japanese Firm: Sources of Competitive Strength (pp. 233–264). Oxford University Press. Jacoby, S. (1985). Employing Bureaucracy: Managers, Unions, and the Transformation of Work in American Industry, 1900–1945. New York: Columbia University Press. Kenyon, P. (1997). Infrastructure Spending and Unemployment: Government Responsibility for Growth and Jobs. Australian Economic Review, 30, 421–432. Merrett, D. T. (1997). Australian firms by employment. Mimeo, University of Melbourne. Norris, K., & Wooden, M. (1996). The Changing Australian Labour Market: An Overview. In: Norris & Wooden (Eds), The Changing Australian Labour Market (pp. 1–14). Economic Planning and Advisory Council, Commission Paper 11. Sammartino, A. (2002). Human Resource Management Practices and Labour Market Structures in the Victorian Railways 1864–1921. Ph.D. Dissertation, University of Melbourne. Schedvin, C. B. (1970). Australia and the Great Depression. Sydney: Sydney University Press. Seltzer, A. J. (2000). Controlling and Motivating the Workforce: Evidence from the Banking Industry in the Late 19th And Early 20th Centuries. Australian Economic History Review, 40, 219–238. Seltzer, A. J., & Merrett, D. T. (2000). Personnel Practices at the Union Bank of Australia: Panel Evidence from the 1887–1900 Entry Cohorts. Journal of Labor Economics, 18, 573–613. Seltzer, A. J., & Simons, K. L. (2001). Salaries and Career Opportunities in the Banking Industry: Evidence from the Personnel Records of the Union Bank of Australia. Explorations in Economic History, 38, 195–224. The Union Bank of Australia Limited. Instructions to be Observed by Officers. ANZ Group Archive, Melbourne, VIC, U/195/1. The Union Bank of Australia Limited. Register of Officers. ANZ Group Archive, Melbourne, VIC, U/271/1, U/271/2, U/271/3. The Union Bank of Australia Limited (1889). Memoranda for Candidates for Junior Appointments in the Colonial Service of the Bank. ANZ Group Archive, Melbourne, VIC, U/87/9. The Union Bank of Australia Limited, Melbourne Branch (1927). Sub-Inspector’s Report. ANZ Group Archive, Melbourne, VIC, U/93/20. The Union Bank of Australia Limited (1900–1930). Annual Report, Victoria and Tasmania. ANZ Group Archive, Melbourne, VIC, U/218. The Union Bank of Australia Limited (1900–1930). Annual Report, New South Wales. ANZ Group Archive, Melbourne, VIC, U/219.
4259 Ch04 4/12/02 9:46 am Page 133
1931 Australian Wage Cut 1 2 3 4 5 6 7 8
133
The Union Bank of Australia Limited (1900–1930). Annual Report, Queensland. ANZ Group Archive, Melbourne, VIC, U/220. The Union Bank of Australia Limited (1900–1930). Annual Report, South Australia. ANZ Group Archive, Melbourne, VIC, U/221. The Union Bank of Australia Limited (1900–1930). Annual Report, Western Australia. ANZ Group Archive, Melbourne, VIC, U/222. Vamplew, W. (Ed.) (1897). Australians, Historical Statistics. Fairfax, Syme & Weldon Associates, New South Wales. Wright, C. (1995). Australian Studies in Labour Relations. Melbourne: Oxford University Press.
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
133
4259 Ch05 4/12/02 9:58 am Page 135
1 2 3 4 5 6 7 8
DECAYING AT THE CORE: URBAN DECLINE IN CLEVELAND, 1915–1980
9 10 11 12
Fred H. Smith
13 14 15 16
ABSTRACT
17 18 19 20 21 22 23 24 25 26 27 28 29
In this essay I use the Mills-Muth monocentric model to examine the impact of urban decay on land values in Cleveland, Ohio. Using the bid-rent surfaces generated from the estimation of the Mills-Muth model, I calculate aggregate assessed land values for the entire city of Cleveland as well as for several of its neighborhoods. My analysis confirms the story told by historians: Cleveland experienced severe urban decay after World War II. However, while land values fell precipitously between 1950 and 1980, most of the decline occurred during the 1970s. The magnitude and the timing of the decline in land values during the early 1970s are difficult to explain, but my results challenge Cleveland historian Robert Whipple Green’s assertion that dramatic demographic changes in Cleveland’s East Side neighborhoods caused these neighborhoods to decay.
30 31
I. INTRODUCTION
32 33 34 35
In 1630, while traveling from England to Massachusetts, John Winthrop told his fellow travelers that when they reached America “we shall be as a city upon a hill, the eyes of all people are upon us” (Norton, 1994, p. 54). It was
36 37 38 39 40
Research in Economic History, Volume 21, pages 135–184. Copyright © 2003 by Elsevier Science Ltd. All rights of reproduction in any form reserved. ISBN: 0-7623-0993-8
135
4259 Ch05 4/12/02 9:58 am Page 136
136 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
FRED H. SMITH
Winthrop’s hope that his followers would work together to create a better society, a society that others would look at in awe. Nearly three centuries later, Tom Johnson would invoke Winthrop’s “City on a Hill” metaphor once again (Keating, 1995, p. 89). Johnson had been elected the Mayor of Cleveland, Ohio in 1901, and it was Johnson’s hope that Cleveland would become one of the nation’s great cities. Cleveland was already the seventh largest city in the United States, and it had developed into one of the nation’s leading manufacturing centers. Johnson worked hard to ensure that Cleveland would become the “City on a Hill”, and when he left office eight years after he was first elected, Cleveland looked to be on the path to greatness. Tom Johnson’s vision that Cleveland could become a “City on a Hill” would never be realized: Instead, Cleveland would become a symbol of the decaying American city. Cleveland’s experience with urban decay has not been unique. Yet, despite the severity of the urban decay that has afflicted Cleveland and other American cities, urban economists have not studied the changes in these cities’ spatial structure. Specifically, there have been no attempts to quantify the impact of urban decay. In the sections that follow, I take a step towards filling that gap in the economics literature.
20 21
II. A BRIEF HISTORY OF CLEVELAND, OHIO
22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Moses Cleaveland founded the city of Cleveland in July of 1796. Cleaveland, a representative of the Connecticut Land Company, had arrived in what is now the northeast corner of Ohio with the intention of founding a capital for the Western Reserve.1 He established the city that would bear his name to the east of the Cuyahoga River along the shore of Lake Erie.2 This site was extremely well suited for the development of a city. It benefited from close proximity to the river and lake, yet its location on the Erie plain meant that it was well above the swampy lands located adjacent to the river. Furthermore, the flatness of the land facilitated the layout of the city (Chapman, 1981, pp. 2–3). Surveyors in Moses Cleaveland’s founding party established the layout for modern Cleveland. They laid out a large Public Square as well as a grid of north-south and east-west streets that surround and intersect the square (Chapman, 1981, p. 5). It was the founding fathers’ intention that Public Square serve as the center of Cleveland. It has. (Figure 1 shows the layout of the city of Cleveland.) The city developed rather slowly until 1825, but between 1825 and 1850 Cleveland’s strategic location allowed the city to benefit tremendously from the construction of the Erie Canal and the Ohio and Erie Canal (Chapman, 1981,
4259 Ch05 4/12/02 9:58 am Page 137
Urban Decline
137
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Fig. 1.
Cleveland, Ohio.
Source: Case Western Reserve University, Image Gallery.
28 29 30 31 32 33 34 35 36 37 38 39 40
p. 31). Similarly, the city reaped enormous benefits when the railroads came to Cleveland in 1851. Indeed, Cleveland’s role as a transshipment node for canal boats and locomotives fueled a population explosion. Cleveland grew from a sleepy town of 7,000 residents in 1840 to the sixth largest city in the United States with a population of 560,663 in 1910. (Table 1 contains population data for Cleveland from 1900 to 2000.) Between 1910 and 1920 Cleveland’s population swelled from 560,663 to 796,841. As its population increased by 42%, Cleveland went from being the 6th to the 5th largest city in the country. However, the growth rate of Cleveland’s population paled in comparison with the growth rate of its black population. There were 8,448 black Clevelanders in 1910. By 1920 the black population had expanded to 32,451, an increase of 284%. In the next decade the city’s 137
4259 Ch05 4/12/02 9:58 am Page 138
138
FRED H. SMITH
Table 1.
1
The Population of the City of Cleveland, 1900–2000.
2 3 4 5 6 7 8 9 10 11 12 13 14
Year
Population
Rank
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
381,768 560,663 796,841 900,429 878,336 914,808 876,050 750,903 573,822 505,616 478,403
7th 6th 5th 6th 6th 7th 8th 10th 18th 23rd 33rd
Source: U.S. Department of Commerce, Census of Population, various years
15 16 17 18 19 20 21 22 23 24 25 26 27 28
population increased by one hundred thousand (from 796,841 to 900,429), and the black population more than doubled (from 34,451 to 71,889). Ultimately, the migration of African-Americans to Cleveland had the greatest impact on the Cedar-Central neighborhood on the city’s East Side. Table 2 documents the demographic transformation that occurred in Cedar-Central’s five core census tracts (Keating, 1995, p. 284). The blacks who moved to Cleveland during the first decades of the twentieth century sought greater economic opportunity and a more hospitable social atmosphere. Unfortunately, they did not find Cleveland to be the progressive city it had been in the nineteenth century. Black Clevelanders were no longer welcome in every restaurant and every theater. They were not permitted to enter the facilities of the YMCA or the YWCA. Several hospitals restricted service
29 30
Table 2. The Expanding African American Population in the Central-Cedar-Woodland Neighborhood, 1910–1930.
31 32 33 34 35 36 37 38 39 40
Census Tract G-9 H-7 H-9 I-2 I-3 Source: Keating, 1995, p. 284.
Percentage Black in 1910
Percentage Black in 1920
Percentage Black in 1930
13.8 19.6 14.7 19.7 24.4
18.5 45.7 50.8 32.7 62.6
17.4 54.9 88.8 36.6 67.4
4259 Ch05 4/12/02 9:58 am Page 139
Urban Decline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
139
to African-Americans, and the Women’s Hospital began a policy of admitting blacks only on Saturday. African-Americans even faced the specter of discrimination in their own neighborhood, for the Cedar-Central neighborhood suffered from irregular garbage pick-up and erratic streetcar service (Keating, 1995, pp. 268, 273). Cleveland’s population grew rapidly during the early twentieth century: The landscape of the city changed just as dramatically. Tom Johnson, Cleveland’s mayor from 1901 to 1909, oversaw the creation of the Group Plan. Drawn up in 1903, the plan called for a complete transformation of the area just to the north and east of Public Square. Specifically, the plan called for a new city hall, county courthouse, public auditorium, school administration building, and public library (Miller, 1997, p. 108). The construction of the buildings in the Group Plan began the transformation of downtown Cleveland, but the transformation was complete after the construction of the Terminal Tower complex. The Terminal Tower complex, completed in 1930, contains Cleveland Union Terminal (CUT) and the Terminal Tower. When it was completed it was the tallest building outside of New York City and remained so until the 1960s (Toman, 1996, p. 149). There were several other important changes to Cleveland’s urban landscape between 1910 and 1930. The theaters of Playhouse Square, located at the intersection of Euclid Avenue near E. 13th Street, were completed in the early 1920s, and the airport (later to be named Hopkins International Airport) was completed in 1925 (Toman, 1996, p. 132). Fisher Body opened an enormous factory on Coit Road in 1921. And five major department store chains opened or maintained stores near Public Square during this period. The early twentieth century was also a golden age for public transportation in the city. Tom Johnson fought for cheap public transportation, and in 1910 a three-cent fare was put in place. The three-cent fare lasted until 1917. (A ride on New York’s subway during the first two decades of the twentieth century cost five cents.) Riding the streetcar was not only inexpensive; it was convenient. Public Square remained the hub of the streetcar system, and lines extended to both the East Side and West Side on every major thoroughfare leading to downtown (Toman, 1996, p. 78). The streetcar’s popularity was short lived. In 1910 the streetcars carried 228 million passengers, and a decade later ridership had increased nearly 100% to 450 million passengers. Yet by 1930 ridership had declined to 302 million. Public transportation in Cleveland would never again serve 400 million riders in a non-war year.3 As ridership of public transportation decreased, car ownership increased. Americans owned eight million automobiles in 1920; a decade later they owned 24 million. Clevelanders owned 61,000 automobiles 139
4259 Ch05 4/12/02 9:58 am Page 140
140 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
FRED H. SMITH
in 1916, and one decade later this figure had grown to 211,000. It was clear that the automobile was the future (Toman, 1996). The future must have looked very bright to the residents of Cleveland in 1930. While it would have been impossible to predict the changes that Cleveland would undergo in the next three decades, change was to be the predominant feature of the period from 1930 to 1960. The automobile, already an important component of the transportation network in 1930, became the dominant mode of transportation (Case Western Reserve University, s.v. Transportation) This sounded a death knell for the streetcar. Cleveland’s population also changed dramatically. The population of the city was 900,429 in 1930, and three decades later it was 876,050. These numbers, though, do not reveal the whole story. Cleveland’s African-American population grew from 71,889 to 250,889. Thus, while seven percent of all Cleveland residents were black in 1930, 28% of Cleveland’s population was African-American in 1960. Change was an important theme for Cleveland’s economy as well. Clevelanders, like the residents of most cities, suffered tremendously during the Great Depression. Unemployment in Cleveland peaked at 33% in 1933, and many of Cleveland’s great firms were devastated by the economic downturn (Case Western Reserve University, s.v. Timeline). Foremost among this group was Union Trust. Union Trust, as well as Guardian Trust, another of Cleveland’s great banks, lost half of its deposits on February 22, 1933. Union Trust had once been the largest bank outside of New York, but it closed its doors forever during the bank holiday imposed by President Roosevelt in March of 1933 (Case Western Reserve University, s.v. Banking). The city’s auto industry was also a victim of the Great Depression. Cleveland would remain a major producer of automobile parts, but the Depression brought an end to the production of cars in the city. Ford, the last firm to make automobiles in Cleveland, closed its plant in 1932 (Toman, 1996, pp. 159–160). The Great Depression crippled the city’s economy, but the Second World War brought it back to health. The U.S. Army’s need for military equipment meant that iron and steel were in great demand. Cleveland, still a leading center for the production of these materials, reaped the benefits. The immediate post war period was also a time of relative prosperity for Cleveland’s leading industries. Americans were purchasing large numbers of cars, homes, and consumer durables, so there was tremendous demand for Cleveland’s steel and iron, auto parts, and metal products. Although its economy remained relatively healthy, Cleveland faced other challenges between 1940 and 1960. The first buses had been added to Cleveland Railway’s fleet of vehicles in 1925, and trackless trolleys followed ten years later. Over the course of the 1930s and 1940s the streetcar was gradually phased
4259 Ch05 4/12/02 9:58 am Page 141
Urban Decline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
141
out. This process was interrupted by the Second World War, for the wartime demand for public transportation necessitated that streetcars remain in service. The city of Cleveland took control of Cleveland Railway in 1942, and the renamed company, the Cleveland Transit System (CTS), eliminated the last streetcar line on January 23, 1954 (Toman, 1996, p. 243). Although streetcar and rail service declined steadily throughout this period, this happened in the face of increasing commuting times for those traveling by car or bus. In 1939, Cleveland had more cars per mile of paved roadway than any other city in the United States. One result was serious congestion. At a time when it took a streetcar just two minutes to cross the Detroit-Superior Bridge (the principal artery connecting the East Side to the West Side), the same trip by car or bus could take as long as 20 minutes. Because of the growing traffic congestion, planners created blueprints for a highway system. Created in 1944, the blueprints would not be put into use until nine years later when construction on the Innerbelt Freeway started. The Innerbelt may be seen on Fig. 1 (it is the section of the interstate highway system that runs just east of downtown Cleveland). The Innerbelt Freeway took nearly ten years to complete, and its construction radically altered the landscape of the near East Side in Cleveland. Though the Innerbelt Freeway did not open during the transition period from 1930 to 1960, Cleveland had already opened its first “expressway” in 1939 when the Shoreway, a limited access divided highway, opened. It stretched from Gordon Park (centered on East 79th Street) to downtown (East 9th Street) (Toman, 1996, p. 268). A final challenge to the automobile’s dominance was developed during the 1950s. Using rail beds from existing railroads, the CTS opened the city’s first rapid transit line in 1955. The line stretched from the West Side to the Terminal Tower complex, and it continued through to the East Side as far as University Circle. The rapid transit line opened to great fanfare, and, initially, it was hailed as a great success. Public accolades, however, did not translate into increased ridership. The car remained king in the city of Cleveland (Toman, 1996, p. 268). The automobile revolution changed the way that Clevelanders traveled within the city. Automobiles made it easy to travel in any direction, and a family who wished to live in the suburbs was no longer forced to choose a suburb with a streetcar link to Public Square. Moreover, since the 1930s, residents of Cleveland had been making the choice to move to the suburbs. In 1930, 75% of the people living in Cuyahoga County lived in the city of Cleveland. Thirty years later that figure had fallen to 53% (Case Western Reserve University, s.v. Timeline). This process of suburbanization, coupled with the dramatic increase in the number of African-Americans moving to Cleveland, profoundly affected the city. Through the first half of the twentieth century Cleveland’s shopping district had been 141
4259 Ch05 4/12/02 9:58 am Page 142
142 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
FRED H. SMITH
located on and around Public Square. As the hub of the streetcar system, the square had naturally evolved as the retail heart of the city. Department stores, restaurants, and banks were all located in this area. However, once the automobile became the dominant form of transportation in Cleveland, the residents of the city no longer needed to do their shopping downtown. Taking advantage of the flexibility offered by the automobile, five shopping centers opened in the surrounding suburbs during the 1950s (Toman, 1996, p. 268). The abundance of inexpensive housing, open spaces, and clean air pulled upper and middle class Clevelanders into the suburbs, but they were also “pushed” into the suburbs by changes occurring within the city. A second wave of African-Americans migrated to Cleveland from the South during the 1940s and 1950s. Cleveland’s black population stood at 84,504 in 1940; ten years later it had increased to 147,847. Then, during the 1950s, the African-American population of Cleveland increased another 103,000. The black population of 250,889 in 1960 represented an increase of nearly 200% over two decades (U.S. Department of Commerce, 1962). Two problems confronted the black residents of Cleveland during this period. Since the African-Americans arriving from the South were generally not well educated, they faced a difficult time securing high wage jobs. In addition, African-Americans were forced to contend with racism in the housing market. The minutes from a National City Bank board meeting held in 1934 document the institution’s policy of redlining (Case Western Reserve University, s.v. Banking). National City Bank’s redlining policy was not unique, and, when coupled with realtors’ reluctance to show blacks homes located outside of the traditional black neighborhoods, it was difficult for African-Americans to find housing (Keating, 1995, p. 249). Racism in the housing market exacerbated the decline of the housing stock in the Central-Cedar-Woodland neighborhood. Blacks, facing few alternatives, were forced to convert single-family homes into multiple-family dwellings.4 Overcrowding made it difficult for residents to keep up with routine maintenance, which, in turn, led to urban blight. Prior to 1950, Cleveland city officials fought urban blight by building public housing. The first public housing complexes were built in the late 1930s, and more were added in the 1940s. (The East Side contained all but one of these public housing complexes.) The city’s officials began a different course of action to combat blight in the 1950s. Under authorization from the 1949 Housing Act, city officials ordered the demolition of entire neighborhoods. The city purchased dilapidated housing and tore it down. Private developers were then supposed to step in and develop housing units on the newly cleared land. Unfortunately, this did not happen the way that it was designed to. Private developers built 800 housing units after
4259 Ch05 4/12/02 9:58 am Page 143
Urban Decline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
143
the first wave of urban renewal, but they failed to complete 3500 units that were supposed to be built after the second wave. In total, the city cleared over 6,000 acres of land during the entire urban renewal program. Urban renewal’s legacy in Cleveland is not impressive. Hundreds of trash strewn vacant lots remain to this day (Miller, 1997, pp. 141–142). The residents of the Central-Cedar-Woodland neighborhood, displaced by urban renewal and the construction of the Innerbelt Freeway, had to move elsewhere. Nearby neighborhoods absorbed the exodus of black residents from Central, radically altering the racial composition of some of these neighborhoods. Consider the transition of the Hough neighborhood. In 1950, Hough was a working class, white neighborhood. Ten years later, Hough had been transformed into an impoverished, black neighborhood (Keating, 1995, p. 31). The population figures from Hough’s principal census tracts may be found in Table 3. (The census tracts listed in Table 3 are clustered in the Hough neighborhood found in Fig. 1.) The transformation of the Hough neighborhood was not unique. Other East Side neighborhoods, and even some East Side suburbs, would experience similar transformations in the 1960s and 1970s. Although by 1960 Cleveland was on the path of racial transformation and economic decline, many Clevelanders had dreams of a better future. Those dreams would not be realized. Cleveland’s next two decades would prove to be catastrophic. Between 1960 and 1980 the city of Cleveland began to disintegrate. In 1960, it was the 8th largest city in the nation with 876,050 inhabitants. Two decades later the city had lost 302,000 of its residents, making it the 18th largest city in the country. Middle class families, white and black, fled the city for the suburbs. During the early 1970s residents of Cleveland moved out of the city at the rate of 20,000 people per year (Case Western Reserve University, s.v. Timeline). By 1980 the East Side of Cleveland was predominantly black, and many of the black families living on the East Side were poor. Approximately 40% of the families residing in Hough and Central-Cedar received AFDC in 1974 (Olson, 1976, pp. 146–147). Along with the changes in its population, Cleveland’s economy also changed. The manufacturing sector shrunk as manufacturing firms relocated to the suburbs, the Sun Belt, or to foreign countries. Factories, department stores, and theaters closed throughout the city, and two riots during the 1960s devastated the East Side neighborhoods of Hough and Glenville. However, nothing captured Cleveland’s decline more poignantly than the Cuyahoga River Fire. The river caught fire during the summer of 1969 and made Cleveland the laughing stock of the nation. The “Forest City” had been transformed into the “mistake by the lake.” Cleveland’s transformation during the period from 1930 to 1960, and its subsequent decline from 1960 to 1980, can be traced to structural changes in 143
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Source: U.S. Department of Commerce.
14
336 369 283 483 452 225 59
13
5194 9157 6323 4366 3466 8606 10625
6.5 4.0 4.5 11.1 13.0 2.6 0.6
5703 11785 7430 5270 3160 8246 10072
4778 10755 6600 4670 2023 5610 4424
83.8 91.3 88.8 88.6 64.0 68.0 43.9
9
13 14 15 16 17 18 19
10 9.8 28.7 17.5 20.7 –8.8 –4.2 –5.2
Percentage Change in Population
8
Percent Black 1960
1322.0 2814.6 2232.2 866.9 347.6 2393.3 7398.3
Percentage Change in Black Population
6
Black Pop. 1960
5
Pop. 1960
4
Percent Black 1950
3
Black Pop. 1950
2
Pop. 1950
1
Census Tract
12
Population Change in Hough, 1950–1960.
11
Table 3.
4259 Ch05 4/12/02 9:58 am Page 144
144 FRED H. SMITH
7
4259 Ch05 4/12/02 9:58 am Page 145
Urban Decline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
145
the economy and the automobile revolution. Manufacturing employment in Cleveland had reached an all-time high in 1969. Ten years later nearly a third of the city’s manufacturing jobs had been lost to other locations. U.S. Steel started closing plants in Cleveland in the 1960s, and by 1984 it no longer operated in the city (Case Western Reserve University, s.v. Industry). National Screw, a Cleveland mainstay since 1889, closed its doors in 1969 (Miller, 1997, p. 171). Fisher’s enormous Coit Road factory scaled back operations throughout the 1970s and shut down in 1982. The advent of interstate trucking diminished the importance of access to rail stations and waterways. Furthermore, many manufacturers changed their production processes. The multistory factories of the early twentieth century became less practical, and single story structures covering several acres of land replaced them. Taken together, these factors placed a premium on suburban locations near freeway exits. Indeed, when National Screw closed its Cleveland plant in 1969, it moved to a one story plant in Medina, a suburb with excellent freeway access (Miller, 1997, p. 171). Cleveland officials tried to combat the effects of urban decline in the city. Although urban renewal failed to improve the city’s residential neighborhoods during the 1950s, it managed to produce one success in downtown Cleveland: Erieview. Designed by I. M. Pei, it was urban renewal’s signature development. It contains office and residential towers, and it also includes a two story shopping mall (Case Western Reserve University, s.v. Architecture). Erieview sits just to the northeast of Public Square, and when its first buildings were completed in the early 1960s Clevelanders were optimistic about the impact that Erieview would have in reversing the city’s fortunes. Sadly, the completion of Erieview did not prevent the closure of four of the five downtown department stores and the theaters on Playhouse Square. Just as firms took advantage of the expanded freeway network to move to the suburbs, so too did many Cleveland residents. Neighborhoods like Hough, Glenville, and Ohio City lost many of their middle class families to the suburbs. Several factors caused the mass migration to the suburbs. The expansion of the freeway system made it convenient for Cleveland residents to move, but it was the availability of cheap housing, clean air, and better schools that led middle class families to leave the city. The once proud Cleveland school system was in a state of disarray by the 1960s. Insufficient resources led to half days in many of the Cleveland schools. Black children received inadequate training in mathematics and foreign languages, and they were taught in highly segregated schools. Although the segregation of the Cleveland public schools could, in large part, be attributed to the segregation of Cleveland’s neighborhoods, the school system exacerbated 145
4259 Ch05 4/12/02 9:58 am Page 146
146 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
FRED H. SMITH
the problem by permitting “selective” transfers. In other words, white students occasionally received permission to go to a different school if their home district was “too black” (Keating, 1995, p. 274). The availability of cheap housing was a second factor pulling middle class families into the suburbs. Prior to 1950 there had been a serious housing shortage, and the building of new housing units and “filtering” that occurred in the 1950s only served to offset the impacts of urban renewal and freeway construction. However, builders continued to build suburban housing at a rapid rate throughout the 1960s. By the 1970s, so much housing had been built in the suburbs that there was a housing surplus in the Cleveland metropolitan area. When middle class families left the city for the suburbs they left behind low income families and families living on public assistance. These families were ill equipped for the tasks of supporting local businesses and paying the property taxes so desperately needed to fund adequate public services. Property tax delinquency rates skyrocketed. Over twenty percent of the properties in Central and Hough were certified delinquent by 1974 (Olson, 1976, pp. 22–23). The city had settled into a downward spiral of decay. Deteriorating conditions in East Side neighborhoods led to two riots during the 1960s. The Hough Riot occurred in July of 1966, and the Glenville Shootout/Riot occurred two years later in the summer of 1968. The reasons for the Hough Riot and the Glenville Shootout/Riot are too numerous too list, yet at their root they were responses by Cleveland’s East Side African-American community to their exceptionally low quality of life. The Glenville Shootout was particularly devastating. In the wake of the Hough Riot, the voters in Cleveland had elected Carl Stokes to the mayor’s office. Stokes was the first African-American mayor of a major American city. Clevelanders elected Stokes in the hope that he would be able to solve some of the most serious problems facing the black community in Cleveland, thereby avoiding another riot like the one that had occurred in Hough. Due to circumstances beyond his control, Stokes could not prevent the Glenville Shootout in the summer of 1968. Although he ran for a second term after Glenville, Mayor Stokes left office in 1971 after just four years in office. The erosion of Cleveland’s tax base resulted in one of its greatest indignities of this period. The city’s government was unable to meet its yearly obligations during the first half of the 1970s, so it issued bonds to address the revenue shortfall. After repeatedly issuing debt to cover recurring expenses, the city government dug itself into a hole from which it could not escape. In December of 1978, the banks decided not to roll over $14 million worth of short-term debt. When the notes came due on December 15, the city was unable to pay its creditors. Cleveland had become the first major city since the Great
4259 Ch05 4/12/02 9:58 am Page 147
Urban Decline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
147
Depression to go into default. Cleveland’s financial health would not be fully restored for nearly ten years (Keating, 1995, p. 97). Cleveland’s achievements during this period were few and far between. In fact, the only real “achievements” were improvements in the transportation network (Toman, 1996, p. 278). CTS established a rapid transit link between the airport and downtown in 1967, and construction crews completed the East Side freeways. Over the past two decades Cleveland has begun the long journey back to respectability. Downtown Cleveland is the home to the Rock and Roll Hall of Fame, several new skyscrapers, a refurbished warehouse district, a new baseball park and football stadium, a revitalized Playhouse Square, and a beautiful shopping mall built in the Terminal Tower complex. A visitor who confines her stay to downtown Cleveland will see these projects and conclude that Cleveland has made a complete comeback. Yet, as is the case with most of Cleveland’s history, the story is more complicated. The city’s downtown has enjoyed a renaissance, but the residential neighborhoods of the East Side still bear the scars of four decades of urban decay. Vacant lots can be found throughout the area, and abandoned buildings are common. Poverty is the defining characteristic of many of the East Side neighborhoods. As I move to an empirical analysis of Cleveland’s decay in the following sections, it will be essential to keep Cleveland’s history in mind. It is far easier to fully appreciate the magnitude of the city’s decline by remembering that Cleveland was once one of the nation’s great cities. Ultimately, a historical perspective also reminds us why this type of study is important. If it is possible to gain a better understanding of the magnitude and causes of Cleveland’s decay, then it may be possible to prevent urban decay from striking other cities in the twenty-first century.
28 29
III. MODELING URBAN SPATIAL STRUCTURE
30 31 32 33 34 35 36 37 38 39 40
Urban decay is a highly descriptive term, one that evokes images of boarded up stores, vacant lots, and abandoned homes, yet it needs to be defined precisely before it is of use to an economist. At its root, urban decay is a term describing a decline in the productivity of a city’s myriad components. These components include businesses, homes, recreational areas, and infrastructure. Therefore, in order to quantify urban decay, it is necessary to quantify the loss of productivity of the city’s components. The one common factor of production for all of the components of a city is land. Land is used in the production of housing services, recreational services, and in the production and sale of goods and services by the business community. 147
4259 Ch05 4/12/02 9:58 am Page 148
148 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
FRED H. SMITH
As with any other factor of production, land is priced according to the value of its marginal product.5 Measuring the productivity of land is not an easy task. In an urban setting the intensity with which a parcel of land is used serves as a good guide for the parcel’s productivity. The land beneath the Terminal Tower complex is used very intensively, and it is extremely valuable. The intensity of use is seen in the height of the Terminal Tower. If land on Public Square had been inexpensive, then the builders of the Terminal Tower complex would surely have chosen to build a far less expensive one-story structure covering several acres. Similarly, one can surmise that the land in a vacant lot eight miles from Public Square is not very valuable.6 These two examples highlight the important feature that makes urban land productive: location. At the time of its construction in the late 1920s, Terminal Tower was built at the hub of the transportation network in the center of the financial and retail sectors. Ultimately, Terminal Tower was built on the southwest corner of Public Square because that site offered advantages that the parcel of land eight miles away did not. In the 1960s, Edwin Mills and Richard Muth developed the monocentric model, a model that formally establishes the relationship between the value of a parcel of land and its distance from the city center.7 The Mills-Muth model is developed for a generic city on a featureless plain. Thus, the Mills-Muth monocentric model does not differentiate between parcels of land that are equidistant from the city center. This is a very strong restriction; it limits the monocentric model’s ability to accurately describe a city’s spatial structure. In the case of Cleveland, the model predicts that two parcels of land equidistant from Public Square would have the same value per square foot. Ultimately, it is this restriction that limits the model’s usefulness in understanding urban decay. General trends are identifiable, but it is impossible to glean information about how urban decay affected land values in a specific location within the city. The Mills-Muth specification of the monocentric model is firmly rooted in economic theory, but urban economists readily admit that it is overly simplistic. Since the city described by the monocentric model sits on a featureless plain there are no reasons for a city to grow more rapidly in any one specific direction, nor for land values to be higher on one side of a city than another. Yet, in the early twentieth century, R. M. Hurd recognized that cities do tend to grow in specific directions (Hurd, 1905). Several decades later, Edwin Mills also touched upon this topic in “The Value of Urban Land” (Mills, 1969, pp. 233–234). Both authors argued that topographical features and location specific transportation improvements could affect urban land values. Evidence from Section II of this paper supports Hurd and Mills. Cleveland was founded on the eastern bank of the Cuyahoga. Stagnant water in the slow
4259 Ch05 4/12/02 9:58 am Page 149
Urban Decline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
149
flowing Cuyahoga River provided a fertile breeding ground for mosquitoes, so residents who lived near the river’s banks were constantly battling disease. Not surprisingly, residents of Cleveland were eager to escape these unhealthy conditions. The higher ground located to the east of the river provided the ideal escape. The East Side’s appeal was cemented in the late 1800s when the streetcar network was expanded throughout East Side neighborhoods. Streetcar lines did extend into the West Side neighborhoods, but an 1893 route map reveals that the East Side possessed more than a dozen lines while the West Side had only five (Toman, 1996, p. 41). Topography and access to transportation played a critical role in determining land values in nineteenth century cities. This is still true today. Nationwide, people pay substantial premiums for land located near oceans, lakes, and forests. Land values are similarly affected by access to a city’s transportation network. Residents of modern cities must concern themselves with more than just topography and transportation, though. Land values in modern cities are also affected by the quantity and quality of public services that are offered. Ultimately, these factors imply that land values are determined by more than just a parcel’s distance from the center of the city. Direction matters. To date, no comprehensive theoretical model has been developed to fully account for direction specific factors in an urban area. Nonetheless, important modifications to the monocentric model have been developed. For example, John Yinger explored the impact of changing the transportation network in a monocentric city (Yinger, 1993, p. 306). Extending work first done by William Alonso, Yinger demonstrated that a city with a grid of urban streets would grow in different directions than a city with an infinite number of radial streets extending out of the city center. His findings are based on the fact that it is total transport cost, not straight-line distance, that affects a parcel of land’s value (Yinger, 1992). Yinger has also explored the impact of urban subcenters, commonly referred to as “Edge Cities”, on land values in the “monocentric” model (Garreau, 1988). In a different line of research, Yinger, Paul Courant, and Susan RoseAckerman have all explored the impact of racial prejudice on urban spatial structure. In their 1977 paper, Yinger and Courant recognized that their model is inherently limited by their inability to incorporate directional variables: “When multidimensional urban models become available, perhaps the analysis of the relationship between prejudice and urban structure can be given a more satisfactory foundation” (Courant & Yinger, 1977, p. 275). Although the comprehensive multidimensional model Yinger and Courant hoped for has yet to be developed, empirical work by Daniel McMillen has successfully incorporated directional variables in the monocentric model. In his 149
4259 Ch05 4/12/02 9:58 am Page 150
150 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
FRED H. SMITH
article “One Hundred Fifty Years of Land Values in Chicago: A Non-parametric Approach,” McMillen augmented a data set of Chicago land values created by Hoyt and Olcott (McMillen, 1996, p. 100). Using the intersection of Madison and State as his “city center”, McMillen created “north” and “east” variables.8 The “north” variable identified a parcel’s location relative to Madison Street, and the “east” variable identified the parcel’s location relative to State Street. McMillen incorporated his directional variables into a third degree polynomial specification of the monocentric model. Then, McMillen graphed the estimated three-dimensional bid-rent surfaces. His graphs of Chicago clearly illustrate land value depressions on Chicago’s South and West Sides. Furthermore, the images depict an edge city near O’Hare airport in the northwest corner of the city. Although McMillen focused on econometric issues in the article, his work has provided me with a useful blueprint to follow. In the next section I discuss McMillen’s model as well as the data set that I collected to permit estimation of the model for twentieth century Cleveland.
16 17 18
IV. DATA AND REGRESSION MODEL
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
I collected assessed land value data from the Cuyahoga County Treasurer’s Tax Duplicates stored in the county archives. Specifically, I constructed a data set containing observations from the years 1915, 1922, 1930, 1940, 1950, 1960, 1970, and 1980. The data in the tax duplicates are arranged alphabetically before 1947, and by permanent parcel number thereafter. Therefore, I used two different collection procedures. The collection procedure for the data from 1915–1940 was more complicated, so I will describe it first. Prior to 1947, each person’s property is listed in the treasurer’s tax duplicates alphabetically by the owner’s name. Unfortunately, the property is only listed by the street that the parcel is on, and the full address of the parcel is not given. As a result, it is not possible to determine the location of the property directly from the duplicates. In order to pinpoint specific properties to use as observations in the data set, I was forced to first look at surveyor’s plat maps. These maps are drawn for every square foot of property in the city of Cleveland, bound in books, and organized by “map districts.”9 The “map districts” are simply an organizational tool that was developed to divide the city into sections systematically. I began my collection process for the 1915–1940 observations by looking at the plat maps for every map district in Cleveland. (The map districts are numbered so as to indicate their location in relation to the Cuyahoga River (east or west).) Each plat map book covers a certain set of map districts for a
4259 Ch05 4/12/02 9:58 am Page 151
Urban Decline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
151
given time period. The dates on the plat map books are spaced somewhat irregularly, but on average a complete new set of plat map books was created every eight to ten years. In a given plat map book, the surveyor’s plans for those map districts are laid out on the pages of the book. Thus, the individual lots, streets, railroad tracks, canals, rivers, and parks are visible on these maps. The name of the owner of each lot is written in ink on the map. When a property was sold or transferred, the former owner’s name was crossed out with a single pen stroke and the new owner’s name was written below it. The date of the sale or transfer was also recorded next to the new owner’s name. I took two different steps to ensure that the parcels I selected from the plat map books were chosen at random. First, I split the pages of the plat map books into groups. If I planned to take “n-1” observations from a given plat map book, then I would divide the pages in the book into “n” groups and take an observation from the last page of the first “n-1” of these groups. In order to save as much time as possible, I did this by estimating the number of pages in each group. It was necessary to do this, for each of the map books is so cumbersome (approximately three feet tall by three feet wide and weighing twenty pounds) that they are very difficult to work with. Next, I took steps to ensure that I didn’t take every observation from the center of the page. (Cochran, 1977) I divided each page into an imaginary three-by-three grid. For each plat map book I used I selected one of the nine cells of the grid at random, then I took the observations for that book from the chosen cell.10 I encountered several difficulties implementing my randomization procedures. After I had selected the pages that I would work with, as well as one of the nine cell locations, I was supposed to select parcels from the chosen pages and cells. This was not always an easy task. In order to record an observation, the name of the owner needed to be legibly written on the parcel. Furthermore, there could be no doubt about who the owner was for the time period in question. Unfortunately, the property transfers from one owner to the next did not always include the year of the sale/transfer, and there were many instances when no legible observation existed in the chosen cell on the chosen page. If an observation was available in an adjoining cell, then that observation was recorded. When it was necessary to move to another page, I did so.11 In general, the age of the plat maps and the quality of the original drawings led to more difficulty in collecting the 1915 and 1922 observations. The observations collected from the plat maps were not “complete” observations, though. The maps allowed me to identify the parcel’s location, owner, and size, but I still had no information on the assessed value of the land. In order to obtain this information I returned to the treasurer’s tax 151
4259 Ch05 4/12/02 9:58 am Page 152
152 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
FRED H. SMITH
duplicates. I was able to look up the owner’s name and identify the property I had selected by cross-referencing the lot size and the street listing. The collection procedure for the 1950–1980 observations was easier to implement. For each of these years the treasurer’s tax duplicates list properties by their permanent parcel number. The entry for each property includes the owner’s name, the address, the lot frontage, the lot depth (sometimes), and the assessed value of the land. The 1950, 1970, and 1980 data is kept in large books, and the parcels are grouped in the same map districts that are found in the plat map books for the years 1915–1940. In order to obtain a random sample, I once again divided the pages of each book into “n” groups if I intended to collect “n-1” observations. There were twenty parcels of land listed on each page of the tax duplicates. I used a random number generator (picking randomly from 1 to 20) to select from the twenty listings.12 As with the 1915–1940 observations, there were some difficulties with this procedure. Not every listing was complete. Many listings did not include a full address, so these were of no use to me. Similarly, many of the listings did not include a measurement for the depth of the parcel. When the chosen observation was of no use I alternated between using the observation above or below the randomly chosen observation. (The 1960 data was collected in a nearly identical fashion. The 1960 observations are on microfiche, so I randomly selected pages of microfiche and then I used the same procedure I have described above.) Archive data were missing only one key piece of information, the distance from each observation to the city center. Public Square has traditionally been the commercial heart of Cleveland, so I chose Public Square as the city center for my initial distance calculations. Using Microsoft’s Streets 98, I measured the straight-line distance from every observation to the center of Public Square. (These measurements have been recorded in hundredths of miles.) Table 4 contains descriptive statistics for the distance and land value data for each of the years in my sample. The data I collected in Cleveland are, largely, residential data. However, it is impossible for me to know how each parcel of land in the sample was being used. Individuals, not corporations of businesses, owned all of the parcels of land for which I collected data. The property records in the treasurer’s tax duplicates only indicate the name of the owner, but they do not indicate how the parcel of land is used. It is possible, even likely, that some of the parcels of land that I have in the data set contained a business and not a residential property. Ideally, I would have been able to identify how each parcel was being used. However, I have not found any evidence to suggest that industrial or commercial property was assessed differently
4259 Ch05 4/12/02 9:58 am Page 153
Urban Decline 1 2
153
than residential property. I used the data to estimate a modified specification of the Mills-Muth monocentric model:
3 4
ln R = 0 + 1 · m + 2
5 6 7 8 9 10
m2 + 3 100
m3 + · n 4 1000
2 3 + 5 (n · m) + 6 n · m + 7 n · m 1000 100
+ 8 · e
2 3 9 (e · m) + 10 e · m + 11 e · m 100 1000
(1)
11 12 13 14 15 16 17
Table 4 contains a full list of the variables included in Eq. (1). I have only listed the cubic specification of this model, for it fits the data better than any of the other specifications that I tried. Following McMillen’s lead, I will henceforth refer to the model in Eq. (1) as the spatial expansion model.13 The specification contained in Eq. (1) is nearly identical to the specification used by McMillen. I also tried variations of this specification. I created
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Table 4. 1915 Observations 61 Avg. value per 15.60 sq. foot – 1980 dollars Std. dev. of 44.74 value Avg. distance 2.64 from Public Square – Miles St. Dev. Of 1.72 distance
Descriptive Statistics and Variable List. 1922
1930
1940
86 12.14
86 21.87
96 10.50
35.26
92.41
2.92
1.60
1950
1960
1970
1980
121 1.80
99 2.10
116 1.56
125 0.54
36.41
6.67
6.32
3.46
0.68
3.39
3.47
4.00
3.75
3.49
4.05
1.98
2.16
2.29
2.16
2.09
2.11
Variable List LnR = The natural log of the market value of a square foot of land. (In 1980 dollars) For the years 1950 to 1980 the market value was created by multiplying each assessed value by 2.8571 – the average ratio of market to assessed value. m = Straight line distance from the parcel of land to Public Square (Measured in miles.) Mathematically: m = n2 + e2 n = How far north of Public Square the parcel of land is. (Measured in miles.) e = How far east of Public Square the parcel of land is. (Measured in miles.)
153
4259 Ch05 4/12/02 9:58 am Page 154
154 1 2 3 4 5
FRED H. SMITH
specifications with squared and cubed directional variables, and I also created specifications that accounted for a parcel’s distance to the airport. However, when I estimated these specifications they did not fit the data as well as the specification contained in Eq. (1).14 In the section that follows I report the results from estimating Eq. (1).
6 7
V. RESULTS
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
I estimated the spatial expansion model from Eq. (1) using ordinary least squares. The coefficient estimates are presented in Table 5. Several interesting trends are apparent. Distance, distance squared, and distance cubed are highly significant for every year in the data set. The coefficients on the directional variables do not follow this pattern. All but two of the directional terms have statistically significant coefficient estimates in 1950, and all of the coefficients from the 1980 regression are statistically significant. Most all of the coefficients on the directional variables are not significant in the 1915, 1922, 1940, and 1970 regressions. The 1930 regression coefficients for the north variables are significant, but the coefficients from the east variables are not. It is difficult to interpret the overall impact of the directional variables by looking at the coefficient estimates presented in Table 5. Therefore, I have graphed the bid-rent curves generated by these coefficients. The graphs of these bid-rent curves appear in Figs 2 through 10. The bid-rent “curves” contained in these figures are three-dimensional surfaces that I generated with Mathematica. Mathematica allows the user to change the perspective from which one views a Mathematica graph. In Fig. 2 through Fig. 9, I have graphed bid-rent surfaces for each year in the data set. The graphs in each of the figures depict a twelve mile wide (east-west) by eight mile deep (north-south) picture of land values in Cleveland. The dependent variable is the natural log of the “market” value of a square foot of land (in 1980 dollars).15 Each of the eight figures offers two different perspectives of the bid-rent surface. The top graph shows the default Mathematica viewpoint, and the lower graph puts the viewer directly to the south of the city center. To facilitate inter-year comparisons I created Fig. 10. Figure 10 presents one graph for each year in the data set. The viewer is, once again, directly to the south of Public Square. However, unlike the lower graph in Figs 2 through 9, the viewer has been brought to “ground level.” In other words, the viewer is looking at the bid-rent surface “head-on” instead of from a slightly elevated position. I have used the bid-rent functions generated from the spatial expansion model to quantify the impact of urban decay. I have done this on a citywide basis as
4259 Ch05 4/12/02 9:58 am Page 155
Urban Decline
155
Table 5.
1
Spatial Expansion Regressions 1915–1980.
2 m + · n + (n · m) + n · m 100m + 1000 100 m m + n · + · e + (e · m) + e · + e · m 1000 100 1000
3
2
ln R + 0 + 1 · m + 2
4
3
2
3
4
3
5
6
2
7
6
5
8
9
10
3
11
7 8 9
Variable M
10 11 12 13 14
(m2/100) (m3/1000) n
15 16 17 18 19
n*m n*(m2/100) n*(m3/1000)
20 21 22 23 24
e e*m e*(m2/100)
25 26 27 28 29
e*(m3/1000) Constant R2
1915
1922
1930
6.36*
4.58*
5.41*
(1.06) 243.21* (57.04) 286.47* (85.52) 1.25 (0.94) 1.01 (0.88) 22.88 (25.89) 15.99 (23.25) 1.07 (0.79) 0.15 (0.64) 16.88 (17.38) 34.14‡ (17.74) 5.37* (0.41) 0.77
(0.80) 161.55* (37.96) 181.05* (48.71) 1.42 (1.05) 1.71‡ (1.01) 47.63 (30.27) 39.38 (28.23) 0.65 (0.85) 0.81 (0.68) 29.64 (18.76) 32.80‡ (17.37) 4.86* (0.34) 0.73
1940
3.77* (0.58) (0.48) 165.50* 103.81* (22.91) (18.61) 150.91* 88.50* (24.22) (19.46) 1.76† 1.23† (0.71) (0.63) 1.39* 0.74 (0.55) (0.48) 38.64* 18.12 (14.05) (11.65) 33.87* 14.54 (11.40) (9.11) 0.06 0.07 (0.55) (0.49) 0.00 0.05 (0.40) (0.35) 0.86 0.93 (9.75) (8.19) 1.00 0.73 (7.64) (6.22) 5.53* 4.56* (0.33) (0.28) 0.76 0.79
1950
1960
1970
1980
2.07*
3.29*
2.17*
1.84*
(0.28) 44.94* (8.34) 31.31* (7.01) 1.21* (0.46) 0.62† (0.29) 10.73‡ (6.12) 5.82 (4.08) 0.78† (0.39) 0.50† (0.25) 9.57‡ (5.32) 5.82 (3.58) 3.58* (0.27) 0.60
(0.37) 79.75* (11.92) 58.47* (10.37) 0.62 (0.58) 0.54 (0.39) 13.51 (8.58) 9.73 (6.01) 0.40 (0.45) 0.51‡ (0.27) 14.37* (5.63) 11.13* (3.91) 4.35* (0.28) 0.70
(0.27) 52.74* (9.00) 38.44* (8.01) 0.40 (0.38) 0.21 (0.24) 3.74 (4.74) 2.18 (2.98) 0.04 (0.26) 0.09 (0.16) 3.95 (3.20) 3.79‡ (2.15) 3.26* (0.19) 0.65
(0.22) 43.66* (6.56) 31.24* (5.49) 0.88† (0.37) 0.54† (0.22) 9.68† (4.35) 5.37† (2.69) 6.05† (0.28) 0.30‡ (0.17) 6.22‡ (3.43) 4.39† (2.21) 2.44* (0.20) 0.65
30 31 32
Standard Errors are in parentheses. Significant at the 99% level: *, at the 95% level: †, at the 90% level: ‡.
33 34 35 36 37 38 39 40
well as a neighborhood by neighborhood basis. Specifically, I am using aggregate assessed land values as a proxy to indicate the economic health of the city and its individual neighborhoods. Aggregate assessed land values are not an ideal proxy for a city’s economic health. A city could experience a tremendous decline in land values without experiencing urban decay. However, the historical evidence presented in Section II makes it clear that Cleveland did experience severe urban decay. By examining the aggregate assessed land values 155
4259 Ch05 4/12/02 9:59 am Page 156
156
FRED H. SMITH
1 2 3 4 5 6 7 8 6
9 10 11 12
4 LnValue
13
2 2 0
0
14 15
North
–5
16
–2
–2.5
17
0
18
2.5
East
19
–4 5
20 21 22 23 24 25 North
26 27
0 2 –4 –2
28
6
29 30 31 32 33 34
LnValue 4 2 0
35 –5
36 37 38 39 40
Fig. 2.
–2.5
0 East
2.5
Three Dimensional Bid-Rent Curve – 1915.
5
4259 Ch05 4/12/02 9:59 am Page 157
Urban Decline
157
1 2 3 4 5 6 7 8 9
6
10
4
11 12
LnValue
13
2 2 0
14 15
0 North
–5
16
–2
–2.5
17
0
18
2.5
East
19
–4 5
20 21 22 23 24 25 26
North
27
0 2 –4 –2
28 29 30 31
6 LnValue
4
32 33
2
34 35
0
36
–5
–2.5
0
2.5
37 East
38 39
Fig. 3.
Three Dimensional Bid-Rent Curve – 1922.
40
157
5
4259 Ch05 4/12/02 9:59 am Page 158
158
FRED H. SMITH
1 2 3 4 5 6 7 8 6
9 10
4 2
11 12
LnValue
13
2 0
0
14 North
–5
15
–2
–2.5
16
0
17 18
2.5
East
19
–4 5
20 21 22 23 24 25 North
26
0 2 –4 –2
27 28 6
29 30 31 32
LnValue 4 2
33 34
0
35 –5
36 37 38 39 40
Fig. 4.
–2.5
0 East
2.5
Three Dimensional Bid-Rent Curve – 1930.
5
4259 Ch05 4/12/02 9:59 am Page 159
Urban Decline
159
1 2 3 4 5 6 7 8
6
9
4
10 11
2 LnValue
12
2 0
0
13
–5
14
North –2
–2.5
15
0
16
2.5
17
–4
East
5
18 19 20 21 22 23 24 25
North
26
0 2 –4 –2
27
6
28 29 30
LnValue 4
31 32
2
33
0
34
–5
35
–2.5
0
2.5 East
36 37 38
Fig. 5.
Three Dimensional Bid-Rent Curve – 1940.
39 40
159
5
4259 Ch05 4/12/02 9:59 am Page 160
160
FRED H. SMITH
1 2 3 4 5 6 7 6
8 9 10
4 LnValue
11 12
2
2 0
0
13
–5
14
North –2.5
15
–2 0
16
2.5
17
–4
East
5
18 19 20 21 22 23 24 North
25 26
0 2 –4 –2
27
6
28 29
LnValue 4
30 31
2
32 33
0
34 –5
35
–2.5
0
2.5 East
36 37 38 39 40
Fig. 6.
Three Dimensional Bid-Rent Curve – 1950.
5
4259 Ch05 4/12/02 9:59 am Page 161
Urban Decline
161
1 2 3 4 5 6 7 8 6
9 10 11 12
4 LnValue
13
2 2 0
0
14 15
North
–5
16
–2
–2.5 0
17 18
2.5
East
19
–4 5
20 21 22 23 24 25
North
26
0 2 –4 –2
27 28 29 30
6 LnValue 4
31 32
2
33 34
0
35 –5
36
–2.5
0 East
37 38
Fig. 7.
2.5
Three Dimensional Bid-Rent Curve – 1960.
39 40
161
5
4259 Ch05 4/12/02 9:59 am Page 162
162
FRED H. SMITH
1 2 3 4 5 6 7 8
6
9
4
10 11
LnValue
2 2
12 0
13 14
0 North
–5
15
–2
–2.5
16
0
17 18
2.5
East
–4 5
19 20 21 22 23 24 25
North
26
0 2 –4 –2
27 28 29 30
6 LnValue 4
31 32
2
33 34
0
35
–5
36
–2.5
0 East
2.5
37 38 39 40
Fig. 8.
Three Dimensional Bid-Rent Curve – 1970.
5
4259 Ch05 4/12/02 9:59 am Page 163
Urban Decline
163
1 2 3 4 5 6 7 8 6
9 10 11 12
4 LnValue
13
2 2 0
0
14 North
–5
15
–2
–2.5
16
0
17 18
2.5
East
19
–4 5
20 21 22 23 24 25 North
26
0 2 –4 –2
27 28
6
29 30 31 32
LnValue 4 2
33 34
0
35 –5
36
–2.5
37 38
Fig. 9.
0 East
2.5
Three Dimensional Bid-Rent Curve – 1980.
39 40
163
5
4259 Ch05 4/12/02 9:59 am Page 164
164
FRED H. SMITH
1 2 3
1915
6
4 5
4
6 7
LnValue 2
8 9 10
0
11
–5
12
–2.5
13
2.5
5
2.5
5
2.5
5
East
14 15
0
1922
6
16 17 4
18 19
LnValue
20
2
21 22 0
23 24
–5
–2.5
0
25 East
26 27 28
1930
6
29 30
4
31 32 33
LnValue 2
34 35
0
36 37 38 39 40
–5
–2.5
0 East
Fig. 10. Three Dimensional Bid-Rent Curves 1915–1980.
4259 Ch05 4/12/02 10:00 am Page 165
Urban Decline
165
1 2 3
1940
6
4 5
4
6 7
LnValue 2
8 9 10
0
11
–5
12
–2.5
0
13
2.5
5
2.5
5
2.5
5
East
14 15
1950
6
16 17 4
18 19
LnValue
20
2
21 22 0
23 24
–5
–2.5
0
25
East
26 27 28
1960
6
29 4
30 31 32 33
LnValue 2
34 35 36
0 –5
–2.5
0
37 East
38 39
Fig. 10. Continued.
40
165
4259 Ch05 4/12/02 10:00 am Page 166
166
FRED H. SMITH
1 2
1970
6
3 4 5 6
4 LnValue 2
7 8 9
0
10
–5
11
–2.5
12
0
2.5
5
2.5
5
East
13 14
1980
6
15 16 4
17 18
LnValue
19
2
20 21 0
22 23
–5
–2.5
0
24 East
25
Fig. 10. Continued.
26 27 28 29
Table 6.
30 31 32 33 34 35 36 37 38 39 40
Aggregate Assessed Value of Land in Cleveland.
Year
Aggregate Assessed Value
1915 1922 1930 1940 1950 1960 1970 1980
$4,048,030,000 $5,786,810,000 $5,714,130,000 $4,774,080,000 $5,619,050,000 $4,568,460,000 $5,049,060,000 $3,437,740,000
All values are in 1980 dollars.
9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40 467.7 325.2 313.0 327.8 588.6 681.0 714.3 475.7
8,898.7 20,081.6 8,688.5 11,037.8 18,282.3 17,719.1 14,440.3 9,750.1
Ohio City 451,401.0 342,080.0 599,071.0 284,568.0 137,682.0 245,089.0 97,576.9 45,431.8
Downtown 41,311.7 34,122.0 21,361.3 27,581.4 32,870.8 37,566.9 41,024.8 13,709.3
CedarCentral 19,792.3 30,952.3 16,539.1 15,667.0 23,174.6 17,705.3 21,348.4 7,126.4
Hough
5,975.1 5,615.0 12,626.4 8,243.2 9,552.1 5,765.2 7,716.5 2,561.8
Glenville
6,799.3 3,602.3 6,228.2 4,243.2 8,368.0 4,329.6 8,054.1 3,246.0
Shaker Square
1
167
All values are in thousands of 1980 dollars.
0.1 23.0 31,848.0 28,910.3 37,467.9 74,523.5 84,931.2 86,770.0
10
1915 1922 1930 1940 1950 1960 1970 1980
8
Dutch Hill
3
Rocky RiverLorain
2
Year
7
Aggregate Assessed Value of Land in Selected Cleveland Neighborhoods.
6
Table 7.
4259 Ch05 4/12/02 10:00 am Page 167
Urban Decline 167
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
38 0.96 0.67 0.64 0.67 1.21 1.40 1.46 0.97
All values are in thousands of 1980 dollars.
0.00 0.00 0.87 0.79 1.03 2.05 2.33 2.38
37
1915 1922 1930 1940 1950 1960 1970 1980
40 Dutch Hill 1.84 4.15 1.80 2.28 3.78 3.67 2.99 2.02
Ohio City 85.51 65.03 113.88 54.09 26.17 46.57 18.55 8.64
Downtown 2.80 2.31 1.45 1.87 2.23 2.55 2.78 0.93
CedarCentral 1.74 2.72 1.45 1.38 2.04 1.56 1.88 0.63
Hough
1.55 1.45 3.27 2.13 2.47 1.49 2.00 0.66
Glenville
1.69 0.90 1.55 1.06 2.08 1.08 2.01 0.81
Shaker Square
3
Rocky RiverLorain
2
Year
39
Average Assessed Value of a Square Foot of Land in Selected Neighborhoods.
5
Table 8.
4259 Ch05 4/12/02 10:00 am Page 168
168 FRED H. SMITH
1
4
4259 Ch05 4/12/02 10:00 am Page 169
Urban Decline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
169
it becomes possible to gain a sense of how profoundly the decay affected the city.16 Table 6 presents aggregate assessed land values for the city of Cleveland. However, the real advantage of using the spatial expansion model becomes evident when looking at Table 7. Table 7 contains aggregate assessed land values for eight neighborhoods in Cleveland. It would be impossible to do this type of calculation using a bid-rent function generated from the standard specification of the monocentric model, for such a function would not allow me to differentiate between neighborhoods. The spatial expansion model, however, has no difficulty pinpointing land values anywhere in the city. In order to make the aggregate assessed land values for the neighborhoods easier to interpret, I created Table 8. This table transforms the aggregate assessed land values into the average value of a square foot of land in each neighborhood. For example, the average value of a square foot of land in the Dutch Hill neighborhood in 1915 was $0.96. I obtained this value by dividing $467,717, the aggregate assessed value of land in Dutch Hill in 1915, by 487,872, the number of square feet in the Dutch Hill neighborhood.17
18 19 20
VI. DISCUSSION
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
The results presented in Section V illustrate the major changes in Cleveland’s urban spatial structure. Taken out of historical context these results are interesting, but they are not particularly meaningful. Therefore, in this section I will tie the results from Section V together with Cleveland’s history. The evolution of Cleveland’s spatial structure is most clearly seen in the bidrent surfaces depicted in Figs 2 through 10. The 1915 bid-rent surface is particularly fascinating. It shows a remarkable increase in land values at the far eastern edge of the city. The edge of the box that is graphed depicts land values for areas that reside outside of the city of Cleveland. Thus, the increase in land values on the far eastern edge of the box suggests that suburbs were forming in Cleveland. Of course, this was what actually happened. The Van Sweringen brothers, who later developed the Terminal Tower complex, had purchased vast tracts of land roughly 5.5 miles east-southeast of the city. They purchased the land and created a planned community, now known as Shaker Heights. After purchasing the land for Shaker Heights in 1905, the Van Sweringens secured streetcar access for their new community in 1913 (Toman, 1996, p. 109). Shaker Heights quickly became an attractive alternative for city dwellers who wished to move away from hectic city. Therefore, it is not surprising that the land values at the eastern edge of the city were increasing 169
4259 Ch05 4/12/02 10:00 am Page 170
170 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
FRED H. SMITH
in 1915. What is somewhat more surprising is that this trend is somewhat less evident in the 1922 bid-rent surface. By 1922 the Van Sweringens had purchased the Nickel Plate Railroad. They made this purchase so that they could use the Nickel Plate’s right-of-way into downtown (Toman, 1996, p. 109). The right-of-way allowed the Van Sweringens to build a rapid transit line that ran from Shaker Heights directly to downtown Cleveland. Unlike the streetcar that initially serviced Shaker Heights, the rapid transit line was required to make relatively few stops. The rapid transit line went into service in 1920, so I would have expected the 1922 bid-rent surface to exhibit the same properties as the 1915 bid-rent surface. However, two things should be noted. First, the 1922 bid-rent surface does display higher land values east-southeast of the city than it does to the northeast or south-southeast. Thus, the influence of Shaker Heights has not dissipated entirely. Second, the curvature at the eastern edge of the 1915 bid-rent surface is very dramatic and nearly uniform along the entire edge. These facts suggest that the curvature may be an artifact of the functional form underlying the bid-rent surface.18 The 1930 bid-rent surface (different viewpoints are available in Figs 4 and 10) displays two interesting differences from the 1922 bid-rent surface. First, the effects of suburbanization are becoming clear. The value of land on the far western edge of the city increased drastically between 1922 and 1930. In the 1922 bid-rent surface the western edge of the city is clearly visible; land values literally plummet 4.5 miles from the city. By 1930 the western edge of the city is no longer obvious. Second, the land value trough encircling the city in the 1922 bid-rent surface has become shallower in the 1930 bid-rent surface. Both the 1922 and 1930 surfaces differ substantially from the 1915 bid-rent surface in this regard. The 1915 surface exhibits a trough on the near West Side, but the 1922 and 1930 surfaces have troughs that literally encircle the city. The depth of the 1915 trough is of particular interest. The trough is located over an area of the city commonly known as the Flats. The Flats are so named because the land, located on the banks of the Cuyahoga River, is “flat” relative to the steep cliffs that rise a few hundred yards inland from either bank. In 1915, the Flats were the home of countless industrial firms. As a result, the land values I recorded from this area appear to have been very low. The firms located in the Flats produced tremendous amounts of pollution; living near the river would have been very unpleasant. The residents who did live near the Flats were immigrants. In fact, an area just to the west of the Flats, Whiskey Island, was one of the largest Irish enclaves in the city (Miller, 1997, p. 54). It is unlikely that the pollution generated by the firms located in the Flats had decreased substantially by 1922. So, if pollution depressed the near West
4259 Ch05 4/12/02 10:00 am Page 171
Urban Decline 1 2 3 4 5 6 7 8 9 10 11
171
Side land values in 1915, then some other factor caused them to increase between 1915 and 1922. One possibility is that land values increased because of a change in demand for goods caused by the First World War. The land value trough encircling the city in the 1922 and 1930 bid-rent surfaces may be a result of center city residents moving to the outskirts of the city. The neighborhoods close to the city center, Central-Cedar on the East Side and Ohio City on the West Side, contained older homes and were closer to the pollution and noise of downtown Cleveland. Thus, they may have been less attractive places to live than the quieter, less polluted neighborhoods at the edge of the city. There is, however, another factor to be considered. The poet Langston Hughes lived in Cleveland at this time, and he wrote about his experience:
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Rents were very high for colored people in Cleveland, and the Negro district was extremely crowded, because of the great migration. It was difficult to find a place to live. We always lived, during my high school years, either in an attic or basement, and paid quite a lot for such inconvenient quarters. White people on the east side of the city were moving out of their frame houses and renting them to Negroes at double and triple the rents they could receive from others. An eight-room house with one bath would be cut up into apartments and five or six families crowded into it, each two-room kitchenette apartment renting for what the whole house had rented for before (Miller, 1997, p. 123).
Hughes paints a bleak picture of the near East Side of Cleveland. Yet the implications from Hughes’ account for land values are somewhat ambiguous. The price for a parcel of land should be the present discounted value of the stream of rents that it will command. Assuming Hughes’ account of East Side rents is accurate, then housing values and land values should have increased dramatically. However, this analysis fails to account for all of the factors that would affect the stream of rents commanded by the home and land. A house that has been split into several apartments will wear out relatively quickly. In the future, it is unlikely that this house will be able to command the same level of rent. This reduces the present discounted value of the stream of rents earned by the asset, which in turns lowers the price of the house. The value of the land may also fall if the neighborhood becomes a less attractive place to live or work. There is another issue that must be considered, though. If the high rents described by Hughes persist, then even racist homeowners will have an economic incentive to sell or rent their home to a black family. As a result, the supply of housing available to black families would increase and the price of housing would fall. Indeed, this is what ultimately happened in the Cleveland neighborhood of Hough. Realtors engaged in “blockbusting” during the 1950s (Keating, 1995, p. 44). In other words, realtors encouraged a black family to move into a home in Hough so that the white residents on that street would quickly move before their home “lost” all of its value. 171
4259 Ch05 4/12/02 10:00 am Page 172
172 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
FRED H. SMITH
Ultimately, it is hard to predict what the impact of racism on land values “should” be from a theoretical standpoint. As I noted earlier, Yinger and Courant have addressed this issue within the framework of the basic monocentric model. Their model does not apply to the spatial expansion model, so it is of little help in determining what happened to land values in the Cleveland neighborhoods that are predominantly black. The trough in the 1930 bid-surface appears to be a bit deeper southeast of the downtown peak. Central-Cedar lies to the southeast of downtown, so it is possible that racism played a role in depressing land values in this area of the city. The evidence is far from conclusive, though. In fact, a careful examination of Table 8 shows that between 1922 and 1930 the average value of land in Ohio City, a nearly all white neighborhood on the near West Side, fell far more dramatically than the average value of land in Central-Cedar. Cleveland’s pre-eminent expert on urban land values, Robert Whipple Green, addressed the issue of race and property values in 1933. He wrote that “An enormous increase in the number of Negroes in a particular section of the city with their influence on property values is always reflected sooner or later in mortgage finance” (Miller, 1997, p. 135). Mr. Green was certainly correct in asserting that blacks faced serious barriers when they tried to obtain mortgages. As I discussed in Section II, one of the city’s leading financial institutions had already adopted a formal policy of restricting credit to blacks. However, Green’s assertion that black families caused property values to decrease is not borne out by the bid-rent surfaces from any year between 1915 and 1970. The issue of how race may have affected land values takes on new significance in 1980. The 1980 bid-rent surface is tilted to the right. The tilting of the bid-rent surface first becomes visible in the 1970 graph, but by 1980 the trend is very clear. Land values on the East Side of Cleveland were almost uniformly lower than land values on the West Side. This means that land values were lower in black neighborhoods than in white neighborhoods because by 1980 the East Side of Cleveland was largely black. Table 9 shows the percent of the inhabitants who were black in the census tracts of the eight neighborhoods I have selected to examine in this chapter. Ohio City, Dutch Hill, and Rocky River are all West Side neighborhoods.19 Ohio City exhibited a dramatic increase in its African-American population between 1970 and 1980, but otherwise the West Side neighborhoods were almost entirely white between 1915 and 1980. This stands in direct contrast to the experiences of the East Side neighborhoods. By 1980, Central-Cedar, Hough, and Glenville were nearly all black. The one East Side neighborhood that did not become entirely black by 1980, Shaker Square, sits flush against the suburb
4259 Ch05 4/12/02 10:00 am Page 173
Urban Decline 1
Table 9.
173
Percent Black by Neighborhood, 1910–1980.
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Neighborhood (Census Tract) Ohio City C3 C6 C7 Dutch Hill D6 D8 Downtown G1 G6 G7 Central-Cedar G9 H7 H8 H9 I3 I6 I7 I8 Hough L4 L5 L6 R9 Glenville R5 Shaker Square S5 Rocky River – Lorain W1 W2 W4 W5
1910
1920
1930
1940
1950
1960
1970
1980
0.0 0.2 1.0
4.2 0.8 3.1
10.0 0.0 0.0
1.2 0.4 4.1
1.1 0.6 6.9
2.5 0.5 1.1
19.3 1.7 21.3
59.6 5.5 41.7
N.A N.A
N.A N.A
0.0 0.0
0.0 0.2
0.0 0.0
0.0 0.9
0.0 0.2
2.1 0.9
8.3 2.3 5.1
16.2 2.5 3.8
0.0 0.0 0.0
2.0 0.2 0.1
1.0 0.2 0.0
0.0 0.0 1.6
12.9 2.7 0.7
62.4 40.8 40.7
N.A N.A N.A N.A 24.4 5.9 0.4 N.A
18.6 45.7 28.5 50.8 62.6 42.0 38.5 20.9
18.0 55.0 52.0 89.0 67.0 53.0 76.0 76.0
3.9 56.0 65.6 92.9 72.6 56.8 83.5 92.9
13.8 72.1 71.2 96.8 89.2 87.4 97.7 98.6
20.6 83.4 76.2 98.0 68.9 95.4 94.4 98.9
55.1 63.8 63.9 99.4 68.7 3.621 95.9 99.5
80.2 62.3 70.9 98.7 85.9 79.3 98.1 99.4
0.2 0.5 5.6 1.3
1.0 1.1 8.4 1.0
0.0 0.0 9.0 0.0
1.5 0.7 7.7 0.2
4.0 4.5 11.1 0.6
91.3 88.8 88.6 43.9
98.7 97.8 98.4 87.0
98.6 96.4 98.6 90.4
N.A
5.0
0.0
4.8
7.1
88.4
99.1
99.3
N.A
0.0
0.0
1.0
0.4
0.5
3.4
9.7
N.A N.A N.A N.A
N.A N.A N.A N.A
0.0 0.0 0.0 0.0
0.1 0.0 0.0 0.0
0.1 0.0 0.0 0.0
0.0 0.0 0.0 0.0
0.0 0.0 0.0 0.0
0.1 0.1 0.1 2.0
Source: U.S. Department of Commerce.
35 36 37 38 39 40
of Shaker Heights. The residents of Shaker Heights became very involved in the process of integration, for they wished to avoid the dramatic demographic changes experienced in Hough, Glenville and the suburbs of East Cleveland and Warrensville Heights (Keating, 1994). Thus, it is not surprising that the Shaker Square neighborhood did not become entirely black. 173
4259 Ch05 4/12/02 10:00 am Page 174
174 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
FRED H. SMITH
While there appears to be a correlation between the racial composition of the East Side neighborhoods and the reduction of land values that occurred between 1970 and 1980, this correlation raises more questions that it answers. If the racial composition of a neighborhood truly affected land values, then why didn’t land values in Hough and Central-Cedar decline before the 1970s? Furthermore, why didn’t the land values in Ohio City, a West Side neighborhood that contained census tracts that were more than 50% black by 1980, fall as dramatically as the land values in the East Side neighborhoods? One possible explanation for why urban decay hit the East Side of Cleveland harder than the West Side might be that the two sides of the city received very different levels of city services. The city of Cleveland was sued in 1976 for failing to provide integrated schools (Keating, 1995, p. 129). This was the culmination of decades of deterioration in the Cleveland Public School system. As I previously mentioned in Section II, funding problems had caused many Cleveland public schools to reduce their course offerings and to operate on a half-day schedule in the 1960s. These facts suggest that it is likely that the residents of the East Side neighborhoods had schools of poorer quality than their West Side counterparts, but there is not an obvious way to measure this difference quantitatively. Measuring the provision of other city services is an equally difficult task. For example, it is impossible to determine if garbage collection, fire protection and police protection were provided equally to all neighborhoods in Cleveland. Kenneth Kusmer has documented the inadequate city services provided to black Clevelanders early in the century, and one may assume that these types of problems persisted throughout the 1940s, 50s and 60s, but I have not found any direct proof of this (Keating, 1995, pp. 268, 273) Ultimately, the reduction in East Side land values is a topic that requires further study. A trend clearly depicted by the bid-rent surfaces that is far easier to explain is the decay of downtown Cleveland. Table 7 confirms what is evident from the bid-rent surfaces: The value of land in downtown Cleveland fell precipitously. After reaching a peak aggregate value of nearly $600 million in 1930, the land in downtown Cleveland was worth 1/12 of that amount 50 years later. As I discussed in Section II, the reasons for this decline are unambiguous. The emergence of the automobile as the primary form of transportation in Cleveland reduced the importance of being near downtown. Once the retail and entertainment heart of the city, by 1970 the area around Public Square and upper Euclid Avenue was the home to abandoned department stores and theaters. The decline in downtown land values proceeds almost without interruption from 1930 to 1980. The slight surge in downtown land values in 1960 may have been caused by the impending completion of the Erieview project and the
4259 Ch05 4/12/02 10:00 am Page 175
Urban Decline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
175
recent completion of the rapid transit line connecting the Terminal Tower complex to both the East and West Sides of the city. Nonetheless, by 1970 downtown land values were declining once again. There are two additional trends depicted in Table 7 that are worth noticing. First, the value of land in the Rocky River neighborhood increases dramatically between 1950 and 1980. This was undoubtedly caused by the mass migration of residents out of the central city. Rocky River lies within the Cleveland city limits, yet it is as far to the west as one can travel and still be in Cleveland. It is interesting to note that not one of the other eight neighborhoods I have chosen to study exhibits a similar increase in land values. A second trend exhibited in Table 7 is a reduction in land values between 1930 and 1940 in all of the neighborhoods except Dutch Hill, Ohio City, and Central-Cedar. These neighborhoods are older than Glenville, Hough, Rocky River and Shaker Square, and they are also closer to downtown Cleveland. One possible explanation for these neighborhoods escaping the Great Depression relatively unscathed is that the demand for housing in older, well-established neighborhoods may have been less income elastic than the demand for housing on the outskirts of the city. Since it was middle and upper class families leaving the central city for the outskirts of the city and suburbs, it is possible that these families decided against moving out of the central city until their financial situation had improved. This theory is supported by evidence from the census. In 1920, the city of Cleveland contained 85% of Cuyahoga County’s total population. A decade later that figure had dropped to 75%. Then, the figure stayed fairly constant during the 1930s, falling to just 72% by 1940. (U.S. Department of Commerce, 1942) The story told by the bid-rent surfaces and Table 7 is clear: Downtown Cleveland began a losing battle with urban decay in 1930, and by 1970, urban decay had spread from downtown to the entire East Side. Returning to Table 6 allows me to complete the story of urban decay in Cleveland. Citywide aggregate land value reached a peak in 1930 and, with the exception of the dip in 1940 undoubtedly caused by the depression, it remained nearly constant until 1960. In 1960, despite the resurgence in downtown land values, the aggregate value of land in Cleveland was a billion dollars less than it had been in 1950. After a slight rebound in 1970, land values reached a 65-year low in 1980. Over the course of three decades the aggregate value of land in the city of Cleveland fell 39%. The aggregate land values presented in this paper tell a powerful story, but I have not yet addressed the issue of whether these values are truly meaningful. Because the aggregate values were calculated using the estimated bid-rent functions from the spatial expansion model they may be challenged on the 175
4259 Ch05 4/12/02 10:00 am Page 176
176 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
FRED H. SMITH
grounds that the spatial expansion model does not accurately describe urban spatial structure in Cleveland. It is true that the spatial expansion model does not fit my Cleveland data as well as it fits the Hoyt and Olcott Chicago data (McMillen, 1996, p. 123). However, two factors may be influencing this result. First, the Hoyt and Olcott data are not “micro-level” data. The observations in these data sets are average land values from individual square miles (Hoyt) or city blocks (Olcott). As McMillen points out, this will cause an increase in the explanatory power of the model. Second, I have not accounted for the impact of edge cities in my spatial expansion model. This is a concern that I attempted to address by including variables that accounted for each parcel’s distance to each of Cleveland’s two edge cities and Hopkins Airport. However, the coefficient estimates for these variables were not statistically significant when they were included in the regression. Nonetheless, this is an issue that merits further attention. If the spatial expansion model is accepted as an appropriate model for Cleveland’s spatial structure, then my estimates of aggregate land values in Cleveland become more difficult to challenge. This is especially true since I uncovered a piece of supporting evidence in Olson and Lachman’s Tax Delinquency in the Inner City. They report that the assessed value of property in Cleveland in 1960 was roughly 1.7 billion dollars. By translating this number into a market value, my estimate of 4.6 billion compares favorably to their “true” value of 4.85 billion (Olson, 1976, p. 17).
22 23
VII. CONCLUSION
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Urban decay decimated the city of Cleveland in the 1970s. While historians have documented Cleveland’s decay in numerous books and essays, to the best of my knowledge there has been no work attempting to quantify this decay. By assembling an original data set, I have been able to take the first step in closing this gap in the literature. I have estimated a modified specification of the monocentric model in order to describe Cleveland’s spatial structure. The bid-rent surfaces generated by my estimates have allowed me to calculate aggregate assessed land values for the entire city as well as for eight of its neighborhoods. Using the aggregate assessed land values as a proxy for the economic health of the city I have been able to quantify the decay that afflicted Cleveland: The aggregate value of land in Cleveland decreased by 39% between 1950 and 1980, and it decreased a remarkable 32% in the decade from 1970 to 1980. While I have taken a step towards quantifying urban decay in Cleveland, and I am able to offer plausible explanations for the urban decay that afflicted Cleveland’s central city, I am unable to provide an explanation for the causes
4259 Ch05 4/12/02 10:00 am Page 177
Urban Decline 1 2 3 4
177
of the decay in Cleveland’s East Side neighborhoods. This is a subject that merits a great deal of additional research, but it is my hope that the results presented in this paper will be able to serve as a foundation for those efforts.
5 6 7
NOTES
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
1. “Western Reserve” is the title given to lands originally held by the state of Connecticut in the Northwest Territory. The Western Reserve was sold to a group of thirty-five investors in 1795. 2. The city’s name was first spelled “Cleaveland”, but the common spelling “Cleveland” seems to have been adopted in the early 1800s. 3. Ridership consistently exceeded 400 million during the 1940s because of the Second World War. 4. Although this was a phenomenon that occurred citywide, it was a serious problem in the Central neighborhood. 5. This is true in the absence of market failure. 6. One must be careful when making conclusions based on casual observation. A vacant lot in downtown Chicago remained undeveloped throughout most of the 1990s, yet it would be incorrect to assume that the parcel of land was worthless. 7. The Mills-Muth model is one of two models developed in the 1960s. W. Alonso also developed a monocentric model that is virtually identical to the one created by Mills and Muth. 8. McMillen’s selection of State and Madison as his city center is not accidental. Addresses in Chicago also use this as the city center for the purpose of identifying east, west, north, and south. 9. Cleveland annexed West Park in 1922. After 1922 the city’s borders remained constant. 10. I named all nine cells (top right, top center, etc.), and created one notecard for each of them. Before selecting the first plat map book to work with, I drew a single notecard from the shuffled stack. This was the cell location I used throughout that book. I replaced the card after drawing it, and I repeated this process every time I selected a new plat map book to work with. 11. When the chosen cell did not produce an observation I looked at the cell immediately to the right of the chosen cell. If that produced an observation, then I used it. The next time the chosen cell did not yield an observation I looked in the cell below and to the right of the chosen cell. If there was no cell below and to the right of the chosen cell then I moved in a clockwise direction to the first cell that did exist. If no observation could be taken from that page, then I turned to the next page to find an observation. The next time I encountered a page that produced no observation I turned to the preceding page to find an observation. 12. Prior to working with the first book I labeled twenty note cards with numbers from one to twenty. Next, I selected one card at random. The number on the card was the observation that I selected from each of the chosen pages of that book. Before working with a new book I replaced the previously drawn card, then I selected a new card. 177
4259 Ch05 4/12/02 10:00 am Page 178
178 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
FRED H. SMITH
13. This functional form produces biased estimates of land value, but the functional form that generates unbiased estimates can produce unrealistic estimates. Specifically, it can produce estimates that suggest that land has a value of less than zero. 14. McMillen also created variables that account for a parcel’s distance from Lake Michigan, O’Hare Airport, and Midway Airport. He created these variables to determine if land values were impacted by amenities other than the city center. Although I created a variable measuring each parcel’s distance from Hopkins International Airport, I did not create a variable measuring each parcel’s distance from Lake Erie. This decision was partially dictated by practical concerns. The shore of Lake Erie is highly non-linear, thus each parcel’s distance from Lake Erie would have had to have been measured independently. However, from a theoretical standpoint, this is of little concern. The shore of Lake Erie does not provide Cleveland with the same quantity of recreational opportunities that the shore of Lake Michigan provides Chicago. In addition to measuring each parcel’s distance from Hopkins Airport, I also measured each parcel’s distance from edge cities identified by Joel Garreau. Garreau identified an edge city at the intersection of I-77 and Rockside Road as well as at the intersection of I-271 and Chagrin Boulevard. 15. I have recreated “market” values for the years 1950–1980 by multiplying the assessed values by 2.8571. This is an average of the ratio of market value to assessed value over the three decades. 16. The process I used to calculate the aggregate assessed land values using the spatial expansion model is rather complicated. Appendix A, located at the end of this paper, includes a complete explanation of the process. 17. It should be noted that my “neighborhoods” are rectangles entirely of my own construction. I have tried to remain faithful to the actual boundaries of these neighborhoods (as identified by various sources) whenever possible, but it was not possible to do that for every neighborhood. For example, the Hough neighborhood identified in Table 8 is only a portion of the entire Hough neighborhood identified by Cleveland historians. I have created a smaller version of Hough, but I believe that it captures the essential elements of the neighborhood. 18. Although the cubic specification of the spatial expansion model fits the data better than any other specification I tried, it is important to recognize that there are costs associated with using a lower order polynomial to describe the data. Specifically, the use of a lower order polynomial tends to produce severe “boundary effects” such as the one found in the 1915 graph. Boundary effects may have biased my estimates of aggregate land values. However, since lower order polynomials tend to overestimate land values at the edge of the city and underestimate land values at the center of the city the direction of any possible bias remains unclear. 19. Rocky River is a suburb of Cleveland; my Rocky River neighborhood includes census tracts that are with the Cleveland city limits, but which are nestled against the suburb of Rocky River. 20. This configuration captures the most important neighborhoods in Cleveland. 21. There are one foot tall rows in the eighty square mile matrix.
36 37
ACKNOWLEDGMENTS
38 39 40
I would like to thank Jeremy Atack, Bill Collins, Bob Margo, and an anonymous referee for their helpful comments on this paper. I would also like to thank
4259 Ch05 4/12/02 10:00 am Page 179
Urban Decline 1 2
179
Vanderbilt University for awarding me a Dissertation Enhancement Grant in 1999. The grant allowed me to collect the data that I used for this paper.
3 4
REFERENCES
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Alonso, W. (1964). Location and Land Use. Cambridge: Harvard University Press. Anas, A., Arnot, R. et al. (1998). Urban Spatial Structure. Journal of Economic Literature, 36, 1426–1464. Atack, J., & Passell, P. (1994). A New Economic View of American History: From Colonial Times to 1940. New York: W. W. Norton & Company. Atack, J., & Margo, R. A. (1998). Location! Location! Location! The Price Gradient for Vacant Urban Land: New York, 1835 to 1900. The Journal of Real Estate Finance and Economics, 16, 151–172. Campbell, T. F., & Miggins, E. M. (Eds) (1988). The Birth of Modern Cleveland 1865–1930. London: Associated University Press. Case Western Reserve University. The Encyclopedia of Cleveland History. http://ech.cwru.edu Chapman, E. H. (1981). Cleveland: Village to Metropolis. Cleveland: The Press of Western Reserve University. Cochran, W. G. (1977). Sampling Techniques. New York: John Wiley and Sons. Collins, W., & Margo, R. A. (2000). Residential Segregation and Socioeconomic Outcomes: When Did Ghettos Go Bad? Economic Letters, 69, 239–243. Collins, W., & Margo, R. A. (2001). Race, Home Ownership, and Segregation: 1900–1990. Explorations in Economic History, 38, 68–92. Collins, W., & Margo, R. A. Race and the Value of Owner Occupied Housing, 1940–1990. NBER Working Paper Series No. 7749. Colwell, P. F., & Munneke, H. J. (1997). The Structure of Urban Land Prices. Journal of Urban Economics, 41, 321–336. Colwell, P. F., & Sirmans, C. F. (1978). Area, Time, Centrality and the Value of Urban Land. Land Economics, 54, 514–519. Courant, P., & Yinger, J. (1977). On Models of Racial Prejudice and Urban Residential Structure. Journal of Urban Economics, 4, 272–291. Cutler, D. M., & Glaeser, E. L. (1997). Are Ghettos Good or Bad? The Quarterly Journal of Economics, 112, 827–872. Cutler, D. M., Glaeser, E. L. et al. (1999). The Rise and Decline of the American Ghetto. The Journal of Political Economy, 107, 455–506. Edel, M., & Sclar, E. (1975). The Distribution of Real Estate Value Changes: Metropolitan Boston, 1870–1970. Journal of Urban Economics, 2, 366–387. Galster, G. (1977). A Bid-Rent Analysis of Housing Market Discrimination. American Economic Review, 67, 144–155. Galster, G. (1983). Empirical Evidence on Cross-Tenure Differences in Home Maintenance and Conditions. Land Economics, 59, 107–113. Galster, G. (1990). Neighborhood Racial Change, Segregationist Sentiments, and Affirmative Marketing Policies. Journal of Urban Economics, 27, 344–361. Garreau, J. (1988). Edge Cities. New York: Doubleday. Glaeser, E. L. (1998). Are Cities Dying? The Journal of Economic Perspectives, 12, 139–160. Green, R. W. Real Property Inventory. These volumes are held in Cleveland State’s library in their special collections room and in their reference section.
179
4259 Ch05 4/12/02 10:00 am Page 180
180 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
FRED H. SMITH
Haig, R. M. (1926). Toward an Understanding of the Metropolis. Quarterly Journal of Economics, 40, 402–434. Hoyt, H. (1970). One Hundred Years of Land Values in Chicago. New York: Arno Press. Hurd, R. M. (1905). Principles of City Land Values. New York, The Record and Guide. Kau, J. B., & Sirmans, C. F. (1979). Urban Land Value Functions and the Price Elasticity of Demand for Housing. Journal of Urban Economics, 6, 112–121. Kau, J. B., & Sirmans, C. F. (1984). Changes in Urban Land Values: 1836–1970. Journal of Urban Economics, 15, 18–25. Keating, W. D. (1994). The Suburban Dilemma. Philadelphia: Temple University Press. Keating, W. D. et al. (Eds) (1995). Cleveland: A Metropolitan Reader. Kent: Kent State University Press. Kennedy, P. (1996). A Guide to Econometrics. Cambridge: The MIT Press. Kusmer, K. L. (1976). A Ghetto Takes Shape. Urbana: University of Illinois Press. Lemann, N. (1991). The Promised Land. New York: Alfred A. Knopf, Inc. Marks, C. (1989). Farewell – We’re Good and Gone – The Great Black Migration. Bloomington: Indiana University Press. Masotti, L. H., & Corsi, J. R. (1969). Shoot-Out in Cleveland. New York: Praeger. McDonald, J. F., & Bowman, H. W. (1979). Land Value Functions: A Re-evaluation. Journal of Urban Economics, 6, 25–41. McMillen, D. P. (1990). Consistent Estimation of the Urban Land Value Function. Journal of Urban Economics, 27, 285–293. McMillen, D. P. (1996). One Hundred Fifty Years of Land Values in Chicago: A Nonparametric Approach. Journal of Urban Economics, 40, 100–124. McMillen, D. P., Jarmin, R. et al. (1992). Selection Bias and Land Development in the Monocentric City Model. Journal of Urban Economics, 31, 273–284. Miller, Z. L. (1973). The Urbanization of Modern America – A Brief History. New York: Harcourt Brace Jovanovich, Inc. Miller, C. P., & Wheeler, R. A. (1997). Cleveland – A Concise History, 1796–1996. Bloomington: Indiana University Press. Mills, E. S. (1967). Transportation and Patterns of Urban Development – An Aggregative Model of Resource Allocation in a Metropolitan Area. American Economic Review, 57, 197–210. Mills, E. S. (1969). The Value of Urban Land. In: H. Perloff (Ed.), The Quality of the Urban Environment (pp. 231–253). Baltimore: Resources for the Future, Inc. Mills, E. S., & Hamilton, B. W. (1984). Urban Economics. Glenview: Scott Foresman and Company. Mumford, L. (1961). The City in History. New York: MJF Books. Mun, S. I., & Sasaki, K. (1992). Effects of Urban Transportation System Change on Land Prices in the Setting of Owner-Occupied Residence. Journal of Urban Economics, 32, 351–366. Muth, R. F. (1961). Economic Change and Rural-Urban Land Conversions. Econometrica, 29, 1–23. Muth, R. F. (1969). Cities and Housing. Chicago: The University of Chicago Press. Muth, R. F. (1975). Urban Economic Problems. New York: Harper & Row. Norton, M. B. et al. (1994). A People and a Nation. Boston: Houghton Mifflin. O’Sullivan, A. (2000). Urban Economics. New York: Irwin McGraw-Hill. Olson, S., & Lachman, M. L. (1976). Tax Delinquency in the Inner City. Lexington: D. C. Heath and Company. Ricardo, D. (1971). On the Principles of Political Economy, and Taxation. London: Penguin Books. Rose, W. G. (1950). Cleveland – The Making of a City. Cleveland: World Publishing Company. Rose-Ackerman, S. (1975). Racism and Urban Structure. Journal of Urban Economics, 2, 85–103.
4259 Ch05 4/12/02 10:00 am Page 181
Urban Decline 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
181
Rose-Ackerman, S. (1977). The Political Economy of a Racist Housing Market. Journal of Urban Economics, 4, 150–169. Siegel, J. (1975). Intrametropolitan Migration: A Simultaneous Model of Employment and Residential Location of White and Black Households. Journal of Urban Economics, 2, 29–47. Spengler, E. H. (1930). Land Values in New York in Relation to Transit Facilities. New York: Columbia University Press. Tassel, D. D. V., & Grabowski, J. J. (Eds) (1987). The Encyclopedia of Cleveland History. Bloomington: Indiana University Press. Tittle, D. (1992). Rebuilding Cleveland. Columbus: Ohio State University Press. Toman, J. A., & Hays, B. S. (1996). Horse Trail to Regional Rails: The Story of Public Transit in Greater Cleveland. Kent: The Kent State University Press. U.S. Department of Commerce (1913). Thirteenth Census of the United States, 1910. Various volumes. Washington, D.C.: GPO. U.S. Department of Commerce (1922). Fourteenth Census of the United States, 1920. Various volumes. Washington, D.C.: GPO. U.S. Department of Commerce (1933). Fifteenth Census of the United States, 1930. Various volumes. Washington, D.C.; GPO. U.S. Department of Commerce (1942). Sixteenth Census of the United States, 1940. Various volumes. Washington, D.C.: GPO. U.S. Department of Commerce (1952). Seventeenth Census of the United States, 1950. Various volumes. Washington, D.C.: GPO. U.S. Department of Commerce (1962). Eighteenth Census of the United States, 1960. Various volumes. Washington, D.C.: GPO. U.S. Department of Commerce (1972). Nineteenth Census of the United States, 1970. Various volumes. Washington, D.C.: GPO. U.S. Department of Commerce (1982). Twentieth Census of the United States, 1980. Various volumes. Washington, D.C.: GPO. U.S. Department of Commerce. Population of the 100 Largest Cities and other Urban Places in the United States: 1790 TO 1990. www.census.gov Wheaton, W. C. (1974). A Comparative Static Analysis of Urban Spatial Structure. Journal of Economic Theory, 9, 223–237. White, M. J. (1986). Property Taxes and Urban Housing Abandonment. Journal of Urban Economics, 20, 312–330. Yinger, J. (1976). A Note on the Length of the Black-White Border. Journal of Urban Economics, 3, 370–382. Yinger, J. (1976). Racial Prejudice and Racial Residential Segregation in an Urban Model. Journal of Urban Economics, 3, 383–396. Yinger, J. (1992). City and Suburb: Urban Models With More Than One Employment Center. Journal of Urban Economics, 31, 181–205. Yinger, J. (1993). Around the Block: Urban Models with a Street Grid. Journal of Urban Economics, 33, 305–330. Zangerle, J. (1916). Tax Atlas (Document Untitled) Publisher Unknown. This document is held at the Cuyahoga County Archives. Zannes, E. (1972). Checkmate in Cleveland. Cleveland: The Press of Case Western Reserve University. Zhang, Y., & Sasaki, K. (1997). Effects of Subcenter Formation on Urban Spatial Structure. Regional Science and Urban Economics, 27, 297–324.
181
4259 Ch05 4/12/02 10:00 am Page 182
182 1
FRED H. SMITH
APPENDIX A
2 3
CALCULATING AGGREGATE LAND VALUES
4 5 6 7 8 9 10 11 12 13 14
I calculated the aggregate land values for the city of Cleveland, presented in Table 6, and the aggregate land values for eight of Cleveland’s neighborhoods, presented in Table 7, with a program written in C++. Ideally, I would have used Eq. 1 and the estimated values of its coefficients to calculate the volume under the three-dimensional surfaces graphed in Figs 2 through 10. To do this, I needed to evaluate the following integral:
ƒ(e, n) dedn
ƒ(e, n) = Exp (LnR(e, n))
15 16 17 18 19 20 21 22 23 24 25 26 27 28
where
LnR(e, n) = Equation (1)
(A1) and
(A2)
{Remember: m = e2 + n2}
(A3)
Using Mathematica, I tried to numerically approximate the integral in Eq. (A1). Mathematica was unable to complete this task. This forced me to find an alternate method of calculating the volume under the three dimensional surfaces. In the paragraphs that follow I explain the method that I used. The first step in calculating the aggregate land value for the city is selecting a “representative” city. Cleveland is neither circular nor rectangular. In fact, its shape resembles the shadow cast by a bird with its wings extended. This highly irregular shape makes it very difficult to model Cleveland mathematically. Consequently, I decided to model Cleveland as an eight mile by ten mile rectangular city. A general bid-rent function would provide land values for every location within the eighty square mile city:
29 30
LnR(e, n) = g(e, n)
(A4)
31 32 33 34 35 36 37 38 39 40
Equation (A4) is analogous to Eq. (1) from Section IV. If I depict the city as an eight by ten matrix of square mile parcels of land, then it would look like the matrix drawn opposite.20 I have placed a four-point star on the matrix to indicate Public Square. Suppose that I wanted to calculate the value of a square foot of land at the location indicated by the five-point star. I simply plug the parcel’s location into the bid-rent function. Specifically, the value of a square foot of land at the five-point star would be Exp(g(–2, –2)). The coordinates are measured in miles to the east and miles to the north of Public Square. Using
4259 Ch05 4/12/02 10:00 am Page 183
Urban Decline
183
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
this coordinate system, the lower left vertex is labeled (–5, –5) and the upper right vertex is labeled (5, 3). To calculate the aggregate assessed value of land in Cleveland I needed to repeat the process I just described for every single square foot of land in the grid. The grid depicted above contains 2,230,272,000 square feet of land, so calculating this by hand was not an option. Lea Overby created a program in C++ that allowed me to carry out this calculation. The program obtains each square foot’s assessed value by plugging its coordinates into the bid-rent function. However, the bid-rent function takes on different values over the domain spanned by a square foot of land. To ensure consistency, the program calculated the value of each square foot using the coordinates from each square’s lower left vertex. The program started at the lower left vertex in the eight by ten matrix adding values as it moved along the bottom “row.”21 Once the values in that row had been summed, the program returned to the left edge of the next row up and resumed the process. In addition to calculating the aggregate value for the eighty square mile model of Cleveland, the program written by Lea Overby allowed aggregate land values to be calculated for any section of the city. This is a particularly useful feature; it allowed me to calculate aggregate assessed values for specific neighborhoods in Cleveland. For example, suppose that the shaded rectangle on the matrix represents one of Cleveland’s neighborhoods. To obtain the aggregate assessed 183
4259 Ch05 4/12/02 10:00 am Page 184
184 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
FRED H. SMITH
value for this neighborhood I simply modified the starting and ending points used by the program. If the lower left vertex of the shaded rectangle is (2.4, 0.3) and the upper right vertex is (2.8, 0.7), then I would plug these values into the program as its starting and ending points. I calculated the aggregate assessed value for eight neighborhoods using this method, and the results from these calculations are found in Table 7.
4259 Ch06 4/12/02 10:12 am Page 185
1 2 3 4 5 6 7 8 9 10 11 12
WAGE-STICKINESS, MONETARY CHANGES, AND REAL INCOMES IN LATE-MEDIEVAL ENGLAND AND THE LOW COUNTRIES, 1300–1500: DID MONEY MATTER?
13 14 15 16
John H. Munro
17 18 19 20 21 22
ABSTRACT
23 24 25 26 27 28 29 30 31 32
The primary explanation for the marked rise in real wages in both England and Flanders, from the later fourteenth to mid fifteenth centuries, was a combination of institutional wage stickiness and deflation. In both countries, nominal wages had indeed risen after the Black Death (1348), but so had the cost of living, with a rampant inflation that lasted until the late 1370s in England and the late 1380s in Flanders. Thereafter, consumer prices fell sharply but money wages did not – or, in Flanders, not as much as did consumer prices. The other thesis of this paper is that these later medieval price movements were fundamentally monetary in nature.
33 34 35 36 37 38 39 40
Research in Economic History, Volume 21, pages 185–297. Copyright © 2003 by Elsevier Science Ltd. All rights of reproduction in any form reserved. ISBN: 0-7623-0993-8
185
4259 Ch06 4/12/02 10:12 am Page 186
186 1 2 3
JOHN H. MUNRO
I. NOMINAL WAGE STICKINESS AND THE REAL-WAGE PROBLEM: J. M. KEYNES, ADAM SMITH, AND THE EVIDENCE
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Plaguing, so to speak, the current debate about changes in real-incomes in latemedieval western Europe, especially during the so-called “Golden Age of the Laborer,” following the Black Death, is the very troubling issue of “wage stickiness.” It simply means the long-term inflexibility, especially downward inflexibility, of nominal or money wages – and thus most emphatically not “stickiness” in real wages. The real wage is the physical quantity of goods and services that the artisan, craftsman or laborer can purchase with his or her money wage plus all the material fringe benefits that this employee receives in kind as a further payment for his/her labor: in terms of food, clothing, shelter, and other material rewards. In most historical studies, however, the real wage index is presented simply as the quotient of the nominal (money) wage index divided by the consumer price index: i.e. RWI = NWI/CPI; and it thus ignores the presence of any such fringe benefits (Van Zandem, 1999; R. Allen, 2001). Most economists focus much more on the behavior of real wages rather than on money wages in analyzing labor markets and the impact of labor changes upon the economy; and that was even more true of the Classical Economists in the nineteenth and early twentieth centuries. For reasons to be explained later, this study focuses upon the daily money wages paid to building craftsmen in southern England and the southern Low Countries (1300–1500) and upon their cost of living (as measured by “basket of consumables” price indexes). Those who neglect the significance of money wages and the historical problem of nominal wage stickiness have ignored not only John Maynard Keynes but also Adam Smith, the renowned founder of the Classical School. The former commented that: “the Classical Theory has been accustomed to rest the supposedly self-adjusting character of the economic system on an assumed fluidity of money-wages; and when there is rigidity, to lay on this rigidity the blame of maladjustment” (Keynes, 1936, p. 257). Drawing upon historical evidence, Smith (1776/1937, p. 74) did in fact offer supporting evidence for the view that money wages were not so fluid, cogently observing in the Wealth of Nations that: The wages of labour do not in Great Britain fluctuate with the price of provisions, [which] vary everywhere from year to year, frequently from month to month. But in many places the price of labour remains uniformly the same sometimes for half a century . . .. The high price of provisions during these ten years past has not in many parts of the kingdom been accompanied with any sensible rise in the money price of labour.
He was not far off the mark. For, in southern England, the money wage of master masons and carpenters had remained fixed at 24d per day (i.e. 2s 0d)
4259 Ch06 4/12/02 10:12 am Page 187
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
187
for precisely 40 years, while the well-known “basket of consumables” composite price index of Henry Phelps Brown and Sheila Hopkins (1955, 1956, 1981, p. 178) had risen, over the same 40-year period (1733–1773), by 57.4%.1 By far the longest period of nominal wage stickiness recorded in English pricehistory is to be found in the payment records for carpenters and masons at the Oxford colleges. From 1363 to 1536, a period of 174 years, they were paid, with very rare exceptions, a standard wage of 6d per day, or 3s 0d per six-day week, certainly for the summer (Easter-Michaelmas) season. When they then received their first sustained modest increase, to 6.5d and 7.0d per day in 1536–1537, they had already suffered some considerable ravages of inflation: a 50% rise in the price level, from the well-known Price Revolution, which had commenced around 1516–1520.2 If the continental evidence fails to provide evidence quite so impressive, nevertheless long-term nominal wage stickiness was clearly also the prevalent feature of labor markets in the Low Countries for the fifteenth and early sixteenth centuries. Thus, in Bruges, the daily money-wage for a civic policemen was fixed at 5d groot Flemish (summer and winter) from 1398 to 1476, nominally and initially half the rate for a master mason, though on an annual payment basis about 90% as much.3 At Mechelen (a Flemish enclave within the neighboring duchy of Brabant), the predominant daily summer wage for master masons and carpenters in the town’s employ was an inflexible 12d groot Brabant (= 8d groot Flemish) from the Burgundian monetary reform of 1434 to 1489 (Table 14). Such wage stability is all the more surprising in view of the virtual doubling of prices during this period, as measured by a similar “basket of consumables” index constructed by Herman Van der Wee (1975, 1978). From 1490 to 1540, with renewed inflation from about 1515, those Mechelen wages were fixed at a constant 13.5d groot (= 9d groot Flemish). In nearby Antwerp that daily summer wage of 12d groot Brabant had prevailed for these building craftsmen for the rather shorter period, from 1442 to 1486 (Table 13).4 Consider the implications of these data on wage stickiness and prices. For, if we calculate the real wage index by the aforementioned formula (RWI = NWI/CPI), then obviously changes in the price level – the extent of inflation or deflation – would fundamentally determine the changes in real wages. Adam Smith (1776/1937, p. 74) himself commented that, if “the labouring poor can maintain their families in dear years, they must be at their ease in times of moderate plenty, and in affluence in those of extraordinary cheapness.” The most famous example of affluence was in later-medieval England, and indeed in other parts of western Europe, following the Black Death (1348) – 187
4259 Ch06 4/12/02 10:12 am Page 188
188 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
JOHN H. MUNRO
though, as will be seen, some considerable time after the several onslaughts of bubonic plague. As Thorold Rogers (1903, p. 325) stated so long ago, and as so many others have continually reiterated since: “the fifteenth century and the first quarter of the sixteenth were the Golden Age of the English labourer, if we are to interpret the wages which he earned by the cost of the necessities of life.”5 The accompanying Tables 3–5 clearly indicate for England, at least, that most of this so-called Golden Age was also one of prolonged deflation, with only a few, sporadic spikes of high prices. Even the cross-Channel Low Countries, with a much different monetary history, and one plagued by repeated coinage debasements, experienced two prolonged periods of deflation – with rising real wages – from c.1390 to c.1420 and again from c.1440 to c.1470 (Tables 3–4, 8–15). Although southern England’s subsequent experience with real wages, into the early-modern era, was not replicated, for example, in the cross-Channel Low Countries, surely this historic pattern of real-wages is a most remarkable experience, and one that poses serious theoretical problems for both economists and historians. As Keynes has reminded us (1936, p. 5), one of the most basic postulates of Classical Economics is that “the wage is equal to the marginal product of labour;” and of course by that statement he meant the real wage. Are we really to believe, through much of later-medieval and early-modern English economic history, that a combination of wage-stickiness and a fluctuating price level had a more powerful impact on real wages than did changes in the productivity of labor and capital?6 As this study will reveal, however, the impact of wage stickiness proved to be much more powerful during periods of deflation, if only because during prolonged inflations, nominal wages were rather less sticky, especially in the Low Countries. But even there, during the Price Revolution era, the frequently rising money wages often did lag behind consumer prices, and then became far more sticky during the deflationary seventeenth century.7
30 31
Commodity Price Indexes and Their Problems
32 33 34 35 36 37 38 39 40
There are of course several intractable measurement problems with all these price and wage indexes. The prices in the Phelps Brown and Hopkins “basket of consumables” for southern England, and in the other two “baskets” used here, for Brabant (Van der Wee, 1975, 1978/1981) and Flanders (Munro, 1984a), both modelled on the English index, are generally those paid by institutions, often wholesale for bulk lots, and not retail prices paid by the typical artisan consumer in the market place. Furthermore, these indexes are far more heavily weighted with the prices of primary products than of finished goods. Thus, prices for
4259 Ch06 4/12/02 10:12 am Page 189
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
189
farinaceous goods and drink are those of just their constituent grains (wheat, rye, barley, peas), because, for this period, we simply lack adequate price series for flour, bread, cereals, and beer. Indeed, in medieval England, local governments regulated the sales of both bread and beer (ale), ostensibly fixed in price, by quantity (size, weight, capacity), which officially was to vary inversely with the price of their component grains.8 As Nef (1937) and more recently R. Allen (2001) have noted, the price changes for bread, flour, and beer were undoubtedly of a much lower amplitude than were those for grains, principally because labor and marketing costs accounted for a much higher proportion of their prices than for those of the raw grains. Consequently, price indexes heavily weighted with grains – 42.50% in the Phelps Brown and Hopkins index – exaggerate the rise of real wages during periods of falling grain prices, as in the deflationary fifteenth century, and the fall of real wages during periods of rising grain prices, as in the inflationary sixteenth century. But wage stickiness also meant that labor costs accounted for a higher proportion of the total prices for these finished food products during deflationary than inflationary periods. Furthermore, one might reasonably object that these indexes are too heavily weighted with foodstuffs, even if the other products – such as fish, meat, butter, and cheese – are “finished’ consumer goods. As Table 1 indicates, foodstuffs account for 80% of the component weights in the English and Flemish price indexes, and for almost 75% in the Brabant index. But Phelps Brown and Hopkins justified their index weights, especially for the fifteenth century base period (1451–1475 = 100), with the expenditure records of a contemporary Dorset household; and Van der Wee (1966) found similar records to justify his weighting for foodstuffs, closely approximating the English index weights.9 If these price indexes are vastly preferable to the older ones based on grains alone, one may yet complain that the range of commodities contained is still too narrow. No allowance, for example, has been or can be made for rents and related household costs. There are also other unfortunate omissions, because of lacunae in the extant data. The Phelps Brown and Hopkins index, for example, ceases to use prices for dairy products after 1430, so that thereafter meat and fish prices (mutton and herrings) must carry the entire weight for that sub-index (37.5%).10 While the Flemish and Brabant indexes do contain butter and cheese prices throughout, the former lacks any prices for meat and fish, while the latter lacks prices for wheat and peas; and in the Flemish index, the only industrial prices are for woolen fabrics.11 Perhaps the chief criticism to be made is that, for the period of this study, the fixed weights for the components in all three “baskets” do not properly reflect the ability of most consumers to respond to changing relative prices by making product substitutions. At the same time, the method of assigning such 189
4259 Ch06 4/12/02 10:12 am Page 190
190 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
JOHN H. MUNRO
weights in the Brabant and Flemish “basket” differs significantly from that chosen for the English “basket,” even though it served as the model for the two Low Countries’ price indexes. For Phelps Brown and Hopkins constructed their composite price index with fixed, unvarying expenditure shares: i.e. 42.5% for grains; 37.5% for meat, fish, dairy products; and 20% for industrial products. In that respect, their “basket” resembles a geometric index, though it was not computed in this fashion. The actual quantities of the component commodities that they specify in their article (1956/1981, p. 20) for the later-medieval and early modern eras are in fact valid only for the base period 1451–1475. But those specific commodities listed in the Phelps Brown and Hopkins “basket” for this period provide an unvarying quantity base for the two Low Countries’ “baskets,” which are thus much more the standard Laspeyres-type price indexes. The annual values of the basket are computed by multiplying the mean annual prices (in pence groot of Brabant and of Flanders respectively) for each commodity in the basket by that quantity; and the base 1451–1475 = 100 was calculated by dividing the sum of those annual basket values by 25. The expenditure shares for the various commodity groups in three baskets are roughly comparable, as noted in Table 1, for this base period. But thereafter, while the expenditure shares for each commodity group in the English “basket” remained constant over time, those in the Brabant and Flemish “baskets” changed with the rise or fall in the relative prices of the component commodities. Thus, for example, as Table 8 indicates for the Flemish price indexes, grains account for 55% of the total value of the “basket” during the Black Death era (1348–1350) and again during the war-torn and plague-ridden period 1436–1440, but for only 37% during the peaceful, grain-abundant, and also deflationary years 1496–1500. In that respect, the two Laspeyres baskets for the Low Countries provide a better approximation of expected consumer expenditures. As R. Allen (2001, pp. 423–425) has recently noted, however, no alternative price indexes are likely to change the basic picture and resolve all of these problems.12
31 32 33 34
II. PLAGUE, PRICES AND REAL-WAGES ACCORDING TO THE DEMOGRAPHIC SCHOOL: THE POSTAN-ABEL MODEL
35 36 37 38 39 40
Avoiding these statistical price- and wage-index problems, most medieval historians have instead found a far simpler and more compelling explanation for the remarkably high level of real wages of later-medieval western Europe: the impact of demographic catastrophes upon agrarian productivity. By far the best known rendition of this view is to be found in the Ricardian-Malthusian
4259 Ch06 4/12/02 10:12 am Page 191
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
191
models of Michael Postan (1950, 1951, 1952, 1966, 1972, 1973), Wilhelm Abel (1978, 1980), and Georges Duby (1962/1968).13 Even the few opponents of their model must now accept as an established fact the very drastic nature of the late-medieval depopulations, especially in England, whose population fell from a peak of at least 4.5 to 5.0 million in the 1290s (if not from 7.2 million, as some have suggested) to about 2.5 or perhaps 3.0 million in 1377, finally reaching a nadir of about 2.25 million in the early 1520s.14 Familiar though the Postan-Abel-Duby model may be, its key elements must be briefly reiterated, if only to clarify the arguments that follow. The model’s basic supposition is that the drastic decline in population during the fourteenth century, especially after the Black Death and subsequent attacks of bubonic plague, and the consequent alteration in the land-labor ratio, sharply increased the marginal productivity of agrarian labor. In a fundamentally agrarian economy, such changes presumably took place fairly quickly, with the abandonment of many high-cost marginal lands that, with prior population growth, had been subject to diminishing returns. Thus, arable husbandry would have become concentrated on much better quality, higher-yielding lands that produced much more grain and livestock products with proportionately much less labor. At the same time, agricultural labor itself probably became even more scarce, as former itinerant laborers took up deserted tenancies or sought better employment in the towns. Furthermore, this agrarian reorganisation served to increase real-incomes even more by reducing the relative cost of cereal-grains and other basic foodstuffs, and perhaps also housing costs, with so much more available land. Artisans and laborers, finding that they were enjoying greater disposable real incomes, after meeting the basic necessities of life, evidently chose to increase their spending on meat, dairy products, industrial goods, and semiluxuries, thus driving up the relative prices of such goods. Prices for manufactured goods should have risen even more, because the continued rise in industrial wages would have accounted for an even greater share of total production costs than did wages in producing grain and livestock products. Such is the classic demographic model for the late-medieval English and indeed West European economy, one that certainly seems both plausible and reasonable. This emphasis upon the behavior of relative prices – i.e. a change in the price of wheat relative to a change in the price of, say, linen cloth – is crucial for any consideration of the Postan model in particular and of the analytical discussion that follows. In Postan’s strongly pronounced views, these demographic changes were almost entirely responsible for late-medieval price movements. Contending that the fall in prices was almost entirely confined to grains, so that they diverged from other prices, many of which were rising, indeed in accordance with the model, Postan (1972, p. 239) concluded that: 191
4259 Ch06 4/12/02 10:12 am Page 192
192 1 2 3 4
JOHN H. MUNRO
monetary factors could not have been the sole or the main cause of the price changes, for the pure logic of the monetary explanation demands that the effects of changes in the circulating medium should be felt throughout the economy, i.e. in the prices of all the goods bought and sold, since changes in money must be, so to speak, ‘neutral’ as between different commodities.15
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Such views about the behavior of late-medieval price-movement are, however, quite misleading. First, as the accompanying Tables 5, 8, 9 and Figs 2–3 on English, Flemish, and Brabantine prices indicate, the component price-indexes for grains, livestock-products, and industrial goods generally tended to rise or fall together, in distinct phases of inflation and deflation, though by no means exactly in tandem, and with varying amplitudes and differing short-term oscillations. Second, no respectable economist would ever contend that, while monetary forces were at work in the economy, demographic and other “real” forces would remain suspended or frozen, in time, i.e. without interacting together, in altering the relative prices of many individual commodities. Third, monetary changes are not “neutral” in the manner that Postan has suggested. Suppose, for example, that a country’s supply of money had increased. The regional distributions of such increased stocks would have benefited some economic sectors more than others, thus allowing some groups or socio-economic strata to gain relatively greater increases in money incomes. The consequent changes in their savings and expenditure patterns, while possibly producing changes in the income velocity of money, would also have altered the prices of a wide variety of individual goods and services through their impact on the price and income elasticities of demand for various commodities and thus also on their elasticities of supply. Furthermore, if those changes in money stocks (even with a possible reduction in the income velocity of money) led to inflation, and if the money incomes of the lower strata of society did not rise correspondingly, many would have been faced with severe budget constraints, thus forcing them to spend proportionately more of their disposable incomes on foodstuffs and necessities and less on other commodities. Such demand shifts would likely have lowered the relative if not the nominal prices of the latter (especially industrial goods).16 Suppose further, that this country – or, for the present analysis, most of western Europe after the Black Death – experienced a rapid reduction in its population without, initially, experiencing any corresponding reduction in its aggregate money supply, and, not at least initially, in the income velocity of money. The result would really have been the same: a relative increase in money supplies, relative that is, to a much contracted economy, in which land, so much of which was abandoned, and labor were the chief inputs. Consider this situation in terms of the modernized version of the Fisher Identity: MV = Py, in which
4259 Ch06 4/12/02 10:12 am Page 193
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
193
the variable y represents NNI (real net national income, in place of “T,” which cannot be measured). Thus, if y fell to a greater extent than did any possible reduction in the volume of money payments (i.e. the product of MV), then obviously prices had to rise. The same conclusions are to be drawn in using the much preferable Cambridge “cash balances’ approach: so that M = kPy (and thus, k = 1/V). As will be seen later in this study, the Black Death was indeed followed by rampant inflation throughout most of western Europe, for about three decades.17 Postan, to his credit, certainly did understand the distinction between changes in relative prices and shifts in the overall price-level. But he did not observe the now readily available evidence for oscillating price-movements of general inflation and deflation in late-medieval Europe (see Figs 1–3). Many of his disciples, however, subsequently did become quite aware of these longer term price-movements, yet they still incorrectly attributed those price movements primarily to demographic changes. They evidently did so by confusing microeconomics with macro-economics: i.e. by assuming that a perfectly valid explanation for grain price-changes can be applied to the entire economy and its overall price-level. A growth in population that encounters diminishing returns in agriculture will likely cause relative grain prices to rise; but, acting alone, population growth cannot cause all or even most prices to rise. And conversely, a fall in population, while it may well have led to a sharp fall in relative grain prices, could not, by itself, have caused a general fall in prices – and certainly not following the Black Death, when, as just noted, prices instead rose sharply.18 In short, demographic forces can influence long-term price movements only by their interaction with other real economic forces, and by their consequences in inducing changes in both money stocks and money flows, including changes in credit. They may also have done so through their impact in inducing or stimulating a growth in real net national output and national income; and such was evidently the case during the “long” thirteenth century (c. 1180 – c. 1320) and again during the sixteenth-century Price Revolution. Ultimately, real economic growth – in turn promoting further population growth and increased settlement and urbanization – might have reduced or even eliminated elements of “slack” (elastic supplies) in the agricultural and natural-resource (forestry, mining) sectors of the economy; and thus the economy might well have failed to sustain a rate of growth comparable to that on the ongoing monetary expansion (in stocks and/or flows). Consequently, under such circumstances, as Keynes himself observed, most prices were bound to rise.19 For the wage question itself, there remains one further problem with the Postan model, and indeed with the usual presentation of Classical Theory. For, 193
4259 Ch06 4/12/02 10:12 am Page 194
194 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
in a frictionless market economy, without long-term wage contracts or other institutional impediments, the real wage for any particular occupation or task, defined by time and place, should be determined not just by the marginal product but by the marginal revenue product of labor (MRP): i.e. the extra revenue that the employer derives by selling the last unit of output produced by the last unit of labor added to his fixed stock of land and capital. If, according to the Postan model, the marginal productivity of agricultural labor rose, with the demographic, agrarian, and price shifts so posited for the late-medieval post-Plague English economy, then presumably the marginal revenue from the sale of grain and other arable products so produced on the now chiefly more productive lands should have fallen, with falling prices (or falling relative grain prices). The result, therefore, may have been just a “wash” in real wage changes. If, however, contrary to the Postan model, relative grain prices did not consistently fall, then real agricultural wages may well have risen in later-medieval England. Though much evidence has been published that purports to show such a continual rise in agricultural wages during the later fourteenth century (in decennial means),20 the accounts of several Winchester manors do not consistently validate that view. As Beveridge has noted (1936, p. 27), wages on most Winchester manors, after beginning to rise from 1362–1363, were forced down in 1367–1368, evidently at the behest of the new Bishop William of Wykeham; but then the higher rates were restored between 1370 and 1375. The manorial records of Taunton (Somerset) provide the most striking deviation from the presumed norm in the behavior of laborers’ wages in later fourteenth-century (Table 7). To be sure, with the onslaught of the Black Death, casual farmlaborers on the Taunton manor had enjoyed a doubling in their money wage, with a substantial increase as well in their real wage, from 1349 to 1356. But thereafter their money and real wages both fell, and from 1362–1363 (i.e. before the election of Bishop William) fell sharply to pre-Plague levels even in money terms. Their real wages fell even more steeply below such levels, recovering only from 1378; but their money wages did not recover before 1412–1413.21 Second, recently published research on fourteenth-century agricultural production does not fully support the standard Malthusian-Ricardo-Postan model, which purports to demonstrate that the marginal product of labor had to rise with continued depopulation. To be sure, labor productivity did rise in pastoral farming on many manors: proportionally fewer men were required to manage larger flocks on an increased acreage. But, in several Winchester, Glastonbury, and Ramsey Abbey manors, labor productivity generally fell on the arable, in the seven decades from 1341 to 1421.22 Indeed, the Malthusian-Ricardian model fails to demonstrate why any rise in industrial productivity should have occurred, especially in the building and textile trades. In woolen-textile manufacturing,
4259 Ch06 4/12/02 10:12 am Page 195
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
195
productivity in fact remained quite unchanged from the early fourteenth to the late eighteenth centuries.23 Some medieval industries did benefit, to be sure, from applications of more complex forms of water-powered machinery; but, in the leading industries – textiles, mining, and metallurgy – most were instituted either long before or a full century after the Black Death (except the blast smelter; Reynolds, 1983; Holt, 1988; Munro, 2003a). For late-medieval industries in general, convincing evidence is lacking to indicate that the forces of later-medieval depopulation led to any positive qualitative changes in the composition, structure, and institutional utilization of the surviving labor force. Classical economic theory would suggest, however, that any rise in real wages in the agrarian sector would necessarily be translated to the other sectors, lest the latter lose labor to or from the agrarian sector (i.e. in reduced migration to towns).24 But without real productivity increases, presumably the employment of urban industrial labor would have been restricted to those crafts in which WL = MRPL. Real urban industrial wages might have risen through a rise in relative industrial prices, if that increased the employer’s marginal revenue; and they would have risen further through any fall in the cost of living, chiefly in foodstuffs. The first lesson to be learned in this study is that real-wage changes in the late-medieval economy were very complex and confusing to the observer, then and now. The other lesson, to be demonstrated later in this study, is not to generalize about prices and real wages from the experience of late-medieval England.
24 25 26 27
III. SOME PROBLEMS WITH THE DATA ON LONG-TERM WAGE STICKINESS
28 29 30
Are These Wage Data Truly Representative: Do “Sticky” Wages Really Reflect Labor Markets?
31 32 33 34 35 36 37 38 39 40
Many historians, of course, have pointed out an even more serious problem: the nature of the available wage data, which seem to be very unrepresentative. For the medieval era, by far the most familiar wage data are those published late in the last century by James E. Thorold Rogers (1866, 1882) and made even more accessible in the famous wage and price indexes of Henry Phelps Brown and Sheila Hopkins (which also includes other sources, from Buckinghamshire, Hampshire, Sussex and Kent). They themselves expressed a few reservations about the validity of these data in truly depicting long-term, downward wage stickiness (1955, 1981, pp. 7–8): 195
4259 Ch06 4/12/02 10:12 am Page 196
196 1 2 3 4 5
JOHN H. MUNRO
Certainly our method of taking a representative rate is biased towards stability . . . since for most of our period these payments were made not by employers to wage-earners but by customers to craftsmen working on their own account, and these customers were generally institutions and not private persons who had to put their hands into their own pockets. But after due allowance for these things, the absence of sustained falls and of falling trends remains remarkable. It has been called the elbow-joint or ratchet effect.
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Unconvinced by this argument, Lindert (1985, p. 618) accused them of deliberately seeking wage rates that could be regarded as “representative because they were recurrent,” and then of constructing their wage series “in such a way as to overstate wage stickiness.” Nevertheless, while pleading for historians to produce “other long-term wage series to ease the strain of over-reliance on the classic Phelps Brown-Hopkins series,” Lindert himself necessarily still used them, as have Wrigley and Schofield (1981, 1989), and many other historians. In medieval economic history, however, beggars seeking statistical data cannot be choosers; and neither Thorold Rogers nor Phelps Brown and Hopkins had much choice. Nor have other historians in subsequent publications. For the only readily available and suitable data are those provided by institutions willing to record them over long periods of time: medieval manors, town governments, hospitals, universities and colleges, guild houses, etc. Apart from the manorial and abbey accounts, which certainly do include some agricultural occupations (not considered in this study), most of the available institutional records provide wage data that are limited to building craftsmen, including carpenters, masons, tilers, plasterers, but also thatchers and street-pavers. By no coincidence, these are the very occupations whose wages were ostensibly regulated by the famous Ordinance (June 1349) and Statute of Labourers (1350–1351), and by most other late-medieval English wage legislation (except the 1388 Statute of Cambridge).25 Such wages, it must be noted are time-based, initially by the week and then more frequently by the day. In the later-medieval and early-modern European economies, however, the great majority of laborers and craftsmen, and most especially those in agriculture, textiles, leather-goods, metal wares, etc., received instead piece-work wages, which, when available, are often difficult to interpret, since we usually cannot measure output per day.26 In their presentation of daily-wage data, Phelps Brown and Hopkins (1955) differed significantly from not only the older studies on wages, by Thorold Rogers (1866, 1882) and Steffen (1901–1905), but also from those published since by several other historians: in particular, Beveridge (1936–1937, 1955–1956), Knoop and Jones (1933, 1967), Schreiner (1954), and Farmer (1983, 1988, 1991). Together, those studies may convince many readers that medieval wage stickiness was just a myth. For their tables, whether annual or decennial, reveal almost continuous oscillations in wage-rates, sometimes small
4259 Ch06 4/12/02 10:12 am Page 197
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
197
but often significant. Seemingly inexplicable, these oscillations disappear when one reads the original sources (manorial and town records) and then realizes that these historians have computed their annual data and decennial means by taking averages of the wage payments or wage rates in each record, thus producing “compositional errors.” Consider, as an example, a building project or manorial repairs employing a dozen carpenters in the year: with three senior master carpenters earning 4.5 and 5.0d per day, seven ordinary master carpenters earning 4.0d per day and two junior, less experienced master carpenters earning 3.5d per day.27 Their mean wage would then be 4.250d unweighted and 4.083d weighted. If, in the following year, only ten carpenters were employed, with only one senior carpenter, earning 5.0d per day, six earning the standard 4.0d per day, and three earning 3.5d per day, the unweighted mean would now be 4.167d and the weighted mean, 3.950d. An examination of these accounts year after year would reveal that, for each class or status of master carpenters, the wage rates were in fact unchanging – and thus very sticky. Phelps Brown and Hopkins, to their great credit, selected for each year what appeared to be the standard or representative wage, which, admittedly, becomes a more difficult task during periods of transitional wages. Further proof that the Thorold Rogers’ and Phelps Brown and Hopkins’ published wage data are indeed representative of the contemporary labor markets may be found in the wage-payment records for craftsmen in such medium-sized towns as Canterbury, Exeter, Dover, Winchester, and York.28 Certainly, during the later fourteenth and the fifteenth centuries, their wage rates did not significantly differ from those that Thorold Rogers (1866, 1882) found in the Oxford and Cambridge records, with generally the same degrees of long-term wage stickiness, if not for as long a period as for the Oxford craftsmen. Together all these town wage-data for master building craftsmen were or came to be about midway between those for master building craftsmen recorded, at the lower end, in the accounts of the Bishop of Winchester’s estates, of the Bishop of Bury St. Edmunds (Redgrave), and of Battle Abbey, and, at the upper end, the wages recorded for London artisans. Thus, during the second half of the fourteenth century, the standard wage for master craftsmen on the Battle Abbey and Winchester rural estates was 4d per day. At Oxford and Cambridge, as noted earlier, the wages for such craftsmen had risen from 5d to 6d per day about 1363, while the wages for such building craftsmen in the other medium-sized towns listed above generally remained at 5d per day until about 1410–1415 (then increased as well to 6d), just before the wage rates for most London building craftsmen had risen from 7.0d or 7.5d to 8.0d per day. About that same time, the manorial wage rates for these building craftsmen also rose, from 4d to 5d per day. 197
4259 Ch06 4/12/02 10:12 am Page 198
198 1 2
JOHN H. MUNRO
Was Medieval Wage Stickiness Confined to Long-term, Contractual Institutional Employment?
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
The best answer to this question, and one that further reinforces the previous answer on the representative nature of such wage data, can be derived from a comparison of wage payments in various London records, in particular those of: the Tower Bridge Authority (London Bridge Master); the various London guilds (Brewers, Cutlers, Carpenters, Grocers, Ironmongers); Westminster Abbey and its manors; and, finally, the Bishop of Winchester’s London manor of Southwark. The Bridge Masters, in maintaining their various buildings and residential tenements, employed many craftsmen who appear to have been virtually “tenured,” i.e. continuously employed for many years or decades; but the other three employers generally hired their masons, carpenters and other such craftsmen only on a purely occasional basis, generally for just a few days at a time, on what appears to be virtually a “spot market” for labor. For almost all such building craftsmen, in those fifteenth-century years for which a quadripartite comparison is possible, the wage data are virtually identical, and thus indicate that all four London-based institutions were compelled to pay the same market wages, for both short and long-term employment.29 It is equally instructive to note that during the early fifteenth century, master carpenters and masons employed on the Southwark manor, across the Thames from the heart of the old City of London (and near the various guild halls), were earning exactly the same daily wage (8d) as were those hired by the nearby guild halls and the London Tower Bridge Authority, and thus from 60% to 100% more than those employed in the Bishop of Winchester’s rural manors in southern England.30 Such evidence therefore also does not support the commonly expressed view that daily wage rates were higher for casual short-term employment than for guaranteed annual employment. It also casts doubt on the view that late-medieval laborers, agricultural workers, and craftsmen came to prefer both the greater individual freedom and higher (real) rates for the former type of discontinuous employment.31 For the Low Countries, similar institutional sources, but most especially town government accounts – for Bruges, Ghent, Ypres, Mechelen, Leuven, Antwerp, Aalst (Alost) – and some urban-based abbeys, churches, and hospitals provide even more voluminous daily-wage data to support all of these conclusions: for the very same sets of building craftsmen, and also for some textile artisans and even policemen (in Bruges).32 The wage data for such craftsmen and policemen throughout the late-medieval Low Countries exhibit patterns very similar to those of the English artisans, as suggested earlier, most especially during periods of stable coinage, even though their governments were sensible enough never
4259 Ch06 4/12/02 10:12 am Page 199
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12
199
to impose wage ordinances, certainly none as restrictive as the Statute of Labourers. In most, if not in all cases, the craftsmen receiving such wages were similarly not long-term privileged employees, but those working for short periods – often just a few days or weeks – for many and various employers. In the woolen draperies of the late-medieval Low Countries, their predominant manufacturing industry, daily wages are available only for the fullers, since most other textile workers received piece-work wages or fees (the latter, for shearers and dyers). Most fullers similarly worked for not one but a variety of different master-weaver drapers, for short and discrete periods of time. They were free to change their employers – and they were also capable of going on strike to achieve their collective guild-wage demands, if not always successfully.33
13 14
Was Late-medieval Wage Stickiness the Consequence of State Intervention?
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Many historians, especially British economic historians, believe that, in so far as we may discern such wage stickiness, it was both the consequence of the “wage freeze” imposed by the infamous Ordinance (1349) and Statute of Labourers (1351) and/or an artifice to evade it: i.e. payment records to simulate compliance, when employers were in reality paying more than the recorded amounts.34 The short answer in response to such charges is that the patterns of long-term wage stickiness were virtually identical in late-medieval England and the Low Countries. In England itself, furthermore, such wage stickiness, and most especially downward wage stickiness, continued to be the prevalent feature for the employment of day laborers and artisans for centuries after such legislation – from the Ordinance of Labourers to the 1563 Statute of Artificers – had become dead letters (Table 2).35 Such answers will, nevertheless, not satisfy those historians who believe that the late-medieval records of wage-payments reflect, in some fashion, compliance with these Statutes, all the more so since Putnam (1908), Ritchie (1934), Penn and Dyer (1990), and Hatcher (1994) have all provided convincing evidence that the crown and local authorities did seek to enforce these statutes, at least until the late 1380s. Their arguments and evidence, especially Hatcher’s, must be taken very seriously, even if such evidence does not really extend beyond the late fourteenth century. Yet the observer may just as strongly question the longer-term success of such enforcement of wage-rate ceilings by examining more closely the manorial wage records – and thereby acquire greater confidence in their validity. For virtually all of these manorial accounts demonstrate that most of the recorded wages were substantially above those permitted by the 1349 Ordinance 199
4259 Ch06 4/12/02 10:12 am Page 200
200 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
JOHN H. MUNRO
and the 1351 Statute of Labourers (Table 2); and a fair number of these accounts indicate that some senior, or more highly skilled, craftsmen earned daily wages that were from 25% to 50% in excess of those permitted by statute.36 Evidently those who maintained these accounts (bailiffs and stewards) had not entertained any fears of prosecution in recording such wages; and it would be difficult to prove that those few who did record wages within the statutory limits did so because of such fears. Possibly the lower wages on the smaller number of “compliant” manors are to be explained by a higher proportion of labor supplied by still customary (servile) tenants and by more isolated rural labor markets (such as at Taunton, in Somerset). Certainly the later-medieval English towns, with presumably higher living costs, readily permitted wages above those prescribed in the Ordinance and Statute of Labourers, even though these royal writs in no way exempted the towns or tolerated higher rates (Table 2). In blatantly ignoring the 1349 Ordinance, London’s civic government issued its wage ordinance the following year (1350) to fix the maximum wage for building craftsmen at 6d per summer and 5d per winter day, i.e. from Easter to Michaelmas (29 September), to Easter. Both were one pence higher than the maximum wages permitted in the previous London ordinance of 1290, but double that permitted in the 1349 royal Ordinance and the 1351 national statute.37 Not until 1495 did Parliament recognize London’s special status within the kingdom, and its much higher cost of living, with legislation to authorize these very same rates of 6d and 5d per day, respectively (but only 4d daily with food and drink), with some minor exceptions (Table 2).38 In any event, long before then, and certainly by the 1370s, most employers of building craftsmen in London were ignoring not only the Statute of Labourers but also the local civic ordinances on maximum wages, as the evidence just cited clearly shows.39 Nevertheless the impact of such labor legislation cannot be fully ignored in that it may have served to restrict the full impact of those forces driving up wages, and in particular may have led to the elimination of some seasonal wages, to be discussed in the answer to the following question.
32 33 34 35
Does Institutional Wage Stickiness Reflect Adjustments in the Real Wage by Other Means: By Adjusting Hours of Work and Thus by Substituting More Leisure for Paid Employment?
36 37 38 39 40
Since medieval and early-modern employers, not only in England, but throughout Europe, paid those employees receiving time-work wages commonly by the week and then more commonly by the day, though not by the hour, employers and employees might have bargained to adjust real wages, according
4259 Ch06 4/12/02 10:12 am Page 201
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
201
to changing economic circumstances, by altering the length of the work-week or day. Some historians have suggested that, in pre-Industrial societies especially, many artisans, craftsmen, and laborers had a “backward bending supply curve of labor.” Thus, such workers, on finding that their real wages had risen to permit them to acquire some desired level of sustenance and comfort, would have preferred to enjoy increased leisure time over further increases in money income; and thus they would have chosen to work less, or to refuse to work for traditionally long hours. Hatcher (1994, pp. 13–19) has suggested that such a substitution of leisure for paid work may have been a component of rising real wages after the Black Death. Indeed, several years earlier, Blanchard (1978) had provided some evidence for such a “backward bending” supply curve for labor in the late-medieval English mining industry, many of whose rural workers were seasonal, primarily engaging in the agrarian economy. Obviously such circumstances would vary by time, region, and occupation. Yet there is no convincing evidence that such a backward-bending supply curve for labor was a significant feature of employment in the late medieval economy, certainly not in the building and textile trades. Some research done on this very question may be found in a recent publication on seasonal wages and leisure in late-medieval England and the Low Countries (Munro, 1994d). For neither region was there any convincing evidence that urban craftsmen sought to increase their leisure time, even after real wages had peaked in the mid-fifteenth century, by reducing either the work day or the work week, which was usually six full days. There are, of course, only scant, sporadic data on the actual number of days worked per year, except for the invaluable data that Herman Van der Wee (1963) has provided on employment in the building trades in the Antwerp-Lier region, for each and every year from 1436 to 1600. For the base period employed in this study, 1451 to 1475, the mean number of days so worked was 210 days, precisely the same average in the late sixteenth century. The fewest number so worked, 191 days, occurred during the civil war era (towns vs. Maximilian) from 1483 to 1492 (91.01% of this mean); the highest number, 260 days (124.0% of the mean), during the 1540s and early 1550s. If the scope of these data is restricted to the first full century, 1436–1535, we find absolutely no correlation between real wages and the number of days worked.40 Some idea of the length of the normal working day, in this era, may be gleaned from data on seasonal wages. For in medieval and early-modern Europe, such wages were paid in accordance with the basic principle that “man works from sun to sun.” Indeed the stipulation that both men and women were required to work at least twelve hours during summer months can be found in the 1495 parliamentary statute on maximum wages. From mid-March to mid-September 201
4259 Ch06 4/12/02 10:12 am Page 202
202 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
JOHN H. MUNRO
all English “artificers and laborers” were to work from 5:00 a.m. to 7:30 p.m., with 30 minutes for breakfast, 30 minutes for nonemete (or sleep) and one hour for dinner; and in the other half of the year, they were to work from sunrise to nightfall.41 Certainly that customary working day and week, if not the law, prevailed in Great Britain, until the enactment of Fielden’s Law or the TenHour Day Act in 1847.42 In the Low Countries, numerous fifteenth-century guild records affirm that the normal paid working day was then identical to the English, with a full six-day working week.43 Far more so than in England, furthermore, the wage data for this region reveal a clear distinction between summer and winter wages (November to March); and, since the number of hours of daylight during winter months at this latitude (50° N) is about eight hours, the winter wage was generally from two-thirds to three-quarters the summer wage, despite the much higher cost of living during these cold, dark months. In London, and some other English towns, however, seasonal wages generally disappeared after the Black Death.44 Quite possibly the payment of a uniform wage rate, i.e. a winter wage equal to the summer wage, came to be an acceptable method of circumventing the Statute of Labourers’ wage controls, since the Statute and its enforcement really focused only on the summer maxima.45 Possibly the combination of the consequently higher wage throughout the entire year may have permitted some such substitution of greater leisure, at least in the summer months, for paid work. In the mid-fifteenth century, however, the London Bridgemaster did introduce a slight differential in seasonal payments (1441), and one that effectively raised the annual wage.46 In the Low Countries, south and north, where seasonal wages prevailed up to the French Revolution, Bruges was exceptional in generally having a uniform wage rate throughout the year, but only until the 1430s. Thereafter, rather surprisingly, lower winter wages were introduced into the building trades, while summer wages were raised.47
30 31 32 33
Does Institutional Wage Stickiness Reflect Adjustments in the Real Wage by Other Means: By Adjusting the Proportion of the Total Wage Package Made “In Kind’?
34 35 36 37 38 39 40
According to very commonly held views about medieval and early modern wages, most artisans and laborers supposedly received a significant portion of their total wage package “in kind’ or “in truck,” so that employers adjusted the real wage by altering the amount paid in the form of food, drink, and clothing amounts. Thus, according to this same view, nominal money-wage stickiness is illusory and/or irrelevant. That some medieval wages were indeed paid partly
4259 Ch06 4/12/02 10:12 am Page 203
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
203
in kind, especially in the agrarian sector, is incontestable. Most wage-payment records, including all the English statutes, are perfectly clear, however, in indicating whether the wage was in coined money alone, or “with food’ (or other benefits); and for the latter, the money wage was always proportionally less.48 Most economic historians in publishing wage data have been equally clear in specifying that the payments were in money alone. In manorial accounts, especially the Winchester accounts, for later medieval England, wage payments in kind were quite common up to the eve of the Black Death, often accounting for 50% of the total wage package; but after the Black Death the component in food and drink fell to a third or less of the total pay package and they became less common in the following generations, except principally at Battle Abbey. The Battle Abbey manorial accounts generally provide two series of wage payments for the same classes of craftsmen and agricultural workers: those paid in both money and kind, and those paid in coined money alone. The sum of the former seems to equal the latter; and the rates for those paid in money alone at Battle Abbey were identical or virtually identical to those for unspecified but presumably “money-alone” wage payments on other manors.49 One may surmise that when the price of foodstuffs fell sharply in the late fourteenth century (see Table 5), laborers and artisans would have resisted having any substantial portion of their pay package supplied in kind. There are very few available urban wage data before the Black Death; but the various accounts for later-medieval London and other smaller English towns (Canterbury, Dover, Exeter), do not indicate any significant evidence of payments in kind, other than some allotments of clothing (chiefly for a few master-masons and carpenters in long-term employ at Westminster Abbey).50 For these reasons, one may conclude that wage payments in kind did not appreciably alter the picture of real wages constructed by using money payments alone, at least not for the later Middle Ages, even if such views are not endorsed in Hatcher’s recent and otherwise persuasive article (1994) discussed above. Across the Channel, in the towns of later-medieval Flanders and Brabant, the town accounts provide virtually no evidence of any significant wage payments made in kind; and, for the rare instances in which they do occur in the records as a drincgeld, they account for no more than 1% of the wage packet, and seem to be for special workers hired outside the town.51 In fifteenth-century Leiden, the fullers’ guild ordinances (keuren), authorized by the town government (gerecht), are quite explicit in forbidding any payments in “truck” (kind), stipulating that payments be made each Saturday in silver coin alone. A remnant of some former partial payment in drink evidently continued in sporadic, occasional payments of drincgeld; but this too was paid in silver coin (a stuiver, 203
4259 Ch06 4/12/02 10:12 am Page 204
204 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
JOHN H. MUNRO
worth 2d groot).52 Similarly, in their analyses of medieval Flemish and Brabantine wages, Verlinden and Scholliers (1959–1965) and Van der Wee (1963) found little evidence of payments in kind; and thus they also provided data only on money wages. Nevertheless, the Flemish and Brabantine town accounts, which generally provide the full names and occupations of the craftsmen so paid, do provide evidence that many master masons, carpenters, slate-tilers, and street-pavers were also industrial entrepreneurs who made substantially more money than their daily wages by selling their supplies – bricks, stones, sand, lumber, nails, wire, etc. – to the city government.53 Indeed, some masters were also drapers, cloth merchants, brewers, or minor civic officials, many of whom so prospered in these roles that they ultimately relinquished their original status as dailywage earners. Nevertheless many other minor masters and virtually all journeymen (knapen) and other assistants did not seem to earn any such extra incomes, though some of them may have had their own small agricultural holdings, either within or just outside the town walls. Those circumstances obviously make a complete and proper estimate of real incomes difficult. As Van Zanden (1999, p. 178) has recently contended: This [artisan’s household] budget is made up of different sources of income, of which wage income is only one. But we assume it was an important source (and not a marginal one) for the European working classes of the early modern period. Moreover, the wage rate is exogenous for the household: it cannot influence its level in the short or the long run. This means that a rationally acting household will adapt its strategy to this given wage-level.
24 25 26 27
IV. THE KEYNESIAN INFLEXIBLE-WAGE MODEL AND THE NATURE OF MEDIEVAL LABOR CONTRACTS
28 29 30 31 32 33 34 35 36 37 38 39 40
Even if these recorded nominal money-wage data reflect market conditions, and are not institutional artifices, such wages would, however, clearly not be determined in the short-run by the Classical equation: i.e. that WL = MRPL – not then and just as certainly not now. As most contemporary macro-economists would agree, if only because of its obvious truth, “nominal wage rates are not determined on a daily basis according to the law of supply and demand . . . even though it is the real wage that determines the demand for and supply of labor;” but not necessarily all of them would find the full solution in “the modern interpretation of the Keynesian inflexible-wage model [which] focuses on the role played by labor contracts,” explicit or implicit.54 My own current research in the late-medieval era has revealed only one form of explicit long-term wage-contracts produced by formal and arbitrated collec-
4259 Ch06 4/12/02 10:12 am Page 205
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
205
tive bargaining between a guild and an association of entrepreneurial employers: those for the aforementioned fullers of Flanders, Brabant, and Holland, whose analyses require a completely separate article (Munro, 2002). As Knoop and Jones (1933, 1936, 1962, 1967), more recently Swanson (1988, 1989) and Blair and Ramsay (1991) have shown, from extensive investigations, medieval English building craftsmen did not enjoy any similar form of guild protection that would have permitted collective bargaining and labor contracts. Nevertheless Knoop and Jones have also demonstrated that for major building projects in later medieval England, France, and Italy, the chief master mason, acting more as a contractor-entrepreneur than as an artisan, often required or obtained written contracts that clearly specified not only the costs of building materials but also the rates of pay for his workers for the duration of the project. Sometimes, however, the contract was in the form of an opus ad tascam whose specified sum amounted to pay for piece-work accomplished, rather than daily rates. Some other older master masons rose in status to become wardens or foremen for large building projects, in the service of a secular or ecclesiastical lord, receiving a contract for many years, or even a lifetime, specifying an annual salary, with clothing and other material benefits. Many other labor contracts may have been unwritten, i.e. the “implicit contracts” (by “invisible hand-shakes,” etc.) about which modern macro- and labor-economists and some economic historians have published so extensively. Their analyses, however, cannot be readily applied to the often intractable medieval sources on artisans.55 We may at least surmise that the long-term and virtually permanent artisan-employees of the London Tower Bridge Authority probably did enjoy at least the privileges of an implicit contract. But, it seems hardly credible that short-term, part-time, itinerant craftsmen, working for a variety of employers (while frequently mixing occupations), for discrete but indeterminate periods, would have benefited from any such form of implicit contracts, even if their wage patterns demonstrated the same degree of wage stickiness. In any event, fully enforceable written contracts would have succeeded, at the most, in fixing wages for only a few years at a time – or for medieval building projects, for their duration. Thus, they can hardly explain these many examples of very long-term wage stickiness lasting for over half-a-century, or as much as 174 years, in that singular Oxford example just cited. Nevertheless, in the absence of good data and established theory, one may still offer credible (rather than merely plausible) hypotheses to explain at least some forms of wage stickiness. During periods of inflation, most employers were reluctant to award nominal wage increases that matched rising consumer prices and living costs, especially if the prices of their own 205
4259 Ch06 4/12/02 10:12 am Page 206
206 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
products were not rising as steeply as were food prices. Despite the emphasis to be given to deflationary periods in this analysis, certainly continental western Europe, during the Hundred Years War era especially (1337–1453), experienced frequent, if short term, inflations from periodic war-induced coinage debasements. With so many jagged oscillations in prices, we, as modern computer-assisted observers, are better able to detect those trends than were medieval employers. If periods of late-medieval inflation were relatively brief in duration – in contrast to the sixteenth-century Price Revolution era, most employers would probably have not encountered any strong pressures to raise money wages, despite the short-term decline in real wages. As Keynes noted (1936, pp. 12–13), in citing one of his chief objections to one of the prime “postulates” of Classical theory: “A fall in real wages due to a rise in prices, with money-wages unaltered, does not, as a rule cause the supply of labor available on offer at the current [money] wage to fall below the amount actually employed prior to the rise of prices.” Indeed modern macro-economic models convincingly demonstrate that under such inflationary conditions wage stickiness “results in an increase in employment and output,” and usually, though not necessarily, in employer profits (Wilton & Prescott, 1987, pp. 216–217). Conversely, during periods of deflation, most wage-earning employees, whether or not they were truly suffering from a “money-illusion,” were and still are extremely hostile to any attempts to reduce their nominal wages, even if, with some inevitable wage reductions, they still ended up better off in real terms. Workers surely had and still do have some legitimate concern that wage rates and other benefits once cut will not readily be restored, without some considerable struggle, once market conditions improve and prices begin to rise again. As Keynes also cogently remarked (1936, p. 15): “Every trade union will put up some resistance to a cut in money-wages, however small.”56 In late-medieval Flanders and Holland, that is precisely what the fullers did, in the form of organized strikes (uutgangen), when faced with wage reductions during periods of monetary reform and deflation.57 If other latemedieval wage-earning craftsmen, in both England and the Low Countries, did not enjoy similarly strong guild protection, undoubtedly many, with or without less formal organizations, were still able to convey their opposition with sufficient conviction or hostility to prevent any nominal-wage reductions. Obviously those unprotected, such as itinerant or part-time agricultural laborers, were much more likely to suffer nominal-wage cuts during periods of deflation, as seems to have been the case in both late-medieval England and Flanders (see below, pp. 226–227).
4259 Ch06 4/12/02 10:12 am Page 207
Wage Stickiness and Monetary Changes 1 2
207
V. MONEY, PRICES, AND WAGES BEFORE THE BLACK DEATH: ENGLAND, 1280–1348
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Far more so than any other region in late-medieval Europe, England and the Low Countries (the latter from at least the 1350s) both offer an abundant amount of price and wage data, in virtually unbroken annual series, to investigate the nature of long-term wage-stickiness, and in particular its most interesting aspect: namely, downward wage-stickiness during periods of deflation, and thus what Phelps Brown and Hopkins called the “rachet effect” (see p. 196 above). But not until the later fourteenth century does the evidence support the view that wages exhibited a greater degree of rigidity or “stickiness” during deflations than during intervening periods of inflation. Therefore, the proper place to commence this aspect of study on historical real wages is in a previous era of deflation, which took place during the second quarter of the fourteenth century (at least, in England). That deflation had followed the era of the Great Famine (1315–1322), which had marked the culmination of the long, generally sustained inflation of the so-called “long thirteenth century.” For Postan and many other historians, the end of this era also marked the onset of an incipient Malthusian crisis of “overpopulation,” deteriorating living conditions for the lower strata of society, and indeed malnutrition and higher mortalities, even before the onslaught of the Great Famine.58 It must be noted, however, that the wage and price data compiled by Thorold Rogers (1866, 1882) and Phelps Brown and Hopkins (1955, 1956, 1981) do not indicate any deterioration in the real wages of manorial building craftsmen before the Great Famine; and such evidence shows indeed that nominal wages were rising, and keeping pace with prices until that calamity struck (Tables 5–6). Thereafter, in the decades following the Great Famine, we witness one of the most striking and puzzling phenomena in all of England’s recorded monetary and price history: a severe, indeed dazzling, plunge in English mint outputs, still entirely in silver; and, following that, an almost equally drastic deflation, reflected in the 35% fall in the Phelps Brown-Hopkins price index, from 138 in 1321–1325 to just 90 in 1341–1345, on the eve of the Black Death (Fig. 1; Tables 3–5). Postan, of course, evidently wanted to attribute the fall in at least the grainprice index to the demographic consequences of the Great Famine, or rather to a more general Malthusian crisis of overpopulation in a relatively primitive agrarian economy: “when the poorer lands, no longer new, punished the men who tilled them with failing crops and with murrain,” so much so as “to send the population figures tumbling down.”59 Lawrence Poos has recently provided evidence to indicate significant population decline in rural Essex after the Great 207
4259 Ch06 4/12/02 10:12 am Page 208
208 1 2 3 4 5 6 7 8 9 10 11 12 13 14
JOHN H. MUNRO
Famine and on into the post-Plague era.60 For Europe more generally one can cite evidence for regional depopulations in early-fourteenth century Provence and Tuscany, though evidently related to the horrendous warfare then afflicting these regions.61 Campbell (1983a, 1983b, 1984), however, has provided equally compelling evidence of continued population growth in post-Famine Norfolk, while also demonstrating that late-medieval English agriculture was far less primitive and far less prone to Malthusian pressures than Postan had suggested. Smith (1991) has provided a masterful survey of demographic developments in rural England, from 1300 to 1348, which, while inconclusive, documents complex regional variations in demographic decline, with continued if slow growth, or stagnation; and thus it does not lend any support to Postan’s drastic views.62 Nor does the recorded fall in nominal wages, from 1337–1338 (see below, p. 209) logically support a depopulation hypothesis. Furthermore, the decline in prices was general, and not limited to just grains (see Table 5).
15 16 17
Some Monetary Causes of Deflation
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
This prolonged fall in prices was instead genuine deflation, for which monetary reasons must be sought. The true explanation for an evidently stark monetary contraction and consequent deflation, apparent also from the Tuscan price data,63 remains a mystery that cannot satisfactorily be resolved. Possibly it was due to a relative scarcity of precious metals, if, as several historians have asserted, the major German and Central European silver mines had begun to experience not just diminishing returns but serious physical depletion by the early fourteenth century, while the European economy continued to grow, and with it, the aggregate demand for coined money.64 Indeed, as Mayhew (1974) has also demonstrated, contrary to another of Postan’s assertions, coined money is perishable to some considerable degree: from wear, tear, and normal loss in circulation, from shipwrecks, unrecovered hoards, conversion into jewellery and plate, etc., so that the money supply will indeed contract if not continually replenished with fresh minting.65 For England itself, some historians have also suggested that the crown’s foreign military expenditures (under both Edward II and Edward III) had led to major outflows of bullion, though the fall in mintoutputs and the onset of deflation seems to precede the evidence for any such drastic bullion outflows.66 Finally, since England in this era was minting only silver, and no gold before January 1344 (none since 1257), the very dramatic rise in the bimetallic ratio, from about 12.0 : 1 in the 1290s to 14.2 : 1 in the mid-1320s may have instigated a large outflow of silver coinage to acquire the higher valued gold. Indeed, such bullion movements may have been necessary
4259 Ch06 4/12/02 10:12 am Page 209
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
209
to permit England’s inauguration of an effective gold coinage in the period 1344–1352, though with a then falling bimetallic ratio.67 Some very general indication of possible bullion outflows from England and a relative scarcity of specie during the second quarter of the fourteenth century may be found in the coinage-output statistics (Tables 3–4). The mean annual values of those outputs (all in silver) fell from a peak of £125,836 sterling in 1306–1310 to a nadir of £381 in 1326–1330; recovering to a mean of only £7,091 in 1346–1350, at the outbreak of the Black Death. Such mint-accounts provide, however, only a very general and very tenuous guide to current monetary conditions. They can be of some value in that years with very low mint outputs were generally followed by prolonged deflation (marked as well by complaints about the scarcity of specie); and, conversely, years of very high mint outputs generally coincide with eras of sustained inflation. But extrapolating a nation’s current money supply from these accounts is an enterprise fraught with great dangers, for many complex reasons, the most important of which is that mint-accounts combine stocks and flows in unpredictable and unquantifiable fashions.68 Nevertheless, some brave historians believe that a 30-year running average of such outputs may provide an acceptably reliable indication of the coined money stock. Recently, two economic historians have used a combination of mint accounts and coin hoards to show that the English money supply probably contracted by over one half in this era.69 What is especially striking and peculiar about this deflation, though affording further evidence that it was a genuine deflation (Tables 4–5), was the marked decline in nominal wages shown in Table 6. From about 1337 to 1340, the mean money wage of a master building craftsmen in southern England fell from 4d per day to 3d per day, a decline of 25%; and that very low mean wage-rate was maintained until early 1351, i.e. several years after the Black Death (Fig. 4).70 As Phelps Brown and Hopkins have indicated (1955/1981, pp. 11–12), the only other period in recorded English price-history with any such decline in the nominal wages of building craftsmen, came almost six-centuries later: the post-World War I slump of 1920–1923 (31.3% decline) and the early depression years of 1929–1934 (8.3% decline). In the earlier part of this early-fourteenth century period, before this fall in nominal wages, i.e. with continued nominal wage rigidity,71 real wages did rise, though the brief rise appears to be dramatic only because of the recovery from the drastic nadir of the Great Famine years. When the real-wage of those master craftsmen peaked in 1334–1335, it was not appreciably higher than in the very early years of the century (1303–1307). Then their real wages suffered a sharp fall, as also shown in Table 6; indeed they fell quite steeply before the Black Death, with the initial 209
4259 Ch06 4/12/02 10:12 am Page 210
210 1 2
JOHN H. MUNRO
recovery of the price level from 1344, and especially with the inflation that soared immediately after the Black Death, one that endured for a full generation.
3 4 5 6
VI. ENGLAND FOLLOWING THE BLACK DEATH (1348): MONEY, WAGES, AND THE STATUTE OF LABOURERS
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
That is why the Ordinance (1349) and Statute of Labourers (1350–1351) were so very unreasonable and cruel, but also so difficult to enforce: in attempting to fix money-wages at the pre-Plague level, when both money and real wages had been so unusually low (Tables 2, 5, 6).72 Thus, specifically forbidding anyone to offer or accept any wages higher than those prevailing in 1346, the Statute of Labourers set the maximum summer wages (Easter to Michaelmas) for master masons, carpenters, and tilers, “without meat or drink,” at 3d per day; for their servants and laborers, at 1.5d a day; but it also permitted a rate of 4d per day for master free-masons.73 This harsh statute remained in force (reconfirmed numerous times), ostensibly on a national basis, until the parliament of 1444–1445, which finally raised the maximum daily rate for such craftsmen to 5.5d for summer and 4.5d for the winter season, without food and drink (or: 4d and 3d, respectively, with food and drink).74 As noted earlier in this study, London had independently, in 1350, established its own maximum wage of 6d per summer and 5d per winter day for these same craftsmen (Table 2);75 and such rates were already in force at Westminster Abbey, for master building craftsmen in 1349.76 Thereafter, but not immediately after the Black Death, money wages did rise, certainly for these urban-based craftsmen.77 At Oxford, where most building craftsmen had not suffered the nominal wage cuts in the 1330s, the prevailing daily wage rate rose from 4d to 5d during the 1350s; in other small-sized towns the rate rose from 3d to 5d by the end of the decade. Not until 1363, as noted earlier (p. 187), did Oxford masons and carpenters gain the daily rate of 6d, at least for summer work, that London had authorized in 1350; and not until about 1407–1409 did the mean daily rate for craftsmen in the other smaller towns achieve the same 6d level, i.e. the mean rate for the Phelps Brown and Hopkins index, unchanged until 1536.78 In many, indeed evidently most, of the Winchester manors and at Battle Abbey, the daily wage for such building craftsmen, while soon rising back to the 4d level that had prevailed from about 1310 to 1337, remained fixed at that truly low level – even if 25% above that stipulated by the Statute of Laborers, with some occasional exceptions for senior craftsmen at 5d daily, until about c.1410–c.1425. 79 Such evidence may lend support to the oft-cited contention
4259 Ch06 4/12/02 10:12 am Page 211
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
211
of Holmes (1975, p. 115), with some support from Britnell (1990), that by the eve of the Peasants’ Revolt, in the later 1370s, the landlords, and especially the great landed magnates, had succeeded in obtaining a greater share of the national income than they had enjoyed before the Black Death. Some post-Revolt judicial records do provide evidence for the payment of higher wage rates in some Essex manors, though largely for agricultural workers. From these records, Nora Kenyon Ritchie concluded that aggressive new leaseholders of former domain (demesne) lands were more willing to offer higher wages to attract labor than were traditional manorial lords. Rather surprisingly, no one has pursued or investigated this intriguing thesis since the publication of her article, in 1934.80 London’s own 1350 wage ordinance was soon if not immediately allowed to lapse. For when the Tower Bridge accounts commence in 1381, the prevailing daily rate then ranged from 7.0d to 7.5d, and indeed for winter months as well as for the summer. At Westminster Abbey, however, the rate was less, at 6.667d per day, though with some extra material benefits not given to the Tower Bridge craftsmen. In the Bishop of Winchester’s Southwark manor, from as early as 1406, and in most London guild houses by 1420, that rate had increased to 8d daily (without food and drink); and that same rate had become uniform in the Tower Bridge Master’s accounts from June 1425, though not at Westminster Abbey until 1439–1440 (perhaps because of those extra benefits given to some craftsmen).81
23 24 25
VII. MONETARY CHANGES, INFLATION, AND REAL WAGES, AFTER THE BLACK DEATH, 1348–1375
26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
These rather complex data indicate that the commonly expressed views about post-Plague wages in England are much oversimplified, in particular the comment in a recent article by Penn and Dyer (1990, p. 356): that “the evidence for a rise in both cash wages and real wages . . . coinciding with the sudden and sustained population decline after the Black Death of 1348–1349 has been well established.” What must be challenged in this statement is the verdict on real wages, for the very simple reason that the Black Death, throughout western Europe, was followed by a horrendous inflation that endured for the ensuing quarter-century in England, and for an even longer period in Flanders and Tuscany.82 Thus, as the Tables 4–10 clearly indicate, its initial consequence was to swamp the rise in nominal wages for most workers, and most especially for Flemish craftsmen. Note from Table 8 that the Flemish price level soared from a mean of 50.57 in 1349–1350 to a peak of 124.72 in 1386–1390, an overall rise of 146.6%. Table 10 provides both the nominal and real wage 211
4259 Ch06 4/12/02 10:12 am Page 212
212 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
JOHN H. MUNRO
indexes (with the same base: 1451–1475 = 100) for master building craftsmen in Bruges, in quinquennial harmonic means, and also the mean number of “commodity baskets” that a master mason or carpenter could have purchased with his estimated annual money-wage income (for an estimated working year of 210 days). Despite more than a doubling in nominal wages over this period (in Bruges, from 5d in 1349 to 12d per day in 1387), the real wage index for Bruges building craftsmen thus fell from a harmonic mean of 89.88 in 1349–1350 (equivalent to 16.440 commodity baskets a year) to one of just 62.31 in 1356–1360 (11.397 commodity baskets), then rose, only to fall again to 65.90 in 1381–1385 (12.053 baskets), but partially recovering to 77.38 in 1386–1390 (14.152 baskets), with the final increase in money wages. Even so, real wages for Flemish building craftsmen in the late 1380s were still below those of the Black Death era itself (Fig. 5).83 In England, urban craftsmen in the small to middle-sized towns, and certainly most of the manorial craftsmen and other artisans, also did not fare that well immediately following the Black Death, though certainly they did not fare as badly as the Flemish craftsmen. As Table 6 indicates, the real wage index for master masons and carpenters at Oxford, Cambridge and other small to medium size towns in southern England (Canterbury and Exeter, etc., but excluding London) did indeed fall in the immediate aftermath of the Black Death: from a harmonic mean of 55.52 in 1341–1345 to one of just 44.56 in 1351–1355. Shortly after, however, real wages did recover, reaching an index of 59.01 in 1356–1360, but then fell again to 56.83 in 1361–1365, while rising thereafter. But even in 1371–1375, the real wage index of 61.61 was still below the mean index of 63.26 achieved in 1336–1340. Indeed the significant, sustained rise in this real wage index did not really commence until the later 1370s (Fig. 4). In many manorial estates, however, especially those of the far-flung rural holdings of the Bishop of Winchester, building craftsmen had to wait until the early fifteenth century to achieve a significant gain in their real incomes (at least those measured in terms of wages given without food and drink): at Hinderclay, Ecchinswell (Itchingswell), Overton, Taunton, Esher, Wycombe; and also Battle Abbey, Redgrave manor (Bishop of Bury St Edmonds), and at Winchester College, to name only a few.84
34 35
Monetary Factors in the Post-Plague Inflations
36 37 38 39 40
The post-Plague European inflation was again clearly a monetary phenomenon. Depopulation, after all, according to some disciples of the Postan school, is supposed to cause deflation, not the opposite. The monetary reasons – if we properly relate monetary and real variables – are not difficult to seek. As Herlihy
4259 Ch06 4/12/02 10:12 am Page 213
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
213
(1967, p. 125) so aptly commented, “men were dying, but coins were not” (at least not so quickly). Thus, whatever the current status of western European precious-metal mining, the effect of such drastic depopulations, perhaps as much as 40% of the total inhabitants, from bubonic and pneumonic plagues, if not so much from warfare, was undoubtedly to augment dramatically the per capita supplies of coined money. Secondly, the fiscal consequences of warfare in western Europe (including the concurrent Italian wars), of increased taxes and other levies, evidently also induced some considerable dishoarding. At the same time the French, Flemish, Brabantine, Aragonese, and various Italian governments, to mention only a few, sought both to finance and to facilitate the necessary cash flows for warfare by engaging in drastic coinage debasements, some severe enough to promote a veritable “flight from coinage.” The English crown, however, was a singular exception, by not engaging in these monetary manipulations, which indeed were so drastic in both France and Flanders: for it undertook only one minor weightreduction in its silver coins, in 1351, thereafter maintaining a perfectly stable coinage, in both metals, until 1411–1412 (Tables 3–4; Fig. 1).85 Thirdly, some historians have suggested that the socio-psychological consequences of both plague and warfare, especially with such devastating and arbitrary death tolls, was to foster a fatalistic yet hedonistic spending spree, facilitated all the more by suddenly inherited cash balances. In support of that thesis they cite the evidence from paintings, architecture, adornments in dress, and examples from Italian literature – e.g. Boccaccio’s Decameron.86 The overall consequence, for England and Flanders, as demonstrated in the Tables 3–4, was to produce a truly momentous increase in coinage outputs, whose inflationary consequences can hardly be disputed. One will note from Tables 5 (England) and 8 (Flanders) that all three major price series, and not just those for grains, rose during this quarter-century period following the Black Death: to the late 1370s in England; and to the 1380s in Flanders, whose far more drastic and regrettably frequent debasements continued until the 1389–1390 monetary reform.
32 33 34 35
VIII. “BULLION FAMINES” AND PERIODIC DEFLATIONS: IN THE LATER FOURTEENTH AND FIFTEENTH CENTURIES
36 37 38 39 40
Thereafter, during the final quarter of the fourteenth-century, much of western Europe – certainly including England, the Low Countries, Tuscany, and AragonNavarre – experienced an equally dramatic deflation that lasted until well into the fifteenth century (Tables 4, 5, 8, Figs 2–3).87 Then with the resumption of 213
4259 Ch06 4/12/02 10:12 am Page 214
214 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
intensive warfare, from the Battle of Agincourt in 1415 to the dénouement of the Hundred Years’ War in the 1440s, another era of inflation ensued in much of north-western Europe, especially in France and the Low Countries, which consequently experienced severe coinage debasements. With the end of the Hundred Years War, and indeed even earlier, from the early 1440s, that warinduced inflationary era was succeeded by yet another era of deflation, this time as prolonged, and by the 1460s, as pronounced as that of the late fourteenth century. Though temporarily disrupted by warfare, by other “supply shocks,” and again by severe coinage debasements, especially in the Low Countries during the 1480s, this deflationary trend lasted until about 1515. Its subsequent reversal marked the beginning of the Price Revolution era. Again, as in all the previous inflationary and deflationary cycles, the three major price-indexes in both England and the Low Countries rose and fell more or less together, if by no means exactly in tandem, during all these fifteenth-century cycles (Tables 4, 5, 8).88 As Tables 3–4 and Figs 1–3 clearly, show these periods of very pronounced deflation, in the later fourteenth and mid-fifteenth century in particular, were accompanied by severe slumps in the gold and silver coinage outputs in both England and Flanders, almost as severe as those for early fourteenth-century England, while the brief intervening periods of inflation were similarly accompanied by upsurges in mint outputs.89 If these brief, periodic bouts of late-medieval inflation can be attributed to a combination of expansionary war-induced fiscal and monetary policies, and especially, in the Low Countries, by severe coinage debasements, the longer periods of deflation are much more difficult to explain. Complicating any such analysis is the now large, contradictory, and often confusing literature on the so-called “Bullion Famines” of this era, from the 1370s to the 1470s. The most favoured theories advanced by proponents of the “bullion famine” thesis are the following three. First, that western Europe experienced a severe worsening of the late-medieval mining slump, so that even the opening of some new mines in Serbia and Sardinia failed to compensate for the sharp decline in outputs elsewhere, and indeed a veritable cessation of silver mining in some regions. Second, that western Europe suffered a steadily worsening of balance-of-payments deficit with Asia, with consequent bullion outflows via the Levant and the eastern Baltic; and third that this deficit was aggravated by a severe diminution, though not a complete cessation, in European gold imports from the Italian trade with North Africa.90 Nevertheless, if Ashtor has provided some impressive evidence for Venice’s large silver exports to the Levant in the later fifteenth century, there is no concrete evidence to show that an overall European balance-of-payments deficit had been worsening from a full century earlier.91 Indeed, a major factor that helped to end
4259 Ch06 4/12/02 10:12 am Page 215
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
215
the so-called “bullion famine” era was the Central European silver-copper mining-boom, which began during the very nadir of deflation in the 1460s, when the consequently high value of silver induced a veritable revolution in mining technology (in chemical and civil engineering). That boom augmented European silver production by over five-fold by the 1540s; and without such large increases in its silver stocks, Venice would never have been able to conduct such an increased volume and value of trade with the Levant in the 1490s.92 The fact, moreover, that even before that mining boom had commenced, mints in England, France, and the Low Countries had all succeeded in reviving, if not in fully recovering, former levels of coinage outputs, in the 1420s and early 1430s, suggests that there was no downward linear trend indicating any general drainage of European bullion to the “East.” Thus, additional, if not alternative, explanations for periodic bullion scarcities should be sought for this era. Such an explanation may be found in examining the behavior of bullion flows rather than of monetary stocks: specifically, in two sets of adverse changes in the income velocity of money, or in the demand for idle cash balances, that may be related to the pernicious effects of warfare and plagues, from the 1370s. First, the now chronic and even more devastating warfare throughout so much of Europe, combined with drastic depopulations, produced severe dislocations to established patterns of international trade, while sharply raising transaction costs in that commerce, thus even more reducing flows of both commodities and bullion. Worse, responses to the ancillary manifestations of that warfare, in terms of commercial blockades, confiscations, and especially coinage debasements, radically reduced bullion flows even more. In particular, most west European rulers, in defending themselves against aggressive debasements by their neighbors, necessarily banned the domestic circulation of most foreign coins, especially silver coins (all the more subject to surreptitious debasements); and such bans also stipulated that all or most foreign coin be surrendered as bullion to the ruler’s mints. More important, in seeking to attract more bullion to their own mints, to increase coinage outputs and their seigniorage revenues, virtually all rulers banned its export. Even when enforcement of those bans failed to prevent international exchanges of precious metals, they still depressed monetary and trade flows by raising transaction costs. Finally, the noxious combination of such warfare, famines, plagues, the imposition of higher taxes and such monetary policies, defensive or aggressive, and the consequent commercial disruptions led to periodic but often severe economic depressions, certainly commencing by the late 1370s and 1380s. Such conditions also bred a more general climate of insecurity and pessimism that further discouraged spending and investment, increased hoarding, and further aggravated those depressions. Thus, by the 1370s, that post-Plague social climate 215
4259 Ch06 4/12/02 10:12 am Page 216
216 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
JOHN H. MUNRO
of hedonistic spending-sprees had given way to much more pervasively gloomy and pessimistic outlooks amongst the populace in general, one that increased their demand for idle cash-balances, i.e. elevated their tendency to hoard.93 While fourteenth-century England has provided an unusual number of extant coin hoards, their survival may be accidental and their evidence is inconclusive.94 Better evidence of late-medieval hoarding, especially following the Black Death, can be found in the greatly increased use of gold and silver ornamentation in the form of plate, jewellery, brocaded textiles, belt buckles, wall hangings, and furniture. Such ostentatious artistic display provided not only aesthetic satisfaction but also a feasible means of storing precious metals that could readily be restored to coinage, when so needed.95 Hoarding is obviously a self-justifying deflationary phenomenon; for, as prices fell from all combined circumstances of monetary contraction, the rational response of the money-holding populace would have been to save rather than to spend in the present, in anticipation of even lower prices in the future. Furthermore, the general forces of more gentle, less disruptive population decline, with smaller communities and smaller families, may also have acted to curb the income-velocity of money, i.e. to increase the relative demand for real cash balances. Finally, for England itself, one may cite a particularly significant monetary change: the fact that the overwhelming bulk of the coinage, sometimes as much as 95% by value, was struck in gold rather than silver – and gold coinages are necessarily undertaken at the expense of silver (Table 3).96 For if the smallest circulating gold coin, the quarter-noble, worth 20d or 20 current silver pence could purchase 20 gallons of ale or loaves of bread, one can readily imagine that its velocity in circulation would have been very low, compared to that of a silver pence or halfpence.
28 29 30
IX. THE QUESTION OF LATE-MEDIEVAL CREDIT: AND THE EFFECTIVE MONEY SUPPLY
31 32 33 34 35 36 37 38 39 40
Nevertheless, according to many historians, a supposed growth in the use of credit during the later Middle Ages should have fully offset or counteracted those deflationary forces: particularly through the agency of deposit-and-transfer banking and bills-of-exchange banking. Yet neither was an innovation in this era, and both saw their most rapid initial diffusion in the last half-century of the preceding Commercial Revolution era (c.1180–c.1320). The reasons why credit instruments largely failed to provide a sufficient remedy for periodic monetary contractions in late-medieval Europe are very complex, but may be briefly summarized here.
4259 Ch06 4/12/02 10:12 am Page 217
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
217
First, late-medieval Europe experienced very few additional innovations; and most credit instruments were still far from being effective substitutes for coined money, with the possible though still dubious exceptions to be found in a few West-European towns whose commerce was dominated by Italian merchantbankers. Second, much more widespread and more powerful forces, economic and political, involving increased hostility from both Church and state, seriously impeded the employment or circulation of credit, with a multitude of examples to be cited in England, France, the Low Countries, and many Hanse towns, if not so much in Italy itself. Indeed, in both England and the Low Countries, late-medieval nationalist monetary policies effectively prevented the emergence of deposit-banking in the former, and virtually closed down such banks in the latter, following the Burgundian unification of the Low Countries, in a series of ever more severe ordinances (1433–1435, 1467, 1480 and 1489). Furthermore, various Burgundian ordinances, both ducal and municipal, also sought to restrict pawnbroking (1442, 1451, 1457, 1473, 1477).97 Third, because of those increasingly hostile attitudes from state-dominated legal institutions, despite the growth of a more independent and international Law-Merchant, the enforcement of debt repayments, especially those involving (non-notarized) holograph documents, became even more costly and ineffective, thus restricting credit instruments to a small circle of merchants, chiefly Italian, who knew and trusted each other. Fourth, most European states and principalities, even in Italy, had failed to provide the legal institutions and sanctions for true negotiability: the conversion of a debt instrument into cash or goods, or to be used in place of cash, as the fundamentally necessary condition to permit credit instruments effectively to augment the money supply (Munro, 1994b, 2000). Fifth, consequently and in sum, credit instruments were far from being divorced from the use of coined money. Spooner (1972, pp. 3, 53–71), Mueller (1984), Spufford (1986, pp. 346–348), and Nightingale (1990) have all effectively demonstrated that credit either expanded or contracted with the coined money supply in the late-medieval and early-modern economies, usually in a non-proportional fashion. As Mayhew (1987, p. 121) so aptly commented as well: “credit reflected the supply of coin rather than compensated for it.”
33 34 35 36
X. THE LATE-MEDIEVAL DEFLATIONS, WAGE STICKINESS, AND THE COURSE OF REAL WAGES: ENGLAND, 1376–1500
37 38 39 40
With the onset of this general and very widespread deflation,98 wage stickiness now assumed a significance never before revealed in European economic history. For in England, at least, unlike its experience in the second quarter of 217
4259 Ch06 4/12/02 10:12 am Page 218
218 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
the fourteenth-century, most money wages did not decline (Fig. 4). Beveridge (1936, p. 31) was the first to call attention to this phenomenon, commenting that “from 1290 to 1379, money wages and the price of wheat move decade by decade always in the same direction, rising and falling together; thereafter they part company completely.” To explain this striking dichotomy, he argued that the socio-economic consequences of the Black Death and subsequent plagues were to transform money wages from their original institution “as a substitute either for customary service or for allowances in kind,” both tied to living costs, into true payments determined by the actual labor market.99 The previously noted fall in the proportion of the wage typically paid in kind, from about one-half before the Black Death to one third thereafter, provides some support to Beveridge’s thesis. In his now classic article on the Black Death, however, Bridbury (1973) came to a strikingly different conclusion: namely, that the initial mortality from that disaster “was quite incapable of altering the social and economic relationships because so much of the population was surplus by the fourteenth century;” that the Plague was “more purgative than toxic;” and that not until the later 1370s did cumulative depopulation succeed in fundamentally altering the land : labor ratio to raise labor productivity and thus real wages. Though his arguments should not be dismissed out of hand, they are not really supported by the wage and price evidence, nor by economic theory and logic.100 In fact, what one now witnesses, from the late 1370s, is the nominal and real-wage pattern that will prevail, certainly in Western Europe, for the next six centuries.101 With a few but significant exceptions, to be noted later (pp. 226–230), real wages in England, and then subsequently in the southern Low Countries as well, began a century long rise that was fundamentally if not exclusively due to wage stickiness with prolonged deflation. Again with just a few exceptions, the only significant retreats in this onward march of rising real wages, were similarly the result, and sole result, of some periodic and generally brief inflations, which were far more prominent in the Low Countries than in England (Tables 5–6, 8–14). Yet it would be wrong to rule real factors entirely out of court; and one may justifiably postulate that the Classical theory does retain some merit. Possibly the high real wages recorded in England and Flanders during the mid-fifteenth century, during these prolonged conditions of both deflation and wage stickiness, were justified, in “clearing labor markets,” so to speak, if there had been some accompanying rise in the marginal revenue product of the artisans enjoying such high wages. How in particular are we to explain the undoubtedly widespread rise in nominal wages during the early to midfifteenth century, in both England and then in Flanders, despite generally
4259 Ch06 4/12/02 10:12 am Page 219
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
219
prevailing low prices, thus further increasing the real wages, without some productivity increases? In the southern Low Countries real-wages for building craftsmen reached their medieval peak in the mid-1460s (Tables 10, 13, 14, Figs 5–6) and in southern England that peak occurred a decade later: in 1477, when RWI = 123.76 (harmonic mean of 111.04 for 1476–1480: Table 6, Fig. 4). Shortly after that peak, however, those English real wages fell quite sharply, to a harmonic mean of 78.51 in 1481–1485, because of a sharp rise in the price level (to 127.38: Table 5). Why it did so, however, is more difficult to explain than the contemporaneous though more prolonged rise in the Flemish and Brabantine price level (Tables 8–9), which was primarily due to drastic coinages debasements and supply shocks from warfare (see below, pp. 224–226). Thereafter, English prices fell sharply to a mean of just 96.70 in 1496–1500, thus producing a rise in the craftsmen’s real wage index to 103.41. Not until 1516–1520, with the swift onset of the Price Revolution in England, did they sustain a sharp fall, to just 83.03 (harmonic mean); and then, during even the initial phases of the Price Revolution, their real wages continued to plunge swiftly (to 50.78 in 1551–1555). But that phenomenon is beyond the scope of this current study, which must terminate in 1500.
20 21 22 23
XI. DEFLATION, WAGE STICKINESS AND THE COURSE OF REAL WAGES IN FLANDERS: THE CASE OF THE FULLERS, 1375–1430
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
As just suggested, late-medieval Flanders and Brabant provide both fascinating comparisons and contrasts with the contemporary English experience in real wages. The first and certainly the most interesting comparison concerns the fate of Flemish textile artisans, with a case study of the fullers, and then of building craftsmen, during the equally sharp deflation of the very late fourteenth and early fifteenth centuries. In Flanders itself, as Tables 3–4 and 8 clearly indicate, the steep decline in coinage outputs and fall in the price-level had commenced about a decade after such phenomena had become apparent in England, from the very late 1380s. Despite a recoinage, with full-scale monetary reform in the early 1390s, coin outputs fell throughout that decade, ending in 1402 with a complete closure of the mints (Table 3). At the same time, Flanders’ composite price level (1451–1475 = 100) fell from a peak of 133.93 in 1387 to just 82.12 in 1405, a fall of 38.7% (Table 5). Aggravating this steep deflation, though by no means solely responsible for it, was Duke Philip the Bold’s dramatic monetary reform of 1389–1390, a renforcement that strengthened the silver coinage by 31.6% – i.e. in reminting the current debased coinage into fewer 219
4259 Ch06 4/12/02 10:12 am Page 220
220 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
coins, but with a much higher silver content. That ordinance also altered the mint ratio in favor of silver (from 10.41 : 1 to 9.68 : 1), thereby reducing the value of the Flemish gold noble from 102d to 72d groot.102 Almost immediately, in 1390, the Burgundian ducal government, along with many town governments, stipulated a proportional reduction in wages and some other money-of-account payments, in order to accommodate these changes, and in particular to compensate those merchants, textile industrialists, and financiers who had depended upon payments in gold nobles. Indeed, a wage reduction inversely proportional to the silver renforcement of 31.6% would have been exactly 24.0%.103 Thus, in Bruges, the daily money-wages of building craftsmen, which had risen during the inflationary 1380s, though not by as much as the extent of general inflation, were abruptly and arbitrarily cut by 25%: from 12d to 9d groot, for masters; and from 6.0d to 4.5d, for journeymen.104 According to some contemporary reports, such wage reductions did produce riots and considerable social unrest in Bruges and other Flemish cities.105 In retrospect, civil disturbances in Flanders were rather less pronounced than might have been expected. But presumably only the guild artisans, a minority, were in position to voice strong objections; and evidently many, if not all of them, grudgingly accepted the argument that their new wages, while nominally lower, were being paid in the new “strong” rather than the old “weak” money.106 Quite evidently wage reductions were imposed much more easily under conditions of a publicly promulgated coinage renforcement, when the public could readily perceive the physical differences between the new and old coinages, than under deflationary conditions without any such coinage changes.107 Nevertheless the various attempts by Flemish weaver-drapers to impose similar wage cuts during the early 1390s did produce some labor strife in the Flemish cloth industry. Only the fullers, however, have left us any concrete records of concerted opposition, and only in three drapery towns: Ghent, Wervik, and Kortrijk. That opposition of course depended upon the relative strength of their ambachten or guilds. The weakest were evidently the Ghent fullers, who, in 1361, had lost both their representation in the town government and the right to select their own leaders.108 In 1373, their weaver-draper employers had brutally crushed a fullers’ strike; but Count Louis de Male – no friend of the weavers – nevertheless awarded them a wage increase: to 45d groot Flemish, for a master and two journeymen, in fulling a maerclaken broadcloth in three days (the earlier wage is unknown). Thereafter, unlike the building craftsmen in Bruges, they failed in all attempts to gain further wage increases.109 Then, with the implementation of the 1390 monetary reform, they evidently accepted, without protest, a 29% reduction in their combined money wage, to 32d groot.110 Since the inflationary
4259 Ch06 4/12/02 10:12 am Page 221
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
221
conditions of the 1380s still prevailed, that meant that their real wage was 41.4% lower than what they had earned in 1373 – but only for that year. As Tables 8 and 11 reveal, rapid deflation soon came to their rescue, so that their mean real wage – as measured in the number of Flemish commodity baskets acquired with their annual money wage income – in 1396–1400 (6.913 baskets) was 7.2% higher than a decade earlier (6.451 baskets in 1386–1390), though still 9.3% lower than what they had enjoyed in 1373–1375 (7.620 baskets). Not until 1423, after a renewed, debasement-induced inflation and another prolonged strike, would the Ghent fullers obtain any further raise in their money wages, to 40d groot per cloth. Even so, their real wage in 1426–1430 (mean of 6.589 baskets) was still below that earned in the late 1390s, and 13.5% below that of 1373–1375 (Table 11; Munro, 2002). Thus, during the later fourteenth and early fifteenth centuries, the Ghent fullers experienced not a Golden Age but an almost continuous decline in their real wages, apart from some stabilization during the very deflationary 1390s. Somewhat more successful were the fullers in the small drapery town of Wervik, now a major leader of the younger, more aggressive, so-called nouvelles draperies, which were faring rather better than the traditional draperies of the drie steden – those of Ghent, Ypres, Bruges – especially in Mediterranean markets (Melis, 1962; Munro, 1999). Soon after the 1390 renforcement, the seigneur and bailiff of Wervik’s feudal seigneury decreed an immediate reduction of 27% in the fullers’ pay: from 48d to 35d groot per cloth, which still left them better off than the Ghent fullers. Nevertheless, in May 1392, the Wervik fullers’ guild, after threatening to go on strike, appealed this decree to the ducal Council of Flanders. The ducal council, however, ruled in favor of the Wervik drapers and their bailiff, forcing the fullers to accept this new wage rate, which was subsequently ratified by the new Wervik drapery keuren in October 1397.111 Of the three recorded disputes, the Kortrijk fullers, employed by a nouvelle draperie that rivalled Wervik’s in importance, fared the best, perhaps because they had responded much earlier and much more vigorously in their protests. Like the Ghent fullers, they also had not received any increase in money wages since 1374, despite the subsequently severe inflation (Tables 8, 12).112 When the Kortrijk weaver-drapers attempted to impose a cut in their wages, though a more modest one of 22%, from 41d to 32d groot per broadcloth (the same rate imposed at Ghent), the fullers guild appealed to Duke Philip’s councillors for arbitration. In imposing a new wage of 36d groot, and thus a relatively minor 12% reduction in the nominal wage, Duke Philip’s councillors split the difference between the weaver-drapers and the fullers.113 The real winners of this Kortrijk labor contract were surely the fullers, who, despite that reduction, soon found themselves much better off because of the 221
4259 Ch06 4/12/02 10:12 am Page 222
222 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
JOHN H. MUNRO
stark deflation that quickly ensued, from the early 1390s (Tables 8, 12). As indicated earlier, the fall in the mean Flemish price index over the 1390s was somewhat greater than the extent of the renforcement: i.e. 29.0% vs. 24.0%.114 Thus, by 1396–1400 the mean purchasing power of wages for the Kortrijk journeymen fullers, expressed in terms of commodity baskets, was 26.2% higher than it had been during the debasement years, in the quinquennium 1386–1390. In 1399, their real-wage was 58.4% above what they had been earning in 1390 when their contract had been imposed (i.e. 8.479 commodity baskets a year vs. 5.352 baskets); and their mean real wage in the 1390s was the highest that they had achieved (at least since the Black Death). But they, too, experienced a deterioration in the real wages after renewed coinage debasements and the consequent inflation from 1416. Thus, in 1426–30, despite having gained a money wage increase in 1429 (from 36d to 40d groot per cloth), these Kortrijk fullers were receiving a real wage (mean of 6.432 baskets) that was also lower than that earned during the early 1390s, by 10.6% (Table 12).
16 17 18 19
XII. DEFLATION, INFLATION, AND REAL WAGES: OTHER URBAN CRAFTSMEN IN FLANDERS AND BRABANT, 1370–1500
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
In Flanders, the building craftsmen of Ghent and Bruges, and the policemen in the latter town, did fare considerably better during the very end of the fourteenth and early fifteenth centuries; though clearly, even for them, the fifteenth century as a whole was also far from being a consistent Golden Age. Only the Bruges wage data are shown, in Table 10 and Fig. 5, because the craftsmen’s wages in the two towns were virtually identical throughout the fifteenth century (or to the late 1480s, when the accounts cease recording daily wages for individual craftsmen); and they of course demonstrate the same relative degree of wage stickiness. Initially, with the abrupt 25% cut in nominal wages with the 1390 monetary reform, from 12d to 9d groot, the Bruges building craftsmen did suffer a very brief if substantial drop in real income, of 30.2%, before the aforementioned deflation became more pronounced. During the 1390s, their real wage rose by 76.0%; and in the latter part of the decade it did so because, in 1396–1397, these craftsmen received a partial restoration of their former nominal wage: masters, to 10d groot; and journeymen, from 4.5d to 5.0d groot. Their real wage index peaked in 1399, at 110.37 (= 20.187 commodity baskets); but that was only 6.5% more than the amount earned in 1388 (18.953 baskets). However, the quinquennial wage for 1396–1400, measured in such commodity baskets (18.241 = RWI 99.73) was substantially higher, by 28.89%, than the
4259 Ch06 4/12/02 10:12 am Page 223
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
223
mean number of 14.152 baskets (RWI = 77.38) earned a decade earlier in 1386–1390.115 During the early 1390s, the Bruges policemen fared somewhat better than did the building craftsmen and fullers because their wages had not been reduced (Table 11, Fig. 5) – after all, they were expected to maintain social order with the arbitrary pay cuts. Furthermore, though paid only 6d per day, they were paid for the full year (i.e. 365 days), rather than for the estimated mean of 210 days for a master building craftsmen. Thus, their annual real wage in 1390 was substantially higher, with an estimated purchasing power of 13.288 baskets, compared to 11.468 baskets for a master carpenter or mason. Finally, in 1398, they did suffer a delayed reduction in their money wages, though only 16.7% (to 5d); but unlike the craftsmen their money-wages were never again restored (for the duration of the wage records, to 1476). In 1399, their annual real wage income, having peaked in 1393, was also now just 87% of that earned by master masons and carpenters: 17.037 vs. 20.187 baskets; but that was still well more than double the amount earned that year by journeymen fullers (7.537 baskets in Ghent; and 8.479 baskets in Kortrijk). The real wages for Bruges policemen continued to remain fairly high, until just before the Battle of Agincourt (14.613 baskets in 1414, but 18.161 baskets in 1411). Thereafter, with Burgundian involvement in the resumption of more serious warfare in northern France, and with renewed debasements in and after 1416 (Munro, 1973, pp. 65–126), their real wages and those for Bruges’ master building craftsmen once more began to fall, and then fell sharply. The reason is quite simple to comprehend: prices rose with the coinage debasements (1416–1433), and then, during the 1430s, with costly warfare as well, from a mean index of 95.31 in 1411–1415 to one of 140.166 in 1436–1440, while their money wages remained generally fixed, at 10d for building craftsmen and 5d for policemen. From 1432, however, some craftsmen began receiving higher wages of 11d and 12d; but that was hardly enough to offset the ravages of inflation, so that, over this entire period, from 1411–1415 to 1436–1440, the real wage index of Bruges craftsmen fell from a harmonic mean of 95.38 (17.446 commodity baskets) to one of 69.45 (12.703 baskets). Indeed, in the worst year, 1439, with plague, famine, and war, their real wage index (47.57 = 8.702 baskets) was only 75.87 per as much as in the previous annus horribilus, 1390 (RWI = 62.70 = 11.468 baskets). According to many (if not all) economic historians, the ensuing three decades were an era of very severe commercial, financial, and agrarian depression for both England and the southern Low Countries. To be sure, the latter still remained by far the wealthiest and most economically advanced region in northern Europe – a veritable “promised land,” in the eyes of many contemporaries; 223
4259 Ch06 4/12/02 10:12 am Page 224
224 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
but at this very time, the southern Low Countries were also experiencing, and largely independently of the current depression, the rapid decline of their traditional luxury woolens industry, in the face of relentless English competition in European cloth markets.116 Nevertheless, in this same period, the real wage index for Bruges building craftsmen soared once again, by a remarkable 169.75%: from 45.57 in 1439 (8.702 baskets) to reach a high of 128.33 (23.472 commodity baskets) in 1464 – just before the next round of debasements. Again real wages had risen so strongly, principally, because this was a period of stark deflation, and, as noted, of very low mint outputs, sometimes with only a bare trickle of coinage – one factor, though not the only one, that precipitated the renewal of debasements in 1466 (Tables 3–4, 8–10; Figs 3, 5). But this time, the Bruges craftsmen had also gained from an increase in their mean nominal wage: by about 1440, most craftsmen were receiving a summer wage of 12d – and some throughout the year – but a reduced winter wage of 10d, so that the median wage had risen from 10.0d to 10.8d and then to 11d groot. The subsequent era of war-inspired and debasement-induced inflations again brought real wages tumbling down, as Table 10 indicates; but unfortunately the Bruges wage payment series terminate in the mid-1480s.117 In the mid-sized town of Aalst, to the east of Ghent, however, the daily wage rates for master masons and carpenters remained fixed at 6d groot per day until well into the early sixteenth century, though some senior craftsmen were earning 7d to 8d daily; but after 1510–1511, more and more such craftsmen were earning these higher rates of 7d and 8d (up to the 1530s).118 In the neighbouring duchy of Brabant, to the east, the price and wage history of the towns was, not surprisingly, similar to that of Flanders, following the Burgundian monetary unification and reform of 1434–1435. But before the Burgundian acquisition of Brabant in 1430, and that monetary reform, their monetary, price, and thus wage history had been quite different, as Tables 9, 13–14 and Figs 5–6 show (employing the same base, 1451–1475 = 100, for prices and wages). With far more drastic coinage debasements in early fifteenth-century Brabant (under Dukes Antoine, Jan IV, and Philip van Sint-Pol, 1406–1430), the real wages of Antwerp’s master masons and carpenters had plunged by about one-third, in the three decades from 1410 (Table 13): from a mean of 100.30 in 1401–1405 (10.262 commodity baskets per work year of 210 days, seasonally adjusted for the winter wage) to a mean of just 66.437 in 1436–1440 (6.840 commodity baskets per year; with a nadir of 5.281 such baskets in 1437).119 Thereafter, as in Flanders, their real wage index soared strongly, well more than doubling, to peak at a harmonic mean of 109.81 in 1461–1465 (11.156 commodity baskets for both Antwerp masons and Mechelen carpenters; but 11.404 baskets for Mechelen masons, with a higher winter wage: Tables 13–14).
4259 Ch06 4/12/02 10:12 am Page 225
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
225
The mid-century deflation reached its very nadir in this same quinquennium, which, as noted earlier, also marks the peak period of real wages for master building craftsmen in Bruges (RWI = 112.733); but the latter continued to enjoy a substantially higher level of real wages, which would have purchased a mean of 20.619 Flemish commodity baskets in that quinquennium (Table 10). Subsequently, as Tables 13–14 and Fig. 6 show, the real wages for master building craftsmen in Antwerp and Mechelen plunged severely, with the series of inflationary coinage debasements undertaken by Archduke Maximilian of Austria during his wars with France and the very destructive civil wars in the now Habsburg Netherlands (from 1477 to 1493: see Tables 3–4). Consequently the real-wage index for Antwerp’s master masons fell to a low of 61.883 in 1486–1490 (a harmonic mean of 6.316 commodity baskets), even though they were now earning a mean daily summer wage of 12.9d; for Mechelen’s masons, earning a lesser mean daily wage of 12.3d, the real wage index fell to 58.587 (for an equivalent annual income of 6.099 commodity baskets).120 The Peace of Senlis in May 1493 (with France), permitting a much need coinage renforcement, brought monetary stability and renewed deflation, though this time with somewhat greater economic prosperity, certainly for Brabant (Van der Wee, 1963, Vol II, pp. 113–142). Real wages thus recovered, and this time even more strongly in Mechelen, to 97.53 in 1496–1500 (= 10.276 commodity baskets), because in 1490 its town masons had received a more substantial rise in their daily summer wage: to 13.5d. The daily summer wage for Antwerp masons was increased to only 12.5d, and only in 1497, so that in 1496–1500 their RWI was only 89.522 (or the seasonally adjusted annual equivalent of 9.039 commodity baskets). Thereafter, the real wages of building craftsmen in Brabant drifted slowly downwards, and then, with the onset of the Price Revolution, from c. 1515, rather more rapidly. Thus, the RWI for Antwerp masons in 1521–1525 was just 71.548 (harmonic mean). Nevertheless, during the subsequent history of the Price Revolution, lasting until the 1640s, building craftsmen in the major towns of Brabant did not, in general, experience the same drastic fall that their brethren did in southern England, as noted earlier (see pp. 188, 219). That history, of course, lies entirely beyond the scope of this study. But if the wage experience of the Brabantine building craftsmen differed so markedly from the English during the Price Revolution, so it did as well for much of the fifteenth century, thus illuminating again the dangers of viewing the changing fortunes of late-medieval European craftsmen merely through an English lens. Only during the middle decades of the fifteenth century were the experiences of such craftsmen reasonably similar in both England and the Low Countries. But if prolonged deflation thus allowed so many craftsmen in both countries to experience a Golden Age, those who benefited from wage-stickiness, 225
4259 Ch06 4/12/02 10:12 am Page 226
226 1 2
JOHN H. MUNRO
one must not forget that it was also an era of severe economic depression (see above, pp. 223–224).
3 4 5 6 7
XIII. SOME FIFTEENTH-CENTURY ANOMALIES: WAGE FLEXIBILITY AND FALLING REAL WAGES IN (CHIEFLY) RURAL AREAS IN THE LOW COUNTRIES AND ENGLAND
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Yet even these decades were far from being a “Golden Age” for all craftsmen, artisans, and laborers. We may doubt that many artisans and workers in the now rapidly declining woolen cloth industries were enjoying much prosperity; but we lack adequate wage data for any textile artisans from the 1440s. We do, however, possess abundant wage data for craftsmen and laborers, in the smaller towns and rural areas of both countries; and they also do not indicate that all of them shared in this so-called Golden Age. The most illuminating are those to be found in some Flemish villages to the east of Ghent (Table 15). By the 1420s, the daily summer wage rates for many carpenters, masons, and thatchers, in Afsne, St. Denijs (Maalte), and Destelbergen, had risen to 10d groot Flemish, as in Bruges and Ghent; and those rates prevailed for the next three decades, to surpass indeed the wages earned by similar craftsmen in Antwerp and Mechelen (from 1434: 10d groot Flemish = 15d groot Brabant). But during the deflationary 1450s and 1460s, the wage-rates for many of these village craftsmen fell. At Afsne, St. Denijs (Maalte), and Destelbergen, those for carpenters and thatchers fell by 30%: to 8d, then 7.5d, and finally to 7.0d, by 1470. The decline in nominal wages for thatchers at Zaffelare was about 25%, from the late 1430s to the late 1450s (rising thereafter); and at Zevergem, only about 20% (Table 15), and only from the later 1450s. Despite the strong deflation of this era, their real-wages also finally did decline, though of course to a much lesser extent. Measured by the number of commodity baskets purchased with the annual money wage, carpenters and thatchers of Afsne and St. Denijs suffered the sharpest fall: from 15.12 baskets in 1446–1450 to 12.06 baskets in 1466–1470. For all craftsmen in all six villages, this real-wage measure fell from a mean 12.492 baskets in 1431–1435 to 10.857 baskets in 1436–1440, then rose to 13.893 baskets in 1451–1455, and again fell sharply to 10.025 baskets in 1456–1460, recovering to just 12.222 baskets in 1466–1470. Such evidence may support Van der Wee’s contention (1963, Vol. II, pp. 61–73) that the agricultural sector suffered the most severely from this mid-fifteenth century “depression.” But even in the urban sector, some building craftsmen enjoyed rather less wage-stickiness and suffered some reduction in at least their nominal wages,
4259 Ch06 4/12/02 10:12 am Page 227
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
227
as can be seen in Table 14, for Mechelen. Thus, the master carpenters employed by Mechelen’s Onse Lieve Vrouw Hospital, after having received a raise in their daily summer wages, from 9d to 12d (= 8d groot Flemish) in 1441, had those wages cut back to 9d in 1448; and the former rate of 12d rate was restored only in 1458 (and then prevailed until 1498). Day labourers working for this same hospital suffered by far the worst wage cuts in this deflationary era: from a peak of 8d groot Brabant in 1444 to a low of 5d in 1458, a decline of 38%. Measured in commodity baskets, their mean real wage declined from a peak of 6.348 units per year in 1446–1450 to a mean of 4.724 units in 1466–1470, a fall of 25.6%. Nor were such nominal wage cuts in the midst of the fifteenth-century deflationary Golden Age limited to the Low Countries. For England during this era, Mate (1986) has provided some similar evidence for rural workers in Kent and Sussex: with nominal wage-reductions up to 30% for mowers at Barton, Kent, in the 1440s (enduring to the 1470s); cuts of 20% for ploughmen at Wye, in Sussex, in the 1440s; and of 20% for carpenters at Lullington, Sussex, in the 1450s. Furthermore, even earlier, during the deflation of the later fourteenthcentury, the various decennial means of English wages published by Thorold Rogers (1866, 1882, 1903), Steffen (1901–1905), and Beveridge (1936, 1955) indicate that many rural workers then also evidently suffered some reductions in their daily wages from 1380 to 1400: as much as 29.4% for wheat threshers in S.E. England and 26.9% for Westminster farm laborers. But doubts about the statistical methods used to compile these means, especially when direct comparisons are made with the actual manorial accounts, prevent one from asserting that the nominal wages for each class and grade of worker fell, or fell as much as portrayed in these decennial indexes.121
27 28 29
XIV. SOME CONCLUSIONS ON LATER-MEDIEVAL WAGE STICKINESS AND REAL WAGES
30 31 32 33 34 35 36 37 38 39 40
From this study of the behavior of both nominal and real wages in late-medieval England and the Low Countries, one may draw a number of conclusions that significantly alter some standard perceptions about living standards for English and indeed west European craftsmen and laborers during this era. First, the Black Death did not immediately usher in a “Golden Age of the Laborer.” Even in England, which escaped the ravages of war that plagued so much of latemedieval Europe, and which enjoyed greater economic, social, and monetary stability than most regions, there is no evidence for a pronounced and sustained rise in real wages for craftsmen and laborers until about a quarter-century after the Black Death; indeed many wage earners had suffered some fall in real 227
4259 Ch06 4/12/02 10:12 am Page 228
228 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
wages in the decade following the Black Death. In the Low Countries, a similar pronounced and sustained rise in real wages did not commence until the 1390s, after an even more drastic fall in real wages after the Black Death (Table 10). Second, the prime if not the complete explanation for the sustained rise in real wages, in both England and the Low Countries, that did commence in the later fourteenth century was a combination of institutional wage-stickiness with prolonged, deep deflation. Third, institutional wage-stickiness itself became a prevalent feature of labor markets in both England and the Low Countries only during and from that final quarter of the fourteenth century. Thus, during the steep deflation that struck England in the quarter-century before the Black Death, many if not all money wages did in fact fall, though most fell less than did commodity prices, so that many building craftsmen then did enjoy some increase in real wages. After the Black Death, furthermore, in both England and the Low Countries, nominal wages did of course rise; and they were thus not particularly sticky during the severe inflation that ensued for about a quarter century (for four decades in Flanders), even if money wages did not fully keep pace with prices for much of that era. In the fifteenth century, however, wages for many building craftsmen, in both England and the Low Countries, were quite sticky even during the two major periods of inflation, which was generally more severe in the latter region. But, in a much longer historical perspective, from the late fourteenth to late nineteenth centuries, money wages were generally much more sticky during deflationary than inflationary periods, because of the so-called “ratchet effect.” Fourth, even if nominal wage stickiness became all the more pronounced and widespread during the fifteenth century, in both England and the Low Countries, it never became universal, and was to be found much more frequently in large towns than in villages and the countryside. Not surprisingly, urban craftsmen in the building trades, and especially those employed by town governments or large institutions, were far more likely to enjoy conditions of downward wage stickiness than were the presumably less well organized and less well protected rural and especially agricultural workers. Even so, some urban exceptions are to be noted: in particular, as shown in this study, employees of Mechelen’s Onse Lieve Vrouw hospital. Why wage stickiness became so much more pronounced from the later fourteenth century but never became universal are subjects for further research. Finally, if the central conclusion of this study remains valid, namely that the most pronounced rise in real wages occurred during strongly deflationary eras and chiefly by those enjoying such nominal wage-stickiness, we cannot totally rule out the role of productivity changes in elevating those real wages, even if they are difficult to discern.
4259 Ch06 4/12/02 10:12 am Page 229
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
229
The more fitting conclusion would be to return to the Keynesian view of such wage rigidity during periods of inflation, at least according to some standard macroeconomics textbooks, namely a situation that “results in a higher real wage and a reduction in the demand for labor,” so that “employment declines and so does output”; and so very often, “a decline in demand leads to a prolonged recession” (Wilton-Prescott, 1987, pp. 216–218). The available evidence does not permit us to ascertain whether such conditions, especially in the depression era of the mid-fifteenth century, led to increased unemployment – apart from evident unemployment in the declining Flemish and Brabantine cloth industries (Munro, 1983b, 1994, 1995, 1999a). Some evidence, discussed above, does suggest that employment for most craftsmen, was normally discontinuous or “discrete” in this late-medieval era; and such conditions may well have reduced the number of days in which they secured paid employment. On the other hand, as suggested either, such unemployment may have been more voluntary than involuntary, if they were satisfied to accept more days of leisure.122 One should also note that, under such deflationary conditions, the typical employer faced and still faces not only a potential rise in his real labor costs, but even more certainly a rise in the cost of borrowing capital. For the effect of such deflation is to increase real interest rates (i.e. the nominal rate minus the inflation rate, or plus the deflation rate), all the more so, since the historical evidence rarely if ever indicates a commensurate fall in nominal interest rates. Rash, however, would be the economic historian who would contend that every deflationary period that also experienced downward wage stickiness for much of the labor force was also one of recession or depression. Two such periods in the modern era with these characteristics were the so-called “Great Depressions” of 1873–1896 and 1929–1939. But if economic historians now doubt that such a label can justifiably be fixed to the first of these, despite some cyclical downswings, certainly the latter was a true Great Depression. Interpreting real-wage trends is also fraught with other dangers. Thus, even if the amount of labor (L) was diminishing, relative to the employed stock of land and capital (whose sum is K), as evidently did occur in late-medieval Europe, and even if labor then enjoyed a rise in real wages, the European labor force may not have gained an increased share of the national income. That would have occurred only if the elasticity of substitution between L and K, as the ratio of the marginal products of land + capital and of labor (yK/yL) had been less than unity:123 that is, if: yK/yL = SKL < 1. Finally, a word of caution and a plea for some historical perspective. As important and as vexing as is the question of real-wage trends in the latemedieval economy, we must remember that real-wages, and most especially 229
4259 Ch06 4/12/02 10:12 am Page 230
230 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
evidence on wage-rates, do not allow us to deduce levels of real income for even the lower orders of society. Penn and Dyer (1990, p. 356) have commented that “at least one-third of the population of late medieval England gained all or a part of their livelihood from earning wages.”124 This statement might be more persuasive if the word order were changed to read: “gained part or all,” or better “part if not all” of their livelihood. Surely, those who lived strictly by money-wages alone, and who were unable to supplement their annual incomes from agricultural holdings, or from private gardens with urban dwellings, and especially by practicing more than one craft, were still a minority of late-medieval English society, if undoubtedly a rather larger proportion of the economically more advanced Flemish society.
10 9
11
12
13
14
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
35
36
37
38
39
40
34 bu bu no. no. lb. lb. bu lb. pt yd yd yd
4.500 7.917
1.500
231
40.000 10.000 10.000
4.250 2.750 0.500 0.667 0.500 0.333
154.567 1.247 0.284 0.610 0.457 0.304
40.000 4.536 4.536
1.500
163.659 287.917
45.461 36.369 18.184 24.243 124.257
100.00%
20.00%
37.50%
22.50% 42.50%
20.00%
Percent
l. kg m. m.
1.800 1.125
kg no. kg kg
l. l.
l.
l.
Unit
162.000 1.333
23.500 40.000 4.800 4.700
162.000 288.000
126.000
126.000
Amount
Brabant
232.524
24.844 60.020
17.000
10.568 7.608
54.704 9.988 19.728 5.968 90.388
39.712 82.116
42.404
42.404
Value in d gr. Brabant
10.68% 25.81%
7.31%
4.54% 3.27%
23.53% 4.30% 8.48% 2.57% 38.87%
17.08% 35.32%
18.24%
18.24%
Percent
155.016 100.00%
16.563 40.013
11.333
7.045 5.072
36.469 6.659 13.152 3.979 60.259
26.475 54.744
28.269
28.269
Value in d gr. Flemish
1.225
13.610 13.610 27.220
163.659 287.917
36.369 18.184 24.243 124.257
m.
m.
l. kg
kg no. kg kg kg
l. l.
l. l. l. l.
1.
45.461
126.295
25.276 25.276
36.087 8.578 44.665
25.805 56.354
7.062 2.867 7.341 30.549
13.279
Value in d gr. Flemish
Flanders Unit
Amount
100.00%
20.01% 20.01%
28.57% 6.79% 35.37%
20.43% 44.62%
5.59% 2.27% 5.81% 24.19%
10.51%
Percent
Wage Stickiness and Monetary Changes
Notes: bu = bushels; lb = pound; pt = pint; yd = yard; l. = liter; kg = kilogram; m. = meter. Sources: Phelps, Brown and Hopkins (1955, 1981), Van der Wee (1963, 1975, 1978); Munro (1984a), Verlinden and Schoilliers (1959–1965, Vol. I, pp. 44–45, 52–53; Vol. II:i, pp. 3–70); Stadsarchief Gent, Stadsrekeningen, 1350–1351 to 1499–1500.
bu bu bu bu bu
1.250 1.000 0.500 0.667 3.417
Metric Measure
England
3
TOTAL
8
Unit
2
Wheat Rye Barley Peas Sub-total Drink: Barley (or malt) Total Farinaceous Meat, Fish, Dairy: Sheep Beef Herrings Butter Cheese Sub-total Industrial: Charcoal Candles Lamp Oil Canvas/Linen Shirting Coarse Woolens Sub-total
7
Amount
1
Farinaceous:
6
Commodity
5
Basket of Consumables Commodity Price Indexes for England, Brabant, and Flanders: mean of 1415–1475 = 100.
4
Table 1.
15
TABLES AND FIGURES
4259 Ch06 4/12/02 10:12 am Page 231
231
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40 6.0 6.0 6.0
5.0 4.0 6.0
Summer Wages in coin only 2.0 1.5
Summer Wages with food
5.0 5.0 5.0
3.0 4.0 5.0
Winter Wages in coin only
London
1.0 1.5
Winter Wages with food
5.5 6.0 7.0
3.0 4.0 4.0
Summer Wages in coin only
4.0 4.0 5.0
Summer Wages with food
4.5 5.0 7.0
[not stated]
[less]b
Winter Wages in coin only
England
3.0 4.0 5.0
Winter Wages with food
b
1290: from Michaelmas (29 September) to Martinmas (12 November) and from Candlemas (Purification: 2 February) to Easter. from Michaelmas: “less according to the rate and discretion of the justices.” c Rates of 4d per day for master free-masons; and 3d per day for other master masons and carpenters. d 4d for the chief master masons and carpeners; but 3d or 2d per day for others, “according as they be worth.” e The higher rate of 7d per day for those master masons and carpenters having charge of six or more men; 5d daily with food and drink, summer and winter (for both rates). Sources: Tomlins and Raithby (1810–1822, Vol. I, pp. 311–312; Vol. II, pp. 337–339, 585–587); Riley (1860, Vol. I, pp, 99–100; Vol. II, pp. 541–543; 1868, pp. 253–255); Sharpe (1905, pp. 148, 301; 1907, p. 184).
a
4
1290 1290a 1350–1351 1350–1351c 1360d 1362 1372 1382 1444–1445 1495 1495e
3
Year
2
Wage Regulation in Medieval England. Official Maximum Wages for London and for the Kingdom of England (1290 to 1515) in pence (d) sterling.
1
Table 2.
4259 Ch06 4/12/02 10:12 am Page 232
232 JOHN H. MUNRO
1
2
3
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
1281–1285 1286–1290 1291–1295 1296–1300 1301–1305 1306–1310 1311–1315 1316–1320 1321–1325 1326–1330 1331–1335 1336–1340 1341–1345 1346–1350 1351–1355 1356–1360 1361–1365 1366–1370 1371–1375 1376–1380 1381–1385 1386–1390 1391–1395 1396–1400
Years
240.01 675.84 1,939.78 1,726.70 2,415.24 1,729.03 802.61 235.33 161.84 504.81 626.55 391.14
England Gold in kg.
9,859.48 27,123.30 83,567.73 74,406.84 104,077.76 74,507.35 34,586.02 10,140.85 6,973.80 21,753.33 26,999.15 16,855.14
England Gold in £ ster. 21,913.31 17,280.60 1,552.35 12,071.42 16,017.47 40,226.55 10,706.71 7,275.68 1,780.11 121.86 209.06 429.49 5,077.46 1,991.05 17,442.91 4,423.02 1,630.81 293.82 316.97 356.90 317.41 247.51 193.49 175.60
England Silver in kg. 68,548.73 54,056.78 4,856.03 37,761.55 50,105.48 125,835.83 33,492.50 22,759.61 5,568.49 381.19 665.13 1,551.60 17,710.47 7,090.87 67,245.28 17,081.46 6,298.11 1,134.73 1,224.11 1,378.32 1,225.83 955.89 747.25 678.14
England Silver in £ ster. 68,548.73 54,056.78 4,856.03 37,761.55 50,105.48 125,835.83 33,492.50 22,759.61 5,568.49 381.19 665.13 1,551.60 27,569.96 34,214.17 150,813.01 91,488.31 110,375.86 75,642.08 35,810.13 11,519.17 8,199.63 22,709.22 27,746.40 17,533.29
England Total in £ ster. 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 35.76% 79.28% 55.41% 81.33% 94.29% 98.50% 96.58% 88.03% 85.05% 95.79% 97.31% 96.13%
England Percent Gold 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% 100.00% 64.24% 20.72% 44.59% 18.67% 5.71% 1.50% 3.42% 11.97% 14.95% 4.21% 2.69% 3.87%
England Percent Silver
Table 3. Aggregate Mint Outputs of England and the Low Countries, 1290–1520. Gold and Silver Coinage Outputs in kilograms of fine metal and in current pounds sterling (England) and groot (Flanders/Low Countries).*
4259 Ch06 4/12/02 10:12 am Page 233
Wage Stickiness and Monetary Changes 233
233
15
16
19
20
21
22
23
24
25
26
27
29
30
31
38
39
40 256.22 40.91 4,483.34 3,882.48 14,765.09 31,785.11 37,350.66 4,527.86 605.71 2,397.68 6,769.09 6,558.02 18,067.35 29,938.35 14,034.25 4,835.25 5,765.30 5,368.79 7,361.88 14,429.82
7,524.61 3,014.48 94,003.24 53,445.55 137,209.46 60,488.18 47,921.84 10,861.16 4,952.17 5,478.10 9,810.71 7,837.31 47,798.68 113,202.34 48,852.80 30,979.88 19,950.05 13,755.52 24,748.40 32,459.06
14
66.34 10.59 967.48 837.76 3,186.02 6,858.61 8,059.55 977.03 130.70 517.37 1,460.64 1,415.09 3,432.92 5,168.09 2,422.65 834.68 995.23 926.79 1,270.84 2,490.94
13
7,268.39 2,973.57 89,519.90 49,563.08 122,444.37 28,703.07 10,571.18 6,333.30 4,346.47 3,080.42 3,041.63 1,279.29 29,731.33 83,263.99 34,818.55 26,144.62 14,184.75 8,386.73 17,386.53 18,029.24
12
168.67 69.01 1,870.67 1,035.15 2,557.31 599.48 220.79 132.27 90.78 64.34 63.53 26.72 488.12 1,288.16 538.67 404.48 219.45 129.75 268.98 278.93
9 96.59% 98.64% 95.23% 92.74% 89.24% 47.45% 22.06% 58.31% 87.77% 56.23% 31.00% 16.32% 62.20% 73.55% 71.27% 84.39% 71.10% 60.97% 70.25% 55.54%
England Percent Gold 3.41% 1.36% 4.77% 7.26% 10.76% 52.55% 77.94% 41.69% 12.23% 43.77% 69.00% 83.68% 37.80% 26.45% 28.73% 15.61% 28.90% 39.03% 29.75% 44.46%
England Percent Silver
234
1401–1405 1406–1410 1411–1415 1416–1420 1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470 1471–1475 1476–1480 1481–1485 1486–1490 1491–1495 1496–1500
11
England Total in £ ster.
8
England Silver in £ ster.
7
England Silver in kg.
4
England Gold in £ ster.
3
England Gold in kg.
2
Years
18
Continued.
17
Table 3.
4259 Ch06 4/12/02 10:12 am Page 234
JOHN H. MUNRO
1
5
6
10
28
32
33
34
35
36
37
15
16
19
20
21
22
23
24
25
26
27
29
30
31
38
39
40 3,641.11 176.76 5,553.49 5,178.95 8,820.73 3,992.17 10,030.19 2,215.76 915.62 2,816.88 1,787.71 3,676.06 5,791.31 691.66 1,113.70 2,484.27 3,124.47 12,143.55 7,999.91 6,609.82 5,015.22 102.68 5.91
4,873.43 310.86 11,138.59 11,397.25 21,251.52 11,141.97 32,269.76 8,315.15 3,648.74 11,467.50 7,792.29 14,958.40 23,507.52 2,826.54 3,887.99 8,665.85 15,052.70 58,804.34 43,326.04 34,252.10 25,788.39 527.55 40.79
13
3,975.69 26.60 6,596.36 24,811.55 80,870.03 77,350.49 50,200.53 32,921.28 10,555.07 22,941.63 20,865.91 14,458.24 12,731.42 1,236.90 636.25 196.76 181.63 2,195.70 69,470.41 115,353.24 28,534.39 6,466.29 148.08
12
32
33
34
35
36
37 266.77 1.32 315.97 1,096.66 3,191.83 2,629.89 1,586.50 825.21 261.20 529.81 423.11 368.61 324.59 31.54 19.03 5.88 4.31 41.06 1,105.07 1,774.87 511.94 111.93 2.55
8,849.12 337.46 17,734.96 36,208.81 102,121.54 88,492.46 82,470.29 41,236.42 14,203.81 34,409.13 28,658.20 29,416.64 36,238.94 4,063.44 4,524.24 8,862.61 15,234.33 61,000.04 112,796.44 149,605.34 54,322.78 6,993.84 188.87
9 44.93% 7.88% 37.19% 68.52% 79.19% 87.41% 60.87% 79.84% 74.31% 66.67% 72.81% 49.15% 35.13% 30.44% 14.06% 2.22% 1.19% 3.60% 61.59% 77.11% 52.53% 92.46% 78.41%
Flanders/ Low Countries Percent Gold
8
1336–1340 1341–1345 1346–1350 1351–1355 1356–1360 1361–1365 1366–1370 1371–1375 1376–1380 1381–1385 1386–1390 1391–1395 1396–1400 1401–1405 1406–1410 1411–1415 1416–1420 1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450
11
Flanders/ Low Countries Total in £ groot
7
Flanders/ Low Countries Silver in £ groot
6
Flanders/ Low Countries Silver in kg. 55.07% 92.12% 62.81% 31.48% 20.81% 12.59% 39.13% 20.16% 25.69% 33.33% 27.19% 50.85% 64.87% 69.56% 85.94% 97.78% 98.81% 96.40% 38.41% 22.89% 47.47% 7.54% 21.59%
Flanders/ Low Countries Percent Silver
4
Flanders/ Low Countries Gold in £ groot
3
Flanders/ Low Countries Gold in kg.
2
Years
18
Continued.
17
Table 3.
4259 Ch06 4/12/02 10:12 am Page 235
Wage Stickiness and Monetary Changes 235
1
5
10
14
28
235
15
16
20
21
22
23
24
25
26
27
29
30
31
38
39
40 51,582.01 15,922.23 404.22 44,268.42 64,119.24 96,844.75 61,553.58 102,460.86 20,857.44 88,067.29
98.29% 97.44% 100.00% 37.05% 29.52% 30.16% 8.47% 23.56% 6.41% 50.49%
1.71% 2.56% 0.00% 62.95% 70.48% 69.84% 91.53% 76.44% 93.59% 49.51%
Flanders/ Low Countries Percent Silver
236
1
5
6
10
28
32
33
34
35
36
37
* Flanders only until 1420; from 1421–1425 until 1496–1500: Flanders, Brabant, Holland and Zealand, within the Burgundian Low Countries. Sources: England: Brooke, G.C., and E. Stokes (1929); Crump, C. G., and C. Johnson (1913); Challis 1992; Munro (1973, 1981). Flanders and the Low Countries: Munro (1973, 1981, 1983a, 1984a, 1984b, 1988a, 1991b.)
880.32 408.31 0.00 27,867.69 45,191.72 67,636.25 56,337.18 78,323.90 19,521.10 43,603.01
14
164.61 64.07 0.00 4,628.96 7,313.98 9,341.50 6,534.30 6,803.60 2,780.07 5,109.49
13
50,701.69 15,513.92 404.22 16,400.73 18,927.51 29,208.50 5,216.39 24,136.96 1,336.34 44,464.28
12
827.29 253.14 6.60 253.59 261.20 380.05 58.54 144.64 20.32 474.63
19 Flanders/ Low Countries Percent Gold
9
1451–1455 1456–1460 1461–1465 1466–1470 1471–1475 1476–1480 1481–1485 1486–1490 1491–1495 1496–1500
11
Flanders/ Low Countries Total in £ groot
8
Flanders/ Low Countries Silver in £ groot
7
Flanders/ Low Countries Silver in kg.
4
Flanders/ Low Countries Gold in £ groot
3
Flanders/ Low Countries Gold in kg.
2
Years
18
Continued.
17
Table 3.
4259 Ch06 4/12/02 10:12 am Page 236
JOHN H. MUNRO
1
2
3
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
1281–1285 1286–1290 1291–1295 1296–1300 1301–1305 1306–1310 1311–1315 1316–1320 1321–1325 1326–1330 1331–1335 1336–1340 1341–1345 1346–1350 1351–1355 1356–1360 1361–1365
Years
84,628.07 66,736.77 5,995.10 46,619.19 61,858.62 155,352.87 41,348.77 28,098.28 6,874.68 470.60 807.36 1,658.66 29,951.44 36,812.55 150,952.61 91,488.31 110,375.86
England: Coinage Outputs in constant £ sterling of 1351–1411 values* 68,548.73 54,056.78 4,856.03 37,761.54 50,105.48 125,835.83 33,492.50 22,759.61 5,568.49 381.19 665.13 1,551.60 27,569.96 34,214.17 150,813.01 91,488.31 110,375.86
England: Coinage Outputs in current £ sterling
104.80 80.52 107.45 102.34 92.35 109.81 115.33 161.91 137.97 111.07 114.12 94.32 90.06 102.70 132.18 129.46 146.64
England: Phelps Brown and Hopkins CPI
25,557.41 739.59 35,062.86 67,258.22 171,607.60 128,744.99
Low Countries: Coinage Outputs in constant £ sterling of 1351–1411 values*
8,849.12 337.46 17,734.96 36,208.81 102,121.54 88,492.46
Low Countries: Coinage Outputs current in £ groot
60.65 87.54 94.43
Low Countries: CPI CPI Flanders Brabant
Table 4. England and the Low Countries: Mint Outputs and Consumer Price Indexes. Mint Outputs in Constant Pounds Sterling and in Current Pounds Sterling and Pounds Groot Flemish 1281–1285 to 1496–1500: in quinquennial means. Price Indexes: 1451–1475 = 100.
4259 Ch06 4/12/02 10:12 am Page 237
Wage Stickiness and Monetary Changes 237
237
11
12
13
15
16
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
38
39
40 107,101.72 44,117.12 14,791.69 33,709.23 25,136.55 30,081.11 36,352.95 4,030.06 5,120.88 9,847.65 12,252.19 48,666.93 78,515.09 102,009.51 41,428.85 5,219.91 132.73 36,285.47 11,155.69
9
146.10 135.26 110.62 112.90 102.53 106.33 110.84 114.84 111.23 108.11 113.40 101.48 112.27 108.48 122.01 92.53 100.90 100.25 97.06
82,470.29 41,236.42 14,203.81 34,409.13 28,658.20 29,416.64 36,238.94 4,063.44 4,524.24 8,862.61 15,234.33 61,000.04 112,796.44 149,605.34 54,322.78 6,993.84 188.87 51,582.01 15,922.23
8
75,642.08 35,810.13 11,519.17 8,199.63 22,709.22 27,746.40 17,533.28 7,524.61 3,014.48 94,003.24 53,445.55 137,209.46 60,488.18 47,921.84 10,861.16 4,952.17 5,478.10 9,810.71 7,837.31
7
75,642.08 35,810.13 11,519.17 8,199.63 22,709.22 27,746.40 17,533.28 7,524.61 3,014.48 84,347.37 47,842.16 122,504.18 52,320.35 40,639.61 9,473.19 4,416.58 4,770.45 8,378.37 6,616.38
6
107.40 115.22 111.66 119.19 124.72 88.51 89.80 88.53 105.26 95.31 107.38 112.18 117.77 123.51 140.17 113.50 109.98 100.90 117.86
64.27 68.55 73.97 80.54 90.19 100.15 102.76 125.43 105.48 99.58 98.54 114.58
Low Countries: CPI CPI Flanders Brabant
238
1366–1370 1371–1375 1376–1380 1381–1385 1386–1390 1391–1395 1396–1400 1401–1405 1406–1410 1411–1415 1416–1420 1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460
10
Low Countries: Coinage Outputs current in £ groot
5
Low Countries: Coinage Outputs in constant £ sterling of 1351–1411 values*
4
England: Phelps Brown and Hopkins CPI
3
England: Coinage Outputs in current £ sterling
2
England: Coinage Outputs in constant £ sterling of 1351–1411 values*
1
Years
18
Continued.
17
Table 4.
4259 Ch06 4/12/02 10:12 am Page 238
JOHN H. MUNRO
14
36
37
12
13
15
16
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
38
39
40
14
36
37
239
404.22 44,268.42 64,119.24 96,844.75 61,553.58 102,460.86 20,857.44 88,067.29
88.71 96.52 96.02 117.21 156.85 184.51 144.98 100.26
91.07 96.95 98.85 120.69 155.75 174.10 133.22 115.35
* Value of 1 kg pure gold in the English mints, in coin, 1351-1411 = £43.0921 sterling; value of 1 kg pure silver in the English mints, in coin, 1351–1411 = £3.862 sterling. Sources: See sources for Tables 1 and 3.
284.22 28,804.70 39,501.95 52,453.54 27,757.60 32,508.04 11,612.12 40,185.51
11
102.73 106.75 97.76 90.06 127.38 102.77 106.80 96.70
9
47,798.68 113,202.34 48,852.80 30,979.88 19,950.05 13,755.52 24,748.40 32,459.06
8
34,291.77 75,468.23 32,568.53 20,653.25 13,300.03 9,170.35 16,498.93 21,639.37
7
Low Countries: CPI CPI Flanders Brabant
6
1461–1465 1466–1470 1471–1475 1476–1480 1481–1485 1486–1490 1491–1495 1496–1500
10
Low Countries: Coinage Outputs current in £ groot
5
Low Countries: Coinage Outputs in constant £ sterling of 1351–1411 values*
4
England: Phelps Brown and Hopkins CPI
3
England: Coinage Outputs in current £ sterling
2
England: Coinage Outputs in constant £ sterling of 1351–1411 values*
1
Years
18
Continued.
17
Table 4.
4259 Ch06 4/12/02 10:12 am Page 239
Wage Stickiness and Monetary Changes 239
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
80.00 84.00 129.00 102.20 110.80 79.20 137.80 107.80 91.60 116.80 114.20 189.80 160.00 102.20 112.00 85.20 81.00 102.00 125.40 116.40
20.00
weight
40
Farinaceous: Wheat, Rye, Barley, Peas
95.00 64.10 97.00 99.80 92.80 78.40 75.00 89.80 88.20 105.40 116.60 127.00 106.40 102.60 100.20 88.20 85.00 87.80 110.80 102.00
25.00
Meat & Fish*
74.00 101.60 95.80 102.80 94.00 96.80 96.80 95.20 93.60 101.70 134.40 142.00 153.40 119.00 113.00 112.40 110.40 120.20 123.40 130.80
12.50
Butter & Cheese
80.00 104.80 131.00 118.20 154.30 100.40 157.60 138.80 119.00 134.80 126.40 238.80 174.80 133.40 148.20 97.00 98.60 120.00 165.60 141.60
22.50
Drink: Barley Malt
164.20 152.20 147.00 149.00 139.00 147.00 158.60 165.60
7.50
Fuel and Light
74.03 48.40 36.40 36.20 40.90 34.80 44.20 60.20 52.60 70.60 75.60 68.60 68.40 69.40 65.40 65.40 49.60 58.40 118.60 160.40
12.50
Textiles
80.00 95.01 130.06 110.67 133.83 90.42 148.28 124.21 106.11 126.33 120.66 215.74 167.84 118.72 131.16 91.45 90.32 111.53 146.68 129.74
42.50
Farinaceous [with drink]
88.00 76.60 96.60 100.80 93.20 84.53 82.27 91.60 90.00 104.17 122.53 132.00 122.07 108.07 104.47 96.27 93.47 98.60 115.00 111.60
37.50
Meat, Fish Dairy
74.03 48.40 36.40 36.20 40.90 34.80 44.20 60.20 52.60 70.60 75.60 68.60 104.33 100.45 96.00 96.75 83.13 91.63 133.60 162.35
20.00
82.44 81.25 103.84 96.61 104.80 80.52 107.45 102.34 92.35 109.81 115.33 161.91 137.97 111.07 114.12 94.32 90.06 102.70 132.18 129.46
100.00
Fuel and Composite Textiles (Adjusted)
1
Year
240
1264–1265 1266–1270 1271–1275 1276–1280 1281–1285 1286–1290 1291–1295 1296–1300 1301–1305 1306–1310 1311–1315 1316–1320 1321–1325 1326–1330 1331–1335 1336–1340 1341–1345 1346–1350 1351–1355 1356–1360
3
Price-Relatives in England, 1266–1270 to 1496–1500. The Phelps Brown & Hopkins “Basket of Consumables” (Adjusted version) in quinquennial means: 1451–1475 = 100.
2
Table 5.
4259 Ch06 4/12/02 10:12 am Page 240
JOHN H. MUNRO
16
19
20
21
22
23
24
25
27
28
29
30
31
32
33
34
35
36
38
39
40 168.40 182.40 163.60 162.40 152.40 139.00 133.00 124.40 123.20 120.20 119.00 111.00 111.20 124.60 120.00 114.20 118.40 106.80 101.00 100.60 94.40 108.60 93.40 92.00
127.40 126.00 138.00 100.50 96.10 90.00 85.40 85.80 92.40 102.00 108.20 104.00 99.20 108.40 98.80 98.00 99.40 97.40 97.20 100.80 101.30 108.40 101.60 107.20
14
200.60 167.20 128.00 113.00 123.60 108.40 123.60 127.40 136.40 119.80 118.80 126.60 118.00 133.60 120.80 141.80 83.60 96.40 104.00 90.80 106.00 99.40 99.60 83.80
13
125.80 128.40 129.60 118.80 107.00 101.80 102.80 114.20 110.00 114.20 117.40 115.30 97.57 110.42
12
122.80 128.00 136.40 105.60 110.20 108.40 102.80 106.40 105.80 105.60 102.60 103.60 92.60 98.80 101.40 106.80 98.80 106.20 97.40 100.80 100.00 111.80 96.00 79.20
168.60 161.46 130.45 105.00 114.19 96.54 110.89 117.42 126.71 114.81 106.66 121.80 106.80 119.95 115.53 143.87 80.40 96.21 103.53 92.02 107.04 101.47 98.94 94.25
9
132.60 155.00 133.20 96.00 103.60 83.20 96.60 106.20 115.80 109.20 93.00 116.40 94.20 104.60 109.60 146.20 76.80 96.00 103.00 93.40 108.20 103.80 98.20 106.00
10 123.80 128.13 134.13 110.00 109.13 106.20 102.80 109.00 107.20 108.47 107.53 107.50 94.26 102.38 101.40 106.80 98.80 106.20 97.40 100.80 100.00 111.80 96.00 79.20
Meat, Fish Dairy
8
1361–1365 1366–1370 1371–1375 1376–1380 1381–1385 1386–1390 1391–1395 1396–1400 1401–1405 1406–1410 1411–1415 1416–1420 1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470 1471–1475 1476–1480
11
Farinaceous [with drink]
142.78 147.15 147.60 123.71 117.21 108.38 103.25 100.28 103.95 108.83 112.25 106.63 103.70 114.48 106.75 104.08 106.53 100.93 98.63 100.73 98.71 108.48 98.53 101.50
146.64 146.10 135.26 110.62 112.90 102.53 106.33 110.84 114.84 111.23 108.11 113.40 101.48 112.27 108.48 122.01 92.53 100.90 100.25 97.06 102.73 106.75 97.76 90.06
Fuel and Composite Textiles (Adjusted)
7
Textiles
6
Fuel and Light
5
Drink: Barley Malt
4
Butter & Cheese
3
Meat & Fish*
2
Farinaceous: Wheat, Rye, Barley, Peas
1
Year
18
Continued.
17
Table 5.
4259 Ch06 4/12/02 10:12 am Page 241
Wage Stickiness and Monetary Changes 241
15
26
37
241
16
19
20
21
22
23
24
25
27
28
29
30
31
32
33
34
35
36
38
39
40
15
26
37 145.47 97.84 104.36 95.61
120.00 105.80 111.80 95.80
102.78 107.58 102.58 100.70
127.38 102.77 106.80 96.70
Fuel and Composite Textiles (Adjusted)
* From 1431–1435 the weight assigned to meat and fish becomes 37.5, in the absence of price data for butter and cheese. Sources: See Table 1, and Archives of the British Library of Political and Economic Science, the Phelps Brown Papers Collection.
105.40 118.00 114.80 111.20
14
98.40 90.20 82.20 83.20
13
143.40 88.80 99.00 91.00
12
120.00 105.80 111.80 95.80
9
147.80 108.00 110.40 100.80
10 Meat, Fish Dairy
8
1481–1485 1486–1490 1491–1495 1496–1500
11
Farinaceous [with drink]
7
Textiles
6
Fuel and Light
5
Drink: Barley Malt
4
Butter & Cheese
3
Meat & Fish*
2
Farinaceous: Wheat, Rye, Barley, Peas
1
Year
18
Continued.
17
Table 5.
4259 Ch06 4/12/02 10:12 am Page 242
242 JOHN H. MUNRO
1
2
3
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
1266–1270 1271–1275 1276–1280 1281–1285 1286–1290 1291–1295 1296–1300 1301–1305 1306–1310 1311–1315 1316–1320 1321–1325 1326–1330 1331–1335 1336–1340 1341–1345 1346–1350 1351–1355 1356–1360 1361–1365 1366–1370 1371–1375 1376–1380 1381–1385
Years (5)
3.00 3.00 3.00 3.00 3.00 3.00 3.00 3.30 3.60 4.00 4.00 4.00 4.00 4.00 3.60 3.00 3.00 3.60 4.60 5.00 5.00 5.00 5.00 5.00
Nominal Day Wage in d for a Master
1.50 1.50 1.50 1.50 1.50 1.50 1.50 1.65 1.80 2.00 2.00 2.00 2.00 2.00 1.80 1.50 1.50 1.80 2.60 3.00 3.00 3.00 3.00 3.00
Nominal Day Wage in d for a Laborer
50.00 50.00 50.00 50.00 50.00 50.00 50.00 55.00 60.00 66.67 66.67 66.67 66.67 66.67 60.00 50.00 50.00 60.00 76.67 83.33 83.33 83.33 83.33 83.33
Master Nominal Wage Index 1451–1475 = 100 [= 6d daily]
37.50 37.50 37.50 37.50 37.50 37.50 37.50 41.25 45.00 50.00 50.00 50.00 50.00 50.00 45.00 37.50 37.50 45.00 65.00 75.00 75.00 75.00 75.00 75.00
Laborer Nominal Wage Index 1451–1475 = 100 [= 4d daily]
81.249 103.838 96.605 104.800 80.519 107.454 102.341 92.346 109.814 115.330 161.908 137.970 111.070 114.120 94.315 90.060 102.700 132.185 129.460 146.635 146.100 135.260 110.618 112.898
Aggregate Price Index (Adjusted) 1451–1475 = 100
61.823 48.707 51.810 48.163 62.930 47.222 49.108 59.568 55.244 58.322 44.849 48.869 60.816 59.552 64.226 55.775 49.138 46.514 59.311 57.053 58.129 62.258 77.272 73.892
Real Wage Index Master 1451–1475 = 100 arithmetic mean 46.368 36.530 38.858 36.122 47.198 35.417 36.831 44.676 41.433 43.742 33.637 36.651 45.612 44.664 48.169 41.832 36.853 34.886 50.161 51.348 52.316 56.032 69.544 66.503
Real Wage Index Laborer 1451–1475 = 100 arithmetic mean 61.539 48.152 51.757 47.710 62.097 46.532 48.856 59.286 54.807 57.805 41.176 48.320 60.022 58.418 63.259 55.519 48.685 44.557 59.007 56.830 57.039 61.610 75.335 73.813
Real Wage Index Master 1451–1475 = 100 harmonic mean 46.155 36.114 38.818 35.782 46.573 34.899 36.642 44.464 41.105 43.354 30.882 36.240 45.017 43.814 47.444 41.639 36.514 33.418 49.413 51.147 51.335 55.449 67.801 66.432
Real Wage Index Laborer 1451–1475 = 100 harmonic mean
Table 6. Wages of Master Building Craftsmen and Their Laborers in Southern England in Pence (d) Sterling per Day, with Wage- and Price-relatives from the Phelps Brown and Hopkins indexes in Quinquennial Means, 1266–1270 to 1496–1500: 1451–1475 = 100.
4259 Ch06 4/12/02 10:12 am Page 243
Wage Stickiness and Monetary Changes 243
243
13
14
15
16
20
21
22
23
24
25
26
27
28
29
30
31
32
33
38
39
40
34
35
36
37 81.311 79.526 75.589 75.286 87.562 92.633 89.126 98.867 90.998 92.358 85.413 108.574 99.228 100.064 103.115 99.009 93.723 102.590 111.678 80.745 98.059 93.931 103.510
73.180 71.573 68.030 71.288 85.887 92.633 89.126 98.867 90.998 92.358 85.413 108.574 99.228 100.064 103.115 99.009 93.723 102.590 111.678 80.745 98.059 93.931 103.510
81.277 78.372 75.187 73.717 87.067 92.503 88.181 98.546 89.074 92.187 81.960 108.079 99.108 99.751 103.034 97.340 93.681 102.297 111.043 78.505 97.305 93.637 103.413
73.149 70.535 67.668 68.693 85.527 92.503 88.181 98.546 89.074 92.187 81.960 108.079 99.108 99.751 103.034 97.340 93.681 102.297 111.043 78.505 97.305 93.637 103.413
Real Wage Index Laborer 1451–1475 = 100 harmonic mean
Sources: Phelps Brown and Hopkins (1955, 1956, 1981); Archives of the British Library of Political and Economic Science, the Phelps Brown Papers Collection.
102.530 106.330 110.835 114.840 111.235 108.105 113.403 101.476 112.267 108.475 122.010 92.525 100.900 100.250 97.055 102.733 106.745 97.755 90.055 127.380 102.770 106.795 96.700
12
75.00 75.00 75.00 80.00 95.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00
9
83.33 83.33 83.33 85.00 96.67 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00
10
3.00 3.00 3.00 3.20 3.80 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00 4.00
19 Real Wage Index Master 1451–1475 = 100 harmonic mean
7
5.00 5.00 5.00 5.10 5.80 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00 6.00
8
Real Wage Index Laborer 1451–1475 = 100 arithmetic mean
244
1386–1390 1391–1395 1396–1400 1401–1405 1406–1410 1411–1415 1416–1420 1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470 1471–1475 1476–1480 1481–1485 1486–1490 1491–1495 1496–1500
11
Real Wage Index Master 1451–1475 = 100 arithmetic mean
6
Aggregate Price Index (Adjusted) 1451–1475 = 100
5
Laborer Nominal Wage Index 1451–1475 = 100 [= 4d daily]
4
Master Nominal Wage Index 1451–1475 = 100 [= 6d daily]
3
Nominal Day Wage in d for a Laborer
2
Nominal Day Wage in d for a Master
1
Years (5)
18
Continued.
17
Table 6.
4259 Ch06 4/12/02 10:12 am Page 244
JOHN H. MUNRO
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
39
40
1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340
6
245 1.00
1.00 1.00
1.00
2.00
1.50
1.00
1.50 1.50
1.00
2.00
1.50
Maximum Wage in d sterling
1.000
1.137 1.170
1.000
2.000
1.500
131.87 215.92 215.14 153.76 118.57 106.16 121.05 140.53 164.80 137.03 126.45 124.03 95.50 96.48 118.93 120.43 134.05 131.08 110.58 99.38 95.53 100.75 111.48 84.55 78.68 96.13 25.000
28.427 29.242
25.000
50.000
37.500
Mean PB & H Nominal Wage Price Index Wage Index Weighted 1451–1475=100 1451–1475=100 in d 4.0d
26.008
25.501 34.586
20.157
42.170
28.438
Real Wage Index NWI/CPI
Wage Stickiness and Monetary Changes
23.5
62.0 82.5
31 73 3
289.0
5.0
1
289
36.0
4
Minimum Wage in d sterling
3
Labor Days Number
2
Laborers Number
1
Year Harvest
5
Wages of Hired Day Laborers on the Taunton Manor, in pence per day, 1315–1415.
4
Table 7.
4259 Ch06 4/12/02 10:12 am Page 245
245
38
22
23
24
25
26
28
29
30
31
32
33
34
39
40
21 3.00 3.00 2.00 2.17 2.00 2.00 3.00 2.00 1.00 1.00 1.00
1.500 1.000 1.000 1.000 1.000 1.000 1.488 1.000 2.458 1.374 1.875 2.000 2.000 2.478 2.572 2.667 2.077 2.000 2.167 2.000 1.727 1.095 1.013 1.000 1.000 1.000
85.70 85.10 84.28 97.00 98.23 87.83 109.35 116.25 97.60 102.48 133.58 159.63 138.95 116.80 111.98 114.43 132.83 139.15 125.73 135.18 130.95 153.43 154.95 151.23 142.63 121.15
37.500 25.000 25.000 25.000 25.000 25.000 37.188 25.000 61.444 34.343 46.875 50.000 50.000 61.943 64.305 66.667 51.923 50.000 54.175 50.000 43.182 27.381 25.316 25.000 25.000 25.000
43.757 29.377 29.665 25.773 25.452 28.466 34.008 21.505 62.955 33.514 35.093 31.323 35.984 53.034 57.428 58.262 39.091 35.932 43.090 36.989 32.976 17.846 16.338 16.532 17.528 20.636
Real Wage Index NWI/CPI
246
2.50 2.00 2.00 2.17 2.00 1.00 1.00 1.00 1.00 1.00 1.00
20
15.0 52.0 1.5 3.0 0.5 11.0 42.0 39.5 18.0 11.0 41.0
19
3 5 1 1 1 5 42 38 18 11 21
16
1.00 1.50 1.00 2.50 2.00 2.00 2.00 2.00 3.00
15
1.00 1.00 1.00 1.00 1.11 1.50 2.00 2.00 2.00
14
15.5 521.0 5.0 71.0 99.0 8.0 2.0 2.0 325.5
13
3 509 1 70 93 7 4 2 7
12
1.50 1.00 1.00 1.00
11
1.50 1.00 1.00 1.00
9
2.5 10.0 6.0 16.0
10
1 1 2 2
8
1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
7
Mean PB & H Nominal Wage Price Index Wage Index Weighted 1451–1475=100 1451–1475=100 in d 4.0d
6
Maximum Wage in d sterling
4
Minimum Wage in d sterling
3
Labor Days Number
2
Laborers Number
1
Year Harvest
18
Continued.
17
Table 7.
4259 Ch06 4/12/02 10:12 am Page 246
JOHN H. MUNRO
5
27
35
36
37
38
23
24
25
26
28
29
30
31
32
33
34
39
40 17.0
22
7
247 1.00
1.00 1.00 1.00 1.00 1.00 1.00 1.00
1.000 2.054 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
136.83 138.90 149.83 183.80 164.20 132.03 130.50 124.78 124.80 146.08 111.58 94.98 94.24 106.23 118.19 111.13 108.23 116.20 110.75 104.10 100.28 102.45 100.28 105.55 132.85 104.40
25.000 51.351 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000
18.272 36.970 16.686 13.602 15.225 18.936 19.157 20.036 20.032 17.114 22.406 26.323 26.529 23.535 21.153 22.497 23.100 21.515 22.573 24.015 24.931 24.402 24.931 23.685 18.818 23.946
Real Wage Index NWI/CPI
Wage Stickiness and Monetary Changes
1.00
21
1.00 1.00 1.00 1.00 1.00 1.00 1.00
20
305.0 13.0 31.0 20.0 142.0 717.0 60.0
19
227 7 6 3 126 715 20
16
1.00
15
1.00
14
235.0
13
219
12
1.00 3.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
11
1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
9
319.0 92.5 156.0 131.0 2653.0 151.0 37.0 219.0 78.0 102.0 61.0 71.0 15.0 70.0
10
12 43 2 57 87 41 15 207 44 95 17 65 10 8
8
1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392
7
Mean PB & H Nominal Wage Price Index Wage Index Weighted 1451–1475=100 1451–1475=100 in d 4.0d
6
Maximum Wage in d sterling
4
Minimum Wage in d sterling
3
Labor Days Number
2
Laborers Number
1
Year Harvest
18
Continued.
17
Table 7.
4259 Ch06 4/12/02 10:12 am Page 247
247
5
27
35
36
37
38
25
26
28
29
30
31
32
33
34
39
40 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
1.00 1.00 1.00 1.00 2.00 3.00 2.83 4.00 4.50 3.00
1.00 1.00
1.00 1.00 1.00 1.00 1.00
1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.086 1.379 1.196 1.070 1.457 1.500
100.50 100.65 93.25 99.18 116.18 121.43 112.93 104.48 130.03 127.28 118.85 99.33 98.73 99.70 99.35 107.23 119.78 130.13 105.95 102.83 108.33 108.13 115.30
25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 25.000 27.160 34.464 29.911 26.756 36.429 37.500
24.876 24.839 26.810 25.208 21.519 20.589 22.139 23.929 19.227 19.643 21.035 25.170 25.323 25.075 25.164 23.315 20.872 20.873 32.529 29.089 24.700 33.691 35.524
Real Wage Index NWI/CPI
248
10
15
27
35
36
37
38
Sources: Archives of the British Library of Political and Economic Science, the Beveridge Price and Wage History Collection.
337.0 571.0 162.0 1157.0 28.0 60.5 35.0 24.5
24
275 235 41 960 12 20 32 2
23
272.0 36.0
22
99 24
21
1.00 1.00 1.00 1.00 1.00
20
122.5 135.5 131.0 10.0 109.0
19
17 55 38 8 27
16
1.00 1.00 1.00 1.00
14
1.00 1.00 1.00 1.00
13
34.5 195.0 62.5 71.0
12
8 170 7 28
9 Nominal Wage Index 1451-75=100 4.0d
8
1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415
11
PB & H Price Index 1451-75=100
7
Mean Wage Weighted in d
6
Maximum Wage in d sterling
4
Minimum Wage in d sterling
3
Labor Days Number
2
Laborers Number
1
Year Harvest
18
Continued.
17
Table 7.
4259 Ch06 4/12/02 10:12 am Page 248
JOHN H. MUNRO
5
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
1346–1350* 1351–1355 1356–1360 1361–1365 1366–1370 1371–1375 1376–1380 1381–1385 1386–1390 1391–1395 1396–1400 1401–1405 1406–1410 1411–1415 1416–1420 1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460
3
Year
2
63.868 76.593 110.558 119.255 135.641 145.519 141.024 150.534 157.514 111.784 113.407 111.810 132.939 120.370 135.616 141.680 148.741 155.989 177.022 143.350 138.904 127.434 148.845
Basket Consumables Total Value in d groot Flemish 62.076 68.466 102.100 108.636 126.994 121.901 105.597 110.799 132.745 82.803 92.733 95.190 115.682 93.652 110.755 112.756 122.830 132.917 172.289 111.205 107.703 95.302 131.873
Grains Group Index 1451–1475= 100 56.354 groot Flemish 46.777 63.048 93.151 98.228 101.825 112.359 121.366 135.714 122.655 99.235 92.132 80.675 91.056 92.417 104.677 114.392 114.511 115.130 109.153 113.067 110.051 102.660 107.281
Dairy Group Index 1451–1475= 100 44.665 groot Flemish 31.624 38.968 45.160 56.023 73.568 105.388 108.038 108.711 110.470 82.282 79.118 87.565 107.127 104.114 104.636 106.998 112.262 117.353 123.350 119.403 114.952 110.282 105.288
Textiles Group Index 1451–1475= 100 25.276 groot Flemish 50.571 60.646 87.540 94.425 107.401 115.222 111.662 119.193 124.719 88.510 89.796 88.531 105.261 95.309 107.381 112.182 117.773 123.512 140.166 113.504 109.984 100.902 117.855
54.77% 50.37% 52.04% 51.34% 52.76% 47.21% 42.20% 41.48% 47.49% 41.74% 46.08% 47.98% 49.04% 43.84% 46.02% 44.85% 46.54% 48.02% 54.85% 43.72% 43.70% 42.14% 49.93%
Commodity Grains Basket Index as percent 1451–1475= of total 100 basket 126.295 by value groot Flemish 32.71% 36.77% 37.63% 36.79% 33.53% 34.49% 38.44% 40.27% 34.78% 39.65% 36.29% 32.23% 30.59% 34.29% 34.48% 36.06% 34.39% 32.97% 27.54% 35.23% 35.39% 35.98% 32.19%
Dairy as percent of total basket by value
12.52% 12.86% 10.32% 11.87% 13.71% 18.31% 19.36% 18.25% 17.73% 18.60% 17.63% 19.79% 20.37% 21.86% 19.50% 19.09% 19.08% 19.02% 17.61% 21.05% 20.92% 21.87% 17.88%
Textiles as percent of total basket by value
132.71% 108.59% 109.61% 110.60% 124.72% 108.49% 87.01% 81.64% 108.23% 83.44% 100.65% 117.99% 127.05% 101.34% 105.81% 98.57% 107.27% 115.45% 157.84% 98.35% 97.87% 92.83% 122.92%
Ratio of Grain to Dairy Price Indexes
196.29% 175.70% 226.09% 193.91% 172.62% 115.67% 97.74% 101.92% 120.16% 100.63% 117.21% 108.71% 107.99% 89.95% 105.85% 105.38% 109.41% 113.26% 139.67% 93.13% 93.69% 86.42% 125.25%
Ratio of Grain to Textile Price Indexes
Prices in Flanders: The Value of a “Basket of Consumables” in d groot Flemish and in price-relatives in quiquennial means, 1346–1350 to 1496–1500: 1451–1475 = 100.
1
Table 8.
4259 Ch06 4/12/02 10:12 am Page 249
Wage Stickiness and Monetary Changes 249
249
14
15
16
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
38
39
40 88.705 96.520 96.017 117.213 156.853 184.511 144.981 100.255
41.78% 43.16% 44.81% 47.83% 56.53% 46.14% 48.27% 36.55%
36.18% 37.08% 36.14% 35.71% 29.75% 35.86% 29.80% 32.92%
10
37 22.05% 19.76% 19.05% 16.46% 13.72% 18.00% 21.93% 30.54%
91.53% 92.24% 98.27% 106.17% 150.63% 101.96% 128.38% 88.01%
Ratio of Grain to Dairy Price Indexes
84.99% 97.95% 105.49% 130.32% 184.80% 114.94% 98.74% 53.68%
Ratio of Grain to Textile Price Indexes
* data for 1349-50 only. Sources: See sources for Table 1 and Munro (1984a, Table B: 5–6, pp. 104–109); Verlinden and Scholliers (1959–1965, Vol. I, pp. 44–45, 52–53; Vol. II:i, pp. 3–70); Stadsarchief Gent, Stadsrekeningen, 1350–1351 to 1499–1500.
97.721 95.304 91.406 96.410 107.537 165.979 158.841 152.966
13
90.737 101.206 98.116 118.347 131.927 187.098 122.174 93.309
12
83.052 93.351 96.422 125.644 198.728 190.773 156.841 82.119
9
112.030 121.900 121.264 148.034 198.097 233.028 183.104 126.617
8
1461–1465 1466–1470 1471–1475 1476–1480 1481–1485 1486–1490 1491–1495 1496–1400
7
Textiles as percent of total basket by value
6
Year
11
Dairy as percent of total basket by value
5
Commodity Grains Basket Index as percent 1451–1475= of total 100 basket 126.295 by value groot Flemish
4
Textiles Group Index 1451–1475= 100 25.276 groot Flemish
3
Dairy Group Index 1451–1475= 100 44.665 groot Flemish
2
Grains Group Index 1451–1475= 100 56.354 groot Flemish
1
Basket Consumables Total Value in d groot Flemish
18
Continued.
17
Table 8.
4259 Ch06 4/12/02 10:12 am Page 250
250 JOHN H. MUNRO
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
251
149.440 159.400 155.882 164.113 168.089 179.277 175.173 194.440 163.507 154.360 152.760 177.613 141.173 150.293 153.240 187.093 241.440 269.880 206.507 178.813
50.952 63.593 67.612 68.342 86.293 112.962 103.463 159.238 103.877 102.562 99.031 132.520 77.451 93.770 97.228 131.083 201.349 229.456 157.387 109.723
Sources: Van der Wee (1963, Vol. I; 1975, 1978).
149.440 159.400 172.000 187.280 209.720 232.880 238.940 291.660 245.260 231.540 229.140 266.420 211.760 225.440 229.860 280.640 362.160 404.820 309.760 268.220
6
1401–1405 1406–1410 1411–1415 1416–1420 1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470 1471–1475 1476–1480 1481–1485 1486–1490 1491–1495 1496–1500
5
69.235 69.744 76.426 84.259 92.734 93.397 100.345 102.646 98.531 95.654 95.367 109.638 97.004 97.823 100.168 117.892 134.819 148.006 117.715 114.595
75.008 73.542 78.974 91.636 91.703 92.802 105.432 113.496 118.127 101.400 102.666 97.468 100.766 100.000 99.100 110.696 124.892 137.654 123.492 124.192
64.269 68.552 73.971 80.542 90.193 100.153 102.759 125.432 105.477 99.577 98.545 114.577 91.070 96.953 98.854 120.693 155.752 174.098 133.216 115.352
Composite Index in Brabant gr 232.524d 1451–1475 = 100
96.403 102.828 100.559 105.868 108.433 115.651 113.003 125.432 105.477 99.577 98.545 114.577 91.070 96.953 98.854 120.693 155.752 174.098 133.216 115.352
28.00% 32.76% 32.28% 29.97% 33.79% 39.83% 35.56% 44.83% 34.78% 36.37% 35.49% 40.85% 30.03% 34.16% 34.73% 38.36% 45.65% 46.54% 41.72% 33.59%
41.88% 39.55% 40.16% 40.67% 39.97% 36.25% 37.96% 31.81% 36.31% 37.34% 37.62% 37.20% 41.41% 39.22% 39.39% 37.97% 33.65% 33.05% 34.35% 38.62%
30.13% 27.69% 27.56% 29.37% 26.24% 23.92% 26.48% 23.36% 28.91% 26.28% 26.89% 21.96% 28.56% 26.62% 25.88% 23.67% 20.70% 20.41% 23.93% 27.79%
73.59 91.18 88.47 81.11 93.05 120.95 103.11 155.13 105.43 107.22 103.84 120.87 79.84 95.86 97.07 111.19 149.35 155.03 133.70 95.75
Composite Grains as Meat-Fish Industrial Ratio of Index in percent Dairy as as Grain Flemish gr of total percent percent Prices to 155.016d of total of total Meat-Fish1451–1475 Dairy Prices = 100 Price relative 67.93 86.47 85.61 74.58 94.10 121.72 98.13 140.30 87.94 101.15 96.46 135.96 76.86 93.77 98.11 118.42 161.22 166.69 127.45 88.35
Ratio of Grain Prices to Industrial Prices: Price relative
2
Basket of Basket of Grain Price Meat-Fish Industrial Consumables Consumables Index Dairy Index Goods Index Total Value Total Value 82.116d 90.388d 60.020d in d groot in d groot 1451–1475 1451–1475 1451–1475 Brabant Flemish = 100 = 100 = 100
1
Year
4
Prices in Brabant (Antwerp-Lier-Mechelen-Brussels region) in pence groot of Brabant and price-relatives. Index: mean 1451–1475 = 100.
3
Table 9.
4259 Ch06 4/12/02 10:12 am Page 251
Wage Stickiness and Monetary Changes 251
1
2
3
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
Basket of Consumables Values in d groot Flemish
63.868 76.593 110.558 119.255 135.641 145.519 141.024 150.534 157.514 111.784 113.407 111.810 132.939 120.370 135.616
Year
1349–1350 1351–1355 1356–1360 1361–1365 1366–1370 1371–1375 1376–1380 1381–1385 1386–1390 1391–1395 1396–1400 1401–1405 1406–1410 1411–1415 1416–1420
5.000 5.200 6.000 6.850 8.000 8.000 8.800 8.800 10.867 9.000 9.850 10.000 10.000 10.000 10.000
45.455 47.273 54.545 62.273 72.727 72.727 80.000 80.000 98.788 81.818 89.545 90.909 90.909 90.909 90.909
Wages Bruges of Master Nominal Building Wage Index Craftsmen 1451–1475 = in Bruges 100 Estimated (11d) Median Wage in d gr Flemish 89.884 78.889 63.230 68.577 68.565 63.684 72.090 68.151 80.335 93.500 100.219 103.150 88.279 96.376 86.463
Real Wage Index 1451–1475 = 100 NWI/CPI arithmetic
89.884 77.572 62.309 65.366 67.716 63.120 70.520 65.898 77.375 92.439 99.731 102.687 86.366 95.384 84.660
Real Wage Index 1451–1475 = 100 NWI/CPI harmonic
16.440 14.429 11.565 12.543 12.541 11.648 13.186 12.465 14.694 17.102 18.331 18.867 16.147 17.628 15.814
Real Wage in Units of Commodity Baskets per year of 210 days arithmetic mean
16.440 14.188 11.397 11.956 12.386 11.545 12.898 12.053 14.152 16.908 18.241 18.782 15.797 17.446 15.485
Real Wage in Units of Commodity Baskets per year of 210 days harmonic mean
252
50.571 60.646 87.540 94.425 107.401 115.222 111.662 119.193 124.719 88.510 89.796 88.531 105.261 95.309 107.381
Commodity Basket Index 1451–1475 = 100 (126.295d)
Table 10. Prices and Wages in Flanders, 1349–1350 to 1496–1500: The “Basket of Consumables” Price Index and Its Components and the Wages of Master Building Craftsmen in Bruges in Quinquennial Means, Arithmetic and Harmonic (Mean of 1451–1475 = 100).
4259 Ch06 4/12/02 10:12 am Page 252
JOHN H. MUNRO
16
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
38
39
40 81.037 77.190 79.378 71.344 88.102 90.922 99.106 84.850 112.733 103.605 104.148 85.315 63.754
10 9
15
37
253
14.868 14.178 14.652 14.112 16.454 16.810 18.167 15.592 20.796 18.970 19.193 16.092 12.607
14.822 14.118 14.519 13.049 16.114 16.630 18.127 15.519 20.619 18.950 19.049 15.605 11.661
Real Wage in Units of Commodity Baskets per year of 210 days harmonic mean
Sources: See sources for Table 1, and Stadsarchief Brugge, Stadsrekeningen 1350–1351 to 1499–1500; Algemeen Rijksarchief, Rekenkamer, reg. nos. 32,461–32,532.
81.287 77.518 80.106 77.154 89.960 91.907 99.326 85.247 113.700 103.714 104.933 87.978 68.927
14
90.909 90.909 98.182 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000
13
10.000 10.000 10.800 11.000 11.000 11.000 11.000 11.000 11.000 11.000 11.000 11.000 11.000
12
112.182 117.773 123.512 140.166 113.504 109.984 100.902 117.855 88.705 96.520 96.017 117.213 156.853 184.511 144.981 100.255
8
141.680 148.741 155.989 177.022 143.350 138.904 127.434 148.845 112.030 121.900 121.264 148.034 198.097 233.028 183.104 126.617
7
Real Wage in Units of Commodity Baskets per year of 210 days arithmetic mean
6
1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470 1471–1475 1476–1480 1481–1485 1486–1490 1491–1495 1496–1400
11
Real Wage Index 1451–1475 = 100 NWI/CPI harmonic
5
Real Wage Index 1451–1475 = 100 NWI/CPI arithmetic
4
Wages Bruges of Master Nominal Building Wage Index Craftsmen 1451–1475 = in Bruges 100 Estimated (11d) Median Wage in d gr Flemish
3
Commodity Basket Index 1451–1475 = 100 (126.295d)
2
Basket of Consumables Values in d groot Flemish
1
Year
18
Continued.
17
Table 10.
4259 Ch06 4/12/02 10:12 am Page 253
Wage Stickiness and Monetary Changes 253
1
2
3
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40 5.250 5.250 5.250 4.947 3.733 3.733 3.733 3.733 3.733 3.733 4.293 4.667
Nominal Wage in d groot Flemish journeymen per day
5.00 5.20 6.00 6.00 6.00 5.40 5.00 5.00 5.00 5.00 5.00 5.00
Bruges Policeman's Daily Wage
1102.500 1102.500 1102.500 1038.800 784.000 784.000 784.000 784.000 784.000 784.000 901.600 980.000
Income for 210 days per fuller journeyman in d groot Flemish
4.594 4.594 4.594 4.328 3.267 3.267 3.267 3.267 3.267 3.267 3.757 4.083
Income for 210 days per fuller journeyman in £ groot Flemish
144.679 141.024 150.534 157.514 111.784 113.407 111.810 132.939 120.370 135.616 141.680 148.741
Value of Commodity Basket in d groot Flemish
114.556 111.662 119.193 124.719 88.510 89.796 88.531 105.261 95.309 107.381 112.182 117.773
7.680 7.834 7.388 6.669 7.094 6.955 7.044 6.028 6.581 5.904 6.374 6.617
Price No. of Index Commodity 1451–1475 = Baskets 100 for fullers' annual wage arithmetic mean
7.620 7.818 7.324 6.451 7.014 6.913 7.012 5.897 6.513 5.781 6.299 6.589
No. of Commodity Baskets for fullers' annual wage harmonic mean
* data available only for 1373–1375. M = master fuller; J = journeymen fullers (2); d = 12d per shilling and 240d to the pound groot Flemish. Sources: Espinas and Pirenne (1906–1924, Vol. II, No. 492, pp. 535–537), Algemeen Rijksarchief België, Trésor de Flandre, Series I, no. 2208; Rijksarchief van Oost Vlaanderen te Gent, Oostenrijks Fonds, layette 2, Sources for Tables 1 and 8.
45.00 45.00 45.00 42.40 32.00 32.00 32.00 32.00 32.00 32.00 36.80 40.00
Nominal Wage in d groot Flemish 3 days: M + 2J
254
1373–1375* 1376–1380 1381–1385 1386–1390 1391–1395 1396–1400 1401–1405 1406–1410 1411–1415 1416–1420 1421–1425 1426–1430
Years
Table 11. Ghent Fullers’ Wages (Flanders): Nominal and Real, 1373–1430. Wages for Master and Two Journeymen To Full a Woolen Broadcloth in Three Days: and Wages for Policemen in Bruges: in d groot Flemish, with annual wages expressed in terms of the value of the Flemish “basket of consumables”, with pricerelatives in quinquennial means: 1373–1375 to 1426–1430.
4259 Ch06 4/12/02 10:12 am Page 254
JOHN H. MUNRO
1
2
3
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
1349–1350 1351–1355 1356–1360 1361–1365 1366–1370 1371–1375 1376–1380 1381–1385 1386–1390 1391–1395 1396–1400 1401–1405 1406–1410 1411–1415 1416–1420 1421–1425
Years (5)
15.167 15.167 15.167 15.167 15.167 25.500 41.000 41.000 40.000 36.000 36.000 36.000 36.000 36.000 38.400 38.400
Fullers' Fee: in d groot Flemish
5.308 5.308 5.308 5.308 5.308 8.925 14.350 14.350 14.000 12.600 12.600 12.600 12.600 12.600 13.440 13.440
Journeymen’s men's Pay in d groot Flemish
1.769 1.769 1.769 1.769 1.769 2.975 4.783 4.783 4.667 4.200 4.200 4.200 4.200 4.200 4.480 4.480
Journeymen’s men's Pay per day in d groot Flemish
371.583 371.583 371.583 371.583 371.583 624.750 1004.500 1004.500 980.000 882.000 882.000 882.000 882.000 882.000 940.800 940.800
Income for 210 days in d groot Flemish
63.868 76.593 110.558 119.255 135.641 145.519 141.024 150.534 157.514 111.784 113.407 111.810 132.939 120.370 135.616 141.680
Value of a Basket of Consumables in d groot Flemish
50.571 60.646 87.540 94.425 107.401 115.222 111.662 119.193 124.719 88.510 89.796 88.531 105.261 95.309 107.381 112.182
Flemish Price Index 1451–1475 = 100
5.818 4.915 3.411 3.234 2.774 3.352 7.138 6.731 6.280 7.981 7.824 7.924 6.782 7.404 7.127 6.668
Journeymen Fuller’s Annual Income in baskets of consumables arithmetic mean 5.818 4.851 3.361 3.116 2.739 3.442 7.123 6.673 6.197 7.890 7.777 7.888 6.635 7.327 6.854 6.595
Journeymen Fuller’s Annual Income in baskets of consumables harmonic mean
Table 12. Nominal and Real Wages of Kortrijk (Courtrai) Fullers, 1349–1450. Kortrijk Fullers' Wages, Nominal and Real: 1349–1450. Wages for Master and Two Journeymen to Full a Woolen Broadcloth in Three Days: in d groot Flemish, with annual wages expressed in terms of the value of a Flemish “basket of consumables,” with price relatives in quinquennial means, 1351–1355 to 1446–1450.
4259 Ch06 4/12/02 10:12 am Page 255
Wage Stickiness and Monetary Changes 255
255
15
16
19
20
21
22
23
24
25
26
27
28
29
30
31
32
34
35
36
38
39
40 148.741 155.989 177.022 143.350 138.904
117.773 123.512 140.166 113.504 109.984
6.446 6.671 5.987 6.981 7.132
6.432 6.616 5.536 6.836 7.055
Journeymen Fuller’s Annual Income in baskets of consumables harmonic mean
1
6
33
37
Sources: Espinas and Pirenne (1906–1924, Vol. II, no. 492, pp. 535–537), Algemeen Rijksarchief België, Trésor de Flandre, Series I. Rijksarchief van Oost Vlaanderen te Gent, Oostenrijks Fonds.
960.400 1033.900 980.000 980.000 980.000
14
4.573 4.923 4.667 4.667 4.667
13
13.720 14.770 14.000 14.000 14.000
12
39.200 42.200 40.000 40.000 40.000
9
Journeymen Fuller’s Annual Income in baskets of consumables arithmetic mean
10
1426–1430 1431–1435 1436–1440 1441–1445 1446–1450
11
Flemish Price Index 1451–1475 = 100
8
Value of a Basket of Consumables in d groot Flemish
7
Income for 210 days in d groot Flemish
5
Journeymen’s men’s Pay per day in d groot Flemish
4
Journeymen’s men’s Pay in d groot Flemish
3
Fullers’ Fee: in d groot Flemish
2
Years (5)
18
Continued.
17
Table 12.
4259 Ch06 4/12/02 10:12 am Page 256
256 JOHN H. MUNRO
1
2
3
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
1401–1405 1406–1410 1411–1415 1416–1420 1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470 1471–1475 1476–1480
Years (5)
7.75 8.00 8.00 8.00 8.00 8.00 9.70 10.00 11.40 12.00 12.00 12.00 12.00 12.00 12.00 12.00
Masons: Masters d groot Summer Wage
6.00 6.00 6.00 6.00 6.00 6.00 7.00 8.00 9.00 9.00 9.00 9.00 9.00 9.00 9.00 9.00
Masons: Masters d groot Winter Wage
7.313 7.500 7.500 7.500 7.500 7.500 9.025 9.500 10.800 11.250 11.250 11.250 11.250 11.250 11.250 11.250
Masons: Masters Mean annual wage*
149.440 159.400 172.000 187.280 209.720 232.880 238.940 291.660 245.260 231.540 229.140 266.420 211.760 225.440 229.860 280.640
64.269 68.552 73.971 80.542 90.193 100.153 102.759 125.432 105.477 99.577 98.545 114.577 91.070 96.953 98.854 120.693
64.583 66.667 66.667 66.667 66.667 66.667 80.833 83.333 95.000 100.000 100.000 100.000 100.000 100.000 100.000 100.000
Value of Brabant Masons: Brabant Price Nominal Commodity Index Wage Basket 1451–1475 Index in d groot = 100 1451–1475 Brabant 232.524d = 100 [12d] 100.301 97.250 90.126 82.772 73.916 66.565 78.130 66.437 89.287 100.425 101.477 87.277 109.805 103.142 101.159 82.855
Masons: Real Wage Index NWI/CPI harmonic mean 10.870 10.540 9.767 8.971 8.011 7.214 8.467 7.200 9.677 10.884 10.998 9.459 11.900 11.178 10.963 8.979
Annual Wage in Commodity Baskets (210 days) harmonic mean 10.262 9.881 9.157 8.410 7.510 6.763 7.858 6.840 9.192 10.203 10.310 8.868 11.156 10.480 10.278 8.418
77.48% 75.00% 75.00% 75.00% 75.00% 75.00% 71.64% 80.00% 79.36% 75.00% 75.00% 75.00% 75.00% 75.00% 75.00% 75.00%
Annual Winter Wage in Wage as Commodity percent Baskets of summer (seasonal)* wage harmonic mean
Table 13. Wages of Antwerp Masons: Daily Summer and Winter Rates in pence (d) groot of Brabant with estimated annual money wage incomes expressed in equivalent “baskets of consumables” with price-relatives, in quinquennial means: 1401–1405 to 1496–1500.
4259 Ch06 4/12/02 10:12 am Page 257
Wage Stickiness and Monetary Changes 257
257
16
19
20
21
22
23
24
25
26
27
28
29
31
32
33
34
35
36
38
39
40 100.000 107.500 100.000 103.333
15
30
37 64.205 61.883 75.066 89.522
6.958 6.707 8.135 9.702
6.523 6.316 7.627 9.039
75.00% 76.67% 75.00% 72.60%
Annual Winter Wage in Wage as Commodity percent Baskets of summer (seasonal)* wage harmonic mean
* Seasonal: the summer wage prevailed for nine months and the winter wage for three. Sources: Van der Wee (1963, Vol. I: Statistics, pp. 333–389: Synoptic Tables of Wages and Appendices 27–30; 1975, 1978).
155.752 174.098 133.216 115.352
14
362.160 404.820 309.760 268.220
13
11.250 12.150 11.250 11.550
11
9.00 9.90 9.00 9.00
9
12.00 12.90 12.00 12.40
10
Annual Wage in Commodity Baskets (210 days) harmonic mean
8
1481–1485 1486–1490 1491–1495 1496–1400
12
Masons: Real Wage Index NWI/CPI harmonic mean
7
Masons: Nominal Wage Index 1451–1475 = 100 [12d]
6
Brabant Price Index 1451-75= 100 232.524d
5
Value of Brabant Commodity Basket in d groot Brabant
4
Masons: Masters Mean annual wage*
3
Masons: Masters d groot Winter Wage
2
Masons: Masters d groot Summer Wage
1
Years (5)
18
Continued.
17
Table 13.
4259 Ch06 4/12/02 10:12 am Page 258
258 JOHN H. MUNRO
2
3
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470 1471–1475 1476–1480 1481–1485 1486–1490 1491–1495 1496–1500
209.72 232.88 238.94 291.66 245.26 231.54 229.14 266.42 211.76 225.44 229.86 280.64 362.16 404.82 309.76 268.22
Value of Basket of Consumables in d groot Brabant
10.00 10.00 10.80 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.30 13.50 13.50
8.00 8.00 8.80 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.40 12.00 12.00
Masons Masons Masters Master summer winter wage wage Town Town in d groot in d groot Brabant Brabant
92.395 83.206 86.642 79.724 94.807 100.425 101.477 87.277 109.805 103.142 101.159 82.855 64.205 58.587 84.449 97.528
Masons: Real Wage Index 1451–1475 = 100 harmonic mean 10.013 9.018 9.390 8.640 10.275 10.884 10.998 9.459 11.900 11.178 10.963 8.979 6.958 6.349 9.152 10.570
Masons: Annual Wage in Commodity Baskets 210 days harmonic mean
9.513 8.567 8.947 8.280 9.847 10.430 10.539 9.065 11.404 10.712 10.506 8.605 6.668 6.099 8.898 10.276
Masons: Annual Wage in Commodity Baskets 210 days harmonic mean
10.00 10.00 10.40 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00 12.00 13.50 13.50
Carpenters Masters summer wage Town in d groot Brabant
7.00 7.60 8.00 8.00 10.00 9.40 9.00 9.00 9.00 9.00 9.20 10.00 10.00 10.00 12.00 12.00
9.262 8.476 8.572 7.920 9.847 10.301 10.310 8.868 11.156 10.480 10.323 8.605 6.668 5.966 8.898 10.276
Carpenters Carpenters Master Annual Wage winter in Commodity wage Baskets Town 210 days in d groot harmonic Brabant mean
1
Years (5)
40
Table 14. Wages of Building Craftsmen in Mechelen in d groot of Brabant: for masons, carpenters, street-pavers and laborers: employed by the town government and the Onse Lieve Vrouw Hospital with estimated annual money incomes expressed in terms of equivalent “baskets of consumables,” with price- and wage-relatives in quinquennial means, 1421–1425 to 1496–1500.
4259 Ch06 4/12/02 10:12 am Page 259
Wage Stickiness and Monetary Changes 259
259
15
16
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
38
39
40 5.00 5.60 6.40 6.80 7.60 7.20 6.00 7.60 8.00 8.00 8.00 8.00 8.00 8.00 8.20 8.00
12
7.677 6.913 7.549 7.202 9.430 9.485 9.271 8.362 11.404 10.712 10.506 8.605 6.668 5.966 7.796 8.673
11
8.00 8.00 9.00 10.60 11.60 11.00 10.60 11.20 12.00 12.00 12.00 12.00 12.00 12.00 12.00 11.60
9
7.510 7.548 7.623 6.840 8.990 9.523 9.623 8.276 10.413 9.781 9.593 7.773 5.944 5.706 7.780 9.787
4.798 4.746 5.376 4.554 6.115 6.178 5.270 5.638 7.603 7.142 7.004 5.737 4.446 3.977 5.291 5.879
8
6.00 6.80 7.20 8.00 9.00 9.00 9.00 9.00 9.00 9.00 9.00 8.60 8.00 8.00 9.40 11.00
7
9.00 9.00 9.40 10.60 11.40 10.20 10.00 11.20 12.00 12.00 12.00 12.00 12.00 12.00 12.00 11.20
Carpenters Master OLV Hospital summer wage in d gr Br
6
8.00 8.90 9.20 10.00 11.00 11.00 11.00 11.00 11.00 11.00 11.00 11.00 11.00 12.00 12.40 13.00
10
Masons Servants Annual Wage Commodity Baskets seasonally adjusted harmonic mean 8.637 7.778 7.883 7.202 9.172 8.783 8.783 8.362 11.404 10.712 10.506 8.605 6.668 5.966 7.796 8.363
Carpenters Annual Wage in Commodity Baskets seasonally adjusted harmonic mean
260
1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470 1471–1475 1476–1480 1481–1485 1486–1490 1491–1495 1496–1500
14
Masons Servants OLV Hospital summer wage in d gr Br
5
Masons Annual Wage in Commodity Baskets seasonally adjusted harmonic mean
4
Pavers Masons Anual Wage Masters in Commodity OLV Baskets Hospital Seasonally summer adjusted wage harmonic in d gr Br mean
3
Pavers Masters winter wage Town in d groot Brabant
2
Pavers Masters summer wage Town in d groot Brabant
1
Years (5)
18
Continued.
17
Table 14.
4259 Ch06 4/12/02 10:12 am Page 260
JOHN H. MUNRO
13
37
20
21
22
23
24
25
30
31
32
33
34
38
39
40
26
27
28
29
35
36
37
261
Seasonal = 157.5 days for the summer wage and 52.5 days for the winter wage. d = pence (d): 12d to the shilling and 240d to the pound groot. Sources: Verlinden and Scholliers (1959–1965, Vol. II, part ii, pp. 1244–1299).
10.430 10.539 7.940 9.504 9.223 10.506 8.605 6.668 5.966 7.796 9.004
16
12.00 12.00 10.60 10.00 10.40 12.00 12.00 12.00 12.00 12.00 12.00
15
6.717 6.369 6.738
14
7.00 7.40 8.00
19 3.00 3.70 4.40 5.30 6.80 7.30 6.40 5.40 5.60 5.40 6.00 6.00 6.00 6.00 6.00 6.00
2.879 3.161 3.649 3.657 5.444 6.348 5.577 4.061 5.247 4.724 5.253 4.303 3.334 2.983 3.898 4.502
Laborers Annual Wage in Commodity Baskets seasonally adjusted harmonic mean
6
1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470 1471–1475 1476–1480 1481–1485 1486–1490 1491–1495 1496–1500
13
Laborers OLV Hospital summer wage in d groot Brabant
5
Pavers Annual Wage in Commodity Baskets seasonally adjusted harmonic mean
4
Pavers Masters OLV Hospital summer wage in d groot Brabant
3
Years (5)
18
Continued.
17
Table 14.
4259 Ch06 4/12/02 10:12 am Page 261
Wage Stickiness and Monetary Changes 261
1
2
7
8
10 9
11
12
1
2
3
4
5
6
7
8
10 9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
151.011 120.370 135.616 141.680 148.741 155.989 177.022 143.350 138.904 127.434 148.845 112.030 121.900
10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 10.00 9.60 7.80 7.40 7.00
Value of Afsne Flemish Carpenters Basket of wage in Consumables in d groot in d groot (daily) Flemish
14.055 17.446 15.485 14.822 14.118 13.463 11.863 14.649 15.118 15.727 11.011 13.861 12.059
Afsne Carpenters wage in units of the Flemish commodity basket (210 days) harmonic mean 8.00 8.60 8.60 8.60 10.00 10.00 10.00 10.00 10.00 9.40 6.00 6.40 7.00
Afsne Thatchers wage in d groot (daily)
11.244 15.030 13.339 12.695 14.118 13.463 11.863 14.649 15.118 15.230 8.465 11.884 12.059
Afsne Thatchers wage in units of the Flemish commodity basket (210 days) harmonic mean 8.00 8.60 9.00 9.20 10.00 10.00 10.00 10.00 10.00 9.60 7.80 7.40 7.00
St. Denijs Thatchers wage in d groot (daily)
11.244 15.030 13.936 13.629 14.118 13.463 11.863 14.649 15.118 15.727 11.011 13.861 12.059
8.00 8.60 9.00 9.20 10.00 10.00 10.00 10.00 10.00 9.20 8.00 8.00 7.80
St. Denijs Destelbergen Thatchers Thatchers wage in wage units of the in d groot Flemish (daily) commodity basket (210 days) harmonic mean
11.244 15.030 13.936 13.629 14.118 13.463 11.863 14.649 15.118 14.980 11.287 14.996 13.415
Destelbergen Thatchers wage in units of the Flemish commodity basket (210 days) harmonic mean
262
1406–1410 1411–1415 1416–1420 1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470
Years (5)
Table 15. Daily Summer Wages for Craftsmen in the Small Towns and Villages of Eastern Flanders, 1406–1470: For carpenters and thatchers in Afsne, St. Denijs (Maalte), Destelbergen, Zaffelare, and Zevergem: in d groot Flemish and annual wage income in units of the Flemish “basket of consumables” in quinquennial means (arithmetic and harmonic).*
4259 Ch06 4/12/02 10:12 am Page 262
JOHN H. MUNRO
15
16
19
20
21
22
23
24
28
29
30
31
38
39
40 10.758 15.702 11.485 8.893 9.025 11.679 9.490 11.720 12.095 13.183 9.778 11.619 13.401
11.163 14.784 12.863 12.094 12.373 12.492 10.857 13.185 13.506 13.893 10.025 12.944 12.222
Small Towns: estimated means of annual wage in Flemish commodity baskets
Wage Stickiness and Monetary Changes
1
6
7
12
13
14
25
26
27
32
33
34
35
36
37
263
* The quinquennial means for the values of the Flemish “basket of commodities” and of the money wages are arithmetic; those for the quantities, or number of baskets that could have been purchased with the annual money wage (210 days), are in harmonic means. Sources: Verlinden and Scholliers (1959–1965: Vol. II.i, pp. 520–524 for Afsne; pp. 524–528 for Destelbergen; pp. 556–558 for St. Denijs/Maalte; pp. 568–572 for Zaffelare; pp. 574–577 for Zevergem.
7.50 9.00 7.50 6.00 6.40 8.80 8.00 8.00 8.00 8.00 7.00 6.20 7.80
11
8.433 10.468 8.998 8.893 8.738 9.424 8.200 8.790 8.468 8.511 8.596 11.444 10.336
9
6.00 6.00 5.80 6.00 6.20 7.00 6.90 6.00 5.60 5.20 6.10 6.10 6.00
10
1406–1410 1411–1415 1416–1420 1421–1425 1426–1430 1431–1435 1436–1440 1441–1445 1446–1450 1451–1455 1456–1460 1461–1465 1466–1470
8
Zevergem Thatchers wage in units of the Flemish commodity basket (210 days) harmonic mean
5
Zevergem Thatchers wage in d groot (daily)
4
Zaffelare Thatchers wage in units of the Flemish commodity basket (210 days) harmonic mean
3
Zaffelare Thatchers wage in d groot (daily)
2
Years (5)
18
Continued.
17
Table 15.
4259 Ch06 4/12/02 10:12 am Page 263
263
4259 Ch06 4/12/02 10:12 am Page 264
264
JOHN H. MUNRO
1 2 3 4 5
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Sources: See Tables 3, 4, 5 and 8.
8
Fig. 1.
7
Mint Outputs & Prices (Logs), 1281–1500: England, Flanders, Low Countries.
6
4259 Ch06 4/12/02 10:12 am Page 265
Wage Stickiness and Monetary Changes
265
1 2 3 4 5 6 7
English Prices, 1281–1500. Phelps Brown & Hopkins Composite Index.
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Fig. 2.
31 32 33 34
Source: See Table 5.
35 36 37 38 39 40
265
4259 Ch06 4/12/02 10:12 am Page 266
266
JOHN H. MUNRO
1 2 3 4 5 6
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 35 36 37 38 39 40
Source: See Table 8.
34
Note: * 1349–1350 only.
9 10
Fig. 3.
8
Flemish Prices, 1351–1500. Composite Price Index: 1415–1475 = 100.
7
4259 Ch06 4/12/02 10:12 am Page 267
Wage Stickiness and Monetary Changes
267
1 2 3 4 5 6
English Masons’ Wages, 1266–1500. CPI, Nominal and Real Wage Indexes.
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Fig. 4.
31 32 33
Source: See Table 6.
34 35 36 37 38 39 40
267
4259 Ch06 4/12/02 10:12 am Page 268
268
JOHN H. MUNRO
1 2 3 4 5
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Sources: See Tables 8 and 10.
8
Fig. 5.
7
Bruges Masons’ Wages, 1351–1500. CPI, Nominal and Real Wage Indexes.
6
4259 Ch06 4/12/02 10:12 am Page 269
Wage Stickiness and Monetary Changes
269
1 2 3 4 5 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Fig. 6.
32 33 34 35 36 37 38 39 40
269
Sources: See Tables 9 and 13.
Antwerp Masons’ Wages, 1401–1500. CPI, Nominal and Real Wage Indexes.
7
4259 Ch06 4/12/02 10:12 am Page 270
270 1
JOHN H. MUNRO
NOTES
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
1. Wages for master building craftsmen were 22d to 24d per day from 1730–1736; 24d from 1736–1773; and rising to 29d per day from 1773 to 1776. I have also revised the values of the sub-indexes and of the composite price index from tabulating the data from their working sheets, in: Archives of the British Library of Political and Economic Science [hereafter BLPES], the Phelps Brown Papers Collection. 2. See Rogers (1866, Vol. II, p. 317; and 1882, Vol. III, p. 628; Vol. IV, p. 520). In this transition period, wages were unstable: some masons and carpenters were paid 6d, 6.5d, 6.67d, 7.0d, and a few even 8d; but the mean appears to be 6.5d in this and the following years. With a base of 1451–1475 = 100, the Phelps Brown and Hopkins “basket of consumables” index stood at 96.90 in 1510, and about 110.0 in 1515–1517; but by 1536 this had risen to 164.25. The summer daily wage for master masons and carpenters rose to 7d in 1542, and thereafter by 1/2d increments to 10d in 1559, remaining at that rate until 1574. 3. Data from Stadsarchief Brugge, Stadsrekeningen 1331–1332 to 1475–1476; and Algemeen Rijksarchief, Rekenkamer, doc. nos. 32,461–32,566 (1406–1513). The policemen were paid this daily rate for the full 365 (or 366) days in the year, while the average master mason or carpenter worked at most about 210 days a year. For wages in Bruges, see also Sosson (1977, pp. 225–260, and graphs, pp. 300–309). 4. Wage data for Brabant taken from Verlinden and Scholliers (1959–1965, Vols. I–II); Van der Wee (1963, Vol. I, Appendices 27–30, pp. 333–389; Van der Wee, 1975, pp. 413–447). 5. See also Steffen (Stuttgart, 1901–1905). 6. See Van Zanden (1999, p. 192), commenting “. . . that there was no clear cut relationship between economic development and real wage growth” (p. 192); and that his research on real-wages “throws doubt on some of the conclusions of the optimists,” referring in particular to De Vries (1994). 7. See Van der Wee (1975, pp. 413–447). His data show that Antwerp craftsmen did not suffer the same deterioration in real wages as did English craftsmen during the Price Revolution era (c.1520–c.1640), even with substantial demographic growth and inflation in both countries. 8. On the English “Assize of Bread and Ale,” see Bland, Brown and Tawney (1914, pp. 155–157); Ross (1956–1957); and Jones (2000, pp. 136–138). The earliest assize dates from the twelfth century (reign of Henry I) and became the statute Assisa Panis et Cervisie evidently in 1266 (51 Henrici III): in Tomlins and Raithby, 1810–1822, Vol. I, pp. 199–200; and it was not abolished until 1819. 9. For the Savernak household, in Bridgport Dorset (1450s), for which foodstuffs accounted for 80% of total expenditures, see Phelps Brown and Hopkins (1956/1981, p. 14) and Wood-Legh (1956). See also Dyer (1983, 1988, 1989, Chapters 5–6) for evidence on consumption patterns in late-medieval England, especially on the rising consumption of meat and dairy products. For medieval and early-modern Flanders, Blockmans and Prevenier (1977, p. 22) endorse weights of 70–80% for foodstuffs, 5–15% for house rents, 5–10% for lighting and heating, and 5–10% for clothing. For the 70% for foodstuffs: 44% for bread grains, 15% for meat and fish, and 11% for vegetables. For Brabant, see Van der Wee (1975/1978 and 1966/1993, especially p. 284). The budget for the Lier Infirmary, from 1526 to 1602, was allocated on average: 44% for bread,
4259 Ch06 4/12/02 10:12 am Page 271
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
271
16% for beer, 1% for wine, 3% for fish, 20% for meat, and 10% for dairy products. The original Dutch-language version of the article (1975, pp. 417, 436) indicated the weights for dairy products as 4.7 kg for each of butter and cheese; but the table provides a weight of just 4.5 kg. for each. In the English translation (1978, Table II, p. 62) the weights are given instead as 4.8 kg for butter and 4.7 kg for cheese; and this version was also presented in the subsequent republication of his essays (1993, p. 225). Nevertheless these weights evidently have no bearing on the statistical values (i.e. in d groot Brabant) provided for this sub-index. 10. A gap from 1430 to 1573, as ascertained from their working papers contained in the Archives, BLPES, the Phelps Brown Papers Collection, Box Ia.324; and J.IV.2a. 11. Van der Wee’s reliance on just rye prices for the farinacous group (while using barley for the drink group) may be justified on the medieval Low Countries’ preference for rye over wheat (see n. 8 above). The omission of prices for lamp oil and shirting in the industrial sub-index is hardly serious. A far wider range of industrial product prices is given in Doughty (1975); but it commences only in 1401. 12. R. Allen (2001) has constructed a somewhat different basket, which may be more useful for modern Europe, but not for medieval north-west Europe. It is also based on commodity values and wages expressed in terms of grams of pure silver, a method that permits international comparisons of commodity prices and wages, but one that can be misleading, since it is based on the unrealistic assumption that commodity price changes are proportional to changes in the silver contents of the coinages (with debasement) or in market values for silver. He also compared standard Laspeyres arithmetic indexes with geometric indexes, using the same data, and found no significant differences in trends. As he notes, “the geometric index is a weighted geometric average of the price relatives in which the weights equal budget shares;” but Phelps Brown and Hopkins (1956), in basing their index on budget shares, did not use the required technique to produce a true geometric index: i.e. by multiplying the price-relatives (P1/P0) and taking the nth root of the product, in which n represents the number of years in the series; or, more simply, by taking the sum of the natural logarithms of the individual price-relatives in a series and dividing that sum by the number of years. The obvious implication of an index based on fixed expenditure shares is that the quantity of grains consumed during a period of rising relative grain prices must, in theory, fall to a greater extent than the quantity of industrial goods, an historically unrealistic assumption. 13. For a more recent exposition, see Hatcher (1977, pp. 11–73). 14. For a general overview, see Hatcher (1977, 1986, 1994); Blanchard (1970), Campbell (1981); Miller and Hatcher (1978, pp. 27–63). For an estimate of 4.5 million in the 1290s, see Nightingale (1996, pp. 89–106); for the estimate of 7.2 million in 1292, see Hallam (1988, p. 536); for an estimate of 2.5–3.0 million in 1377, see Miller (1991, p. 6). 15. The following quotations also succinctly express his fundamental views: (1) “And price changes which are not ‘general’ but are mainly confined to grain, point to a factor which has already been shown to have operated in the opposite direction in the early centuries of the Middle Ages, i.e. population” (Postan, 1952, p. 213); (2) “A fall in population would also have, so to speak, a selective effect on prices, in that it would tend to lower the prices of agricultural products, which were previously being produced at high and ever rising cost . . . under steeply diminishing returns. . . but would have little effect on commodities not greatly subject to diminishing returns, i.e. most industrial products.” (Postan, 1952, p. 214); (3) “. . . the movements of agricultural and 271
4259 Ch06 4/12/02 10:12 am Page 272
272 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
industrial prices did not synchronize” (Postan, 1951/1973, p. 9); (4) “Butter happened to be a semi-luxury entering into popular consumption. It is not necessary to know what [Alfred] Marshall said about the elasticity of the demand for bread in order to conclude that agricultural labourers were now better able to indulge in a little butter, however expensive . . . It is therefore highly significant that the price of butter and the price of grain diverged more widely than the prices of any other commodities [during] the fourteenth and fifteenth centuries. [And furthermore] the prices for animal products in current coinage actually rose [continuing from 1351] the rising trend of the previous two centuries for another 125 years.” (Postan, 1952, pp. 209–210). 16. See Schwartz (1974, p. 253), for a review of Spooner (1972), in which she comments that: “the author subscribes to a familiar fallacy, namely that a monetary explanation to be valid requires that all prices move in unison.” Her verdict would be equally valid if directed against the injudicious criticisms that Postan (1959, pp. 77–82) directed against Robinson (1959). 17. See below, pp. 211–213 On the abandonment of land in so much of medieval western Europe, the so-called “Wüstungen,” see in particular Abel (1966/1980, pp. 49–79, and especially pp. 80–95) and Slicher van Bath (1963, pp. 160–94). 18. As noted above, for the pre-modern era, population growth, by increasing pressures on relatively fixed stocks of land and capital, would – in the absence of technological changes – likely have induced diminishing returns, producing rising marginal costs and thus rising grain prices. But such sharp rises in grain prices might well have forced many consumers, faced with budget constraints – as argued above in the context of monetary changes – to reduce their expenditures on other, less necessary commodities, thus driving down their relative prices. For a similar argument, see McCloskey (1972), in reviewing Ramsey’s collection of essays on the Price Revolution (1971). 19. Keynes (1936, p. 300): “It is probable that the general level of prices will not rise very much as output increases, so long as there are available efficient unemployed resources of every type. But as soon as output has increased sufficiently to begin to reach the “bottle-necks,” there is likely to be sharp rise in the prices of certain commodities . . . [But] the elasticity of supply partly depends upon the elapse of time.” For further arguments on these issues, see Munro (1991, pp. 119–183); Munro (1999), in a review of Fischer (1996); and Munro (2003b). 20. See in particular Farmer (1983, pp. 117–155; 1988, pp. 760–778, 811–817, with Table F; and 1991, pp. 467–490, 516–524, with Tables H and I); Clark (2001). 21. The data in Table 3 are for “hired’ and not workers ad tascam; and without food or other payments in kind. I have not, however, found comparable examples in other manorial accounts. Beveridge (1936, p. 30), commented that, at Taunton, wages “are affected by the greater importance there of customary services and the best method of presenting the [wage] series has not been determined.” But that does not satisfactorily explain why building wages at Taunton were generally higher before the Black Death than elsewhere. Yet if Beveridge is correct, then wages on many late-medieval English manors would have been – to a greater or lesser degree – affected by the relative supplies of servile labor, before and after the Black Death. 22. Farmer (1996, pp. 214–220, and Table 11.4). On the selected Winchester manors, labor productivity on the arable fell from 34.3 acres per famuli ploughman in 1305 only marginally to 32.3 acres in 1382, but then more precipitously to 27.9 acres per ploughman in 1421, a 15.8% decline. In animal husbandry, the number of sheep under the care of
4259 Ch06 4/12/02 10:12 am Page 273
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
273
a single shepherd (again chiefly famuli) rose from 231 in 1341 to 342 in 1421, a 48.1% increase. For some corroborative evidence on Glastonbury and Ramsey manors, see also Raftis (1996, pp. 191–206). 23. In late-medieval Flanders, weaving a standard fine woolen broadcloth of 42 ells by 3.5 ells (29.4m by 2.45m = 72.0m2), containing 84 lb. or 38.2 kg of wool (16.3 kg of warp and 21.8 kg of weft), typically required about 12–14 days, with two weavers and a boy. Another dozen days of labor were expended in wool-beating, wool-greasing, carding, combing, spinning, reeling, and warping the yarns for the same cloth, involving about 26–30 artisans and helpers; and at least another 6–9 days in the finishing processes of foot-fulling (three or four days per broadcloth), napping, shearing, and dyeing. According to a Parliamentary report of the 1790s, weaving a superfine broadcloth of 34 yards (i.e. before fulling), with 80 lb. of wool (36.2 kg), then required 364 man-hours (= 14.5 days, with two weavers and a boy), and a further 888 man-hours in wool preparation, spinning, reeling, and warping, and fulling (74 days). Fulling had been mechanized in England (water-wheels) from the thirteenth century, and remained the only significantly powered manufacturing process before the Industrial Revolution. See Endrei (1981, pp. 253–262; 1983, pp. 108–119); Van Uytven (1981, pp. 283–294); Munro (1988, pp. 693–715); Lipson (1921, Appendix I, pp. 258–259, citing Parliamentary Papers, Vol. 23, 1840, pp. 439–441, for the years 1781–1796). 24. Such a process would have been facilitated by the increased labor mobility that ensued from the later fourteenth century, with the breakdown of the manorial domain economies, the consequent leasing of domain lands to peasant tenants, and the decay of villeinage (serfdom). See Raftis (1964, pp. 129–204; 1996, pp. 291–206). 25. The Ordinance of Labourers, decreed on 14 June 1349, is restated and reissued, with some modifications, as the Statute of Labourers, 25 Edwardi III stat 2. c. 3 (1351). See also the so-called 1388 Statute of Cambridge (Statute 12 Ricardi II, c.3-c.5), which stipulated the annual stipends for agricultural servants and laborers, ranging from the Bailiff of Husbandry, at 13s 4d sterling per year, with clothing, to 6s 0d (i.e. 72d) for “Swineherds, Women Laborers, and Deyrie Women.” Both in Tomlins and Raithby (1810–1822, Vol. I, pp. 307–308; Vol. II, pp. 56–57). For other legislation, see also, in Vol. I, pp. 311 (1351), 327 (1352), 350 (1357), 366 (1361), 373–375 (1362), 388 (1368); in Vol. II, pp. 11 (1378), 32 (1383), 63 (1390), 137 (1402), 157–158 (1406), 176 (1414), 196 (1416), 225 (1423), 227 (1425), 233–235 (1427), 244 (1429), 337 (1446), 585 (1495), 637 (1497). On such legislation, see in particular Given-Wilson (2000, pp. 85–89). 26. Some manorial records, however, periodically record mowers’ and threshers’ wages both by day and by output, thus permitting some such comparisons; and other records provide wages of sawyers and street-pavers by both quantity (piece-work) and by day. For one exercise in using such medieval data, see Clark (2001). 27. See Statute 11 Henry VII c. 22 (1495), which authorized the payment of a higher wage to those master masons and carpenters who supervised six or more men; the statute was reissued in 1514–1515 as 6 Henry VIII c. 3. Both in Tomlins and Raithby (1810–1822, Vol. II, pp. 585–587; Vol. III, pp. 124–126). 28. In the list of references, see the English archives consulted for this study, especially the Archives, BLPES, Beveridge Price and Wage History Collection; The London Guildhall Manuscripts Library; and The Corporation of London Record Office. 29. See the previous note. Beveridge (1936, p. 34) notes that in the early thirteenth century Southwark wages had been about the same as those paid elsewhere on the Winchester estates. 273
4259 Ch06 4/12/02 10:12 am Page 274
274 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
30. At 4d or 5d per day, at Esher, in Surrey; Witney, in Oxfordshire; Wycombe, in Buckinghamshire; Overton and Ecchinswell (Itchingswell), in Hampshire; Taunton, in Somerset. For a list of the manors and their locations, see Titow (1972, pp. 28–39, with map). 31. See Hatcher (1994, pp. 13–19); Penn and Dyer (1990, pp. 356–376); Ritchie (1934, pp. 91–112; and n. 3). 32. In the appended list of References, see the archival sources for Belgium, chiefly town accounts; and also the published wage data in Verlinden-Scholliers (1959–1965), and Van der Wee (1963, Vol. I, Appendices 27–30, pp. 333–392; and Synoptic Tables of Wages, pp. 457–475). 33. Fullers, organized into guilds, also received piece-work wages, but since they were required to full a cloth in three to five days, according to the size and quality, a daily-wage rate can be extrapolated. Master weavers, who were actually industrial entrepreneurs, earned their incomes as profits, while dyers and shearers, also guild-protected, earned professional fees. See Munro (1988b, pp. 693–715; 1994a, pp. 377–388; 2002, pp. 153–206). For the fullers’ strikes, see below, pp. 219–222. 34. See in particular Hatcher (1994); and below, pp. 211–212. 35. See nn. 20, 22 above. For the Statute of Artificers, Statute Elizabeth I c.4 see Tomlins and Raithby (1810–1822, Vol. IV: Part i, pp. 414–422) and Bland, Brown and Tawney (1914, pp. 325–336). 36. Discussed in much greater detail below, on pp. 210–212, 217–219 Hatcher himself admits (1994, p. 23) that “the vast majority of the greater landlords openly paid cash wages in excess of the unrealistic maxima specified by the law.” 37. Riley (1860, Vol. I, pp. 99–100; Vol. II, pp. 541–543; 1868, pp. 253–255); Sharpe (1905, pp. 148, 301; 1907, p. 184). By 1349 the higher rate was already in force at Westminster Abbey (BLPES, Beveridge Price and Wage History Collection: Westminster Abbey). 38. Statute 11 Henry VII c. 22 (1495), in Tomlins and Raithby (1810–1822, Vol. II, pp. 585–587). The major exception was a maximum daily rate of 7d, summer and winter, for those senior or chief master masons and carpenters employing or supervising six or more men; or a rate of 5d daily, with food and drink. See n. 27 above. 39. See pp. 197–198 and nn. 28–29 above. 40. Van der Wee (1963, Vol. I, Appendix 48, pp. 540–544): interpolating 208 days for 1436). For this regression, R2 = 0.00002943; adjusted R2 = –0.01017; F = 0.002885. 41. Statute 11 Henry VII c. 22, repeated in 6 Henry VIII c. 3 (1514–1515), in Tomlins and Raithby (1810–1822, Vol. II, pp. 585–587; Vol. III, pp. 124–126). The framers of the 1495 statute, however, evidently believed that too many wage-earners were stealing leisure time during working hours by “late commyng unto their worke, erly departing therfro, long sitting at ther brekfast, dyner, and nonemete, and long type of sleping at after none.” 42. See Campbell (1747/1969, pp. 331–341). In specifying hours of work for 380 crafts, this treatise indicated that the typical working day still remained a very long one: from 6:00 a.m. to 8:00 p.m. (or during all “daylight” hours). Subsequently, the 1833 Factory Act had stipulated a maximum working day of nine hours for children, aged nine to thirteen; and for those aged thirteen to eighteen, a maximum of 69 hours a week, with no more than 12 hours per day. The 1844 Factory Act limited the working day for women to 12 hours per day, and for children aged eight to thirteen, to 6.5 hours. Fielden’s Act, imposing a limit of ten hours per day for both women and men (implicitly), was
4259 Ch06 4/12/02 10:12 am Page 275
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
275
passed at a time of great labor unrest, on the eve of the 1848 Chartists’s Revolt. See Clapham (1964, pp. 572–578); Checkland (1964, pp. 244–249); Derry (1963, pp. 124–137, 151–154). In the U.S., New Hampshire’s state legislature also enacted, in the same year 1847, a maximum ten-hour day; but the legislation applied only to women, and proved to be unenforceable. The effective state legislation dates only from 1887. See Goldin (1990, pp. 189–192, and Table 7.1); Atack and Passell (1994, pp. 542–544). 43. See Sosson (1977, p. 149, n. 41), citing guild statutes in Rijksarchief Brugge, Ambachten, no. 1, fo. 62: “dat hi sculdich es te werkene van nuchtens toten avonde alzo wel tsaterdaechs up vighelie avond als up anderen daghen . . .” Similarly the Bruges carpenters guild forbade them “te weerkene . . . by avonde of bi nachte met keersen [candles].” See other guild records that specify working hours in the textile trades (forbidding work by night), in Delepierre and Willems (1842); Espinas and Pirenne (1906–1924); Joossen (1935). 44. See n. 37 above: for the London ordinance of 1350, setting maximum summer and winter daily rates (at 6d and 5d respectively). 45. See Table 1. The subsequent Statutes of 1444 and 1495 did more clearly specify a winter maximum wage, lower than the stipulated summer wage. In London, and possibly also in Bruges, seasonality reappeared, after the 1440s in the form of a slightly higher summer wage. See sources in nn. 27, 37–38, 41, and 46 above. 46. In 1441: by reducing the uniform daily rate to 7.5d (or even to 7.0d for some) for the winter season (three months), while raising it to 8.5d for the rest of the year. (Corporation of London Record Office, London Bridgemaster’s Accounts, Weekly Payments, First Series, Vol. IV. Unfortunately these accounts cease in 1445). See also Knoop and Jones (1967, pp. 105–106). 47. See below, p. 224. For accounts of wage payments in Bruges, see: Stadsarchief van Brugge, Stadsrekeningen 1424–1425 to 1467–1468. Before the 1440s, there are only occasional and sporadic indications of seasonal wages. On the other hand, some wage differences were clearly not seasonal in nature: thus some master masons and carpenters were paid 12d per day in December 1435, February 1436, December 1437, while some were paid 10d daily in June and September 1437, May and September 1438 (and thereafter to April 1444), in May and September 1446. Subsequent accounts, however, are explicit that winter wages were “paid before Easter” (in particular, the accounts of 1458–1459 and 1460–1461, specifying a winter wage of 10d rate for street-pavers). For seasonal wages Bruges from this period, or ca. 1450, see Sosson (1977, pp. 225–228, and 300–304, Figs 12–16). For Holland, see in particular De Vries (1978, pp. 79–97; and 1994, pp. 47–63). For England, see Thorold Rogers (1866, Vols I–II; and 1882, Vols III–IV); Phelps Brown and Hopkins (1955/1981, pp. 1–12). 48. See pp. 196–197 above; and the other archival sources listed in References. On this question, see in particular Beveridge (1936, pp. 36–37): “that in the Winchester manors there is no reason for suspecting any general practice of supplementation,” not disclosed in the records, which record separately payments in cash (stipendium) and in kind (corredia, allocationes). He also notes that “vadia (originally cash payment in place of allowance) is used as a term interchangeable with stipendium (cash payment exclusive of allowance) and is invariably at or about the same level . . ..” 49. Archives, BLPES, Beveridge Price and Wage Collection, Battle Abbey: Alciston Manor, 1336–1487 (Boxes H.10–11). 50. Archives, BLPES, Beveridge Price and Wage Collection, Westminster Abbey: 1393–1541 (Box P.10). 275
4259 Ch06 4/12/02 10:12 am Page 276
276 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
51. See the archival sources for Belgium in the list of References; and also Van der Wee (1963, Vol. I, Appendices 27–30, pp. 333–392; and Synoptic Tables of Wages, pp. 457–475); Verlinden and Scholliers (1959–1965). 52. See, for example, in Posthumus (1910, Vol. I, pp. 90–91, 161, 621–622, 653–666) documents nos. 74. 111–113 (March 1415); no. 132.VI.9 (1436/1437); no. 508: 10–13 (Nov. 1478); no. 525: 23–225 (Dec. 1478). Inter alia, these ordinances stipulated that: “ende die meester sal gehouden wesen sijn volle gelt te nemen van elc laken . . . dat die drapenieres hoir volres hoir volle loen geven sullen, ende alle Saterdages hoir gelt geven sullen, sonder horen volres enigerhande dinck te vercopen, te weten kairden, groff wolle laken, bier, broot, sout, gort, kairssen, seep, butter noch kase” (no. 508:13, p. 623); and, “die betalingen sel wesen mit ghelde ende mit gheerehande wair of andere goeden . . .” (no. 525:24, p. 653). 53. See the archival sources for Belgium in the list of References. On this issue, for early modern England, see also Woodward (1981, pp. 28–46; 1994, pp. 11–21; and 1995). 54. For both quotations, see Wilton and Prescott (1987, p. 214). 55. See Huberman (1986, 1991 and 1996). 56. See also Wilton and Prescott (1987, p. 215), in commenting on “the very deep recession of 1982,” when the International Woodworkers “resisted wage cuts at sawmills in British Columbia. The union took the view that if the weakest firms succeeded in getting wage concessions in order to maintain employment, then all firms would demand the same concessions and all workers would suffer. Moreover, the union would have to fight battles already won in order to restore wage and benefit levels once market conditions improved;” and thus “wage cuts were resisted and many sawmills were closed.” 57. See Munro (2002); Boone, Brand and Prevenier (1993); Boone and Brand (1993); Brand and Stabel (1995). 58. See in particular Postan (1950; 1951; 1972, pp. 31–35); and especially Postan and Titow (1959/1973, pp. 150–185, which presents evidence to indicate rising death rates from heriots, or “death-duties” levied on customary peasant holdings (usually in the form of livestock). For other evidence on rising demographic pressures, especially fragmentation of holdings, though not conclusive evidence for a Malthusian crisis, see Miller and Hatcher (1978, pp. 53–63); Campbell (1984), Smith (1984, 1991, pp. 25–77); Bailey (1998); Poos (1985; 1991, pp. 9–57, 89–130); and for peasant difficulties in paying increased royal taxes, see Maddicott (1975, pp. 1–75). For the “Great Famine” itself, see Lucas (1930); Kershaw (1973), and especially Jordan (1996, pp. 24–150). 59. Postan (1951, p. 14; and also Postan, 1950, pp. 186–214; 1972, pp. 27–39). 60. Poos (1985; 1991, pp. 89–130). 61. Herlihy (1967, pp. 55–77); Herlihy and Klapisch-Zuber (1985, pp. 232–279); Munro (1991, pp. 139–148); Harvey (1991, pp. 1–24). 62. Smith (1991, pp. 1–23); see also Harvey (1966, pp. 23–42), with a much stronger attack on Postan’s thesis, denying any evidence of population decline before the Black Death; and also Hallam (1988, pp. 508–593), casting further doubts on general population decline before 1348. 63. Herlihy (1967, pp. 122–125). 64. That is: if the aggregate decline in West European net national income was greater than any change in the volume of money payments: if y > (M.V) ⇒ P ↓. See Robinson (1959; and Postan’s “Note,” pp. 77–82, which is not a valid response);
4259 Ch06 4/12/02 10:12 am Page 277
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
277
Spufford (1988, pp. 267–282) ; Nef (1952, 1987); Kovacevic (1960, pp. 248–258); Braunstein (1983, pp. 573–591); Westermann (1986, pp. 187–211); Munro (1991, 2003b). See also pp. 214–215 below. 65. See also Patterson (1972); Munro (1983, pp. 97–126). 66. Ames (1965), Prestwich (1977), Mate (1975), Robinson (1959), Spufford (1988, pp. 267–288). 67. The fall in the bimetallic ratio may have been due to both declining silver outputs from Bohemian mines and increasing supplies of West African (Sudanese) and Hungarian gold. See Lane (1977, pp. 52–59); Spufford (1986, graph 3 and Table II, pp. li–lxiii; and 1988, pp. 267–288, 340–342). Spufford’s dates have been adjusted by those of Lane (1977). See also Mate (1978). 68. An absence of minting may indicate only that the nation’s mints were offering a mint price for bullion uncompetitive with those of neighbouring foreign mints. With coinage debasement in a bimetallic system (in England, after 1344), a coinage debasement in, say, silver might succeed in recoining much of the current silver monetary stock, in inducing dishoarding, and in attracting foreign bullion, but at the expense of losing the now “disfavoured’ gold to foreign mints. See the essays in Munro (1994). See Fig. 1; and Munro (1984, Table A-1, p. 86): regressing quinquennial mean current prices against corresponding mint output values (current moneys) in these two regions, see for 1350–1409: R2 = 0.4327 (t statistic: 2.762, significant at the 2.00% level); for 1350–1499: R2 = 0.2697 (t-statistic: 1.455, significant only at the 15.71% level). 69. Mayhew (1987, Table I, p. 125): indicating that the coined money supply contracted from about £1,100,00 sterling in 1311–1324 to just £500,000 in the 1340s. See Mayhew (1995, 1974); and sources in nn. 67–68 above. For a more recent estimate, see: M. Allen (2000; 2001), indicating (Table 1 in M. Allen, 2001, p. 603) a larger estimated coined silver stock of £1,900,000 – £2,300,000 in 1319, falling to about £700,000 – £900,000 in 1351. See also Table 2, p. 607, providing, for 1470, an estimate of just £350,000 to £450,000 in silver, £400,000 to £500,000 in gold, and thus a total of no more than £750,000 to £950,000. 70. Phelps Brown and Hopkins (1955, p. 11); Thorold Rogers (1866, 1882); and the list of References for the Beveridge Price and Wage History collection (see nn. 48–50 above). Wages for masons and carpenters at the Oxford colleges did not fall, however; but remained at the daily rate of 4d set from at least 1300. 71. Money wages for master masons and carpenters were typically 3d per day from before 1264 to 1302; for many, 3.5d until 1310; and then 4d until 1338. See nn. 2, 20–21, 28–29, and n. 78 below. 72. For other analyses of medieval English wages before and after the Black Death, see in particular Farmer (1981, 1983, 1988, 1991), whose conclusion do not always match my own. For the more general economic and social consequences of the Black Death, see: Hatcher (1977, 1986, 1994); Platt (1996, pp. 1–47, 177–192); Bolton (1996, pp. 17–78); Putnam (1908); Poos (1991, pp. 218–221). 73. Statute 25 Edwardi III stat. 2 c. 3, in Tomlins and Raithby (1810–1822, Vol. I, pp. 311–312). See also n. 25 and pp. 196, 199–200 above. Wages for the winter season from Michaelmas to Easter were not specifically stipulated, except that they were to be “less according to the rate and discretion of the justices.” For a discussion of this Statute, and the subsequent labor legislation, see Farmer (1991, pp. 483–490); Given-Wilson (2000, pp. 85–90); Putnam (1908). 277
4259 Ch06 4/12/02 10:12 am Page 278
278 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
74. Statute 23 Henrici VI c. 12, in Tomlins and Raithby (1810–1822, Vol. II, pp. 337–339). See above p. 210. 75. See above p. 200. As also noted there, these rates were just 1d per day higher than those of 1290. 76. Archives, BLPES, Beveridge Price History Collection, Box P9. 77. See Beveridge (1955–1956, pp. 26–28): commenting on “the failure of the Black Death to cause any immediate change of [wage] rates” in the Winchester manorial accounts of 1349–1350. 78. For this and the following, see: Archives, BLPES, Beveridge Price and Wage History Collection, Box H:10–11). At Battle Abbey, carpenters and masons had their pay raised from 4d to 5d daily in 1425; at Ecchinswell (Itchingswell) (Box A.33:159432), from 1433; but at Overton (Box A.33:159406), and Wycombe (BoxA.33:159407) first carpenters and then masons enjoyed the same increase from 4d to 5d as early as 1401–1405. At the urban Winchester College (Box F.8), wages for master carpenters had risen to 6d per day without food by 1398 (4d daily with food); but for master masons, the mean rate did not reach 6d daily (without food) until 1409. See also the raw-wage data published in Thorold Rogers (1866, Vol. II, pp. 272–334; and 1882, Vol. III, pp. 583–663). 79. As also noted above, on p. 194, the wages of occasional day laborers, chiefly agricultural, on Taunton manor, after doubling in the years following the Black Death (1349–1356), then fell to the pre-Plague level of 1d daily, remaining there until 1412–1413. See Table 7. 80. Ritchie (1934): data from presentments before Justices of the King’s Bench at Brentwood in November 1389, following the 1388 Statute of Cambridge. As she also notes (p. 102), “eight hundred men who were receiving illegally high wages is not a large number for a country the size of Essex.” She also noted daily wage payments of 4d with food, for tasks in the Winchester manorial accounts that are specifically listed as “without food’. See also Poos (1991, pp. 218–221); Penn and Dyer (1990). 81. Corporation of London Record Office, Bridgemaster’s Accounts, Weekly Payment Series, Vols III–IV; Corporation of London Record Office, London Bridgemaster’s Accounts, Weekly Payments, First Series, Vol. IV. Subsequently, from 1441, the craftsmen employed by the Tower Bridgemaster received a very minor seasonal adjustment in wages: a reduction in the daily rate for a now shorter winter season, to 7.5d but an increase to 8.5d for the rest of the year, thus slightly increasing the annual wage payments. Unfortunately these accounts cease in 1445. See also Knoop and Jones (1967, pp. 105–106). 82. For Flanders, see Munro (1981, 1983, 1984a, b); for Tuscany, see Herlihy (1967, pp. 122–125). 83. For this study, for both England and the Low Countries, the tables on real-wages present data in quinquennial means, which are calculated as both arithmetic and harmonic means. For both the annual data, within each quinquennium, are computed in the standard fashion as: RWI = NWI/CPI (nominal wage index divided by the consumer price index for each year). The arithmetic mean is calculated by computing the sum of the five annual data in each quinquennium, and not by dividing the quinquennial mean of the NWI by that for the CPI (which would produce an entirely false result). The harmonic means is the reciprocal of the arithmetic mean of the reciprocals of the individual values (index numbers of commodity baskets) in each quinquennium (or series); and it is the better measure of the average number of units that could be purchased with a given sum
4259 Ch06 4/12/02 10:12 am Page 279
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
279
of money. Its formula is: H = 1/[3 (1/r1 + 1/r2 + 1/r3 + . . . 1/rn)]/N. Note that the harmonic mean is always less than the arithmetic mean (and also less than the geometric mean). See Mills (1956, pp. 108–109 and 395–324, for index numbers using these means; and also n.12 above). See above, however, on pp. 189–190 some reasons why these indexes probably exaggerate the rise in real wages during periods of falling grain prices and their fall during periods of rising grain prices, particularly though with sticky wages. 84. Archives, BLPES, Beveridge Price and Wage History Collection, boxes A.31 (Taunton), A.32 (Esher), A. 33 (Ecchinswell/Itchingswell, Overton, Wycombe), F.8 (Winchester College), G.14 (Hinderclay and Redgrave), H.10–11 (Battle Abbey). 85. See Feavearyear (1963, pp. 15–45); Munro (1979, 1981, 1983a, 1983b, 1984a, 1984b, 1988a, 1992); Fournial (1970); and Cazelles (1966). 86. See in particular: Miskimin (1975, pp. 25–32); Herlihy (1967, pp. 55–71, 180–212); Lopez (1962, pp. 29–52); Munro (1983a, pp. pp. 115–120); Boccaccio (1353/1921: introduction, esp. p. 7); Cassell (1983, pp. 277–290). 87. Herlihy (1967, pp. 128–130); Hamilton (1936, statistical appendices). 88. Unfortunately, the Flemish “basket of consumables” index contains only one industrial commodity, in this period (Table 1): cheap strijptelaken (rayed cloth) from Ghent (supplemented with cheap voeringlaken from 1401). Graphs of some prices for building materials during this period are contained in Sosson (1977, pp. 289–293): no. 1, for paving stones: very few data, but with a sharp decline in the later 1390s; no. 2, for bricks: decline in late 1380s, rise in early 1390s, then decline (to 1405); no. 3: for slates, decline in early 1390s, then rise, then decline, from 1400 to c. 1410; no. 4, for lime: sharp decline in early 1390s, then stable; no. 5, for solder: stable; and iron: severe fluctuations. For a similar graph, though only for solder and iron prices, see also Thoen (1988, Vol. I, Fig. 16, p. 255). 89. The sources of the mint data used in these tables, for England and the Low Countries, may be found in: Munro (1973, Appendix I, pp. 188–211; 1981, pp. 71–116; 1983, pp. 127–158; 1984, pp. 71, 86–111); Crump and Johnson (1913); Brooke and Stokes (1929); Challis (1992, pp. 83–178, 179–397; and Appendix 1, pp. 673–698). 90. The arguments are summarized in Nef (1941, 1952); Graus (1951); Miskimin (1964; 1975, pp. 25–72, 132–157); Lopez, Miskimin, and Udovitch (1970); Day (1978); Munro (1983a, pp. 97–112); Spufford (1988, pp. 267–288, 340–342). 91. Ashtor (1971; 1976, pp. 319–331). See also Day (1978); Munro (1983a); and nn. 93–94, 96. 92. The two sets of technological revolutions were, first, in mechanical engineering: adits and mechanical pumps to permit much deeper, well drained, mining shafts; and then in chemical engineering: the Seiger- or Saigerhüttenprozess, for smelting argentiferous-cupric ores with lead to separate the two metals. See Nef (1941; 1952, pp. 691–761); Braunstein (1983); Westermann (1972, 1986); Spufford (1988, pp. 363–377); Munro (1991, 2003a, b). 93. See Spufford (1988, p. 347): “many contemporary European observers believed that ‘thesaurisation’ [hoarding, the accumulation of plate] was the main cause of the bullion famines” during the later fourteenth and early fifteenth centuries; but he also comments (pp. 346–347) that: “In retrospect it appears that it was itself in part a response to the famine. Nevertheless it made that shortage worse . . ..” 94. See Thompson (1956, pp. xxxvi–xlix and p. 163), stating that “the reigns of the three Edwards [1272–1377] are, with the exception of the Civil Wars of Charles I, the most prolific in coin-finds since the Romano-British era. This is due primarily to a 279
4259 Ch06 4/12/02 10:12 am Page 280
280 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
period of continual unrest and bad economic conditions which encouraged an abnormal amount of hoarding.” But elsewhere (p. xvi) he suggests that it is “not a safe assumption” to attribute all coins hoards to such reasons. Many of the fourteenth-century hoards are related to the continuing Scottish wars. Of the 394 coin hoards in the British Isles from c.600 to 1500, 85 can be dated to the reigns of Edward II, Edward III, and Richard II (1307–1399), with 14 in the period 1377–1399. See M. Allen (1999) for an updated list of fourteenth-century coin hoards: 36 for 1330–1351 (and another 40 in Scotland); 16 for 1351–1412; and only 12 for 1412–1464. Such coin hoard evidence seems therefore to provide better evidence for the deflation of c. 1327–1343, weaker evidence for the deflation of 1377–1410, and virtually none for the deflation era of c.1440–1470. Nevertheless, while the absolute number of extant hoards declines in the late fourteenth and early fifteenth centuries, population and coinage outputs declined to a much greater extent, thus reducing the likely incidence of hoards. The survival of specific hoards may be accidental; and money hoarded in one period was undoubtedly dishoarded and spent in some ensuing period (thus eliminating the evidence of the hoard). For the fifteenthcentury Low Countries, see Spufford (1970, pp. 55–73; Appendix D, pp. 203–213). 95. See Miskimin (1975, pp. 92–104; 134–144); Huizinga (1926, pp. 140–152); Piponnier (1970, Chapters. 7–10); Geijer (1979, pp. 141–155); Stuard (1999, pp. 215–242); Lopez (1962, pp. 19–32). In or about 1500, a Venetian visitor wrote a memoir about England, in which he stated, with some considerable wonder, that “there is no small innkeeper, however poor and humble he may be, who does not serve with silver dishes and drinking cups; and no one, who has not in his house silver plate to the amount of at least £100 sterling, . . . is considered by the English to be a person of any consequence.” Sneyd (1847, pp. 28–29). 96. See also Mate (1978), Mayhew (1995), and M. Allen (2001, Table 2, p. 607). 97. See Munro (1979, pp. 194–196; 1994, pp. 147–195 and 204–208; 2000, 2001); De Roover (1948, pp. 130, 236–246, 331–357, esp. pp. 339–342); Van der Wee (1963, vol. II, pp. 85–86, 333–340, 355–358; 1977, pp. 302, 312, 323–324, 361–362; 2000, pp. 87–112, 125–133); Nightingale (1990). 98. For deflation in late-fourteenth and early fifteenth-century Tuscany, see also Herlihy (1967, pp. 125–130). 99. See also Beveridge (1955, pp. 18–35); Bridbury (1973, pp. 582–586). 100. Bridbury’s thesis evidently rests upon the assumption that England was so overpopulated before the Black Death that the entire agrarian economy suffered from massive disguised unemployment, with a negative marginal product for labor. Not only is that thesis untenable, but it is contradicted by the arguments and evidence posed in his subsequent article (Bridbury, 1977). 101. Many of the arguments that follow could just as well pertain to the so-called “Great Depression” era of severe deflation and rapidly rising real wages from 1873 to 1896. See in particular Feinstein (1990), MacKinnon (1994), and Saul (1985). 102. Van Werveke (1938); Munro (1973, pp. 43–74; and appendices, pp. 187–214; 1981, pp. 78–95; 1983, pp. 112–126). See also other monetary studies in Munro (1992). 103. According to the monetary formula for a renforcement, which involves reciprocal changes: [1/(1 + x)] – 1, where x = the percentage change in the silver content of the groot (gros). Thus [1/(1.316) – 1] = 0.760 – 1 = –0.240 or 24.0%. See Munro (1988, pp. 389–392, 417–418). 104. Stadsarchief Brugge, Stadsrekeningen 1388/1389 to 1399/1400: wage payments for building craftsmen in the “werken” accounts.
4259 Ch06 4/12/02 10:12 am Page 281
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
281
105. Gilliodts-van Severen (1871–1878, vol. III, no. 706, pp. 134–135, 140–142: for January 1390); Van Werveke (1931, pp. 1–15; 1938, pp. 336–347); De Roover (1948, pp. 227–229); Van der Wee (1963, Vol. II, pp. 14–18, 29–30). 106. In 1383, during the Artevelde revolt (1379–1385) and just before the initial, abortive renforcement of 1384, the daily wages of Bruges master craftsmen in the building trades had been reduced even more sharply, from 12d to 8d groot, but were raised to 9.33d in 1386 and then fully restored to 12d in 1387. In 1386, just after the suppression of the Artevelde revolt, the daily wages of Bruges policemen had been raised from 6d to 7.67d (7d 16 mites) groot, but were restored to 6d in 1387. Note again that the policemen’s daily wages were paid for a 365-day year. Data extracted from the Bruges municipal accounts: Stadsarchief Brugge, Stadsrekeningen 1382/1383 to 1397/1398. Unfortunately the Ghent wage accounts are far too sparse to permit similar comparisons, and Ypres’ stadsrekeningen now survive only from 1406 in the second copy deposited at the Lille Chambre des Comptes (now in the Algemeen Rijksarchief België). For Bruges policemen, see below pp. 222–223. 107. For some evidence on nominal-wage reductions during this same deflationary era (c.1380-c.1430), and on the difficulties in interpreting the published evidence, see pp. 226–227. 108. See in particular, Fris (1907, pp. 421–459); Nicholas (1987; 1988; 1992, pp. 273–323); Boone and Brand (1993, pp. 168–192); Boone, Brand, and Prevenier (1993, pp. 59–74); Brand and Stabel (1995, pp. 203–224); Munro (1994a, pp. 377–388; and 2002, pp. 153–206). 109. Text of 4 September 1373 in Rijksarchief Van Oost-Vlaanderen, Oostenrijks Fonds, layette 1; provisions also repeated in layette 2 (2 May 1423). See also Espinas and Pirenne (1906–1924, vol. II, pp. 526–527): doc. no. 485, ordinances banning strikes; no. 491, pp. 533–535: letters of the deken of weavers guild submitting the dispute with fullers to arbitration; no. 492, pp. 535–537: Ghent fullers seek the count’s pardon, who then awards a wage of 45d per maerclaken. In January 1386, after their real wage had deteriorated by 20% since just 1382 (as measured by this Flemish price index, in Table 8), the fullers’ peaceful request for another increase now encountered a very hostile reaction from Count Louis’ successor, Duke Philip the Bold. Fully supporting the drapers, he curtly told the fullers “to be content” with their current wage, and furthermore decreed that henceforth any foreign fullers would be free to establish fulleries within Ghent. Duke Philip also rebuffed the fullers’ demands for a change in their guild constitution. Algemeen Rijksarchief, Trésor de Flandre, Series I, no. 2208; also published in Bartier and Van Nieuwenhuysen (1965, Vol. I, no. 88, pp. 123–124). See also Boone (1990, pp. 133–134). 110. Rates deduced from texts in Espinas and Pirenne (1906–1924, Vol. II, no. 492, pp. 535–537); Algemeen Rijksarchief, Trésor de Flandre, Series I, no. 2208; and especially Rijksarchief van Oost Vlaanderen te Gent, Oostenrijks Fonds, layette 2 (for 2 May 1423): “desquelz [desdiz foulons] ils n’avoient et ne leur en vouloit en baillier que trente deux gros . . . et est salaire trop petit . . .” See also Van Werveke (1931, pp. 4–14); Nicholas (1987, p. 130); and the following note. Fullers’s wage data for other Flemish and Dutch drapery towns indicates that, on average, each journeyman received 35% of the wage, leaving 30% for the master (who, of course, received revenues from several similarly manned fulling vats). 111. De Sagher (1951–1966, Vol. III, pp. 445–446, 451–452, no. 553, and p. 468, no. 554:136). The two journeymen were to receive 14d groot each and the master 7d, 281
4259 Ch06 4/12/02 10:12 am Page 282
282 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
for a total of 35d per cloth in three days. The journeymen were also to receive another 1.25d groot (16d parisis) for scrubbing the cloths. The Wervik dickedinnen broadcloth was to be 38 ells by 9.5 quarter ells on the loom, about the same size as the Ghent dickedinnen. 112. Again thanks to intervention from Count Louis de Male. Algemeen Rijksarchief, Trésor de Flandre, Series I, no. 1103; partly published in Espinas and Pirenne (1906–1924, Vol. I, no. 206, pp. 668–669); and in Bartier-Van Nieuwenhuysen (1965, Vol. I, no. 253, pp. 385–386). The only previous wage datum is for a wage increase to 15d 4 mites (15.167d) in 1348; in 1349, that wage represented 31.35 g of silver and 0.238 unit of a basket of consumables (priced at 63.69d gros, in Table 12). The new wage set at 41d groot in 1374, represented 45.67 g silver and 0.304 basket (priced at 134.90d groot). 113. Algemeen Rijksarchief, Trésor de Flandre, Series I, no. 1103; partly published in Espinas and Pirenne (1906–24, Vol. I, no. 206, pp. 668–669; and in Bartier-Van Nieuwenhuysen (1965, Vol. I, no. 253, pp. 385–386): “les diz drapiers disans que ce estoit trop grand salaire et qu’ilz devoient estre contens de xxxii [32] gros, attendu que au temps de la dicte ordonnance [1373–1374, of Louis de Male] la monnoie estoit plus feble que elle n’est de present, car le franc d’or valoit pour lors xxxvii [37] gros ou environ et aujourduy il ne vault que xxxiii [33], et selonc ce que la dicte monnoie estoit plus forte le salaire des diz foulons devoit estre diminue . . .” (In fact the Flemish silver coinage of 1374 had been stronger than that of 1390). The Flemish gold noble had been revalued to 6s 0d or 72d groot, from 8s 6d or 102d. 114. The Flemish price index used here (1451–1475 = 100), fell from the peak of 124.72 in 1386–1390, reflecting the debasements of those years, to a trough of just 88.51 in 1391–1395 and was virtually at that same level (88.53) in 1401–1405. The renforcement reduced the traite – the coined value of a kilogram of pure silver – from £5.337 to £4.050 groot Flemish; expressed in terms of the increase in grams of silver in the coinage (0.315g) the change becomes : [(1.00 + 0.316)/1] – 1 = 0.7599 – 1 = –0.2401 or – 24%. See Tables B-1 and B-3 in Munro (1984a, pp. 96, 100). 115. Stadsarchief Brugge, Stadsrekeningen 1384–1385 to 1399/1400: from the construction werken accounts. 116. See in particular: Hatcher (1996, pp. 237–272); Nightingale (1997, pp. 631–656); Van der Wee (1963, Vol. II, pp. 61–73; 1975, pp. 203–221); Munro (1983, pp. 235–250; 1994; 1995, pp. 37–60; 1999, pp. 1–74. For a contrary view and then more nuanced views see Van Uytven (1961, 1966, 1992, 1995). 117. For an explanation see Sosson (1977, pp. 51–55, 225–227). Unfortunately, there are almost no continuous wage series for Ghent. For the few available for the fifteenth century, see Verlinden and Scholliers (1959–1965, Vol. II, part i, pp. 386–397, 409–428), generally in the same range as those at Bruges. 118. Algemeen Rijksarchief België, Rekenkamer, Stadsrekeningen Aalst, registers nos 31,412–431, 520. 119. The summer wage prevailed for nine months; the winter wage, for three months. In neighbouring Mechelen most master masons and some carpenters had their daily summer wage increased from 8d to 10d groot Brabant by 1420 (Stadsarchief Mechelen, Stadsrekeningen nos. 86–97). In 1434, with the monetary reform, it was increased again to 12d per day; but Antwerp master masons did not receive that same daily wage until 1443. Thus, Mechelen master masons maintained a higher level of real wages than did their Antwerp counterparts during this period: with a decline in the seasonally adjusted annual money wage for Mechelen masons, in terms of equivalent commodity baskets,
4259 Ch06 4/12/02 10:12 am Page 283
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
283
from a harmonic mean of 9.513 baskets in 1421–1425 to 8.280 baskets in 1436–1440 (See Tables 13–14). 120. Note that, for both the Antwerp and Mechelen masons, the number of commodity baskets that could have been purchased with their annual money wage income has been computed by assigning nine months (157.5 days) to the summer wage and thus three months (52.5 days) to the lower winter wage. See Tables 13–14 and their sources. 121. For other examples given by Beveridge (1936, 1955), for the Westminster and Winchester manors (especially Hinderclay) in the 1390s: a fall in the mean wage of threshers from 9.50d to 8.88d per three-rased quarters; and for reapers, from 17.13d to 14.96d per acre, from 1380–1389 to 1390–1399. From the tables presented in all these published sources, I had constructed a working-table with 40 wage groups for the years 1380–1400; and the data suggest that 32 of the 40 groups experienced at least some nominal-wage reductions: 20 in the 1380s and 22 in the 1390s, and thus 12 over both decades. But the unweighted decennial means of the 40 money-wage series fell 3.6% overall. For reasons indicated on p. 197 above, concerning the computation of decennial mean wage rates, this possibly dubious working-table is not being published here. 122. See above pp. 200–202 and Blanchard (1978) for the discussion of a possibly backward-bending supply curve of labor during this late-medieval era. 123. The elasticity of substitution sKL is the ratio of the proportional changes in the marginal products of land+capital (yK) and labor (yL). Conversely, if K increases relatively to L, and if the proportional increase in K/L exceeds the proportional fall in the ratio of the marginal products of land-capital and of labor – i.e. if SKL > 1, then the share of the national income accruing to these increased factors K will also rise. For mathematical proof, with the CES Production Function, see Layard and Walters (1978, pp. 63–68, 270–276). See also Phelps Brown and Hart (1952/1981, pp. 106–130): “when big changes occur in money wage rates, the accompanying changes in the share of wages [of national income] are relatively small;” and “a main cause of the long-period change in the share of wages has been simply the change in the relative number of wage earners.” 124. They cite in particular Hilton (1985, pp. 253–267); and more recently Dyer (2000, p. 24) has made an even bolder claim: that “by the end of the thirteenth century the lowest estimate for those mainly dependent on employment by others is 50 per cent.” But another recent estimate provides quite a contrary view: that in the early fourteenth century, a typical English smallholder peasant, with 18 acres, could find wage-paying employment, outside his holding, for only about 27 days per year; or 80 days for three adult peasants on one holding. See Kitsikopoulos (2000, p. 243); and also Bailey (1998).
30 31
ACKNOWLEDGMENTS
32 33 34 35 36 37 38 39 40
Research for this article was funded, in part, by grants from the Social Sciences and Humanities Research Council of Canada (3-195-156-06; 410-93-0866; 410-96-0306; 410-99-0274). I wish to thank the staff of the Archives of the British Library of Political and Economic Science and of the Algemeen Rijksarchief of Belgium for their invaluable assistance in doing the research for this article; and the members of the Economic History Workshops of the University of Toronto and of Rutgers University for their criticisms, advice, and suggestions for revising this paper, with particular thanks to Gillian Hamilton, Michael Baker, Xiadong 283
4259 Ch06 4/12/02 10:12 am Page 284
284 1 2
JOHN H. MUNRO
Zhu, Eugene White, and James Masschaele; and also thanks for similar advice from the anonymous referees. The usual disclaimer applies.
3 4
REFERENCES
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Archival Sources: Belgium: national archives Algemeen Rijksarchief België, Rekenkamer: treasurer’s accounts of the town governments: reg. nos. 31,412–31,485 (Aalst: 1395–1500), 32,461–32,566 (Bruges:1406–1500); 33,147–33,238 (Kortrijk, 1393–1493), 38,635–38,722 (Ieper/Ypres: 1406–1500). Algemeen Rijksarchief, Rekenkamer, reg. 18,195-200 and Acquits de Lille, liasses nos. 936–937: Ghent mint accounts (with wage data in expense accounts). Algemeen Rijksarchief, Trésor de Flandre, Series I, No. 1103, [n. 131], 2208 [n. 126]: fullers’ guilds. Rijksarchief van Oost Vlaanderen te Gent, Oostenrijks Fonds, layette 2 [n. 126]: fullers’ guilds. Municipal Archives: treasurers’ accounts of the town governments. Bruges: Stadsarchief Brugge, Stadsrekeningen 1350/1351 to 1479/1480. Ghent: Stadsarchief Gent, Stadsrekeningen 1350/1351-–1479/1480. Mechelen (Malines): Stadsarchief Mechelen, Stadsrekeningen 1350/1351 to 1499/1500, Nos 44–180. Leuven (Louvain): Stedelijke Archief Leuven, Stadsrekeningen 1350/1351 to 1499/1500, Nos 4986–5125.
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
United Kingdom: Archives of the British Library of Political and Economic Science, the Beveridge Price and Wage History Collection: Canterbury, 1393–1600 (Box D.3); Dover, 1227–1565 (Box H.1314); Exeter (Exebridge Accounts), 1338–1600 (Box F.1); Westminster Abbey, 1393–1541 (Box P.10); Winchester College, 1354–1513 (Box F.8); York, 1354–1513 (Box I.10); Battle Abbey: Alciston Manor, 1336–1487 (Boxes H.10–11); Downton, 1257–1306 (Box C.157); Esher, 1257–1306, 1270–1308, and 1300–1453 (Boxes C. 157, A.31–32); Hinderclay (Suffolk), 1262–1405 (Box G.14); Itchingswell (Ecchinswell), 1270–1453 (Box A.33); Meon, 1257–1306 (Box C.157); Nailsbourne, 1257–1306 (Box C.157); Overton, 1309–1453 (Box A.33); Redgrave (Abbot of Bury St. Edmunds, Suffolk), 1323–1492 (Box G.14); Southwark (Bishop of Winchester), 1406–1454 (Box A.34); Taunton, 1270–1308, 1309–1453 (Boxes A.31–32); Wargrave, 1257–1306 (Box C.157); Witney, 1257–1306 (Box C.157); Wycombe, 1257–1306, 1309–1453 (Boxes C.147, A.33). Archives of the British Library of Political and Economic Science, the Phelps Brown Papers Collection. Corporation of London Record Office: Bridge Master’s Account Rolls, 1381–1398; Bridge Master’s Accounts: Weekly Payment Series, 1404–1510 (Vols I–III). London Guildhall Manuscripts Library: Armourers’ Company Accounts (1499–1557): MS 12,065, Vol. I; Bakers’ Audit Books (1505–1547), MS 5174, Vol. 1; Brewers’ Guild, Warden’s Accounts (1424–1562): MS 5440; Carpenters’ Guild, Warden’s Accounts (1456–1573): MS 4326, Vols. I and II; Cutlers’ Guild Accounts (1442–1497): MS 7146, roll 1; Grocers’ Guild, Warden’s Accounts (1452–1578): MS 11,570–11,571, Vols I–VI: Ironmongers’ Guild Accounts (1455–1561): MS 11,698: Vols I–II; Pewterers’ Company Accounts (1474–1500): MS 7086, Vol. I.
4259 Ch06 4/12/02 10:12 am Page 285
Wage Stickiness and Monetary Changes
285
1
Printed Sources:
2
Abel, W. (1980). Agrarkrisen und Agrarkonjunktur (3rd ed.). Berlin, 1978. Translated by Olive Ordish as Agricultural Fluctuations in Europe from the Thirteenth to the Twentieth Centuries, London: Methuen and Co. Allen, M. (1999). The Durham Mint: The Control, Organization, Profits, and Output of an Ecclesiastical Mint. Unpublished Ph.D. dissertation, University of Durham. Allen, M. (2000). The Volume and Composition of the English Silver Currency, 1279–1351. British Numismatic Journal, 70, 38–44. Allen, M. (2001). The Volume of the English Currency, 1158–1470. Economic History Review, (2nd Series), 54(4), 595–611. Allen, R. (2001). The Great Divergence in European Wages and Prices from the Middle Ages to the First World War. Explorations in Economic History, 38(4), 411–447. Ames, E. (1965). The Sterling Crisis of 1337–1339. Journal of Economic History, 25, 496–552. Reprinted in: R. Floud (Ed.), Essays in Quantitative Economic History (pp. 36–58). Oxford: Oxford University Press 1974. Ashtor, E. (1976). A Social and Economic History of the Near East in the Middle Ages. London: Collins. Ashtor, E. (1971). Les métaux précieux et la balance des payements du Proche-Orient à la basse époque. Paris: S.E.V.P.E.N. Atack, J., & Passell, P. (1994). A New Economic View of American History: From Colonial Times to 1940 (2nd ed.). New York: Norton. Bailey, M. (1998). Peasant Welfare in England, 1290–1348. Economic History Review, (2nd Series), 51(2), 223–251. Bartier, J., & Van Nieuwenhuysen, A. (Eds) (1965). Les ordonnances de Philippe le Hardi, de Marguerite de Male, et de Jean Sans Peur, 1381–1419. Vol. I: Les ordonnances de Philippe le Hardi et de Marguerite de Male, du 16 octobre 1381 au 31 decembre 1393. Recueil des Ordonnances des Pays Bas, première série: 1381–1506. Brussels: Palais des Academies. Beveridge, W. (1936–1937). Wages in the Winchester Manors. Economic History Review, (1st Series), 7, 22–43. Beveridge, W. (1955–1956). Westminster Wages in the Manorial Era. Economic History Review, (2nd Series), 8, 18–35. Blair, J., & Ramsay, N. (Eds) (1991). English Medieval Industries: Craftsmen, Techniques, Products. London: The Hambledon Press. Blanchard, I. (1970). Population Change, Enclosure, and the Early Tudor Economy. Economic History Review, (2nd Series), 23, 427–445. Blanchard, I. (1978). Labour Productivity and Work Psychology in the English Mining Industry, 1400–1600. Economic History Review, (2nd Series), 31(1), 1–24. Bland, A. E., Brown, P. A., & Tawney, R. H. (Eds) (1914). English Economic History: Select Documents. London: B. Bell and Sons. Blockmans, W. P., & Prevenier, W. (1975). Armoede in de Nederlanden van de 14e tot het midden van de 16e eeuw: bronnen en problemen. Tijdschrift voor geschiedenis, 88, 501–538. Republished in English translation, with some abridgements and additions, as: Poverty in Flanders and Brabant from the Fourteenth to the Mid-Sixteenth Century: Sources and Problems. Acta Historiae Neerlandicae: Studies on the History of the Netherlands, Vol. 10, 1977, 20–58. Boccaccio, G. (1921). The Decameron. J. M. Rigg (Trans.). London: Routledge. Bolton, J. L. (1996). The World Upside Down: Plague as an Agent of Economic and Social Change. In: M. Ormrod & P. Lindley (Eds), The Black Death in England (pp. 17–78). Stamford: Paul Watkins.
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
285
4259 Ch06 4/12/02 10:12 am Page 286
286 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
Boone, M. (1990). Gent en de Bourgondische hertogen ca. 1384 – ca. 1453: een sociaal-politieke studie van een staatsvormingsproces. Verhandelingen van de Koninklijke Academie voor Wetenschappen, Letteren en Schone Kunsten van Belgie, Klasse der Letteren, Jaargang 52, No. 133. Brussels: Paleis der Academien. Boone, M., Brand, H., & Prevenier, W. (1993). Revendications salariales et conjoncture économique: les salaires de foulons à Gand et à Leyde au XVe siècle. In: E. Aerts, B. Henau, P. Janssens & R. Van Uytven (Eds), Studia Historica Oeconomica: Liber Amicorum Herman Van der Wee (pp. 59–74). Leuven: Leuven University Press. Boone, M., & Brand, H. (1993). Vollersproeren en collectieve actie in Gent en Leiden in de 14e en 15e eeuw. Tijdschrift voor sociale geschiedenis, 19(2), 168–192. Bothwell, J., Goldberg, P. J. P., & Ormrod, W. M. (Eds) (2000). The Problem of Labour in Fourteenth-Century England. York: York Medieval Press. Brand, H., & Stabel, P. (1995). De Ontwikkeling van vollerslonen in enkel laat-middeleeuwse textilecentra in de Nederlanden: een poging tot reconstructie. In: J.-M. Duvosquel & E. Thoen (Eds), Peasants and Townsmen in Medieval Europe: Studia in Honorem Adriaan Verhust (pp. 203–224). Ghent: Snoeck-Ducaju & Zoon. Braunstein, P. (1983). Innovations in Mining and Metal Production in Europe in the Late Middle Ages. Journal of European Economic History, 12, 573–591. Bridbury, A. R. (1973). The Black Death. Economic History Review, (2nd Series), 26, 577–592. Bridbury, A. R. (1977). Before the Black Death. Economic History Review, (2nd Series), 30, 393–410. Britnell, R. H. (1990). Feudal Reaction after the Black Death in the Palatinate of Durham. Past & Present, (128), 28–47. Brooke, G. C., & Stokes, E. (1929). Tables of Bullion Coined from 1377 to 1550. The Numismatic Chronicle, (5th Series), 9, 27–69. Campbell, B. M. (1981). The Population of Early Tudor England: A Re-evaluation of the 1522 Muster Returns and the 1524 and 1525 Lay Subsidies. Journal of Historical Geography, 7, 145–154. Campbell, B. M. (1983a). Agricultural Progress in Medieval England: Some Evidence from Eastern Norfolk. Economic History Review, (2nd Series), 36(1), 26–46. Campbell, B. M. (1983b). Arable Productivity in Medieval England: Some Evidence from Norfolk. Journal of Economic History, 43(2), 379–404. Campbell, B. M. (1984). Population Pressure, Inheritance, and the Land Market in a FourteenthCentury Peasant Community. In: R. M. Smith (Ed.), Land, Kinship and Life-Cycle (pp. 87–134). Cambridge: Cambridge University Press. Campbell, B. M. (2000). English Seigniorial Agriculture, 1250–1450. Cambridge Studies in Historical Geography No. 31. Cambridge and New York: Cambridge University Press. Campbell, R. (1747/1969). The London Tradesman. Republished New York: A.M. Kelly (1969). Cassell, A. (1983). Boccaccio, Giovanni. In: J. Strayer et al. (Eds), Dictionary of the Middle Ages (Vol. II, pp. 277–290). New York: Charles Scribner’s Sons. Cazelles, R. (1966). Quelques reflexions à propos des mutations monétaires de la monnaie royale française (1295–1360). Le moyen âge, 72, 83–105, 251–278. Challis, C. (1992). Appendix 1: Mint Output, 1220–1985. In: C. Challis (Ed.), A New History of the Royal Mint (pp. 673–698). Cambridge: University Press. Checkland, S. G. (1964). The Rise of Industrial Society in England, 1815–1885. London: Longmans. Clapham, J. H. (1964). An Economic History of Modern Britain, Vol. II: The Early Railway Age, 1820–1850. Cambridge: University Press. Clark, G. (2001). The Long March of History: Farm Laborers’ Wages in England 1208–1850. UCDavis. http://www.econ.ucdavis.edu/faculty/gclark/papers/long_march_of_history.pdf
4259 Ch06 4/12/02 10:12 am Page 287
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
287
Crump, C. G., & Johnston, C. (1913). Tables of Bullion Coined Under Edward I, II and III. The Numismatic Chronicle, (4th Series), 13, 200–245. Day, J. (1978). The Great Bullion Famine of the Fifteenth Century. Past & Present (79), 1–54. Reprinted in: J. Day (Ed.), The Medieval Market Economy (1978, pp. 1–54). Oxford: Basil Blackwell. De Sagher, H. et al. (Eds) (1961–1966). Recueil de documents relatifs à l’histoire de l’industrie drapière en Flandre, IIe partie: Le sud-ouest de la Flandre depuis l’époque bourguignonne (3 Vols). Brussels: Palais des Academies. De Vries, J. (1978). An Inquiry into the Behaviour of Wages in the Dutch Republic and the Southern Netherlands, 1580–1800. Acta Historia Neerlandicae, 10, 79–97. Reprinted in: M. Aymard (Ed.), Dutch Capitalism and World Capitalism (1982, pp. 37–62). Paris: Editions de la Maison des Sciences de l’Homme, and Cambridge: Cambridge University Press. De Vries, J. (1994). An Employer’s Guide to Wages and Working Conditions in the Netherlands, 1450–1850. In: C. S. Leonard & B. N. Mironov (Eds), Hours of Work and Means of Payment: The Evolution of Conventions in Pre-Industrial Europe (pp. 47–63). Proceedings of the Eleventh International Economic History Congress, Milan, September 1994, Session B3b. Milan: Bocconi University Press. De Vries, J. (1994). The Industrial Revolution and the Industrious Revolution. Journal of Economic History, 54(2), 249–270. De Roover, R. (1948). Money, Banking and Credit in Mediaeval Bruges: Italian Merchant Bankers, Lombards, and Money Changers: A Study in the Origins of Banking. Cambridge, Mass: Medieval Academy of America. Delepierre, O., & Willems, M. F. (Eds) (1842). Collection des keuren ou statuts de tous les métiers de Bruges. Ghent: Annoot-Braeckman. Derry, J. (1963). A Short History of Nineteenth-Century England. Mentor Books on English History. New York: New American Library [British Title: Reaction and Reform, 1793–1868: England in the Early Nineteenth Century. London: Blandford Press.] Doughty, R. (1975). Industrial Prices and Inflation in Southern England 1401–1640. Explorations in Economic History, 12, 177–192. Duby, G. (1962). L’économie rurale et la vie des campagnes dans l’occident médiéval. Paris: Aubier. Republished (1968) in: English translation as Rural Economy and Country Life in the Medieval West. C. Postan (Trans.). Columbia SC: University of South Carolina Press. Dyer, C. (1983). English Diet in the Later Middle Ages. In: T. H. Aston, P. R. Coss, C. Dyer & J. Thirsk (Eds), Social Relations and Ideas: Essays in Honour of R. H. Hilton (pp. 191–216). Cambridge: Cambridge University Press. Dyer, C. (1988). Changes in Diet in the Late Middle Ages: The Case of Harvest Workers. The Agricultural History Review, 36, 21–38. Dyer, C. (1989). Standards of Living in the Later Middle Ages: Social Change in England c. 1200–1520. Cambridge: Cambridge University Press. Dyer, C. (2000). Work Ethics in the Fourteenth Century. In: J. Bothwell, P. J. P. Goldberg & W. M. Ormrod (Eds), The Problem of Labour in Fourteenth-Century England (pp. 21–41). York: York Medieval Press. Endrei, W. (1981). La productivité et la technique dans l’industrie textile du XIIIe au XVIIe siècle. In: S. Mariotti (Ed.), Produttivita e tecnologia nei secoli XII–XVII (pp. 253–262). Atti delle “Settimana di Studi” e altri convegni, No. 3, Istituto Internazionale di Storia Economica “Francesco Datini” Florence: Felice le Monnier.
287
4259 Ch06 4/12/02 10:12 am Page 288
288 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
Endrei, W. (1983). The Productivity of Weaving in Late Medieval Flanders. In: N. B. Harte & K. G. Ponting (Eds), Cloth and Clothing In Medieval Europe: Essays in Memory of Professor E. M. Carus-Wilson (pp. 108–119). Pasold Studies in Textile History no. 2, London: Heinemann Educational Books. Espinas, G., & Pirenne, H. (Eds) (1906–1924). Recueil de documents relatifs à l’histoire de l’industrie drapière en Flandre: Ire partie: des origines à l’époque bourguignonne (4 Vols). Brussels: Librairie Kiessling et Compagnie. Farmer, D. (1983). Crop Yields, Prices and Wages in Medieval England. Studies in Medieval and Renaissance History, 6, 117–155. Farmer, D. (1988). Prices and Wages [1042–1350]. In: H. E. Hallam (Ed.), The Agrarian History of England and Wales (Vol. II: 1042–1350, pp. 760–778, 811–817). Cambridge: Cambridge University Press. Farmer, D. (1991). Prices and Wages, 1350–1500. In: E. Miller (Ed.), The Agrarian History of England and Wales (Vol. III: 1348–1500, pp. 467–524). Cambridge: Cambridge University Press. Farmer, D. (1996). The Famuli in the Later Middle Ages. In: R. Britnell & J. Hatcher (Eds), Progress and Problems in Medieval England: Essays in Honour of Edward Miller (pp. 214–220). Cambridge: Cambridge University Press. Feavearyear, A., Sir. (1963). The Pound Sterling: a History of English Money (2nd ed., revised by E. V. Morgan). Oxford: Clarendon Press. Feinstein, C. (1990). What Really Happened to Real Wages?: Trends in Wages, Prices, and Productivity in the United Kingdom, 1880–1913. Economic History Review, (2nd Series), 43, 329–355. Fischer, D. H. (1996). The Great Wave: Price Revolutions and the Rhythm of History. Oxford and New York: Oxford University Press. Fournial, E. (1970). Histoire monétaire de l’occident médiéval. Paris: F. Nathan. Fris, V. (1907). Les origines de la réforme constitutionelle de Gand de 1360–1369. Annales du XXe congrès de la federation archéologique et historique de la Belgique, 3, 421–459. Geijer, A. (1979). A History of Textile Art. London: Pasold Research Fund and Sotheby Parke Bernet. Gilliodts-van Severen, L. (Ed.) (1871–1878). Inventaire des archives de la ville de Bruges (6 Vols). Bruges. E. Gaillard. Joossen, H. (Ed.) (1935). Recueil de documents relatifs à l’histoire de l’industrie drapière à Malines, des origines à 1384. Bulletin de la Commission Royale d’Histoire, 99. Given-Wilson, C. (2000). The Problem of Labour in the Context of English Government, c. 1350–1450. In J. Bothwell, P. J. P. Goldberg & W. M. Ormrod (Eds), The Problem of Labour in Fourteenth-Century England (pp. 85–100). York: York Medieval Press. Goldin, C. (1990). Understanding the Gender Gap: An Economic History of American Women. New York: Oxford University Press. Graus, F. (1951). La crise monétaire du XIVe siècle. Revue belge de philologie et d’histoire, 29, 445–454. Hallam, H. E. (1988). Population Movements in England, 1086–1350. In: H. E. Hallam (Ed.), Agrarian History of England and Wales (Vol. II: 1042–1350, pp. 508–593). Cambridge: Cambridge University Press. Hamilton, E. (1936). Money, Prices, and Wages in Valencia, Aragon, and Navarre, 1351–1500. Perspectives in European History no. 6 Cambridge, Mass: Harvard University Press [Reissued: Philadelphia: Porcupine Press, 1975]. Harvey, B. (1966). The Population Trend in England Between 1300 and 1348. Transactions of the Royal Historical Society, (5th Series), 16, 23–42.
4259 Ch06 4/12/02 10:12 am Page 289
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
289
Harvey, B. (1991). Introduction: the ‘Crisis’ of the Early Fourteenth Century. In: B. M. Campbell (Ed.), Before the Black Death: Studies in the “Crisis’ of the Early Fourteenth-Century (pp. 1–24). Manchester and New York: Manchester University Press. Hatcher, J. (1977). Plague, Population, and the English Economy, 1348–1530. Studies in Social and Economic History. London: The MacMillan Press. Hatcher, J. (1986). Mortality in the Fifteenth Century: Some New Evidence. Economic History Review, (2nd Series), 39(1), 19–38. Hatcher, J. (1994). England in the Aftermath of the Black Death. Past & Present, (144), 3–35. Hatcher, J. (1996). The Great Slump of the Mid-Fifteenth Century. In: R. Britnell & J. Hatcher (Eds), Progress and Problems in Medieval England (pp. 237–272). Cambridge: Cambridge University Press. Herlihy, D. (1967). Medieval and Renaissance Pistoia: The Social History of an Italian Town, 1200–1430. New Haven: Yale University Press. Herlihy, D., & Klapisch-Zuber, C. (1985). Tuscans and Their Families: A Study of the Florentine Catasto of 1427. New Haven: Yale University Press. Hilton, R. (1985). Some Social and Economic Evidence in Late-Medieval English Tax Returns. In: R. H. Hilton (Ed.), Class Conflict and the Crisis of Feudalism (pp. 253–267). London: The Hambledon Press. Holmes, G. A. (1957). The Estates of the Higher Nobility in Fourteenth-Century England. Cambridge: Cambridge University Press. Holt, R. (1988). The Mills of Medieval England. Oxford: Basil Blackwell. Huberman, M. (1986). Invisible Handshakes in Lancashire: Cotton Spinning in the First Half of the Nineteenth Century. Journal of Economic History, 46, 987–998. Huberman, M. (1991). How Did Labor Markets Work in Lancashire? More Evidence on Prices and Quantities in Cotton Spinning, 1822–1852. Explorations in Economic History, 28(1), 87–120. Huberman, M. (1996). Escape from the Market: Negotiating Work in Lancashire. Cambridge and New York: Cambridge University Press. Huizinga, J. (1926). The Waning of the Middle Ages: A Study of the Forms of Life, Thought, and Art in France and the Netherlands in the XIVth and XVth Centuries. London: Edward Arnold, and New York: St. Martin’s Press. Jones, S. R. (2000). Household, Work and the Problem of Mobile Labour: The Regulation of Labour in English Towns. In: J. Bothwell, P. J. P. Goldberg & W. M. Ormrod (Eds), The Problem of Labour in Fourteenth-Century England (pp. 133–153). York: York Medieval Press. Jones, G. (1952). Building in Stone in Medieval Western Europe. In: M. Postan & E. E. Rich (Eds), The Cambridge Economic History of Europe, Vol. II: Trade and Industry in the Middle Ages (pp. 493–518). Cambridge: Cambridge University Press. Jordan, W. C. (1966). The Great Famine: Northern Europe in the Early Fourteenth-Century. Princeton: Princeton University Press. Kershaw, I. (1973). The Great Famine and Agrarian Crisis in England, 1315–1322. Past & Present, (59), 3–50. Reprinted (1976) in: R. H. Hilton (Ed.), Peasants, Knights, and Heretics (pp. 85–132). Cambridge: Cambridge University Press. Keynes, J. M. (1936). The General Theory of Employment, Interest and Money. London: MacMillan. Kitsikopolous, H. (2000). Standards of Living and Capital Formation in Pre-Plague England: A Peasant Budget Model. Economic History Review, (2nd Series), 53(2), 237–261. Knoop, D., & Jones, G. P. (1933). Masons’ Wages in Medieval England. Economic History, 2, 473–499. Knoop, D., & Jones, G. P. (1936). The Rise of the Mason Contractor. Journal of the Royal Institute of British Architects, (3rd Series), (20).
289
4259 Ch06 4/12/02 10:12 am Page 290
290 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
Knoop, D., & Jones, G. P. (1967). The Mediaeval Mason: An Economic History of English Stone Building in the Later Middle Ages and Early Modern Times (3rd edn.). Manchester: Manchester University Press. Kovacevic, D. (1960). Les mines d’or et d’argent en Serbie et en Bosnie médiévales. Annales: Economies, Sociétés, Civilisations, 15, 248–258. Lane, F. (1977). The First Infidelities of the Venetian Lire. In: H. A. Miskimin, D. Herlihy & A. L. Udovitch (Eds), The Medieval City (pp. 52–59). New Haven and London: Yale University Press. Layard, P. R. G., & A. A. Walters (1978). Microeconomic Theory. London and New York: McGraw Hill. Lindert, P. (1985). English Population, Wages, and Prices: 1541–1913. Journal of Interdisciplinary History, 15, 618–626. Lipson, E. (1921). The History of the Woollen and Worsted Industries. London: A & C Black Ltd. Lopez, R. (1962). Hard Times and Investment in Culture. In: W. Ferguson et al. (Eds), The Renaissance (pp. 29–52). New York: Harper and Row. Lopez, R. S., Miskimin, H., & Udovitch, A. L. (1970). England to Egypt, 1350–1500: Long Term Trends and Long Distance Trade. In: M. A. Cook (Ed.), Studies in the Economic History of the Middle East (pp. 93–128). London: Oxford University Press. Lucas, H. S. (1930). The Great European Famine of 1315, 1316, and 1317. Speculum, 5(4), 343–377. Reprinted (1954–1962) in: E. M. Carus-Wilson (Ed.), Essays in Economic History (Vol. II, pp. 49–72). London: Edward Arnold. MacKinnon, M. (1994). Living Standards, 1870–1914. In: R. Floud & D. McCloskey (Eds), The Economic History of Britain Since 1700 (3 Vols., 2nd ed., Vol. II: 1860–1939, pp. 265–290). Cambridge: Cambridge University Press, 1994. Maddicott, J. R. (1975). The English Peasantry and the Demands of the Crown, 1294–1341. Past & Present, (Supplement No. 1), 1–75. Reprinted (1987) in: T. H. Aston (Ed.), Landlords, Peasants, and Politics in Medieval England, Past and Present Publications (pp. 285–359). Cambridge: Cambridge University Press. Mate, M. (1975). High Prices in Early Fourteenth-Century England: Causes and Consequences. Economic History Review, (2nd Series), 28(1), 1–16. Mate, M. (1978). The Role of Gold Coinage in the English Economy, 1338–1400. Numismatic Chronicle, (7th Series), 18, 126–141. Mate, M. (1986). Population Fluctuations in South-East England, 1400–1530. Unpublished working paper. Mayhew, N. (1974). Numismatic Evidence and Falling Prices in the Fourteenth-Century. Economic History Review, (2nd Series), 27(1), 1–15. Mayhew, N. (1987). Money and Prices in England from Henry II to Edward III. Agricultural History Review, 35(2), 121–132. Mayhew, N. (1995). Population, Money Supply, and the Velocity of Circulation in England, 1300–1700. Economic History Review, (2nd Series), 48(2), 238–257. McCloskey, D. (1972). Review of Peter Ramsey (Ed.), The Price Revolution in Sixteenth-Century England (London, 1971). In: The Journal of Political Economy, 80, 1332–1335. Melis, F. (1962). La diffusione nel Mediterranea occidentale dei panni di Wervicq e delle altre citta della Lys attorna al 1400. In: Studi in onore di Amintore Fanfani (Vol. III: Medioevo, pp. 219–243). Milan: Giuffre. Miller, E., & Hatcher, J. (1978). Medieval England: Rural Society and Economic Change, 1086–1348. Social and Economic History of England. London and New York: Longman.
4259 Ch06 4/12/02 10:12 am Page 291
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
291
Miller, E. (1991). Introduction: Land and People. In: E. Miller (Ed.), Agrarian History of England and Wales (Vol. III: 1348–1500, pp. 1–33). Cambridge: Cambridge University Press. Mills, F. C. (1956). Introduction to Statistics. New York: Henry Holt. Miskimin, H. (1975). The Economy of Early Renaissance Europe, 1300–1460. Cambridge: Cambridge University Press. Mueller, R. (1984). ‘Chome l’ucciello di passegio’: la demande saisonnière des espèces et le marché des changes à Venise au moyen âge. In: J. Day (Ed.), Études d’histoire monétaire, XIIe-XIXe siècles (pp. 192–220). Université de Paris VII. Lille: Presses Universitaires de Lille. Munro, J. (1973). Wool, Cloth, and Gold: The Struggle for Bullion in Anglo Burgundian Trade, ca. 1340–1478. Brussels: Editions de l’Université de Bruxelles. Munro, J. (1979). Bullionism and the Bill of Exchange in England, 1272–1663: A Study in Monetary Management and Popular Prejudice. In: Center for Medieval and Renaissance Studies UCLA (Ed.), The Dawn of Modern Banking (pp. 169–239). New Haven and London: Yale University Press. Reprinted (1992) in: J. Munro, Bullion Flows and Monetary Policies in England and the Low Countries, 1350–1500, Variorum Collected Studies (Series C5355). Aldershot: Ashgate Publishing. Munro, J. (1981). Mint Policies, Ratios, and Outputs in England and the Low Countries, 1335–1420: Some Reflections on New Data. The Numismatic Chronicle, 141, 71–116. Reprinted in J. Munro, Bullion Flows and Monetary Policies in England and the Low Countries, 1350–1500. Munro, J. (1983a). Bullion Flows and Monetary Contraction in LateMedieval England and the Low Countries. In: J. F. Richards (Ed.), Precious Metals in the Later Medieval and Early Modern Worlds (pp. 97–158). Durham: Carolina Academic Press. Reprinted in: J. Munro, Bullion Flows and Monetary Policies in England and the Low Countries, 1350–1500. Munro, J. (1983b). Economic Depression and the Arts in the Fifteenth-Century Low Countries. Renaissance and Reformation, 19, 235–250. Reprinted (1994) in J. Munro, Textiles, Towns, and Trade: Essays in the Economic History of Late-Medieval England and the Low Countries, Variorum Collected Studies (Series CS 442). Aldershot: Ashgate Publishing. Munro, J. (1984a). Mint Outputs, Money, and Prices in Late-Medieval England and the Low Countries. In: E. Van Cauwenberghe & F. Irsigler (Eds), Münzprägung, Geldumlauf und Wechselkurse/Minting, Monetary Circulation and Exchange Rates, Trierer Historische Forschungen (Vol. 7, pp. 31–122). Akten des 8th International Economic History Congress, Section C-7, Budapest 1982, Trier: Verlag Trierer Historische Forschungen. Munro, J. (1984b). Monnayage, monnaies de compte, et mutations monétaires au Brabant à la fin du moyen âge. In: J. Day (Ed.), Études d’histoire monétaire, XIIe–XIXe siècles (pp. 263–294). Études de l’Université de Paris VII et du Centre National des Lettres. Lille: Presses Universitaires de Lille. Reprinted in: J. Munro, Bullion Flows and Monetary Policies in England and the Low Countries. Munro, J. (1988a). Deflation and the Petty Coinage Problem in the LateMedieval Economy: The Case of Flanders, 1334–1484. Explorations in Economic History, 25, 387–423. Reprinted in: J. Munro, Bullion Flows and Monetary Policies in England and the Low Countries. Munro, J. (1988b). Textile Technology and Textile Workers. In: J. R. Strayer et al. (Eds), The Dictionary of the Middle Ages (Vol. XI, pp. 693–715). New York: Charles Scribner’s Sons. Reprinted (1994) in: J. Munro, Textiles, Towns, and Trade: Essays in the Economic History of Late-Medieval England and the Low Countries. Aldershot: Ashgate Publishing. Munro, J. (1991a). Industrial Transformations in the North-west European Textile Trades, c.1290–1340: Economic Progress or Economic Crisis? In: B. M. Campbell (Ed.), Before the Black Death: Studies in the “Crisis” of the Early Fourteenth Century (pp. 110–148). Manchester and New York: Manchester University Press. Reprinted (1994) in: J. Munro,
291
4259 Ch06 4/12/02 10:12 am Page 292
292 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
Textiles, Towns, and Trade: Essays in the Economic History of Late-Medieval England and the Low Countries. Aldershot: Ashgate Publishing. Munro, J. (1991b). The Central European Mining Boom, Mint Outputs, and Prices in the Low Countries and England, 1450–1550. In: E. Van Cauwenberghe (Ed.), Money, Coins, and Commerce: Essays in the Monetary History of Asia and Europe (From Antiquity to Modern Times) (pp. 119–183). Leuven: Leuven University Press. Munro, J. (1992). Bullion Flows and Monetary Policies in England and the Low Countries, 1350–1500, Variorum Collected Studies (Series CS355). Aldershot: Ashgate Publishing. Munro, J. (1994a). Industrial Entrepreneurship in the Late-Medieval Low Countries: Urban Draperies, Fullers, and the Art of Survival. In: P. Klep & E. Van Cauwenberghe (Eds), Entrepreneurship and the Transformation of the Economy (10th–20th Centuries): Essavs in Honour of Herman Van der Wee (pp. 377–388). Leuven: Leuven University Press. Munro, J. (1994b). Patterns of Trade, Money, and Credit. In: T. Brady Jr., H. Oberman & J. Tracy (Eds), Handbook of European History in the Later Middle Ages, Renaissance and Reformation, 1400–1600, Vol. I: Structures and Assertions (pp. 147–195). Leiden: E.J. Brill. Munro, J. (1994c). Urban Wage Structures in Late-Medieval England and the Low Countries: Work Time and Seasonal Wages. In: I. Blanchard (Ed), Labour and Leisure in Historical Perspective, Thirteenth to Twentieth Centuries (Vierteljahrschrift für Sozial und Wirtschaftsgeschichte Beiheft Series, No. 116, pp. 65–78). Stuttgart: Franz Steiner Verlag. Munro, J. (1995). Anglo-Flemish Competition in the International Cloth Trade, 1340–1520. In: J.-M. Cauchies (Ed.), Rencontres d’Oxford (Septembre 1994): L’Angleterre et les pays bas bourguignonnes: relations et comparaisons, XVe–XVIe siècle (Vol. 35, pp. 37–60). Publication du centre européen d’études bourguigonnes. Munro, J. (1999a). The Symbiosis of Towns and Textiles: Urban Institutions and the Changing Fortunes of Cloth Manufacturing in the Low Countries and England, 1270–1570. The Journal of Early Modern History: Contacts, Comparisons, Contrasts, 3(1), 1–74. Munro, J. (1999b). Review of David Hackett Fischer, The Great Wave: Price Revolutions and the Rhythm of History. Oxford and New York: Oxford University Press, 1996. For EH.Net Review [
[email protected]], 24 February. Munro, J. (2000). English Backwardness and Financial Innovations in Commerce with the Low Countries, 14th to 16th Centuries. In: P. Stabel, B. Blondé & A. Greve (Eds), International Trade in the Low Countries (14th to 16th Centuries): Merchants, Organisation, Infrastructure (No. 10, pp. 105–167). Studies in Urban, Social, Economic, and Political History of the Medieval and Early Modern Low Countries. Leuven-Apeldoorn: Garant. Munro, J. (2002). Gold, Guilds, and Government: the Impact of Monetary and Labour Policies on the Flemish Cloth Industry, 1390–1435. Jaarboek voor middeleeuwsche geschiedenis, 5, 153–206. Munro, J. (2003a). Industrial Energy from Water-Mills in the European Economy, Fifth to Eigtheenth Centuries: the Limitations of Power. In: S. Cavaciocchi (Ed.), Economica ed Energia, Seccoli XII–XVIII. Atti delle “Settimane di Studi” e altrie Convegni, Istituto Internazionale di Storia Economica, “Francesco Datini da Prato” (Vol. 34). Florence: Le Monnier. Munro, J. (2003b). The Monetary Origins of the “Price Revolution” Before the Influx of SpanishAmerican Treasure: The South German Silver-Copper Trades, Merchant-Banking, and Venetian Commerce, 1470–1540. In: D. Flynn, A. Giráldez & R. von Glahn (Eds), Global Connections and Monetary History, 1470–1800. Aldershot and Brookfiel, VT: Ashgate Publishing. Nef, J. (1937). Prices and Industrial Capitalism in France and England. Economic History Review (lst Series), 7, 2. Reprinted (1953) in: F. Lane & J. Riemersma (Eds), Enterprise and Secular
4259 Ch06 4/12/02 10:12 am Page 293
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
293
Change (pp. 292–321). London: George Allen and Unwin, and (1954–1962). In: E. M. Carus-Wilson (Ed.), Essays in Economic History (3 Vols, Vol. I, pp. 108–134). London. Nef, J. (1941). Silver Production in Central Europe, 1450–1618. Journal of Political Economy, 49, 575–591. Nef, J. U. (1952). Mining and Metallurgy in Medieval Civilization. In: M. M. Postan (Ed.), Cambridge Economic History of Europe, Vol. II: Trade and Industry in the Middle Ages, 456–469. Cambridge: Cambridge University Press. (Revised ed., Cambridge (1987), pp. 696–734.) Nicholas, D. (1987). The Metamorphosis of a Medieval City: Ghent in the Age of the Arteveldes, 1302–1390. Lincoln and London: University of Nebraska Press. Nicholas, D. (1988). The Van Arteveldes of Ghent: The Varieties of Vendetta and the Hero in History. Ithaca: Cornell University Press. Nicholas, D. (1992). Medieval Flanders. London and New York: Longman. Nightingale, P. (1996). The Growth of London in the Medieval English Economy. In: R. Britnell & J. Hatcher (Ed), Progress and Problems in Medieval England (pp. 89–106). Cambridge and New York: Cambridge University Press. Nightingale, P. (1997). England and the European Depression of the Mid-Fifteenth Century. The Journal of European Economic History, 26(3), 631–656. Patterson, C. C. (1972). Silver Stocks and Losses in Ancient and Medieval Times. Economic History Review, (2nd Series), 25(2), 205–235. Penn, S., & Dyer, C. (1990). Wages and Earnings in Late Medieval England: Evidence from Enforcement of the Labour Laws. Economic History Review, (2nd Series), 43(3), 356–376. Phelps Brown, H., & Hart, P. E. (1952). The Share of Wages in National Income. Economic Journal, 62. Reprinted (1981) in: H. Phelps Brown & S. V. Hopkins (Eds), A Perspective of Wages and Prices (pp. 106–130). London and New York: Methuen. Phelps Brown, H., & Hopkins, S. V. (1955). Seven Centuries of Building Wages. Economica, 22, 87, 195–206. Reprinted (1954–1962) in: E. M. Carus-Wilson (Ed.), Essays in Economic History (3 Vols, Vol. II, pp. 168–178). London: Edward Arnold, and (1981). In: H. Phelps Brown & S. V. Hopkins (Eds), A Perspective of Wages and Prices (pp. 1–12). London and New York: Methuen. Phelps Brown, H., & Hopkins, S. V. (1956). Seven Centuries of the Prices of Consumables, Compared with Builders’ Wage Rates. Economica, 23, 92, 296–314. Reprinted (1954–1962) in: E. M. Carus-Wilson (Ed.), Essays in Economic History (3 Vols, Vol II, pp. 179–196). London: Edward Arnold. And (1981). In: H. Phelps Brown & S. V. Hopkins (Eds), A Perspective of Wages and Prices (pp 13–59 with statistical appendices not previously published). London and New York: Methuen. Piponnier, F. (1970). Costume et vie sociale: la cour d’Anjou, XIVe–XVe siècle. Paris: Mouton. Platt, C. (1996). King Death: The Black Death and Its Aftermath in Late-Medieval England. London: University College London Press, and Toronto and Buffalo: University of Toronto Press. Poos, L. R. (1985). The Rural Population of Essex in the Later Middle Ages. Economic History Review, (2nd Series), 38, 515–530. Poos, L. R. (1991). A Rural Society After the Black Death: Essex, 1350–1525. Cambridge: Cambridge University Press. Postan, M. (1950). Some Economic Evidence of Declining Population in the Later Middle Ages. Economic History Review, (2nd Series), 2, 130–167. Reprinted (1973) with the revised title of “Some Agrarian Evidence of Declining Population in the Later Middle Ages” in: M. Postan, Essays on Medieval Agriculture and General Problems of the Medieval Economy (pp. 186–213). Cambridge: Cambridge University Press.
293
4259 Ch06 4/12/02 10:12 am Page 294
294 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
Postan, M. (1951). The Economic Foundations of Medieval Society. Jahrbücher für Nationalökonomie, 161. Reprinted (1973). In: M. Postan, Essays on Medieval Agriculture and General Problems of the Medieval Economy (pp. 3–27). Cambridge: Cambridge University Press. Postan, M. (1952). The Trade of Medieval Europe: the North. In: M. M. Postan & E. E. Rich (Eds), Cambridge Economic History of Europe, Vol. II: Trade and Industry in the Middle Ages (pp. 119–256). Cambridge: Cambridge University Press. Republished (1987), with a few changes, in the 2nd revised edition: M. M. Postan & E. Miller (Eds) (pp. 168–305). Cambridge: Cambridge University Press, and (1973). In: M. M. Postan, Medieval Trade and Finance (pp. 92–131). Cambridge: Cambridge University Press. Postan, M. (1959). Note. Economic History Review, (2nd Series), 12, 77–82. [See Robinson, W.C. 1959]. Postan, M. (1966). Medieval Agrarian Society: England. In: M. M. Postan (Ed.), Cambridge Economic History, Vol. I: The Agrarian Life of the Middle Ages (2nd rev. ed., pp. 560–570). Cambridge: Cambridge University Press. Postan, M. (1972). The Medieval Economy and Society: An Economic History of Britain, 1100–1500. Cambridge: Cambridge University Press. Posthumus, N. (Ed.) (1910). Bronnen tot de geschiedenis van de Leidsche textielnijverheid, 1333–1795, Vol. I: De Middeleeuwen, 1333–1480. The Hague: Rijks Geschiedenkundige Publicatiën. Prestwich, M. (1977). Currency and the Economy of Early Fourteenth-Century England. In: N. Mayhew (Ed.), Edwardian Monetary Affairs, 1279–1344 (pp. 45–58). BAR International Series, no. 36. Oxford: British Archeological Reports. Putnam, B. H. (1908). The Enforcement of the Statute of Laborers During the First Decade after the Black Death. New York: Columbia University. Raftis, J. A. (1964). Tenure and Mobility: Studies in the Social History of the Medieval English Village. Studies and Texts No. 8. Toronto: Pontifical Institute of Mediaeval Studies. Raftis, J. A. (1996). Peasants and the Collapse of the Manorial Economy on Some Ramsey Abbey Estates. In: R. Britnell & J. Hatcher (Eds), Progress and Problems in Medieval England: Essays in Honour of Edward Miller (pp. 191–206). Cambridge: Cambridge University Press. Ramsey, P. (Ed.) (1971). The Price Revolution in Sixteenth-Century England, Debates in Economic History Series. London: Methuen and Co. Reynolds, T. S. (1983). Stronger than a Hundred Men: A History of the Vertical Water Wheel. Johns Hopkins Studies in the History of Technology, New series No. 7. Baltimore and London: The Johns Hopkins University Press. Riley, H. T. (Ed.) (1860). Munimenta Gildhallae Londoniensis: Vol. II: Liber Custumarum. London: Longman, Brown and Green. Riley, H. T. (Ed.) (1868). Memorials of London and London Life, in the XIIIth, XIVth, and XVth Centuries: From the Archives of the City of London, A.D. 1276–1419. London: Longman, Brown, and Green. Ritchie (née Kenyon), N. (1934). Labour Conditions in Essex in the Reign of Richard II. Economic History Review (1st Series, Vol. 4, p. 4). Reissued (1962) in revised form. In: E. M. CarusWilson (Ed.), Essays in Economic History (3 Vols, Vol. II, pp. 91–112). London: Edward Arnold. Robinson, W. C. (1959). Money, Population, and Economic Change in Late Medieval Europe. Economic History Review, (2nd Series), 12, 63–76. Rogers, J. E. T. (1866). A History of Agriculture and Prices in England From the Year After the Oxford Parliament (1259) to the Commencement of the Continental War (1793) (7 Vols). Oxford: Clarendon Press, 1866–1902. Vols I and II: 1259–1400.
4259 Ch06 4/12/02 10:12 am Page 295
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
295
Rogers, J. E. T. (1882). A History of Agriculture and Prices in England From the Year After the Oxford Parliament (1259) to the Commencement of the Continental War (1793) (7 Vols). Oxford: Clarendon Press, 1866–1902. Vols III and IV: 1401–1582. Rogers, J. E. T. (1903). Six Centuries of Work and Wages: the History of English Labour. London: S. Sonnenschein. Ross, A. S. C. (1956). The Assize of Bread. Economic History Review, (2nd Series), 9(2), 332–342. Saul, S.B. (1985). The Myth of the Great Depression, 1873–1896 (2nd rev. edn.). Studies in Social and Economic History. London: MacMillan. Schreiner, J. (1954). Wages and Prices in the Later Middle Ages. Scandinavian Economic History Review, 2, 61–73. Schwartz, A. J. (1974). Review of Spooner, International Economy and Monetary Movements in France, 1493–1725 (1972) in: Journal of European Economic History, 3(1), 252–255. Sharpe, R. R. (Ed.) (1905). Calendar of Letter-Books Preserved Among the Archives of the City of London at the Guildhall: Letter-Book G. c. A.D. 1352–1374. London J.E. Francis. Sharpe, R. R. (Ed.) (1907). Calendar of Letter-Books Preserved Among the Archives of the City of London at the Guildhall: Letter Book H., c.A.D. 1375–1399. London: J. E. Francis. Slicher-Van Bath, B. H. (1963). The Agrarian History of Western Europe, A.D. 500–1850. Translated by Olive Ordish. London: Edward Arnold. Smith, A. (1776). An Inquiry into the Nature and Causes of the Wealth of Nations, edited with introduction and notes by Edwin Cannan. New York: Modern Library, 1937. Smith, R. (1984). Families and Their Land in an Area of Partible Inheritance, 1260–1320. In: R. M. Smith (Ed.), Land, Kinship and Life-Cycle (pp. 135–196). Cambridge Studies in Population, Economy and Society in Past Time 1. Cambridge: Cambridge University Press. Smith, R. (1991). Demographic Developments in Rural England, 1300–1348: A Survey. In: B. M. Campbell (Ed.), Before the Black Death: Studies in the “Crisis” of the Early Fourteenth-Century (pp. 25–77). Manchester and New York, Manchester University Press. Sneyd, C. A. (Ed.) (1847). A Relation or Rather a True Account of the Island of England with Sundry Particulars of the Customs of These People, and of the Royal Revenues Under King Henry the Seventh, about the Year 1500 (Old Series), 37. Camden Society Publications, London: Camden Society: J. B. Nichols and Son. Sosson, J.-P. (1977). Les travaux de la ville de Bruges, XIVe–XVe siècles: les matériaux, les hommes. Collection Histoire Pro Civitate No. 48 Brussels: Crédit Communal de Belgique. Spooner, F. (1972). The International Economy and Monetary Movements in France, 1493–1725. Cambridge, Mass.: Harvard University Press. Spufford, P. (1970). Monetary Problems and Policies in the Burgundian Netherlands, 1433–1496. Leiden: E. J. Brill. Spufford, P. (1986). Handbook of Medieval Exchange. Royal Historical Society Guides and Handbooks No. 13. London: Royal Historical Society, University College London. Spufford, P. (1988). Money and Its Use in Medieval Europe. Cambridge: Cambridge University Press. Steffen, G. F. (1901–1905). Studien zur Geschichte der englischen Lohnarbeiter mit besonderer Berücksichtigung der Veranderungen ihrer Lebenshaltungen (3 Vols). Stuttgart: Hobbing. Stuard, S. M. (1999). Gravitas and Consumption. In: J. Murray (Ed.), Conflicted Identities and Multiple Masculinities: Men in the Medieval West. Garland Medieval Casebooks (Vol. 25, pp. 215–242). New York: Garland. Swanson, H. (1988). The Illusion of Economic Structure: Craft Guilds in Late Medieval English Towns. Past & Present, (121), 29–48.
295
4259 Ch06 4/12/02 10:12 am Page 296
296 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
JOHN H. MUNRO
Swanson, H. (1989). Medieval Artisans: An Urban Class in Late Medieval England. Oxford: Basil Blackwell. Thoen, E. (1988). Landbouwekonomie en bevolking in Vlaanderen gedurende de late Middeleeuwen en het begin van de Moderne Tijden: Testregio: de kasselrijen van Oudenaarde en Aalst (eind 13d - eerste helft 16d eeuw) (2 Vols). Belgisch Centrum voor Landelijke Geschiedenis No. 90. Ghent: Beschutte Werkplaats Nevelland. Thompson, J. D. A. (1956). Inventory of British Coin Hoards, A.D. 600–1500. Oxford: Royal Numismatic Society Publications. Titow, J. Z. (1972). Winchester Yields: A Study in Medieval Agricultural Productivity. Cambridge Studies in Economic History. Cambridge: Cambridge University Press. Tomlins, T. E., & Raithby, J. (Eds) (1810–1822). The Statutes of the Realm (6 Vols). London: Record Commission. Van der Wee, H. (1963). The Growth of the Antwerp Market and the European Economy, 14th 16th Centuries (3 Vols). The Hague: Martinus Nijhoff. Van der Wee, H. (1966). Voeding en dieet in het Ancien Regime. Spiegel Historiael, 1(2), 94–101. Republished (1993) in English translation as: “Nutrition and Diet in the Ancien Regime.” In: H. Van der Wee, The Low Countries in the Early Modern World (pp. 279–287). Aldershot and Cambridge: Cambridge University Press. Van der Wee, H. (1975a). Prijzen en lonen als ontwikkelingsvariabelen: Een vergelijkend onderzoek tussen Engeland en de Zuidelijke Nederlanden, 1400–1700. In: Album aangeboden aan Charles Verlinden ter gelegenheid van zijn dertig jaar professoraat (pp. 413–447). Ghent and Brussels: Economische Hogeschool Sint-Aloysius. Republished (1978) in English translation (without the tables, but with the graphs) as: “Prices and Wages as Development Variables: A Comparison Between England and the Southern Netherlands, 1400–1700.” Acta Historiae Neerlandicae (Vol. 10, pp. 58–78). And reprinted (1993) in: H. Van der Wee, The Low Countries in the Early Modern World (pp. 223–241). Aldershot and Cambridge: Cambridge University Press. Van der Wee, H. (1975b). Structural Changes and Specialization in the Industry of the Southern Netherlands, 1100–1600. Economic History Review, (2nd Series), 28, 203–221. Van der Wee, H. (1977). Monetary, Credit, and Banking Systems. In: E. E. Rich & C. H. Wilson (Eds), Cambridge Economic History of Europe, Vol. V: The Economic Organization of Early Modern Europe (pp. 290–392). Cambridge: Cambridge University Press. Van der Wee, H. (2000). European Banking in the Middle Ages and Early Modern Period (1476–1789). In: H. Van der Wee & G. Kurgan-Van Hentenrijk (Eds), A History of European Banking (2nd ed., pp. 71–266). Antwerp: Mercator Fonds. Van Uytven, R. (1961). La Flandre et le Brabant: “terres de promission” sous les ducs de Bourgogne? Revue du Nord, 43, 281–317. Reprinted (2001) in: R. Van Uytven, Production and Consumption in the Low Countries, 13th–16th Centuries, Variorum Collected Studies (Series CS 714, No. II). Aldershot: Ashgate Publishing. Van Uytven, R. (1966). La conjoncture commerciale et industrielle aux Pays-Bas: une récapitulation. In: J.-M. Duvosquel, J. Nazet & A. Vanrie (Eds), Les Pays-Bas Bourguignons: histoire et Institutions: Mélanges André Uyttebrouck. Brussels: Archives et Bibliothèques de Belgique (pp. 435–468). Reprinted (2001) in: R. Van Uytven, Production and Consumption in the Low Countries, 13th–16th Centuries, Variorum Collected Studies (Series CS 714, No. III). Aldershot: Ashgate Publishing. Van Uytven, R. (1981). Technique, productivité, et production au moyen âge: le cas de la draperie urbaine aux Pay-Bas. In: S. Mariotti (Ed.), Produttivita e tecnologia nei secoli
4259 Ch06 4/12/02 10:12 am Page 297
Wage Stickiness and Monetary Changes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39
297
XII–XVII (pp. 283–294). Florence: Felice Le Monnier. Reprinted (2001) in: R. Van Uytven, Production and Consumption in the Low Countries, 13th–16th Centuries, Variorum Collected Studies (Series CS 714, No. VII). Aldershot: Ashgate Publishing. Van Uytven, R. (1995). Stages of Economic Decline: Late Medieval Bruges. In: J.-M. Duvosquel & E. Thoen (Eds), Peasants and Townsmen in Medieval Europe: Studia in Honorem Adriaan Verhulst. Ghent (pp. 259–269). Reprinted (2001). In: R. Van Uytven, Production and Consumption in the Low Countries, 13th–16th Centuries, Variorum Collected Studies (Series CS 714, No. V). Aldershot: Ashgate Publishing. Van Werveke, H. (1931). De economische en sociale gevolgen van de muntpolitiek der graven van Vlaanderen (1337–1433). Annales de la Société d’Emulation de Bruges, 74, 1–15. Reprinted (1968) in: H. Van Werveke, Miscellanea Mediaevalia: Verspreide opstellen over economische en sociale geschiedenis van de middeleeuwen (pp. 243–254). Ghent: E. StoriaScientia. Van Werveke, H. (1938). De Vlaamse munthervorming van 1389–1990. Nederlandsche historiebladen, 1(3), 336–347. Reprinted (1968) in: H. Van Werveke (Ed.), Miscellanea Mediaevalia: Verspreide opstellen over economische en sociale geschiedenis van de middeleeuwen (pp. 268–280). Ghent: E. Storia-Scientia. Van Zanden, J. L. (1999). Wages and the Standard of Living in Europe, 1500–1800. European Review of Economic History, 3(2), 175–197. Verlinden, C., Scholliers, E., & Craeybeckx, J. (Eds) (1959–1965). Documents pour l’histoire des prix et des salaires en Flandre et en Brabant/Dokumenten voor de geschiedenis van prijzen en lonen in Vlaanderen en Brabant (2 Vols). Rijksuniveristeit te Gent: Werken Uitgegeven door de Faculteit van de Letteren en Wijsbegeerte Nos. 125 [vol. I], 136 [Vol II.i], 137 [Vol. II.ii]. Bruges: De Tempel. Westermann, E. (1972). Die Bedeutung des Thüringer Saigerhandels für den mitteleuropäischen Handel an der Wende vom 15. zum 16. Jahrhundert. Jahrbuch für die Geschichte Mittelund Ostdeutschlands, 21, 68–92. Westermann, E. (1986). Zur Silber- und Kupferproduktion Mitteleuropas vom 15. bis zum frühen 17. Jahrhundert: über Bedeutung und Rangfolge der Reviere von Schwaz, Mansfeld und Neusohl. Der Anschnitt: Zeitschrift für Kunst und Kultur im Bergbau, 38 (May–June), 187–211. Wilton, D. A., & Prescott, D. M. (1987). Macroeonomics: Theorv and Policy in Canada (2nd ed.). Toronto: Addison Wesley Publishers. Wood-Legh, K. L. (1956). A Small Household of the Fifteenth Century. Manchester: Manchester University Press. Woodward, D. (1981). Wage Rates and Living Standards in Pre-Industrial England. Past & Present, (91), 28–46. Woodward, D. (1994). The Means of Payment and Hours of Work in Early Modern England. In: C. S. Leonard & B. N. Mironov (Eds), Hours of Work and Means of Payment: The Evolution of Conventions in Pre-industrial Europe (pp. 11–21). Proceedings of the Eleventh International Economic History Congress, Milan, September 1994, Session B3b. Milan: University of Bocconi Press. Woodward, D. (1995). Men at Work: Labourers and Building Craftsmen in the Towns of Northern England, 1450–1750. Cambridge: Cambridge University Press. Wrigley, E. A., & Schofield, R. S. (1989). The Population History of England, 1541–1871: A Reconstruction (2nd ed.). Cambridge: Cambridge University Press.
40
297
4259 Ch07 4/12/02 10:19 am Page 299
1 2 3 4 5 6 7 8 9 10
WHAT PROTECTED PEASANTS BEST? MARKETS, RISK, EFFICIENCY, AND MEDIEVAL ENGLISH AGRICULTURE
11 12 13 14
Gary Richardson
15 16 17 18 19 20
ABSTRACT
21 22 23 24 25 26 27 28 29 30 31 32
McCloskey’s model of medieval English agriculture shapes the debate over the origins and persistence of the open fields. This essay deconstructs McCloskey’s often-cited analysis and demonstrates that McCloskey’s assumptions contradict her conclusions. McCloskey’s claims concerning the costs of various methods of mitigating risk are inconsistent with her vision of autarkic villages where scattering protected peasants from adverse agricultural shocks. Instead, McCloskey’s claims corroborate an alternative markets-plus-morals view of medieval English villages. Correcting McCloskey’s erroneous inferences resurrects long forgotten explanations for the persistence of the open fields and suggests a new agenda for the study of medieval English agriculture.
33 34 35 36 37 38 39 40
Research in Economic History, Volume 21, pages 299–356. Copyright © 2003 by Elsevier Science Ltd. All rights of reproduction in any form reserved. ISBN: 0-7623-0993-8
299
4259 Ch07 4/12/02 10:19 am Page 300
300 1
GARY RICHARDSON
INTRODUCTION
2 3 4 5 6 7 8
There are very good explanations about why peasants did not starve when there were good crops . . . charity . . . wages . . . credit . . . . . . I am not sure where economists came up with the idea that peasants did not practice charity . . . [And] there were other ways for peasants to get food, especially in the post Black Death period, that make me very suspicious of the underlying assumptions of the theory Barbara Hanawalt, 1995.1
9 10 11
A most instructive example of the type of specious reasoning by which we often delude ourselves when defending a proposition that, from habit, we have come to believe needs no defense at all
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Joseph A. Schumpeter, 1954 (1986, p. 673).
Deirdre McCloskey paints a bleak picture of life in medieval English villages (McCloskey, 1972, 1975a, 1975b, 1976, 1977, 1979a, 1979b, 1980, 1983, 1989, 1991; McCloskey & Nash, 1984). Famines frequently ravaged rural populations. Peasants lived near the edge of subsistence. Compact calamities such as hailstorms, insect infestations, and wandering herds of hungry cattle obliterated arable fields and pushed peasants into the abyss. Peasants were selfish, individualistic, and opportunistic. Their outlook precluded cooperation and benevolence. The absence of legal and political institutions inhibited collective action. Peasants seldom practiced charity, pooled risk, or helped neighbors in need. Inchoate markets for food, labor, land, and chattels could not offset idiosyncratic shocks. Insurance, credit, and contingent contracts did not exist. Peasants scattered arable holdings throughout villages. Ten percent less grain grew in scattered strips than consolidated counterparts. The loss of one bushel in ten translated into “something under ten percent of national income, but not far under” (McCloskey, 1989, p. 25). Why would society waste 10% of aggregate income? Why would rational peasants throw away 10% of the harvest year after year? Why would such inefficiency persist? McCloskey explains why. First, some of the crops scattered in acre-sized strips throughout villages the size of Central Park would survive local disasters that destroyed all of the crops in consolidated fields. In terms of portfolio theory, scattering diversified farmers’ holdings among many small plots facing different weather, weed, rodent, insect, and soil conditions. Diversification reduced the risk each farmer faced as measured by the variance of total output. If one plot did poorly and another did well, from all of his plots put together a farmer could still harvest enough grain to survive. Second, conditions varied over short distances enough to reduce significantly the variance of portfolios of scattered strips held by prosperous peasants – those
4259 Ch07 4/12/02 10:19 am Page 301
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
301
possessing twenty acres or more. Poorer peasants without large landholdings did not benefit from scattering, but “the share of the land they held was so small that they had little voice in the layout of the open fields” (McCloskey, 1989, p. 6). Third, alternative methods of mitigating risk – such as storing grain – were exorbitantly expensive or did not exist. The annual interest rate exceeded thirty percent, indicating the opportunity cost of storing grain was far higher than the opportunity cost of scattering strips and suggesting that the costs of other alternatives – charity, credit, crop diversification, extended families, kinship networks, and so on – exceeded the costs of scattering furlongs throughout open fields. Finally, peasants enjoyed life, feared death, and therefore, adopted the lowvariance, low-yield methods of cultivating crops that minimized the risk of starvation. McCloskey’s appealing analogy to the investment principles of diversification and trading return for risk shaped her conception of medieval agriculture. But the brilliance of her rhetoric obscured ideas inconsistent with the autarkic image of medieval England that captivated Cliometricians during the 1970s. That dismal place existed only in their minds. Recent research has revealed the true state of affairs. Food, land, labor, and credit markets filled medieval English villages. Victuals were readily available, and staples were cheap enough to be purchased by peasants, who could raise the necessary funds by selling assets, acquiring credit, or earning wages. Local courts enforced an array of intertemporal contracts. These allowed peasants to purchase annuities in cash and kind, conclude agreements that amounted to futures for grain, and lease out land and animals in arrangements that amounted to selling risk. The enforcement of contracts and density of markets permitted prudent peasants with fluctuating incomes to smooth consumption by accumulating savings, substituting labor for leisure, working in other occupations independent of idiosyncratic agricultural shocks, and shifting consumption across time and individuals. Risk-sharing institutions also existed in medieval English villages. Fraternities supported indigent members and defrayed funeral costs (Bainbridge, 1996; Barron, 1985; Blair, 1980; Brigden, 1984; Britton, 1977; Burgess, 1991; Champion, 1980; Cook, 1947; Cornwall, 1970; Crouch, 2000; Duparc, 1975; Fryde, 1985; Hanawalt, 1984; Hanawalt & McCree, 1992; Hicks, 1985; Homans, 1941; Jones, 1974; Mattingley, 1989; McRee, 1993; Palmer, 1902; Richardson, 1999; Rosser, 1988b; J. Smith 1870; Westlake, 1919). Kinship networks also supported those facing hard times (Bossy, 1973; Dinn, 1990; Razi, 1981). Villages coordinated customary poor support. One ubiquitous custom permitted poor villagers to pick peas at any time of the year from any field in the village, 301
4259 Ch07 4/12/02 10:19 am Page 302
302 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
GARY RICHARDSON
even those owned by their neighbors. Another permitted poor villagers to glean grain spilled during the harvest. A third allowed villagers facing hard times to graze more than their share of animals on the commons, wastes, and meadows. A fourth settled able-bodied poor men on small dower plots of land on the holdings of other men, usually relatives (Ault, 1972; Page, 1930; Richardson, 1999; Titow, 1969). Even secular landlords occasionally acted charitably. The account roll of Langenhoe manor reveals “liveries of grain to the poor and other small gifts of grain . . . in elemosina or de dono” (Britnell, 1996, p. 381). A network of charitable institutions operated throughout the realm. The continuing cult of the dead promised spiritual rewards to those who helped the unfortunate. The church espoused the superstition, and ecclesiastic organizations dispensed abundant alms. Monasteries were particularly prolific. Cathedrals, colleges, abbeys, nunneries, hospitals, and almshouses followed close behind. So did parish churches, which established charitable traditions such as ‘days of holy loaves’ when parishioners gave paupers loaves of freshly baked bread and which possessed poor boxes from which patriarchs subsidized the old, infirm, and indigent. Recent studies reveal the scope and scale of charitable activity in medieval England (J. M. Bennett, 1992; Brigden, 1984; Burgess, 1987, 1988; Dyer, 1994; Finucane, 1981; Heal, 1984; Hicks, 1985; McIntosh, 1998; Rubin, 1987; Swanson, 1989; Wright, 1988). These studies build on earlier archival research that illuminates the abundance of the evidence (Ault, 1967, 1970; Clay, 1909; Cobbett, 1935; Gasquet, 1907; Godfrey, 1955; Jones, 1974; Jordan, 1959, 1960, 1961; Mason, 1976; Moorman, 1945; Rosenthal, 1972; Thompson, 1965; Tierney, 1959; Vale, 1976; Wrightson & Levine, 1979). Scattering farm fields did not reduce crop yields. Data from Arthur Young’s tours demonstrates that yields of scattered strips were as high or higher than yields of consolidated fields as late as 1770 (Allen, 1982; Allen & Grada, 1988). Data from the Charity Commission reports reveals that consolidating common fields raised rents by a few percent at most, and before 1720, that increase paid primarily for the costs of conversion and investments in fences, roads, and drains (G. Clark, 1998). Bruce Campbell’s study of manorial accounts from the pre-plague era demonstrates that cereal and legume yields from the thirteenth century, when most arable land lay in open fields, were not surpassed until the seventeenth or eighteenth centuries, when most arable had been enclosed (Campbell, 1983, 1991, 1993a, 1993b, 1995, 1997). Eona Karakacili’s pioneering work with 35 account rolls from the Ramsey estate for the years 1280 to 1348 proves a similar point. “Labor productivity rates in open fields either equal or surpass the figures produced by Clark for nineteenth century workers in English enclosed farms” (Karakacili, 2000, p. 4).
4259 Ch07 4/12/02 10:19 am Page 303
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
303
This essay corroborates those claims. Section I summarizes recent research on mortality, markets, and medieval English villages. The synopsis refutes key assumptions of McCloskey’s model of medieval agriculture by proving two points. First, peasants mitigated adverse shocks in many ways including the storage of grain and the pooling of risk. No method was perfect. When combined together however, they were surprisingly effective. Prosperous peasants seldom starved in medieval England, and when they did die from hunger, they did not die alone. Famines occasionally ravaged rich and poor alike, but idiosyncratic agricultural calamities had little observable impact on the lives and deaths of peasants who tilled the arable lands of medieval England. Second, the market was an important mechanism for mitigating risk. The volume of transactions in markets for food, labor, land, chattels, and credit could have mitigated many times over all of the idiosyncratic shocks that transcend McCloskey’s model of medieval agriculture. Patterns of exchange suggest that markets did so. The frequency of exchange increased during difficult times, when unlucky individuals traded assets and labor for the necessities of life, and fell during typical years, when individuals gradually accumulated reserves. Section II demonstrates that even if McCloskey’s assumptions were accurate, her conclusions would remain incorrect. Her estimates of the cost of storing grain were high, but not high enough to justify her claims. A mathematical model proves this point. It examines the relationship between the costs of storage and the optimal arrangement of arable holdings. This exercise reveals the quantitative miscalculations that led McCloskey to her incorrect conclusions. Section III extends the model by examining the behavior of peasants who simultaneously stored grain and pooled risk. This exercise reconciles recent research on agricultural institutions with recent research on agricultural productivity and costs. Section IV concludes with an observation and an inference. McCloskey’s Coasian insight – her realization that inefficient institutions persist only when market failures persist and that dense markets engender efficient institutions – was correct but misapplied. The conventional historical wisdom of the 1970s, which formed the foundation of McCloskey’s historical arguments, has been revised. Historians no longer believe that open fields were inefficient, markets were absent, villages were autarkic, and peasants were inhumanly individualistic. Now historians know markets were ubiquitous. Villages were embedded in a regional economy. Open fields were efficient. And peasants were ordinary people who on some occasions cooperated and on others competed. These new facts fit Coase’s theorem better than the anachronistic notions of the 1970s, and they open new avenues for the study of institutions and English economic development. 303
4259 Ch07 4/12/02 10:19 am Page 304
304 1
GARY RICHARDSON
I. MORTALITY AND MARKETS
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
The scarcity of folklore concerning hunger and famine and the affluence of medieval England (compared to developing nations today) inspired Alan Macfarlane’s search for The Origins of English Individualism (1979, p. 4). Anglo-Saxon individuality, Macfarlane concluded, arose from the labor, land, asset, commodity, and credit markets that flourished from an early date in the British Isles, the property rights, legal institutions, and strong central government that permitted those markets to thrive, and the social mobility, occupational specialization, nuclear families, acquisitive attitudes, and extraordinary affluence that those markets created. Recent research substantiates Macfarlane’s suppositions. Famines were rare. Markets were everywhere. This section reviews the essential evidence. Historical demography generates clear conclusions. Medieval Englishmen seldom experienced starvation. Nationwide famines occurred only during the century before the Black Death when the growth of the population and deterioration of the climate destabilized the demographic balance. By far the worst was the Great Famine (Kershaw, 1973; Lucas, 1930). Between 1315 and 1322, extreme and unpredictable weather destroyed much of the harvest. Sheep scab, cattle plagues, and wars with the Scots exacerbated the crisis. Prices of foodstuffs rose to unheard of heights. Heriots paid on Winchester manners indicate crude death rates approached 10% with the deaths concentrated among peasants with little or no land (Postan & Titow, 1973). “The bodies of paupers, dead of starvation, littered the streets” (Kershaw, 1973, p. 11). Chronicles mention no comparable famine since the Dark Ages. Regional subsistence crises occurred, of course, from time to time. Chronicles mention 12 regional famines and nine cattle plagues between 1066 and 1300. The murrains of 1086, 1103, 1201, and 1258 accompanied crop failures. The others, 1111, 1131, 1225, 1277, and 1283 did not (Kershaw, 1973, p. 26). The manorial accounts of the Winchester manors reveal three years between 1400 and 1450 when low grain yields coincided with high grain prices. The years run from harvest 1401 to 1402, 1416 to 1417, and 1438 to 1439 (Farmer, 1991b, pp. 501–508). The north in particular suffered during the 1430s and again during the following century. In 1597, dearth afflicted Cumberland and Westmoreland. In 1623, hunger haunted Lancashire. In 1649, dearth again afflicted the far north. But, those local food crises did not discernibly raise mortality in the rest of the realm, and during high-grain-price years of the sixteenth century (the 1520s, 1557, and 1588) when famines ravaged France, mortality increased little, if at all (Appleby, 1978, p. 866). “After 1649, England was free from subsistence crises” forever more (Appleby, 1978, p. 867).
4259 Ch07 4/12/02 10:19 am Page 305
What Protected Peasants Best? 1 2 3 4 5 6 7 8
305
Tables 1, 2, and 3 summarize mortality data derived by demographic historians from parish registers and manorial records. Table 1 illustrates Ron Lee’s conclusion that “great swings in population during this period were due to . . . exogenously varying mortality” (Lee, 1973, p. 605). In other words, virulent diseases such as the Black Death regulated the “real wage and population size of pre-industrial England” (Lee, 1973, p. 606). The logic underlying Lee’s conclusion arises from the relationship between the crude death and mortality rates. Death rates increased when real wages were high, perhaps
9 10 11
Table 1.
12
Population, Mortality, Fertility, and Wages in England 1250 to 1750. Averages Over 50 Year Intervals.
13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
50 Year Interval Begin 1251 1301 1348 1401 1451 1501 1551 1601 1651 1701
End 1300 1347 1400 1450 1500 1550 1600 1650 1700 1750
Population (millions)
Crude Death Rate
Crude Birth Rate
Real Wage Index (1451–1471 = 100)
3.5 4.0 3.1 2.3 2.6 3.2 4.3 5.4 5.8 6.0
0.0323 0.0361 0.0486 0.0470 0.0352 0.0301 0.0271 0.0305 0.0329 0.0297
0.0381 0.0361 0.0367 0.0486 0.0383 0.0351 0.0347 0.0334 0.0329 0.0310
53.9 55.0 67.0 93.2 98.7 75.2 54.5 40.6 49.3 62.2
Source: Lee, 1973, pp. 606–607, Appendix: Basic Data Series.
28 29
Table 2.
30
The Positive Check. Grain Prices and Mortality in England 1546–1870.
31 32
Non-Infant Death Rate
Crude Death Rate
Crude Death Response to 21.4% Grain Price Increase
0.0207 0.0230 0.0188
0.0264 0.0293 0.0249
0.002 0.019 0.005
33 34 35
Period
Percent Variation in CDR Explained by Percent Variation in Grain Price 0.11 0.19 0.08
36 37 38
1546 to 1674 1675 to 1755 1756 to 1870
39 40
Source: Galloway, 1998, p. 281, Table 1, and pp. 292–293, Table 2, Column 12.
305
4259 Ch07 4/12/02 10:19 am Page 306
306
GARY RICHARDSON
1 2
Table 3.
3 4
Demographic Regimes in Late-Medieval and Early-Modern England.
5
Mortality
Famines
6 All Causes
7 8 9
(1)
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Pre-Plague Prosperity Pre-Plague Crisis Plague, Epidemic Plague, Endemic Post-Plague
1100 1275 1276 1347 1348 1450 1451 1670 1670 1750
Starvation Prosperous Poor Peasants Peasants (2) (3)
National
(4)
Regional Core
Periphery
(5)
(6)
to
Low
Absent
Rare
Absent Infrequent Infrequent
to
Medium
Rare
Occasional
to
High
Absent
Rare
Absent
Absent
Infrequent
to
Low
Absent
Rare
Absent
Absent
Infrequent
to
Low
Absent
Rare
Absent
Absent
Infrequent
Repeated Repeated Repeated
Definitions: Mortality refers to deaths of individuals. Mortality rates due to starvation indicate the percent of the population dying during non-famine years from malnourishment or diseases that prey on the malnourished. Famine rates describe the frequency of famines during the indicated eras and regions. Mortality rates during famine years never exceeded 5% nationwide for prolonged periods. Column (1): High indicates annual mortality from all causes exceeded 4.5 per 1000. Medium indicates mortality was between 4.5 and 3.5 per 1000. Low indicates mortality was below 3.5 per 1000. Columns (2) and (3): Absent implies there is no quantitative or anecdotal evidence that peasants starved when their neighbors had enough to eat. Rare indicates that historians and demographers have found a few cases in which individuals – usually paupers or vagrants – died when the surrounding community had food to spare. Occasional indicates that historians have found substantial quantitative and anecdotal evidence of starvation usually of vagrants, paupers, cottars, marginal peasants, the young, and the old in chronicles, personal narratives, coroners’ rolls, manorial accounts, court records, and other government documents. Columns (4), (5), and (6): Absent indicates no records exist of famines encompassing the indicated area. Infrequent indicates subsistence crises occurred once in every 20 years or less often at some point in the indicated region and never encompassing the entire region. Repeated indicates that subsistence crises afflicted the entire region once in every 20 years or more often. Evidence of famines includes narrative and quantitative records of crop failures and their causes, impoverished individuals seeking sustenance at higher than normal rates, correlations between high grain prices and high mortality rates, increases in the price of most foodstuffs, and deaths due to starvation, diseases aggravated by malnutrition, and exposure. Sources: Described in text.
4259 Ch07 4/12/02 10:19 am Page 307
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
307
because extensive trading networks facilitated the spread of infectious diseases. Mortality fell when population growth reduced real wages. The Malthusian positive check would have produced the opposite pattern. Mortality would have increased when overpopulation required the cultivation of ever more marginal land and lowered wages towards subsistence levels. Table 2 illuminates Patrick Galloway’s conclusion about the short-run relationship between grain supplies and human mortality. Variations in the price (and thus the supply) of grain had little influence on year-to-year variation in non-infant and crude death rates. A one-standard-deviation increase in the price of grain coincided with an increase of one-fifth of one percent increase in the crude death rate before 1670, a two percent increase between 1670 and 1756, and a one-half of one percent decline thereafter. “England was unique in being the first country in Europe, and probably the world, to escape the bonds of the positive check” (Galloway, 1998, p. 298). Table 3 compares the conclusions that Lee and Galloway draw from aggregate time series to those that M. M. Postan and J. Z. Titow drew from micro data. When grain prices rose, crude death rates increased with the additional death concentrated among the old, young, and poor. Few prosperous peasants perished, and most of those that died left land and chattels to their descendants, suggesting they died of the epidemics that accompanied dearth rather than poverty or starvation. Hanawalt’s analysis of coroners’ inquests from the counties of Norfolk, Cambridge, Lincoln, Northampton, Bedford, and Wiltshire reaches similar conclusions. Hunger plagued the poor during most years, and starvation stalked the poor when crops failed. “Secondary villagers might find themselves struggling for survival during famine years . . . for the cottars, hunger and high infant mortality precluded more than hoping for work, begging, and stealing” (Hanawalt, 1986, p. 6). But prosperous peasants seldom starved, and when they did during famines, they did not starve alone. The idiosyncratic shocks underlying the Cliometric conception of medieval agriculture, in other words, had no observable influence on the lives (and deaths) of ordinary Englishmen. Food was available in all rural communities. Grain and ale were always for sale. Peasants could buy staples from their neighbors or the victuallers that lived in their village. Perhaps 2 to 4% of all households processed commodities professionally, buying ingredients and selling finished foodstuffs. Another 5 to 15% of all households processed food periodically, usually when they had surpluses to sell. The fines assessed on bakers and brewers for violating the Assizes of Bread and Ale prove those facts. So do occupational data derived from tax, tithe, and court records (J. M. Bennett, 1987, pp. 12, 208–213; Britnell, 1986, p. 382; DeWindt, 1990, fiche 1, p. 79; Dyer, 1989, pp. 151–160). 307
4259 Ch07 4/12/02 10:19 am Page 308
308 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
GARY RICHARDSON
In the unlikely event that individuals could not buy food in their village, they could shop in licensed markets. More than 1,200 weekly markets existed. They were sprinkled throughout the countryside, particularly along coasts, waterways, and principal thoroughfares. Most peasants lived within easy reach of at least one market (J. M. Bennett, 1987, p. 52; Britnell, 1981; Hilton, 1982, pp. 5–6; Waugh, 1991, p. 41). On average, the distance from a peasant’s house to the nearest market was 4.2 miles, less than a two-hour walk. More than nine out of 10 peasants lived within six miles of a market. Most lived within that distance of multiple markets (Farmer, 1991a, pp. 331–341). In addition to licensed markets, there were innumerable unlicensed emporiums, thousands of fairs where merchants sold grain as well as goods shipped from the farthest corners of Europe, and hundreds of market towns where retailers sold victuals as well as manufactured merchandise. Municipal documents and royal records such as the returns from the Poll Tax of 1381 reveal “the distribution of occupations [was] much the same in all of the towns . . . there [was] a clear preponderance of retailers of food and drink and articles of clothing” (Hilton, 1975a, pp. 80, 83, 88). Food was abundant because most of the population lived in the countryside, cultivated crops, and sold surplus produce, especially grain. Shipments of grain traveled hundreds of miles from point of cultivation to point of sale. Evidence appears in innumerable documents. Ault describes the documents available for the study of medieval English villages (1967). Dyer surveys the documentary evidence available for the study of the medieval English countryside (Dyer, 1988). The Domesday Book suggests “60% of net demesne output in 1086 was marketed” (Snooks, 1994, p. 37). Net demesne output was output after deducting seed, fodder, and other intermediate goods. Toll records indicate towns taxed tremendous quantities of grain, eggs, cheese, and other victuals in transit. Judicial records indicate merchants bought large quantities of grain from many sources. For example, during a three-year period, thirteen Tewksbury cornmongers purchased 600 quarters of grain in the neighboring villages of Staverton, Deehurst, Oxenton, and Haw and 3,340 quarters of grain from further afield. They sold some of this grain in their town and shipped the rest down river to Bristol. During the same period, merchants from Gloucester dealt in larger quantities. They bought at least 2,000 quarters in nearby villages and 11,000 quarters from further afield (Hilton, 1975b, p. 89). Studies of estate accounts reveal similar transactions. The Bishop of Winchester’s estates sold significant quantities of grain from the date of the earliest written records (Farmer, 1957, 1958, 1969, 1989; Titow, 1972). Ecclesiastic and secular estates in every corner of the realm sold grain throughout the thirteenth, fourteenth, and fifteenth centuries (Beveridge, 1939;
4259 Ch07 4/12/02 10:19 am Page 309
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
309
Booth, 1981; Gras, 1926; Phelps-Brown & Hopkins, 1955, 1956; Rogers, 1884, 1888; Rogers & Rogers, 1866). From 1288 to 1315, manors in 10 southeastern counties sold 39.8% of their crop yield after subtracting tithe, tax, and seed. From 1375 to 1400, manors from the same counties sold 36.1% of their net crop yield. If fodder and food liveries to farm workers are also netted out, these averages rise to 50%. Counting tithes as marketed grain would add another 10% to the averages. This addition would be logical. Tithes contributed to the volume of crops on the market. Since tithed grain greatly exceeded the consumption requirements of parish clerks and their staffs, most of it was sold (Campbell, 1996). Even the smallest estates sold large quantities of grain. During the fourteenth century in Essex, the minor manor of Langenhoe sold 29.9% of its wheat for cash after saving 21.4% for seed, paying 18.9% to workers as wages, and consuming 22.8% in the manorial household. Some grain sold from the lord’s barns never left the village. It was bought by his tenants, or the landless who grew no grain of their own. Other grain was sold in the nearby town of Colchester. At least half of the sixty plus people who purchased grain from Langenhoe between 1324 and 1348 were townsfolk. The number of small sales to Colchester men shows convincingly that the sergeant put in a regular appearance at the Colchester market. He claimed twopence for his expenses in Colchester in 1344–1345. The sergeant also sold grain on credit and accepted advance payments for large purchases of grain that buyers would retrieve from the manor at a later date (Britnell, 1966). Four smaller estates northeast of Langenhoe also emphasized production for the market. They sold large quantities of grain and relied on the resulting revenues to pay workers, finance investments, and generate profits (Britnell, 1980). This finding supports Kosminsky’s generalization, based on the broad but coarse data contained in the Hundred Rolls, that small estates readily adapted themselves to the rigors of the market (Kosminski, 1956). The grain market’s magnitude can be gauged by the following facts. During the late thirteenth century, at least 10% of England’s population – aristocrats, clerics, bureaucrats, merchants, craftsmen, and other urban residents – did not grow their own food. Another 40% held some arable land but not enough to satisfy their nutritional needs. These small holders bought, on average, half the grain they needed to survive. Everyone had to consume two pounds of grain a day to keep flesh on their bones. That amounted to 12 bushels of grain per individual per year. England’s population was 5,000,000. Thus, to feed those who did not grow enough of their own food, 18,000,000 bushels of grain must have been marketed between one harvest and the next. While grain sustained most men, meat supplemented most diets. Even the poorest possessed chickens and pigs. Prosperous peasants also raised cattle and 309
4259 Ch07 4/12/02 10:19 am Page 310
310 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
GARY RICHARDSON
sheep. Cows and lambs provided manure which fertilized arable land, leather which craftsmen used to make equipment and clothing, wool which weavers used to make cloth, and milk which could be churned into butter, turned into cheese, or drunk from the pail. Dairy products were so popular that landlords often rented animals. Cows were usually leased on a yearly basis; leasing costs doubled if cows were held only for the five months beginning in May and ending in September when milk was plentiful and fodder abundant. Sheep leases lasted similar lengths. Horses were hired at daily rates for purposes of travel and plowing. Usually only the oldest, least productive animals found the way to the dinner table, except in regions where cattle were raised for meat. Ranching clustered in northern and western Britain where poor soil favored husbandry over tillage. Since most of the population lived in southern and eastern England, where fertile soil favored the cultivation of crops, ranchers drove their herds to the midlands, fattened them on grass and grain, and sold them to victuallers. Evidence of the trade in cattle appears in accounts from the manors of the Bishop of Winchester, eighteen manors of the earldom of Norfolk, eleven manors of Merton College, eight manors of Westminster Abbey, seven manors of Crownland Abbey, twenty manors belonging to smaller estates, and the Exchequer Pipe Rolls (Farmer, 1969). Bruce Campbell’s data from 173 manors in the 10 southeastern counties demonstrates that manorial lords sold significant quantities of cattle, sheep, and pigs (Campbell, 1995, p. 163–164). Sheep took a similar journey when they stopped producing worthwhile wool. Shepherds sheared them one last time and sold them for mutton. The sheep population was vast. Estate accounts recorded flocks of up to ten thousand sheep. Customs accounts reported wool shorn from six to nine million sheep being exported annually. If a sheep produced worthwhile wool for twenty years on average, then from 300,000 to 450,000 sheep went to the slaughterhouse every year, yielding millions of pounds of mutton (Beardwood, 1931, p. 160; Carus-Wilson, 1963, pp. 185–189; Grant, 1988, p. 151; Power, 1941, p. 29). While cows and sheep were common animals, most men ate fish far more frequently than meat or mutton. Herring and cod caught in the North Sea and off the Scandinavian coast were common provender. Cod was dried, and herring salted or smoked (Grant, 1988, p. 172–173). Preserved fish was shipped far inland and eaten in large quantities, judging from the heaps of sea-fish bones found in peasants’ rubbish pits. Written records from monastic kitchens also report the consumption of vast quantities of fish and reveal the variety of foods that could be bought including luxuries such as spices, wine, fruit, and vegetables and staples such as cheese, cider, corn, eggs, beans, peas, meat, fish, fowl, and ale (Harvey, 1993, pp. 34–71, 216–230). A similar spectrum of food appears in aristocratic and ecclesiastic account books (Dyer, 1994, p. 56; Wood-Legh, 1965, pp. 45–56).
4259 Ch07 4/12/02 10:19 am Page 311
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
311
So, food was widely available. The market could easily supply the 60 bushels of barley that the typical prosperous peasant needed to feed his family, and the 30 bushels of barley, 10 bushels of wheat, six bushels of peas, and two bushels of oats that he needed to replant his fields. At late-thirteenth century prices – nine pence per bushel of wheat, six and a half pence per bushel of barley, four pence per bushel of peas, and four pence per bushel of oats – those purchases cost nearly £3. The typical peasant did not keep that much cash on hand, but he could raise it by selling assets. Peasants’ most valuable possession was arable land. Values varied, of course, based on factors such as soil fertility, feudal obligations, proximity to markets, and prevailing economic and climatic conditions. Some land sold for less than £1 an acre. Other land sold for more than £3 per acre. Thus, to completely offset calamities that consumed all his crops, an unlucky holder of 20 acres had to sell from one to three acres of arable land – that is, from one to three of their strips scattered throughout the open fields. Selling strips was simple. Finding buyers posed no problem, since land was in great demand. The procedure for transferring title was simple. The seller and buyer simply appeared at a meeting of the manorial court, asked the clerk to include the transfer in the court record, and paid the appropriate fee. Hundreds of manorial court rolls survive today, and scholarly studies of them reveal land markets active enough to have offset all of the shocks predicted by McCloskey’s model. One of the most detailed studies is Judith Bennett’s investigation of late-thirteenth and early-fourteenth century Brigstock, a village with four hundred forty-five households. Several dozen households altered the size or configuration of their landholdings every year. An average of 1.2 land transfers occurred in each court session from 1301 to 1303; an average of 2.6 land transfers in each session during the famine years of 1314 to 1317; an average of 2.3 land transfers in each session from 1331–1333; and an average of 3.4 transfers in every session from 1343 to 1345. Since Brigstock’s court met every third week, those sales could have smoothed all of the idiosyncratic shocks predicted by McCloskey’s model (Bennett, 1987, pp. 33–36, 56–57, 207). Another detailed study is Bruce Campbell’s examination of the small lay manor of Hakeford Hall in Coltishall, a township seven miles north of Norwich. Hakeford Hall’s court rolls recorded an average of 17 inter-vivos land transactions each year from 1275 to 1405. Those transfers could have smoothed the idiosyncratic shocks predicted by McCloskey’s model twice over. In addition, those transfers revealed a harvest-related pattern that began in the last quarter of the thirteenth century and continued until the Black Death. When harvests failed, the number of transactions increased. During the famine years of 1314 to 1317, the number of transactions multiplied three-fold and the number of 311
4259 Ch07 4/12/02 10:19 am Page 312
312 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
GARY RICHARDSON
individuals participating in the market multiplied two-fold. “In most cases peasants were selling land to buy food” (Campbell, 1984, p. 114). Thus, Campbell concluded “the very readiness with which the peasantry was able to liquidize its assets, selling off land to buy food, may have been one of the factors which gave the community so much resilience” (Campbell, 1984, p. 120) and “at Coltishall this [the land] market provided a kind of security against risk, enabling peasants to raise cash or credit at little notice and thereby weather periods of often acute economic hardship” (Campbell, 1984, p. 129). Campbell’s conclusion coincides with other studies that show during famines in the early-fourteenth century, the volume of land sales increased in East Anglian villages and landholding became increasingly unequal as poor people sold parcels of land to their wealthy neighbors to keep themselves alive for another year (Dyer, 1994, pp. 123–125). Campbell’s conclusion also coincides with a study of Martham, a manor in Norfolk with three hundred sixty-four tenants. Martham’s court certified twenty-one to thirty-seven land transfers each year involving as buyers or sellers on average one-tenth of Martham’s households. This percentage could also have alleviated all of the idiosyncratic shocks predicted by McCloskey’s model (Dyer, 1994, p. 123). Selling land was not the only way peasants could raise cash. Renting was another option. Leasing was ubiquitous, because “leasing, as long as the term ran for only a few years, was tolerated on virtually all seigniorial estates” (Bennett, 1987, p.56). Evidence from many manors proves that fact. For example, on the manor of Bishop’s Cleeve at the end of the thirteenth century, an official composing a survey of peasants’ holdings noted that “the bond tenants have under them many who pay them rent” (Bennett, 1987, p. 64). On the royal manor of Havering in 1251, a survey found that 113 out of 367 heads of households (31%) held land as subtenants without holding land directly from the lord. Another 5% held land as subtenants and also held land directly from the lord. A survey in 1352 indicates 301 out of 493 residents (61%) held land as subtenants but not from the lord, and of those who held land from the lord, 69% also held land as a subtenant of someone else on the manor (McIntosh, 1986). Similar surveys show subleasing was common in Bedford and Essex (Dyer, 1994, pp. 120–121; Hilton, 1983; Kerridge, 1969, pp. 48–49). From such evidence, J. Z. Titow concludes, “there can be no doubt that there has always been a great deal of land transference between peasants, either by permanent alienation or through short-term and long-term leases” (Titow, 1972, p. 3). While some peasants lacked land, all possessed chattel that could be converted into cash. Table 4 displays the average number of animals possessed by peasants on four thirteenth-century manors and lower-bound estimates of the value of those beasts. Table 5 presents Dyer’s estimates of peasants’ chattels and market
4259 Ch07 4/12/02 10:19 am Page 313
What Protected Peasants Best?
313
Table 4.
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Number of Animals Per Tax-Payer or Tenant on Four Thirteenth-Century Manors Horses and Oxen Cattle Sheep Pigs Wiltshire, 1225 Blackbourne Hundred, Suffolk, 1283 Liberty of Ramsey, 1291 Mickleover, Derbyshire, 1280
1.8 1.0 2.4 1.7
2.8 3.2 4.5 1.3
15.6 10.5 6.2 12.6
Value in Pence of Animals Per Tax-Payer or Tenant Horses and Oxen Cattle Sheep Wiltshire, 1225 Blackbourne Hundred, Suffolk, 1283 Liberty of Ramsey, 1291 Mickleover, Derbyshire, 1280
124 59 92 103
151 234 321 108
79 162 74 110
0.3 1.4 3.8 1.9
Pigs
Total
12 43 127 67
366 498 614 388
15 16 17 18 19 20
Notes: The price for ‘horses and oxen’ is taken to be the lowest price of oxen, plough-horses, and carthorses in Farmer’s Table 2. The price for sheep is taken to be the lowest of muttons and ewes in the same table. For years which Farmer’s table does not provide prices, the price is taken to be the price in the preceding year. The calculations are rounded to the nearest pence. Sources: Table adopted from Dyer, 1994, p. 124, Table 10; Farmer, 1969; Postan, 1973, pp. 214–248; Walmsley, 1972.
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
values. Before the Black Death, the typical prosperous peasant possessed chattels worth approximately £6 or £7. After the Black Death, the typical prosperous peasant possessed chattels worth £15, approximately double the value of those owned by their pre-plague predecessor (Dyer, 1994, pp. 171–173). In addition to land and chattels, some peasants possessed significant savings, and all peasants kept some cash on hand, since they needed coins to purchase merchandise and pay rents and taxes. Inquests reported lately departed villagers who possessed £2 and £3. Court rolls described thefts of similar sums, often stored in boxes or jars buried on peasants’ property. Inheritance records indicate peasants paid entry fines as high as £5 during the thirteenth century and more at later dates. Serfs paid £5, £10, and even £20 for manumission. Customary tenants paid similar sums to convert tenancies to leaseholds. Even marriages required large outlays of cash. On a lucky couple’s wedding day, they had to pay for the services of priests and wardens and hold a feast for their friends and family. The bride also had to pay a dowry, often between £1 and £3. Brides of servile status also had to pay a merchet, which lords set as high as two marks and as low as a shilling depending on what the newlyweds could afford. In total, “an upper peasant marriage could easily have cost £3 to £4, more than 313
4259 Ch07 4/12/02 10:19 am Page 314
314
GARY RICHARDSON
Table 5.
1 2
Chattels Possessed by Typical Prosperous Peasants and Approximate Market Values.
3 Before Black Death (circa 1300)
4 5
Characteristics
Value (£)
After Black Death (circa 1450) Characteristics
Value (£)
6 7
House
Wooden frame, open 3 hearth, floor of packed earth, walls of wattle and daub, 1 room for family, 1 room for animals, plus garden and yard.
Stone walls, wooden floors, shuttered windows, brick hearths, multiple rooms for family, animals stabled in barn or pastured in garden/yard.
7
Agricultural equipment
Handheld tools, share of a plow, some gear for beasts of burden.
1
Durable tools, personal plow, wagon or cart, tack and harness.
3
Personal possessions
One or two sets of clothing of uncolored cloth, one or two pairs of inexpensive shoes or boots, roughly hewn tables and chairs, limited domestic paraphernalia.
1
Colored clothing, kitchen accoutrements, wooden furniture, cloth bedding possibly bronze pots, copper pans, or pewter vessels,
3
Rights
Pasture animals in commons, meadows, and fields.
1
Pasture animals in commons, meadows, and fields.
2
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Total
6
15
Source: Dyer, 1994, pp. 151–187.
29 30 31 32 33 34 35 36 37 38 39 40
the expense of a new house” (Dyer, 1994, pp. 178–179). To make these payments, peasants might have hoarded coins for years in advance. But many “of the large payments detailed above, or the sums needed to build a home or to buy a cart, must have been raised by means of credit” (Dyer, 1994, p. 179). Credit was widely available because creditors knew they would be repaid. Courts enforced credit contracts and collected delinquent debts. Legal procedures were straightforward. Lenders extended loans in the presence of witnesses, preferably respected men such as village elders, manorial officers, and parish clergy who could confirm the loan’s existence when borrowers defaulted and lenders sought legal assistance. Law and custom required all grown men to attend local courts, which met regularly; thus delinquent debtors had to appear,
4259 Ch07 4/12/02 10:19 am Page 315
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
315
and when they did, their creditors could file suit. Creditors prevailed by presenting concrete evidence such as the witnesses to the transaction, written agreements, and tallies. Tallies were either marked sticks cleft length-ways or indentures torn in two, of which both parties kept half. Concrete evidence could not be countered by compurgation alone, so debtors could not defend themselves by ‘wagering their law’ unless insufficient evidence prevented the creditor from prevailing outright and the case came down to one man’s word against another’s.2 If the judge found for the creditor, the bailiff would confiscate enough of the debtors’ chattels to compel him to pay his debts. Enough usually equaled a few animals which would be corralled in the center of the village where their public incarceration would humiliate the owner, deprive him of their services, and burden him with the cost of their upkeep (Baker, 1995, pp. 360–384; H. S. Bennett, 1967, pp. 203–204, 220–221). The most detailed study of credit within villages is Elaine Clark’s work on the village of Writtle (1981). Writtle’s court rolls survive from 967 meetings between 1382 to 1490 (approximately half of those actually held during those years). These records report 948 cases concerning debts and describe 1189 credit transactions. The number of court cases lagged the number of reported transactions because creditors often brought a single plaint to settle multiple debts. Writtle’s court met every three weeks and at most meetings heard pleas of debt for the recovery of money, pleas of detinue for the recovery of chattels, and pleas of covenant to enforce contracts. These disputed debts stemmed from five types of transactions. The first involved commercial credit such as deferred payments for purchases and advanced payments for future deliveries, usually of goods, grain, and livestock (46.6% of cases on record). The second involved payments for labor, either wages owed to employees or services owed by those who had been paid in advance (23.7%). The third consisted of leases for land and animals (12.1%). The fourth involved loans of cash or equipment (11.1%). The fifth involved obligations stemming from suretyship or special arrangements such as unpaid dowries or unfulfilled promises to contribute to charity (6.6%). In these five kinds of cases, approximately eight of ten plaintiffs benefited from litigation. This ratio can be calculated from the following facts. Juries or oath-helpers settled 63% of all cases. Plaintiffs won two-thirds of the time. Thirty-seven percent of all cases were settled without trail when defendants paid “an amercement of two to three pence for leave to settle (licencia concordi) out of court” (E. Clark, 1981, p. 253). Reasons for out-of-court settlements were not recorded, although debtors may have chosen to discharge obligations when faced with suits that looked like they would succeed. Because disputed debts amounted to only a portion, probably small, of all lending, court records provide only a lower bound for the extent of indebtedness 315
4259 Ch07 4/12/02 10:19 am Page 316
316 1 2 3 4 5 6 7 8 9 10 11
GARY RICHARDSON
and availability of credit. Court records do, however, allow firm conclusions to be drawn about the kinds of agreements that courts enforced. One type was the installment-purchase agreement, which was common for livestock, particularly cows and horses, and other expensive items. A second was contracts for future delivery, which were common for commodities such as wood, wool, wheat, barley, piglets, lambs, and ale, and often included clauses requiring deliveries to be of particular qualities such as wheat that was “good and pure” and barley that was “ably malted” (E. Clark, 1981, p. 260). A third was contingent contracts. A suit from 1397 illustrates this point. In September 1396, John Wells had sold barley for four shillings per quarter to Rosa Borell on the condition that”
12 13 14
if a quarter of barley increased in price between 1 November (1396) and 2 February (1379), Rosa would pay John as much for each quarter of barely as it had increased in value (E. Clark, 1981, p. 257).
15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
By early February, the price had increased by 8 pence per quarter, and Writtle’s court held Rosa to the bargain. A fourth was a contract requiring collateral and resembling the practice of pawnbroking because the payment for reclaiming collateral exceeded the amount of the loan. This surcharge amounted to something akin to interest. Explicitly collecting interest was illegal, although courts enforced several kinds of contracts that allowed creditors to earn returns on capital. Creditors could, for example, charge higher prices for merchandise sold on credit than merchandise sold for cash. Creditors could also hold land as collateral for a loan and cultivate the collateral until the loan was repaid. A suit from 1412 illustrates contracts of this kind: Peter Herolf loaned £10 to a neighbor, Thomas Criour, [who] in turn enfeoffed Herolf with five acres of land upon this condition: if Criour repaid the loan within ten years, the land would revert to him and his heirs. Herolf took for himself the produce and profits (exitus et proficus) of the land and these were assessed at five shillings per annum or 50 shillings over a ten year period; thus the land’s fertility acted as interest on the loan (E. Clark, 1981, p. 263).
Since £1 equaled 20 shillings, Herolf earned a return of 2.5% per year on this loan. The practice of using land as collateral-cum-interest was probably widespread at the local level where “the few known mortgages may in fact be stray examples of a much more widespread practice” (Dyer, 1994, p. 180). It seems to have been widespread in Writtle, where “transactions in land, its transfer or acquisition by charter, occupied one lender in every three” (E. Clark, 1981, p. 263). In addition to describing how peasants procured credit, court records also reveal what peasants did with the credit they acquired. Those in the middle and
4259 Ch07 4/12/02 10:19 am Page 317
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
317
lower tail of the income distribution often borrowed seasonally. They took out loans when grain stores ran low during lean times of early summer and repaid their debts in the fall after the harvest (Dyer, 1994, p. 179; Hilton, 1975b, pp. 46–47; Sayles, 1936, pp. 80–81; Walker, 1982, p. 85). Peasants of all ranks borrowed to pay for marriages, baptisms, funerals, taxes, clothing, footwear, farm equipment, grain, and animals. Creditors also came from all walks of life, although men who practiced crafts loaned money more often than men who did not, and prosperous peasants loaned money more often than their poorer neighbors. Craftsmen sold merchandise to peasants and accepted payments at later dates. Prosperous peasants supplied craftsmen with raw materials and received finished goods to settle the debt. Prosperous peasants also extended credit to local merchants who transported grain, cheese, and ale to markets and fairs and engaged in long-term, repeated relationships with their peasant suppliers. Reciprocal credit created complex webs of financial entanglements. In Writtle’s court rolls, one-third of those who appeared as creditors in one legal case appeared as debtors in another. A similar percentage appears to have been involved in complex chains of credit. One type of chain linked husbandmen to butchers to tanners to cobblers, all of who at some point possessed the cow hides that became leather shoes. Another type of chain linked shepherds to spinners to weavers and to fullers, all of whom at some point possessed the wool that became cloth. A third type connected farmers to brewers to innkeepers, all of whom at some point possessed the barley that became beer sold in taverns. Credit extended through these chains from those who supplied raw materials to those who retailed consumer goods. This facilitated production by permitting everyone involved to pay for inputs after receiving payment for outputs. Even in complex chains of credit, most loans were for small sums. More than half of all disputed loans in Writtle were for five shillings or less. The same was true of loans made on other manors. “Two-thirds of those in the courts at Wakefield in 1331–1332 and 50% in the west midlands in the late fourteenth and early fifteenth centuries, were for five [shillings] or less” (E. Clark, 1981, p. 267). The abundance of small transactions suggests that the costs of extending credit and enforcing debts were low. The low fixed costs suggest peasants could acquire substantial credit at reasonable rates. Large sums could also be borrowed, either by taking out several small or one large loan. More than one-in-ten loans in Writtle exceeded £1 and one-in-twenty-five exceeded £2. Large amounts could also be borrowed from merchants, monasteries, moneylenders, and craftsmen. Peasants seeking large loans often had to rely on these out-of-village expedients, because contemporary interpretations of the Statute of Gloucester reserved all suits of over 40 shillings for the king’s 317
4259 Ch07 4/12/02 10:19 am Page 318
318 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
GARY RICHARDSON
courts. The king’s courts enforced loans in accordance with the Common Law, the Statute of Merchants, and clauses nine, ten, and eleven of the Magna Carta (Baker, 1995; Britnell, 1986). Writtle’s courts escaped this restriction because Writtle was a royal manor. In sum, all available evidence indicates villagers had access to credit. Nothing indicates credit was restricted to particular segments of peasant society, to particular places, or particular times. As Pamela Nightingale concludes in her study of mercantile credit in medieval England: “credit was used at all levels of society, even for the smallest sums” (Nightingale, 1990, p. 562). Credit permitted peasants to smooth consumption after adverse shocks by consuming tomorrow’s income today. Other financial arrangements permitted peasants to smooth income before shocks and consume today’s income tomorrow. Monasteries sold annuities, streams of periodic payments in cash or kind purchased for initial lump sums. Purchasers of cash annuities tended to be “men of means who sought a secure investment for some surplus capital,” while purchasers of annuities in kind tended to be “humble folk who wanted security for their old age” (Woodward, 1966, pp. 24–25). The latter were called corrodies and came in several standard types. A servant’s corrody included a daily ration of bread and ale, and it might include leftover food from monks’ meals, a bed in a dormitory, and clothing to replace worn garments. A monk’s corrody included a daily ration of bread and ale, a cooked dinner similar to that eaten by the monks, a room in the monastery, and a small stipend. Purchasers of corrodies, called corrodians, were often elderly men and women who wanted to exchange their savings for a pension. But, people at the peak of their careers also purchased corrodies
26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
to provide against incapacity to earn at whatever point in the life-cycle this should come about – and under medieval conditions that point was not easily predicted. [And] in many cases . . . the main purpose of the transaction in the eyes of the male corrodian was surely to provide a degree of security for his spouse, were she to be the survivor, and to give her a modest competence that did not depend on support by the children of the marriage, if any. Hence, the remarkably early appearance, in arrangements relating to corrodies, of joint tenancies for husband and wife, with the remainder to the survivor (B. F. Harvey, 1993, p. 208).
While ecclesiastic institutions sold many corrodies, they were not the sole sellers. Annuities in kind could be purchased from peasants. Christopher Dyer’s study of 141 maintenance agreements from 10 counties and Elaine Clark’s study of 200 maintenance agreements from fourteenth-century East Anglia demonstrate the practice was widespread (E. Clark, 1982, pp. 307–320; Dyer, 1994, p. 153). Maintenance agreements usually involved exchanging land for lifelong payments of food, clothing, and shelter. Two-thirds of the agreements in the
4259 Ch07 4/12/02 10:19 am Page 319
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
319
Dyer-Clark samples were made between unrelated individuals; one-third between family members. Carefully written contracts typically covered most conceivable contingencies such as the amount of grain, meat, and ale the pensioner would receive and the size of the room they would live in. Many contracts also specified regular laundering of clothes and access to gardens, riding horses, and firewood. Others made provisions for regular visits from friends, for sickbeds, funeral processions, and postmortem prayers. To facilitate enforcement, maintenance agreements were recorded in court rolls, backed by sureties and pledges, and contained contingent clauses. Breaches of contracts returned possession of the land to the pensioner. If corrodies and credit did not provide sufficient relief, unlucky villagers could always earn the money they needed to survive by working for wages. Agricultural employment was always available. Lords hired people to serve in their houses, supervise subordinates, manage estates, operate mills, and work on the demesne. Peasants with large holdings also hired men to plow fields, sow seeds, pull weeds, dig ditches, chop wood, haul manure, and perform other physical tasks. So did widows and the elderly. All landholders hired help during the harvest, which lasted five to six weeks or one tenth of the year. During the harvest, competition for workers was intense. Wages doubled. Labor mobility peaked as people left their homes and normal occupations in search of good pay in fruiting fields. Most manors provided harvest workers with daily dinners and paid them wages in kind and cash. Some employers offered large sums to those who worked for the entire harvest season. Between one harvest and the next, jobs were available in commerce, manufacturing, and services. Every village contained bakers, brewers, butchers, millers, smiths, carpenters, and tailors who processed foods, shoed horses, repaired houses, and fitted clothes. Some villages also had industrial orientations. In East Anglia and the Midlands, villagers spun wool and wove cloth. In Devon and Cornwall, villagers extracted lead and tin. Near the coast, villagers fished. Near the fens, villagers dug peat, dried salt, hunted fowl, gathered wood, and collected reeds and rushes. Evidence of employment abounds. The Poll Tax of 1380–1381 reveals wageworkers in every village. The “twenty-nine peasant households” of the village of Kempsford included “seven persons employed in ‘service’ occupations, one weaver, and twenty-eight servants and laborers” (Hilton, 1975b, p. 80). In the more industrialized villages of Brentwood and Penkridge, the proportion of farmers and craftsmen was “about half and half” and many residents worked part of the year in the fields and the remainder in other occupations (Hilton, 1975b, p. 86). Tithe records from the hundred of Chewton and hallmote of Ubley – both located in Derbyshire’s lead fields – tell the same 319
4259 Ch07 4/12/02 10:19 am Page 320
320 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
GARY RICHARDSON
tale. Even farmers with thirty-five acre hereditary holdings in the common fields and fully committed to the agricultural life of their village spent one or two months toiling in the local lead quarries and earning 15 to 30 shillings depending on their diligence and the state of the market (Blanchard, 1972). Quarrying began in late April and continued until mid-July when the harvest required men to return to their farms. Court records also report innumerable individuals with multiple occupations. The combinations seem endless: farmer, baker, and tanner; smith, cooper, and cow owner; yeoman and mercer; plowmen and parish priest. Copious suits over unpaid wages reveal the frequency of wage work. In Writtle’s court records, 18.8% of all debt suits pertained to unpaid wages, indicating that on average Writtle’s courts heard four such suits on average every year or one for every 150 residents (E. Clark, 1981, p. 254). Estate accounts also reveal the existence of large labor forces. In 1334 for example, the sergeant of the manor of Kelvedon had under him a hayward, six plowmen, a carter, and a housemaid, all employed for the whole year, a swineherd employed for thirtyfive weeks, an unskilled laborer for twenty-seven weeks, and a gardener for ten weeks (Britnell, 1980, p. 7). Kelvedon had such a substantial staff, paradoxically, because it was such a small manor, only a few hundred acres. Small manors possessed few residents owing labor services, and therefore, depended heavily on hired hands (Britnell, 1980, p. 10; Kosminski, 1956, pp. 269, 275–276). Additional evidence of employment is so numerous that only a minor fraction can be cited. These include enrolled employment agreements, the Statute of Laborers and subsequent prosecutions, coroners’ inquests, sheriff’s purveyance accounts, the Domesday Book, the Hundred Rolls, other censuses, manorial surveys, descriptive appellations in guild membership lists and marriage and burial records, literature including Langland’s Peirs Plowman and Chaucer’s Canterbury Tales, Sir Walter of Henley’s treatise on estate management, and archeological evidence of workshops, mines, forges, and mills (J. M. Bennett, 1987, pp. 55–57; DeWindt, 1972, pp. 91–95; Hanawalt, 1986, pp. 269–274; Kussmaul, 1981; Masschaele, 1993; Postan, 1954). Extant evidence indicates medieval laborers received livable although not luxurious wages. During the thirteenth century, the typical wage was one penny per day for unskilled labor and three pence per day for skilled work. Harvest workers also received approximately 13,000 calories worth of barley bread and pottage: enough food to feed an entire family. During the fifteenth century, typical wages had risen to three or four pence per day for unskilled labor and five to six pence per day for skilled work. The quality and cost (although not caloric value) of food given to harvest workers also increased: approximately 5,000 calories mainly from ale, meat, and bread baked from wheat (Dyer, 1994, pp. 24–26). Such wages could certainly sustain a prosperous peasant temporarily beset by bad luck.
4259 Ch07 4/12/02 10:19 am Page 321
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
321
Proof appears in the distribution of wealth. Poor peasants owned little land and survived on earnings from employment, beasts pastured in the commons, and herbs, fruits, and vegetables grown in their gardens (Waugh, 1991, pp. 24–25). How many smallholders were there? On most manors, one-in-ten peasants possessed no plough land at all. Another 40% owned ten acres or less (J. M. Bennett, 1987, p. 50; Postan, 1975, pp. 142–147; Raftis, 1965, pp. 83–100; Waugh, 1991, pp. 24–25). Once the “prevalence of subletting” is considered “the national figure for smallholdings would probably rise above 50%” (Dyer, 1994, p. 126). The most extensive survey of holding sizes is the Hundred Rolls of 1279–1280, which is summarized in Table 6. The Hundred Rolls suggest nearly a half of all holdings were smaller than a half yardland . . . In the east of the country, from Lincolnshire to Kent, holdings were much smaller and it is not uncommon to find manors where the majority of tenants held 5 acres or less . . . Smallholders were also concentrated in areas of commercial development, such as the outskirts of large towns, where there were more opportunities for employment (Dyer, 1994, p. 119).
The significance of this evidence should be spelled out. McCloskey claims the starvation threshold was half the yearly yield of the typical prosperous peasant’s holding. That holding was about 19 acres, and half of that was nine and a half acres, which seems reasonable. Other research suggests the threshold was about right for the Midlands, where 10 acres provided bare subsistence, and a bit high for East Anglia and Kent, where notably fertile soil allowed families to get by with less land (J. M. Bennett, 1987, pp. 86, 254; Titow, 1972, pp. 64–96). McCloskey intentionally ignores poorer peasants because “the share of the land that they held was so small that they had little voice in the layout of the fields” (McCloskey, 1989, p. 6). That may be true. But, their voices should be heard in this academic debate. Their survival year after year, generation after generation, century after century, proves prosperous peasants had little to fear from idiosyncratic agricultural shocks. Table 6.
Size of peasant holdings in the Hundred Rolls of 1279–1280.
32
Number
Percent
694 4,844 6.807 2,153 6,938
3 23 32 10 32
21,436
100
33 34 35 36 37 38
Over 1 yardland 1 yardland 1/2 yardland 1/4 yardland Smaller than 1/4 yardland Total
39 40
Sources: Dyer, 1994, p. 119; Kosminski, 1956, pp. 216–223.
321
4259 Ch07 4/12/02 10:19 am Page 322
322 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
GARY RICHARDSON
Data on the density of markets also calls into question McCloskey’s notion that open fields had inefficient yields. Ronald Coase has argued that inefficiency persists only when some market’s absence or expense prevents property rights from being reorganized to alleviate the inefficiency. Deirdre McCloskey conjectured the market missing in medieval villages was the market for insurance. If she is wrong, then her theory is wrong, as she acknowledges when she writes: Each explanation of scattering depends on a non-existent or expensive market. If the market presumed not to exist does exist, then the explanation is wrong (McCloskey, 1989, p. 26. Italics are McCloskey’s).
A formal market for insurance is not the only issue. Labor, land, and credit markets also enable individuals to smooth consumption across time and space. Their existence contradicts McCloskey’s conjecture, as McCloskey acknowledges when she writes: The explanation in terms of risk aversion is not immune to the criticism leveled at the other explanations: Was the market missing? The market in question is insurance. Was there no insurance that could have saved peasants the bother of the expensive self-insurance of scattering plots? (McCloskey, 1989, p. 46).
18 19 20 21 22 23 24
And when she writes: If alternative forms of insurance were available and cheap, it would be silly for a medieval peasant to suffer the inconvenience of self-insurance by scattering . . . My explanation must contend with evidence (if any) on cheap substitutes for insurance (McCloskey, 1977, pp. 402–403).
And when she writes:
25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
The explanation from risk depends like all explanations on a missing market (since only missing markets can account for inefficiencies), the market in this case for insurance. Any asset could provide some insurance . . . If land could be sold easily, then it could provide the margin against disaster. It depends on how liquid the market was, that is, and the quantities involved . . . In fact the development of any asset market, including the labor market, increases the liquidity of peasants and therefore their security (McCloskey, 1991, p. 352).
This section has shown that markets for grain, fish, meat, dairy, and other victuals existed within medieval English villages and extended far beyond their borders. Markets for labor, land, credit, assets, and annuities existed within (but perhaps not between) rural villages. The innumerable transactions for small sums suggest transaction costs in local markets were low. The volume of transactions indicates that those markets could have mitigated all of the idiosyncratic shocks that afflicted prosperous peasants. The fact that prosperous peasants never starved alone suggests that markets (and perhaps other mechanisms) did so. To borrow a turn of phrase from one of
4259 Ch07 4/12/02 10:19 am Page 323
What Protected Peasants Best? 1 2 3 4 5 6
323
my favorite economic historians, medieval peasants did have access to cheap relief against idiosyncratic shocks. McCloskey must be wrong. The assumptions underlying her argument are incorrect. Fear of local disasters did not dictate the organization of the open fields. Of course, additional issues should be considered before that conclusion can be encased in a stout fence of fact.
7 8 9 10
II. STORAGE COSTS AND THE OPTIMAL ORGANIZATION OF ARABLE LAND
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
As McCloskey noted in her opus on the open fields, the optimal method of mitigating risk depended upon the relative costs of the available alternatives. One method was the scattering of strips, which smoothed income at a cost of 10% of aggregate output each year. Another method was the storage of grain, which smoothed consumption at a cost approaching 33% per year. Little was known about the costs of other methods of smoothing consumption across individuals, space, and time, but they were likely to be similar to or higher than the costs of storage. Since 33% exceeded 10%, McCloskey concluded, scattering was the most efficient method of mitigating risk. What should be made of McCloskey’s calculations? Are her descriptions of the costs of various methods of mitigating risk accurate? Do her estimates prove her point? No is the answer to the last two questions. This section explains why. It begins by examining the costs of smoothing consumption across individuals (via markets or pooling) and over time (via storage). Consider the costs of markets. Purchasing grain from one’s neighbors or bread from the local baker was quick and painless. Walking down the street, exchanging assets for victuals, and returning home consumed few calories and was often a pleasurable experience. Working for wages consumed both calories and leisure time. Exchanging strips in the open fields involved small payments to manorial courts. So, mitigating idiosyncratic shocks through markets did not have prohibitive costs. Now, consider the cost of pooling. McCloskey claims that operating all of the institutions that regulated cooperative cultivation in medieval villages consumed approximately 1% of aggregate output. Development economists find similar costs in modern villages where repeated interactions reduce the costs of administrating, monitoring, and enforcing reciprocal risk-sharing arrangements to levels so low that they have little impact on economic performance. The sustainability of the agreement, of course, is the key issue. Miles Kimball analyzes this issue in his seminal article, “Farmers’ Cooperatives as Behavior Towards Risk” (1988). He concludes that 323
4259 Ch07 4/12/02 10:19 am Page 324
324 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
GARY RICHARDSON
If utility discount rates in the Middle Ages were anything like they are now – if medieval peasants were perhaps a trifle more impatient because they were poorer and also discounted the future a few percent more per year because of their higher adult mortality rates – then . . . farmers’ cooperatives could have provided a substantial amount of insurance even among farmers with a degree of risk aversion ( = 1) that would not justify scattering. Farmers with higher levels of risk aversion ( > 1) could have formed self-enforcing cooperatives involving substantial sharing of output even at very high discount rates. These cooperatives could not enforce full sharing in very bad times, but in normal times would provide insurance at a much lower cost than scattering of plots (Kimball, 1988, p. 233).
In other words, at all plausible discount and interest rates including those that McCloskey assumes, risk-pooling agreements would have been self-enforcing and would have mitigated risk more effectively than the scattering of strips. Now, consider the cost of storing grain. Storage was the subject of a famous debate between Stefano Fenoaltea and Deirdre McCloskey. In “Risk, Transaction Costs, and the Organization of Medieval Agriculture,” Fenoaltea argued that peasants stored grain after good harvests and survived on stockpiles during difficult times (Fenoaltea, 1976, pp. 138–139). In “Corn at Interest,” McCloskey and Nash replied that peasants seldom stored significant stocks of grain because the cost of storage exceeded the costs of scattering by a wide margin (McCloskey & Nash, 1984). I have argued elsewhere that McCloskey and Nash overestimate the cost of storage (Richardson, 1999). Gregory Clark reaches a similar conclusion. In the years before 1350, the cost of storing grain was 15–20% per year. Between 1400 to 1700, the cost fell to 10% per year, which McCloskey notes is the highest wastage rate observed in modern times when storing grain at temperate climates using traditional methods. In sum, McCloskey’s assumptions concerning the costs of markets, pooling, and storage appear to be exaggerated. If they are, then McCloskey’s deductions from them are also erroneous. If they are not, then McCloskey’s conclusions remain wrong, because her assumptions do not substantiate her claims, as the rest of this section shows by testing her hypothesis. Hypothesis testing is a forgiving task. The null is given the benefit of the doubt and assumed to be true until proven false. In this case, the null hypothesis is McCloskey’s model of medieval agriculture. It depicts a peasant’s choice between cultivating a consolidated field or a set of scattered strips and contains three essential elements. The first describes the consequences of consolidation. It is a probability distribution that indicates the average annual yield of an arable field of 20 acres, the standard deviation of that yield, and the probability of earning each amount. It is summarized by the notation X ~ N (x, 2 ) where N denotes the normal distribution, x = E (X) denotes the average income earned from farming consolidated fields, and 2 = V (X) denotes the variance of that income.
4259 Ch07 4/12/02 10:19 am Page 325
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
325
The second element of the model describes the consequences of scattering. It is a probability distribution indicating the average annual yield of 20 acresized strips scattered throughout a village, the standard deviation of that income, and the probability of earning any particular amount. It is summarized by the notation Y ~ N(x(1 – ), 2(1 – )2) where denotes the cost of scattering as a percentage of crop yield, b is the benefit of scattering as a percentage of the standard deviation of crop yield, x(1 – )=E(Y) denotes the average yield of scattered fields, and 2(1 – )2 = V(Y) denotes the variance of that yield. The third element of the model describes how peasants choose between consolidation and scattering. It is a rule: pick the option that maximizes your family’s chance of survival, or equivalently, pick the option that minimizes your family’s risk of starvation. Implementing that rule requires knowing how much a family must consume in order to survive. Table 7 displays McCloskey’s estimate of that amount and summarizes the parameters of her model (McCloskey, 1976, 1989). Columns (1) and (2) describe crop yields. The average income earned from scattered fields, x (1 ), is set at 1 to indicate 100% of normal output. The variance of that income, (1), is 0.347. The average income earned from consolidated fields, x, is 1.1. The variance of that income, , is 0.484. All of these estimates are relative to the average output from scattered fields. Thus, for example, a consolidated field’s average yield exceeds a scattered field’s average yield by 10%. Column (3) indicates the distance to disaster (also known as the distance to starvation), which is the number of standard deviations by which the average yield exceeds the amount of grain needed to survive, denoted , which equals 1/2. In the consolidated case, the distance to disaster is (1.1–0.5)/0.484 which equals 1.24. In the scattered case, the distance is (1–0.5)/0.347 which equals 1.44. Column (4) indicates the probability of disaster (also known as the probability of starvation). It is calculated by comparing the disaster distance to the cumulative normal distribution, (–). Column (5) indicates the frequency of disaster (also known as the frequency of starvation). It is the expected waiting time between disasters calculated by inverting the distance to disaster. Comparing frequencies of disasters generates the main implication of the basic model. Disasters strike peasants who cultivate consolidated fields more frequently than peasants who cultivate scattered strips. Those comparisons can be summarized in several synonymous sentences. Peasants who possessed scattered strips starved less often than peasants who possessed consolidated fields. Scattered strips were safer than consolidated fields. Yields of diverse portfolios varied less than yields of single plots; diversification reduced risk. Prudent peasants did not place all of their eggs in one basket. McCloskey’s message, it turns out, was a moral often found in 325
4259 Ch07 4/12/02 10:19 am Page 326
326 1
GARY RICHARDSON
Table 7.
2
The Average and Standard Deviation of Crop Yields. On Consolidated and Scattered Holdings
3 4
Average
5 6
(1)
Values Standard Distance deviation from disaster (2) (3)
Probability of disaster (4)
Frequency of disaster (5)
10.75% per year 7.48% per year
once every 9.3 years once every 13.4 years
Probability of disaster
Frequency of disaster
7 8 9
Consolidated
110
48.4
1.24
Scattered
100
34.7
1.44
10 11 Average
Notation and Formulas Standard Distance deviation from disaster
x x (1–)
d = (x – )/ 1– (d) (1–) d = (x (1– ) – )/ (1–) 1– (d)
12 13 14 15 16 17 18 19 20
Consolidated Scattered
f = 1/[1–(d)] f = 1/[1–(d)]
Source: Detailed calculations in McCloskey, 1976, pp. 132–151 and 1989, pp. 37–45. Note: is the standard normal cumulative distribution function. is the survival threshold for consumption.
21 22 23
(x) =
x
1 e 2
z2 2
( ) dz
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
financial theory and fables. In those cases, the moral makes sense. There are, however, many reasons to doubt its sensibility in this case. According to McCloskey’s data, scattering cost a great deal but did little. The dispersion of strips reduced crop yields by 10% while increasing average intervals between harvest failures by a mere four years, from 9 to 13. Crops in scattered strips still failed, and peasants who scattered strips still starved. In villages with scattered strips, 7.5% of the residents died of hunger each year; 54% in ten years; 79% in twenty years. All starved eventually. The corresponding figures for villages with consolidated fields were 10.75%, 68%, and 90%. Thus, scattering saved no lives. Peasants with scattered strips merely lingered a little longer than their cousins with consolidated fields before hunger sent them to an early grave. This point can be put in another way. The mortality rate predicted by McCloskey’s model (7.5% of the population per year plus those that died from other causes) exceeded the sustainable rate. In such circumstances, the population should have fallen, land and output per person should have risen, and the population should have converged to a sustainable level.
4259 Ch07 4/12/02 10:19 am Page 327
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
327
McCloskey’s model also misses the intertemporal aspect of a peasant’s consolidate-or-scatter decision. A peasant’s decision, like all investment decisions, had long-term implications. Intelligent investors think about the future before buying stock with their life savings, and medieval peasants must have thought about the future before making life-or-death decisions about agricultural arrangements. But in McCloskey’s model, the peasants do not. Their shortsightedness suggests her model can be improved by adding an intertemporal element and permitting peasants to smooth consumption through time. McCloskey speculated about the implications of such an exercise. The cost of scattering was 10%. The cost of storage was 32.5%. Thus, she claimed, scattering was cheaper than storage, and given the choice between the two, the prudent peasant would choose the former. But, correctly calculating the costs and benefits of smoothing consumption across time reveals the flaws in McCloskey’s logic. The simplest way to correct the calculations is to extend the model through time by adding additional periods. Multiple periods permit peasants to smooth consumption forwards and backwards in time. Accumulating assets such as cash, land, tools, and textiles shifts consumption from the present to the future. Borrowing shifts consumption from tomorrow to today. For simplicity’s sake, the model permits intertemporal smoothing in only one direction, toward the future, via the storage of grain. The multi-period model begins with a peasant household that exists for many years. In the first year, the household chooses whether to farm a consolidated field or a set of scattered strips. If the household chooses the consolidated field, its first period income is a draw X ~ N (x, 2) as is its income in subsequent periods. Its consumption in year t is indicated by ct, its savings in year t by the xt – ct, and the cost of storing grain from one period to the next as a percentage of stockpiled grain is where 1 – indicates the percent of a bushel of grain stored in period t that is available for consumption in period t + 1. The accumulated stock of savings carried forward into year t is given by the formula
t–1
(xi – ci ) (1 – ) t – i
If the household stores the same amount each year (denoted by the letter c without a subscript), the stock of grain will expand until annual savings equals the annual carry-over cost, and the stock carried over from one period to the next equals
38 39 40
(1)
i=1
(x – c)
1–
327
(2)
4259 Ch07 4/12/02 10:19 am Page 328
328 1 2 3 4 5 6 7 8
GARY RICHARDSON
If the household chooses the set of scattered strips, its first period income is a drawn from Y ~ N (x (1 – ), 2(1 – )2) as is its income in subsequent periods. The notation for consumption and storage remain the same, and its accumulation and steady-state are described by Eqs (1) and (2) after substituting E(Y) for E(X). The steady-state levels of consumption and risk are described by an ordered pair [c, ] whose first element indicates consumption of grain and second the probability of surviving from one year to the next. For peasants with consolidated fields,
9 10
1
11
if (x – c)
12 13 14
=
15 16
1– x + (x – c) –
17
1– ≥ (3)
if (x – c)
1– <
18 19
For peasants with scattered strips,
20 21 22
1
23 24 25 26 27 28
=
x(1 – )+(x(1 – ) – c) (1 – )
if (x – c)
1– –
1– ≥
(4)
1– if (x(1 – ) – c) <
29 30 31 32 33 34 35 36 37 38 39 40
The likelihood of survival contains a conditional statement because grain fields cannot have negative yields. Therefore, a peasant who has stockpiled enough grain to feed his family for an entire year will survive even if his crops fail. A peasant whose stockpile falls short of the survival threshold will survive only if the yield of his land added to his stockpile of grain surpasses the survival threshold. This minor adjustment to McCloskey’s assumptions is necessary because the normal distribution extends to negative infinity. Unless it is truncated, the infinitely negative yield of one peasant’s fields could consume all of the grain in the universe. This no-black-hole assumption need not be offset by
4259 Ch07 4/12/02 10:19 am Page 329
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
329
a no-big-bang assumption because the shape of the upper end of the distribution is neither a choice variable nor a parameter of interest. Figure 1 compares: (a) McCloskey’s original model; and (b) the multi-period framework derived in this essay. Graphs (a) and (b) plot onto a typical twogood (consumption, survival) budget set the options available to prosperous peasants given McCloskey’s assumptions, x = 1.1, x(1 – ) = 1, = 0.484, (1 – ) = 0.347, = 0.325, and = 0.5. Graph (a) displays McCloskey’s original result. For peasants cultivating consolidated fields, the average yield is 1.1, and the likelihood of survival is 89.25% per year. For scattered strips, the average yield is 1 and the likelihood of survival is 92.52%. Graph (b) depicts the options available to peasants who store in each year xt – c units of grain at the intertemporal cost calculated by McCloskey, that is = 0.325. Storing grain expands the set of consumption-survival opportunities by permitting peasants to substitute longevity for consumption. The feasible set for peasants with scattered fields falls below and to the left of the dashed line. The feasible set for peasants with consolidated fields falls below and to the left of the solid line. Since the former falls everywhere below the latter, peasants with consolidated fields could do anything that peasants with scattered fields could do and much more. Peasants with consolidated fields could have higher incomes, higher survival rates, or both. They could consume each year 85% of the average yield, save the rest for a rainy day, and accumulate reserves large enough to offset even the worst idiosyncratic shocks. They could consume the same amount as farmers with scattered fields, store the surplus from consolidation, and survive at much higher rates than peasants with dispersed holdings but without savings. They could survive at the same rates as peasants with scattered strips and consume on average 5% more during each year of their lives. Their choice of course, depended on their preferences for risk and return. If they feared starvation, they could reduce consumption and minimize risk. If they defied death, they could increase consumption and amplify risk. Whatever their preferences, consolidation and storage provided more and better options than scattering with or without storage. In sum, at the costs of storage calculated by McCloskey, consolidation and storage increased consumption and longevity. Consolidation protected peasants best. Figure 2 provides a more accurate image of the situation using Clark’s estimates of storage costs. Graphs (a) and (b) plot onto a typical two-good (consumption, survival) budget set the options available to prosperous peasants given all of McCloskey’s assumptions, x = 1.1, x (1 – ) = 1, = 0.484, (1 – ) = 0.347, and = 0.5, with one exception. Graph (a) depicts the options available to peasants who store in each year xt – c units of grain when the cost of storage was 20% per year (i.e. when = 0.20). That percentage was the top 329
4259 Ch07 4/12/02 10:19 am Page 330
330
GARY RICHARDSON
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Fig. 1.
Risk-Return Combinations Available to Prosperous Peasants.
32 33 34 35 36 37 38 39 40
of the range calculated by Clark for years before 1350. Graph (b) depicts the options available to peasants who store in each year xt – c units of grain when the cost of storage was 10% per year (i.e. when = 0.10). That percentage represents the typical storage cost calculated by Clark for years between 1400 and 1700. These graphs reinforce the message of the first figure. Given McCloskey’s assumptions, consolidation protected peasants best. The conclusion remains valid even after improving the accuracy of the crucial assumptions concerning the costs of storage.
4259 Ch07 4/12/02 10:19 am Page 331
What Protected Peasants Best?
331
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Fig. 2.
Risk-Return Combinations Available to Prosperous Peasants.
32 33 34 35 36 37 38 39 40
Figures 3 and 4 reveal where McCloskey’s verbal speculation about the relative costs of storage and scattering went wrong. Figure 3 plots the minimum risk method of agriculture for various costs of storage and scattering. In the upper region, storage is expensive. Scattering is cheap, and scattering without storage minimizes risk. In the lower region, Scattering is expensive. Storage is cheap, and storage with consolidation minimizes risk. The graph is derived by translating McCloskey’s assertion – that scattering without storage was safer than consolidation with storage given her estimates of the costs of smoothing 331
4259 Ch07 4/12/02 10:19 am Page 332
332 1 2 3 4
GARY RICHARDSON
risk across space and time – into a mathematical formula. Equation (5) defines the values of , , and for which scattering’s likelihood of survival exceeds storage’s likelihood of survival when the yield from scattered fields is consumed and the surplus from consolidation is saved.
5 6 7 8 9 10
x + (x – x(1 – a))
1– –
≤
x (1 – ) – (1 – )
(5)
11 12 13 14
Solving for reveals that when scattering without storage was superior to consolidation with storage.
15 16 17
≥
x 1– · x (1 – ) –
(6)
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Substituting McCloskey’s values for the variables , , x, and reveals scattering minimized risk if and only if exceeded 50.66%. Since McCloskey’s estimate of is 32.5%, her logic and data should have led her to the conclusion opposite of that which she advocated. Consolidation with storage, rather than scattering without storage, protected peasants best. The equations above illuminate the intuition underlying the correct calculations. McCloskey’s argues the cost of storage was 32.5%. The cost of scattering was 10%. The former was larger than the latter. Therefore storage was more expensive than scattering, and scattering protected peasants best. But, percentages cannot be compared directly. One-third of a dime is less than one-tenth of a dollar. One-third of a grape is smaller than one-tenth of a grapefruit. Correct comparisons require knowledge of both the percentage and the base. For scattering, the percentage was 10%. The base was the annual average yield of consolidated fields. Multiplying the percentage and base reveals that scattering cost each year 10% of the average yield of consolidated fields. For storage, the percentage was 32.5%. The base was the stockpile of grain that had to be carried over from one year to the next in order to equate the risk of consolidation and scattering. That stock was 8.86% of the average yield of consolidated fields. Multiplying the percentage and the base reveals that storing enough grain to equate the risks of consolidation and scattering cost each year 2.88% of the average yield of consolidated fields. Thus, the annual cost of storage was less than one-third of the annual cost of scattering and one-tenth of the figure publicized by McCloskey.
4259 Ch07 4/12/02 10:19 am Page 333
What Protected Peasants Best?
333
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Fig. 3.
19
Minimum Risk Method of Agriculture Given Costs of Storage and Scattering.
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Fig. 4.
Consumption With Reserves Large Enough to Offset All Idiosyncratic Shocks. 333
4259 Ch07 4/12/02 10:19 am Page 334
334 1 2 3 4 5
GARY RICHARDSON
Figure 4 examines this issue from another perspective. It plots the levels of consumption that peasants could maintain in the steady state while retaining reserves large enough to offset even the worst idiosyncratic shocks. These consumption levels are derived by equating the grain reserves in the steady state and the survival threshold
6 7
consolidated (x – c)
8 9 10 11 12 13 14 15
scattered (x (1 – ) – c)
1– = 1– =
(7)
(8)
and solving for the level of consumption. consolidated c =
–x –1
(9)
16 17 18
scattered c =
– x (1 – ) –1
(10)
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Equations (9) and (10) elucidate several important points. Given McCloskey’s data on the costs of storage and scattering, consolidation raised the risk-free level of consumption. Scattering lowered that level. This pattern was true regardless of the cost of storing grain. In addition, cultivating consolidated fields and accumulating stocks of grain as large as possible was the optimal strategy for peasants possessing the safety-first preferences posited by McCloskey. Peasants with consolidated fields who consumed the subsistence amount and stored the remainder would accumulate reserves equal to 125% of the average yield of consolidated fields since (110 – 50)(1 – 0.325)/0.325 125. That stockpile enabled them to survive two consecutive years of complete crop failures. The minimum stockpile needed to survive consecutive complete crop failures was 116% since 50 + 50 (1.325) 116. Peasants with scattered strips who behaved in the same manner would accumulate reserves equal to 104% since (100 – 50)(1 – 0.325)/0.325 104. Thus, they could survive one harvest failure but not a second.
35 36 37
III. RISK POOLING AND OPTIMAL ARABLE ORGANIZATION
38 39 40
Miscalculating the costs and benefits of storage was one flaw in McCloskey’s arguments. Another was neglecting institutions that permitted peasants to pool
4259 Ch07 4/12/02 10:19 am Page 335
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
335
risk. Within medieval villages, fraternities protected peasants from the perils of everyday life. Customary poor laws redistributed resources towards villagers beset by bad luck. Charitable customs fed families who could not feed themselves. Adding risk-pooling institutions to the model requires additional notation. The Greek letter denotes the cost of pooling risk as a percentage of the grain transferred from the well off to the worse off. The letter denotes the quantity of grain transferred as a percent of the average yield of consolidated fields. The letter denotes the benefit of pooling risk as a percentage of . The addition of risk-sharing institutions also requires changes in the framework of the model. The sharing of risk occurs within a period of time. So, the risksharing version of McCloskey’s model examines a single year in which peasants base their decisions on expected levels of consumption and risk. The feasible levels are described by an ordered pair [E(c), E( )] whose first element indicates expected consumption of grain and second the probability of surviving from one year to the next. For peasants with consolidated fields,
16
E[c] = E[x – x ]
17 18
19 20
E[ ] =
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
x – x – (1 – )
= x[1 – E()]
=
(11)
x[1 – E()] – (1 – )
(12)
For peasants with scattered strips, E[c] = E[x (1 – ) – x ]
E[ ] =
x (1 – ) – x – (1– )
= x[1 – – E()] =
(13)
x[1 – – E()] – (14) (1 – )
The terms (1 – ) and E() are the gates through which risk sharing enters the model. The former indicates how pooling influences the variability of consumption. The latter indicates how pooling influences average consumption. The gates can be opened by several mathematical methods. This essay relies on assumptions that exaggerate the cost of pooling and conform to the conclusions of Kimball’s essay. Given McCloskey’s estimates of the relevant parameters, Kimball found that self-interest and repeated interaction could sustain any risk pooling arrangement up to and including full pooling of output. Thus, it is safe to assume that selfish peasants who had extra food did not neglect neighbors in need. If enough food existed to feed everyone, then everyone ate. That assumption can be encapsulated in a verbal statement, all 335
4259 Ch07 4/12/02 10:19 am Page 336
336 1 2
GARY RICHARDSON
households earned the average income earned by all other households, and a mathematical statement,
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
n
x¯ =
i=1
xi n
(15)
where n is the number of households, xi is ith household’s income, x¯ is the income of a peasant participating in the risk-pooling arrangement, the average income of those peasants, and the aggregate income of the village as a whole. The multiple interpretations of x¯ exist because it is an average of percentage outcomes across homogenous individuals. In a village with scattered fields, xi = x(1 – ) for all households, and therefore, x¯ = x(1 – ). In a village with consolidated fields, xi = x for all households, and therefore, x¯ = x. The formula for x¯ allows the calculation of (1 – ). The calculation requires several assumptions. Assume, as McCloskey does, that all households, all portfolios of strips, and all correlation coefficients between every pair of portfolios are identical. Then, the relationship between the standard deviation of average output, (1 – ), and the standard deviation of a scattered holding, (1 – ), is summarized in the following formula:
19 20 21 22
(1 – ) = (1 – )
1 + (n – 1)R n
(16)
23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
where R is the correlation coefficient between sets of scattered strips. This formula, which is derived in the appendix, reveals three important points. First, average output varies less than the output of individual households because average output offsets one household’s shortage with another household’s surplus. In mathematical terms, (1 – ) ≤ (1 – ) because n > 1 and 0 ≤ R ≤ 1 and therefore [(1 + (n – 1)R)/n]1/2 – < 1. Second, the variance of average output approaches the variance of individual output as population falls and correlations between crop yields rises, or in mathematical terms, as n → 1, (1 – ) → (1 – ) and as R → 1, (1 – ) → (1 – ). The second point suggests risk sharing’s effectiveness varies with the size of the village and uniformity of land within the village. The larger the village, the more effective village-wide risksharing institutions would be, and the less uniform the land, the more effective village-wide risk-sharing institutions would be. Third, for the typical village where n 50 and R 0.8, Eq. (16) indicates that (1 – ) 0.311 and 0.357. The value of varies little as the number of households in the village falls. remains above 0.350 as long as n exceeds 10. Only below five households does the benefit of pooling fall quickly towards zero.
4259 Ch07 4/12/02 10:19 am Page 337
What Protected Peasants Best? 1 2 3 4 5 6 7 8
The formula for x¯ also allows the calculation of E(). E() represents the percent of aggregate output that must be transferred to fulfill the promise enshrined in Eq. (15). The pledge produces the most expensive pooling arrangement possible. Rather than transferring just enough grain to feed peasants beset by bad luck, it requires transferring enough output to equate consumption among all villagers. The quantity transferred to complete that task is approximately 18% of the average yield of village fields, as the following calculation shows
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
337
E() = x
48.4 *0.3989 = 0.175516 18% (y)y dy = 110
(17)
y =0
Determining the value of is a difficult and may be a moot task. Development economists find the costs of transferring resources among social and geographically concentrated households within modern villages to be negligible. Transportation costs rarely if ever impede transfers of foodstuffs among members of small groups. The sustainability of the agreement is the key issue. Individuals help others only when they believe their compatriots will reciprocate in the future. Kimball proves that self-interest should have sustained sensible reciprocal arrangements as non-cooperative equilibrium given plausible values of the relevant parameters and even for values when scattering was unappealing. His intuition is clear. If discount rates were so high that the expectation of reciprocity in the future did not elicit pooling in the present, than the intertemporal tradeoff inherent in scattering was unattractive. Scattering was merely another method of shifting consumption over time. Peasants with scattered strips consumed less in the present in hopes of increasing the number of years of consumption in the future. Whenever preferences for future consumption were strong enough to induce peasants to shift consumption via scattering, the value of future reciprocity would permit peasants to form sustainable risk-sharing cooperatives and shift smooth consumption effectively and at low cost. An estimate of the elusive value of can be derived from McCloskey’s study of the costs of coordinating communal activities within medieval English villages. She finds the total cost amounted to approximately one percent of village output. Both fraternal covenants and customary poor laws were enforced through the manorial courts that supply McCloskey’s data. Assume all of the costs that she uncovered were due to those risk-sharing activities. Then, the total cost of pooling risk, E() · 0.175516, equaled 0.01, and 0.057. This value implies that for every bushel received by unfortunate peasants, 1.057 bushels had to be taken from prosperous neighbors. Table 8 calculates survival 337
4259 Ch07 4/12/02 10:19 am Page 338
338 1
Table 8.
2
GARY RICHARDSON
Consumption and Survival for Peasants Pooling Risk. For Various Costs of Transferring Grain Within Villages.
3 4
Cost of Transfer
Consolidated Fields Survival Consumption % per year % of x(1 – )
5
Scattered Strips Survival Consumption % per year % of x(1 – )
6 7 8 9
0.0 0.057 0.2 0.4
0.97 0.97 0.96 0.95
1.10 1.09 1.06 1.03
0.95 0.94 0.93 0.91
1.00 0.99 0.96 0.93
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
Notes: The cost of pooling is in terms of percent per bushel received by the needy. Thus, 0.2 indicates that for every 1.2 bushels taken from fortunate peasants, 1.0 bushels reach their unfortunate neighbors. Source: Calculations in text.
and consumption opportunities for peasants pooling risk at costs above, below, and equal to McCloskey’s estimate. The likelihoods of survival and levels of consumption are calculated from Eqs (11), (12), (13), and (14). The cost runs from zero, which is what development economists typically assume it to be, to 0.40, which is so high that peasants cannot benefit from pooling risk. No matter what the cost, the conclusion remains the same. Consolidation and pooling increased consumption and longevity. The question remains, however, whether the costs of pooling risk could have been so high that peasants who collaborated in almost every other aspect of everyday life – planting, harvesting, husbandry, religion, and adjudication to name a few – chose not to cooperate in the crucial endeavor of feeding friends and family. Answering that question involves finding the values of for which peasants with scattered strips who did not pool risk survived at higher rates than peasants with consolidated fields who did pool risk. In mathematical terms, this involves determining the for which
30 31
32 33 34
x(1 – ) –
(1 – )
≥
x – x E[] – (1 – )
(18)
Solving for reveals
35 36 38 39 40
x 1 – (1 – )
37
≥
1– 1–
–
x · E[]
1–
1– 1–
(19)
4259 Ch07 4/12/02 10:19 am Page 339
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
339
Plugging in McCloskey’s values for the variables on the right-hand yields ≥ 78.5%. In other words, scattering without pooling would have been safer than consolidation with pooling only if the cost of transferring a bushel of grain between neighboring houses was more than three-fourths of a bushel. The cost could not have been that high. Estimates of transport costs in medieval England show that shipping grain from one end of the kingdom to another consumed less than a tenth of that amount. The calories expended carrying a bushel of grain between houses in a village amounted to no more than a miniscule fraction of the calories contained within a bushel. Social scientists who study traditional societies know of no place where the cost of transferring resources within villages approaches 78%. Instead, development economists, sociologists, and anthropologists commonly conclude the cost was near zero and thus too small to include in quantitative models, where the burden of additional complexity would outweigh the benefit of additional knowledge. The assumptions underlying the calculation of E[] bias upwards the denominator of Eq. (19) and downwards the break-even value of . In sum, there is no basis for concluding that the costs of pooling could have prevented the pooling of risk. Even if the costs were exorbitant, even if they amounted to one third of total transfers, even if peasants argued for hours on end every time someone requested help and wandered aimlessly for many miles while carrying aid to neighbors and spilled twenty-five percent of the grain along the way, pooling would have been an efficient method of mitigating idiosyncratic agricultural risk. A model combining pooling and storage reinforces that point. The feasible levels of consumption and survival are described by an ordered pair [E(c), E( )] whose elements indicate consumption and survival respectively. For peasants with consolidated fields,
28 29 30
31
1– ≥
if [x (1 – ) –c]
1
32 33 34
E [ ] =
35 36 37 38
1– x (1 – ) + (x (1 – ) – c ) –
(20)
if [x (1 – ) – c]
(1 – )
39 40
339
1– <
4259 Ch07 4/12/02 10:19 am Page 340
340 1
GARY RICHARDSON
For peasants with scattered strips,
2 3 1
4
if
5 6
E [ ] =
7 8
(21)
1– x (1 – )(1 – E []) + (x(1 – )(1 – E[] – c)– c ) –
9
≥
if
<
(1 – )
10 11 12
where equals
13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
[x(1 – )(1 – E[]) – c]
1– .
Figure 5 (a) plots the options available to peasants who stored in each year xt – c units of grain and pooled risk at the costs calculated by McCloskey. Graph (b) depicts the options available to peasants who stored grain and pooled risk at costs commonly calculated by development economists. Transferring resources within villages consumed time and effort but not food. Time was an abundant resource. Farmers had long periods of light work, and individuals on the brink of starvation had nothing more pressing than staying alive. The effort required to transfer food from one villager to another was negligible. Therefore, graph (b) assumes = 0. Medieval peasants lacked modern technology for storing foodstuffs. Stocks of grain diminished due to rotting, rodents, theft, and others natural and human causes. The highest attrition rates observed for storing grain at temperate climates using traditional methods, as McCloskey notes, approximated 10% per year. Average rates appear to be half of that or less. The figure for medieval England must have been similar. Clark’s work shows that was so. Therefore, graph (b) assumes = 0.1. Together, graphs (a) and (b) communicate a consistent message. At the costs of storage calculated by McCloskey, consolidation, storage, and pooling increased consumption and longevity. At the lower cost of storage estimated by Clark and commonly found in traditional modern villages, the increase was much larger. These results concur with evidence established by recent research in demographic, social, and agricultural history. Prosperous peasants successfully walked the tightrope between survival and starvation. Those with 20 acres of arable land seldom starved, and when they did, they did not starve alone. Famine killed many during the terrible transition from the thirteenth to the fourteenth centuries but few after the plague reduced population levels and raised incomes.
4259 Ch07 4/12/02 10:19 am Page 341
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
341
Prosperous peasants – like men and women in most societies past and present – used many methods of mitigating risk. Storage was one. Consistently saving moderate amounts of grain enabled peasants to accumulate substantial stockpiles. Pooling was another. Scattering may have been a third. No single method provided complete protection, but when combined together, these institutions – and perhaps others yet to be identified – effectively mitigated idiosyncratic risk. These points can be put another way. Given the extant evidence, McCloskey’s claims that scattering reduced yields and that scattering reduced risk, cannot coexist. If the former was correct and scattering reduced yields, then scattering increased the risk of starvation for individuals and the risk of famine for society as a whole. Conversely, if the latter was true and scattering reduced risk, then the yields of scattered fields had to equal or exceed those of consolidated fields. So, one of McCloskey’s claims must be wrong. It is important to note six things this result does not depend upon. The first is refuting McCloskey’s estimates of the costs of pooling and storage. McCloskey’s assertions about the superiority of scattering were wrong regardless of the veracity of her estimates. Her estimates were high, but not nearly high enough to corroborate her claims. The second is preferences for risk and return, or the existence of altruistic motives, or customs that compelled cooperative behavior, or in economic jargon, the form of the utility function. As long as peasants preferred consuming more to less and living longer to shorter then whatever the nuances of peasants’ preferences towards consumption and risk, they would have preferred consolidation, pooling, and storage to the methods of mitigating risk advocated by McCloskey. Their preference for consolidation would have been extremely strong if the yields of consolidated fields significantly exceeded the yields of scattered fields, as McCloskey claims. The reason is clear. High-yield fields gave peasants more options than lowyield fields. Peasants with the former could do anything done by those with the later and potentially much more. Readers skeptical of this argument are invited to conduct the following exercise. Apply the utility function of your choice to the opportunity sets defined by Eqs (3) and (4), or (11) through (14), or (20) and (21) and solve for the parameters of interest. None of the subsequent calculations or conclusions change. Repeat the exercise with different utility functions until it exhausts you. Since the utility function has no influence on the outcome of the debate, the principal of parsimony dictates that it be excluded from the essay. Carrying around extra letters and concepts confuses the issue. As long as individuals prefer abundance to poverty and longevity to death, then the mathematical form of the utility function is irrelevant. 341
4259 Ch07 4/12/02 10:20 am Page 342
342
GARY RICHARDSON
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Fig. 5.
Risk-Return Combinations Available to Prosperous Peasants.
32 33 34 35 36 37 38 39 40
The third is the rate of convergence of the dynamic system. The findings for the steady state hold true off the equilibrium path. In almost all situations, the opportunity set for consolidated fields dominated the opportunity set for scattered strips. This mathematical mumbo jumbo means peasants without substantial stocks of grain could quickly accumulate sufficient reserves. Most could do so by cutting consumption below the optimal level for a single year.
4259 Ch07 4/12/02 10:20 am Page 343
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
343
Almost all could do so by acting prudently two years in a row. So, the periods after shocks when peasants possessed insufficient reserves were short, and in those vulnerable moments, consolidation and pooling arrangements protected peasants until reserves returned to optimal levels. The fourth is the discount rate. Like all risk-mitigating strategies, scattering was a method for making intertemporal substitutions. Scattering lowered consumption in the present and extended (one hoped) the number of years of consumption in the future. When discount rates rose and the value of future consumption fell, scattering’s attractiveness decreased. People valued the present and preferred agricultural arrangements that concentrated consumption where it was most rewarding. When discount rates fell and the value of the future rose, scattering’s attractiveness increased. So, in the absence of pooling and storage, low discount rates should have encouraged scattering. But, lower discount rates also enhanced the effectiveness of risk-pooling arrangements and the appeal of storage. People who valued the future highly helped neighbors today because the cost of cooperation today was low and the value of future reciprocation was high. People who valued the future also increased storage because storage was an efficient method of shifting consumption through time. The fifth is risk-reducing mechanisms that were not included in the model but which did exist in medieval England. Peasants had access to markets for credit, assets, labor, and victuals. Charity was widely available. The layout of arable fields mitigated climatic shocks. Ridges and furrows facilitated drainage, as did the orientation of furlongs up the slopes of hills. Crop diversification slowed the spread of pests and moderated seasonal variations in yields due to precipitation, temperature, sunlight, and wind. Some of these methods may have been more efficient than storing grain and pooling risk. Labor markets allowed needy individuals to trade leisure for food – a small inconvenience when survival was at stake. Credit markets enabled individuals to smooth consumption over time in response to adverse shocks rather than incurring the expense of storing grain in good years and bad. Asset markets provided a similar service and an additional advantage. Durable goods that could be sold for foodstuffs during troubled times provided a stream of services during good years. Adding these low-cost high-payoff methods of mitigating risk to the model would strengthen the conclusions of this essay and weaken McCloskey’s claim that scattering was the leastexpensive method of mitigating risk. The sixth is the covariance of yields of scattered fields. When prosperous peasants could pool risk and store grain (or smooth consumption across time and individuals using other methods) even at the exorbitant rates estimated by McCloskey, they could smooth away most (if not all) idiosyncratic risk. They 343
4259 Ch07 4/12/02 10:20 am Page 344
344 1 2
GARY RICHARDSON
need not fear diminutive downturns. Their worry was correlated shocks – calamities that threatened the survival of entire communities.
3 4
IV. DISCUSSION
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
For three decades, McCloskey’s scholarship shaped the research agenda regarding medieval English agriculture. Her prose posed piercing questions. Did markets exist in medieval England? Could peasants participate? Or were peasants bound by tradition and unaffected by market forces? How did markets’ existence or absence influence the lives of the rural majority? Were medieval villagers different from modern descendants? Were they more or less altruistic? Was cooperation widespread? Was charity abundant? Did peasants have access to cheap relief? Why did peasants cultivate strips of land scattered throughout the open fields of their village? Was scattering prudent, rationale, and efficient? What was its purpose? Was it a method of sharing risk? Did other methods of mitigating risk exist? Which protected peasants best? Answering McCloskey’s inquiries took three decades. Only in the last few years have the facts become clear. Markets for food, labor, land, assets, and credit existed within medieval English villages and connected them to the wider world. Peasants had access to those markets. A high percentage of rural residents worked for wages and purchased the necessities of life. Cooperation was widespread. Teamwork occurred in planting, harvesting, husbandry, religious organizations, and courts of law. Repeated interaction and anticipated reciprocation sustained communal arrangements among self-interested individuals. Peasants (like most people in most times and places) employed multiple methods to offset adverse shocks. Mutual-assurance societies and customary poor laws protected villagers from the uncertainties of everyday life. Charitable institutions also mitigated intermittent shocks. So did the storage of grain for future consumption. All of these methods smoothed shocks that scattering could not. These shocks included climatic calamities such as droughts, floods, frosts, insect infestations, and crop diseases that extended over regions wider than a few hundred yards; risks of everyday life such as accidents, illness, infirmity, old age; deaths of family, friends, and farm animals; and unexpected expenses such as court costs, legal fines, home repairs, equipment failures, and new taxes. Many of the alternative methods of mitigating risk also protected groups of individuals and entire villages, in contrast to scattering, which protected households with enough land to make scattering worthwhile, but pushed everyone else towards the brink of subsistence. While this recent research contradicts the assumptions of McCloskey’s model of medieval agriculture, it is consistent with the Coasian logic underlying her
4259 Ch07 4/12/02 10:20 am Page 345
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
345
claims. Coase showed that inefficiencies persist only if missing or prohibitively expensive markets prevent people from rearranging their activities and alleviating the inefficiency. McCloskey combined that logic and the notion that markets did not exist in medieval England – which was the conventional academic wisdom thirty years ago when she wrote the first in her series of seminal essays – and concluded that open-field agricultural operated inefficiently. Recent research reaches the opposite conclusion. The yields of open fields equaled or exceeded those of consolidated fields from the earliest written records until the eve of enclosure. Open-fields operated efficiently, and markets were ubiquitous. These new facts fit the Coase theorem more comfortably than the old facts (which appeared to be consistent with Coase only when viewed through a kaleidoscope of erroneous assumptions and opaque miscalculations) and generate a new implication. Inefficient agricultural arrangements could not have persisted in medieval England, because the extensive markets, stable institutions, and high volume of transactions enabled peasants to rearrange activities and alleviate inefficiencies. What does this finding reveal about McCloskey’s claim that the rationale for scattering was the reduction of risk? Her hypothesis remains consistent with the facts, but it must be modified in several ways. First, scattering could (and did) coexist with alternative risk-sharing institutions such as pooling and storage. These arrangements were complements, not substitutes. Second, scattering was a low cost, low benefit method of mitigating risk. It had little influence on the efficiency of agricultural arrangements. And, scattering did little to mitigate risk; it slightly reduced shocks caused by local calamities, which were a small element in the large array of risks faced by medieval English peasants. Third, all peasants, rather than the small fraction of the peasantry possessing the sizeable portfolios of arable land, benefited from scattering. Fourth, scattering facilitated the operations of alternative risk-sharing arrangements, such as mutual aid societies, by reducing the size and frequency of ex-post transfers and the financial and organizational burdens placed upon insurance institutions. This extended form of McCloskey’s hypothesis leaves room for alternative theories concerning the origins and persistence of the open fields. Many have been advanced over the years. One of these alternatives is that low levels of capital forced farmers to plant, plow, and harvest cooperatively. No one had enough animals and implements to act independently. Lack of resources prevented peasants from constructing sufficient infrastructure (fences, drainage ditches, alleyways, etc.) to enable individuals to farm on their own. So, individuals pooled their resources to work the land. Two, intermingling farm fields facilitated monitoring of labor inputs. Farmers could readily determine if cultivators of neighboring strips expended sufficient effort on weeding and other 345
4259 Ch07 4/12/02 10:20 am Page 346
346 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
GARY RICHARDSON
mundane yet crucial tasks. Three, scattering enabled farmers to allocate labor efficiently during the peak periods of planting and harvesting. For example, crops on hilltops ripened at a different rate than those on valley floors. Possessing fields in both places permitted farmers to work one field and then the other. Four, scattering facilitated diversification among crops. For instance, some grains grew better on dry ridgelines. Others grew better in sheltered hollows. Possessing fields in both places enabled farmers to plant each of their crops in a favorable location. Five, the size of strips was the minimum efficient scale. Plots averaged one acre each. A farmer and his family could sow, weed, or harvest about an acre in a single day. Time was never lost shifting equipment from one plot to the next. Six, the intermingling of strips was a symptom of an active land market, and buying and selling acre-sized plots enabled land to be used as a mechanism for life cycle savings. The young worked for wages and accumulated portfolios of arable land. The middle-aged hired laborers to help them work their land. The old transferred their land to their children or sold their land to the highest bidder to support themselves in old age. These alternative explanations have a common theme. Open fields persisted because they minimized inputs of capital and labor, maximized outputs of food and fodder, and optimized net returns from arable land. Enclosure occurred when scattering lost its efficiency advantage over consolidation. Over the centuries, many forces influenced the relative efficiency of the former and the latter. Declines in the cost of capital made independent cultivation less costly. The same is true for increases in the efficiency of labor and the availability of hired hands. Changes in laws concerning landholding and the composition of Parliament also reduced the costs of enclosure. New species of crops and breeds of animals may have made independent farming increasingly attractive. Which factor was the most influential remains an open question. As with all multi-causal phenomena, distinguishing the primary and secondary causes requires comparative analysis. Some villages enclosed early in the fifteenth century. Other villages enclosed late in the eighteenth century. Villages differed in population, area, soil, climate, non-agriculture employment opportunities, income, distribution of wealth, proximity to transport, towns, markets, and additional attributes. Villages also changed over time. So did society as a whole and many of the factors influencing the value of enclosure. The cost of capital fell. The stock of capital grew. The population rose, fell, stagnated, and then rose again. Technology progressed. Better tools and techniques raised the productivity of labor. New species of crops, breeds of animals, and methods of cultivation raised the productivity of land. Transport costs fell. Mercantile activity increased. Legal changes loosened bonds between landlords and tenants. Innumerable other aspects of society changed over time as well. Determining
4259 Ch07 4/12/02 10:20 am Page 347
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
347
which of these changes triggered enclosure entails detecting correlations between causes and consequences. The straightforward method is to compare dates of enclosure to characteristics of villages and aspects of society that changed over time. Is this feasible? Today, no. Soon, yes. The technical term for the statistical procedure is multiple-hazard duration analysis. The technique requires panel data. At this moment the necessary information is not in electronic form, but more than enough evidence resides in English archives, and the current craze for computerizing historical records provides hope that during the next few years, we will answer the questions posed by this essay. Why during the Middle Ages did the yields of open fields exceed the yields of enclosed fields? Why during early-modern times did the yields of open fields fall below the yields of enclosed fields? Why did villages enclose at different dates? Why did enclosure occur in fits and starts over several centuries? How did enclosure relate to the Agricultural Revolution? Was it a cause? Was it a consequence? Was it merely a symptom of some underlying force? How much of the variation in field structures across time and space is explained in each of the existing theories? Do large portions of the variation remain unexplained? Answering the penultimate question will reveal the relative merits of the existing theories. Answering the final question will reveal their aggregate explanatory power. If it is high, then the questions are answered, and the debate is over. If it is low, then only devising new theories and pondering additional economic, institutional, and exogenous variables will resolve the riddles of the open fields.
24 25
NOTES
26 27 28 29 30 31 32 33 34 35
1. E-mail to the author on 9 October 1995. The original discusses wages and credit at length after discussing charity. 2. Defendants wagered their law by presenting oath helpers who joined them in swearing that their statements were true and that they were trustworthy. This fundamental method of medieval jurisprudence was based on the belief that men who swore over religious relics would be reluctant to lie and condemn themselves to the wrath of the Lord, who would punish them in purgatory, and the wrath of their neighbors, who would ostracize them for blasphemous behavior. Compurgation could be corrupted, particularly when courts allowed oath-helpers to be paid for their services, so creditors usually insisted on documentation.
36 37
ACKNOWLEDGMENTS
38 39 40
I thank UC Berkeley’s Department of Demography, the All-UC Group in Economic History, and the Social Science Research Council for financial 347
4259 Ch07 4/12/02 10:20 am Page 348
348 1 2 3 4
GARY RICHARDSON
support, the editor, and anonymous referees for invaluable advice, and participants in seminars at UC Berkeley, UCLA, UC Davis, and UC Irvine for insightful comments. Shaista and Shagufta Ahmed provided exceptional research and editorial assistance.
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
REFERENCES Allen, R. C. (1982). The Efficiency and Distributional Consequences of Eighteenth Century Enclosures. The Economic Journal, 92, 937–953. Allen, R. C., & Grada, C. O. (1988). On the Road Again With Arthur Young: English, Irish, and French Agriculture During the Industrial Revolution. Journal of Economic History, 48, 93–116. Appleby, A. B. (1978). Famine in Tudor and Stuart England. Stanford, CA: Stanford University Press. Atkinson, T. D. (1898). On the Gilds of Cambridgeshire. Proceedings of Cambridge Antiquarian Society, 9, 385–390. Ault, W. O. (1967). Manor Court and Parish Church in Fifteenth-Century England: A Study of Village By-Laws. Speculum, 42, 53–67. Ault, W. O. (1970). The Village Church and the Village Community in Mediaeval England. Speculum, 45, 197–215. Ault, W. O. (1972). Open-field Farming in Medieval England. London, New York: Allen and Unwin; Barnes and Noble. Bainbridge, V. R. (1996). Gilds in the Medieval Countryside: Social and Religious Change in Cambridgeshire, c. 1350–1558. Rochester, NY: Boydell Press. Baker, J. H. (1995). An Introduction to English Legal History. London; Boston: Butterworths. Banerjee, A. V., Besley, T., & Guinnane, T. W. (1994). Thy Neighbor’s Keeper: the Design of a Credit Cooperative With Theory and a Test. Quarterly Journal of Economics, 109, 491–515. Barron, C. M. (1985). The Parish Fraternities of Medieval London. In: C. M. Barron & C. HarperBill (Eds), The Church in Pre-Reformation Society (pp. 13–27). Woodbridge. Beardwood, A. (1931). Alien Merchants in England, 1350 to 1377. Cambridge, MA: Mediaeval Academy of America. Bennett, H. S. (1967). Life on the English Manor: a Study of Peasant Conditions, 1150–1400. Cambridge: Cambridge University Press. Bennett, J. M. (1987). Women in the Medieval English Countryside: Gender and Household in Brigstock Before the Plague. New York: Oxford University Press. Bennett, J. M. (1992). Conviviality and Charity in Medieval and Early Modern England. Past & Present, 134, 19–41. Besley, T. (1995). Non-market Institutions for Credit and Risk Sharing in Low-income Countries. Journal of Economic Perspectives, 9, 115–127. Besley, T., & Coate, S. (1995). Group Lending, Repayment Incentives and Social Collateral. Journal of Development Economics, 46, 1–18. Besley, T., Coate, S., & Loury, G. (1993). The Economics of Rotating Savings and Credit Associations. American Economic Review, 83, 792–810. Besley, T., & Levenson, A. R. (1996). The Role of Informal Finance in Household Capital Accumulation: Evidence from Taiwan. Economic Journal, 106, 39–59. Beveridge, W. H. B. (1939). Prices and Wages in England From the Twelfth to the Nineteenth Century. London and New York: Longman Green.
4259 Ch07 4/12/02 10:20 am Page 349
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
349
Blair, J. (1980). Religious Gilds as Landowners in the Thirteenth and Fourteenth Centuries: the Example of Chesterfield. In: P. Riden (Ed.), The Medieval Town in Britain: Papers From the First Gregynog Seminar in Local History (pp. 35–49). Cardiff: University College Cardiff Dept. of Extra-Mural Studies. Blanchard, I. (1972). The Miner and the Agricultural Community in Late Medieval England. Agricultural History Review, 20, 93–106. Booth, P. H. (1981). The Financial Administration of the Lordship and County of Chester, 1272–1377. Manchester: Manchester University Press. Bossy, J. (1973). Blood and Baptism: Kinship, Community and Christianity in Western Europe from the Fourteenth to the Seventeenth Centuries. SCH, 10, 129–143. Brentano, L. (1870). On the History and Development of Gilds, and the Origin of Trade-unions. London: Trübner & Co. Brigden, S. (1984). Religion and Social Obligation in Early Sixteenth Century London. Past & Present, 103, 67–112. Briggs, H. M. (1935). Surrey Manorial Accounts; a Catalogue and Index of the Earliest Surviving Rolls Down to the Year 1300. Frome, London: Printed by Butler & Tanner Ltd. for the Surrey Record Society. Brink, R. V. D., & Chavas, J. P. (1997). The Microeconomics of an Indigenous African Institution: the Rotating Savings and Credit Association. Economic Development & Cultural Change, 45, 745–772. Britnell, R. H. (1966). Production for the Market on a Small Fourteenth-Century Estate. Economic History Review, 19, 380–387. Britnell, R. H. (1980). Minor Landlords in England and Medieval Agrarian Capitalism. Past and Present, 89, 3–22. Britnell, R. H. (1981). The Proliferation of Markets in England, 1200–1349. Economic History Review, 34, 209–221. Britnell, R. H. (1986). Growth and Decline in Colchester, 1300–1525. Cambridge: Cambridge University Press. Britnell, R. H. (1996). The Commercialization of English Society, 1000–1500. Manchester, U.K.; New York: Manchester University Press: Distributed exclusively in the USA by St. Martin’s Press. Britton, E. (1977). The Community of the Vill: a Study in the History of the Family and Village Life in Fourteenth-Century England. Toronto: Macmillan of Canada. Burgess, C. (1987). By Quick and By Dead: Wills and Pious Provision in Late Medieval Bristol. Economic History Review, 40, 837–858. Burgess, C. (1988). A Fond Thing Vainly Invented: an Essay on Purgatory and Pious Motive in Late Medieval England. In: S. J. Wright (Ed.), Parish, Church and People: Local Studies in Lay Religion, 1350–1750 (pp. 56–84). London: Hutchinson. Burgess, C. (1991). The Benefactions of Mortality: The Lay Response in the Late Medieval Urban Parish. In: D. M. Smith & Borthwick Institute of Historical Research (Eds), Studies in Clergy and Ministry in Medieval England: Purvis Seminar Studies (pp. 65–86). York, England: University of York, Borthwick Institute of Historical Research. Burgess, C., & Kumin, B. (1993). Penitential Bequests and Parish Regimes in Late Medieval England. Journal of Ecclesiastical History, 44, 610–630. Campbell, B. M. S. (1983). Agricultural Progress in Medieval England: Some Evidence from Eastern Norfolk. Economic History Review, 36, 26–46. Campbell, B. M. S. (1984). Population Pressure, Inheritance and the Land Market in a FourteenthCentury Peasant Community. In: R. Smith (Ed.), Land, Kinship and Life-Cycle (pp. 87–134). Cambridge: Cambridge University Press.
349
4259 Ch07 4/12/02 10:20 am Page 350
350 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
GARY RICHARDSON
Campbell, B. M. S. (1991). Land, Labour, Livestock, and Productivity Trends in English Seignorial Agriculture, 1208–1450. In: B. M. S. Campbell & M. Overton (Eds), Land, Labour, and Livestock: Historical Studies in European Agricultural Productivity (pp. 144–182). Manchester: Manchester University Press. Campbell, B. M. S. (1993a). The Common Fields of England. Economic History Review, 46, 613–614. Campbell, B. M. S. (1993b). A New Perspective on Medieval and Early Modern Agriculture: Six Centuries of Norfolk farming c.1250–1850. Past & Present, 141, 38–105. Campbell, B. M. S. (1995). A Second Handbook of Anglo-Saxon Food and Drink: Production and Distribution. Economic History Review, 48, 818. Campbell, B. M. S. (1996). Matching Supply to Demand: Agricultural Production and Disposal By English Demesnes in the Century of the Black Death. Paper presented at the annual conference of the Economic History Association. Campbell, B. M. S. (1997). Matching Supply to Demand: Crop Production and Disposal by English Demesnes in the Century of the Black Death. Journal of Economic History, 57, 827–858. Campbell, B. M. S., & Overton, M. (1991). Land, Labour, and Livestock: Historical Studies in European Agricultural Productivity. Manchester: Manchester University Press. Campbell, B. M. S., & Overton, M. (1993). A New Perspective on Medieval and Early Modern Agriculture: Six Centuries of Norfolk Farming, c.1250–1850. Past & Present, 141, 38–105. Carus-Wilson, E. M. (1963). England’s Export trade, 1275–1547. Oxford: Clarendon Press. Caselli, F. (1997). Rural Labor and Credit Markets. Journal of Development Economics, 54, 235–260. Champion, B. (1980). The Gilds of Medieval Beverley. In: P. Riden (Ed.), The Medieval Town in Britain: Papers from the First Gregynog Seminar in Local History (pp. 51–66). Cardiff: University College Press. Chibnall, A. C. (1950). Subsidy Role for the County of Buckingham, Anno 1524. Bedford Bedfordshire: Printed for the Buckingham Record Society. Clark, E. (1981). Debt Litigation in a late Medieval English Vill. In: J. A. Raftis (Ed.), Pathways to Medieval Peasants (pp. 274–279). Toronto: Toronto University Press. Clark, E. (1982). Some Aspects of Social Security in Medieval England. Journal of Family History, 7, 307–320. Clark, G. (1998). Commons Sense: Common Property Rights, Efficiency, and Institutional Change. Journal of Economic History, 58, 73–102. Clay, R. M. (1909). The Mediaeval Hospitals of England. London: Methuen & Co. Cobbett, L. (1935). The Hospitals of St. John the Baptist and St. Mary Magdalene at Ely. Proceedings of Cambridge Antiquarian Society, 36, 58–75. Cook, G. H. (1947). Medieval Chantries and Chantry Chapels. London, Phoenix House. Cornwall, J. (1970). English Population in the Early Sixteenth Century. Economic History Review, 23, 32–44. Crouch, D. J. F. (1988). Religious Gilds in Late Medieval Yorkshire, 1400–1550. Medieval Life, 9, 30–32. Crouch, D. J. F. (2000). Piety, Fraternity, and Power: Religious Gilds in Late Medieval Yorkshire, 1389–1547. Woodbridge, Suffolk, U.K.; Rochester, NY: York Medieval Press. Dewindt, E. B. (1972). Land and People in Holywell-cum-Needingworth; Structures of Tenure and Patterns of Social Organization in an East Midlands Village, 1252–1457. Toronto: Pontifical Institute of Mediaeval Studies. Dewindt, E. B. (1990). The Court Rolls of Ramsey, Hepmangrove, and Bury, 1268–1600. Toronto, Canada: Pontifical Institute of Mediaeval Studies.
4259 Ch07 4/12/02 10:20 am Page 351
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
351
Dinn, R. (1990). Baptism, Spiritual Kinship and Popular Religion in Late Medieval Bury St. Edmunds. Bulletin of John Rylands Library, 122, 93–106. Duparc, P. (1975). Confraternities of the Holy Spirit and Village Communities in the Middle Ages. In: F. L. Cheyette (Ed.), Lordship and Community in Medieval Europe: Selected Readings. Huntington, N.Y.: R. E. Krieger Pub. Co. Dyer, C. (1988). Documentary Evidence: Problems and Enquiries. In: G. Astill & A. Grant (Eds), The Countryside of Medieval England. Oxford: B. Blackwell. Dyer, C. (1989). Standards of Living in the Later Middle Ages: Social Change in England, c.1200–1520. Cambridge: Cambridge University Press. Dyer, C. (1994). Everyday Life in Medieval England. London: Hambledon Press. Fafchamps, M. (1992). Solidarity Networks in Preindustrial Societies: Rational Peasants with a Moral Economy. Economic Development & Cultural Change, 41, 147–174. Farmer, D. L. (1957). Some Grain Price Movements in thirteenth-century England. Economic History Review, 10, 218–219. Farmer, D. L. (1958). An Examination of Price Fluctuations in Certain Articles in the Twelfth, Thirteenth, and Early Fourteenth Centuries. History. Oxford: Oxford University Press. Farmer, D. L. (1969). Some Livestock Price Movements in Thirteenth-Century England. The Economic History Review, 22, 1–16. Farmer, D. L. (1989). Two Wiltshire Manors and their Markets. Agricultural History Review, 37, 1–11. Farmer, D. L. (1991a). Marketing the Produce of the Countryside 1200–1500. In: E. Miller (Ed.), The Agrarian History of England and Wales Volume III: 1348–1500 (pp. 324–430). Cambridge: Cambridge University Press. Farmer, D. L. (1991b). Prices and Wages 1350–1500. In: E. Miller (Ed.), The Agrarian History of England and Wales Volume III: 1348–1500 (pp. 431–525). Cambridge: Cambridge University Press. Fenoaltea, S. (1976). Risk, Transaction Costs, and the Organization of Medieval Agriculture. Explorations in Economic History, 13, 129–151. Finucane, R. C. (1981). Sacred Corpse: Profane Carrion: Social Ideas and Death Rituals in the Later Middle Ages. In: J. Whaley (Ed.), Mirrors of Mortality: Studies in the Social History of Death (pp. 40–60). London. Fisher, H. E. S., Jurica, A. R. J., & Clapp, B. W. (1976). Documents in English Economic History. London: G. Bell. Fryde, N. (1985). Gilds in England before the Black Death. Vortrage Und Forschung, 29, 215–229. Fuentes, G. A. (1996). The Use of Village Agents in Rural Credit Delivery. Journal of Development Studies, 33, 188–209. Galloway, P. R. (1998). Basic Patterns in Annual Variations in Fertility, Nuptiality, Mortality, and Prices in Pre-Industrial Europe. Population Studies, 42, 275–302. Gasquet, F. A. (1907). Parish Life in Mediaeval England. London: Methuen. Godfrey, W H. (1955). The English Almshouse, With Some Account of its Predecessor, the Medieval Hospital. London: Faber and Faber. Grant, A. (1988). Animal Resources. In: G. Astill & A. Grant (Eds), The Countryside of Medieval England (pp. 149–187). Oxford: Blackwell Publishers. Gras, N. S. B. (1926). The Evolution of the English Corn Market from the Twelfth to the Eighteenth Century. Oxford: Oxford University Press. Hanawalt, B. (1984). Keepers of the Lights: Late Medieval English Parish Gilds. Journal of Medieval and Renaissance Studies, 14, 21–37. Hanawalt, B. (1986). The Ties That Bound: Peasant Families in Medieval England. New York: Oxford University Press.
351
4259 Ch07 4/12/02 10:20 am Page 352
352 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
GARY RICHARDSON
Hanawalt, B., & Mccree, B. (1992). The Guilds of “Homo Prudens” in Late Medieval England. Continuity and Change, 7, 163–179. Harvey, B. F. (1993). Living and Dying in England, 1100–1540: the Monastic Experience. Oxford: Oxford University Press. Harvey, P. D. A. (1965). A Medieval Oxfordshire Village: Cuxham, 1240 to 1400. Oxford: Oxford University Press. Harvey, P. D. A. (1984). The Peasant Land Market in Medieval England. Oxford: Oxford University Press. Heal, F. (1984). The Idea of Hospitality in Early Modern England. Past and Present, 102, 66–93. Hicks, M. (1985). Chantries, Obits, and Almshouses: The Hungerford Foundations 1325–1478. In: C. M. Barron & C. Harper-Bill (Eds), The Church in Pre-Reformation Society (pp. 123–142). Woodbridge, Suffolk ; Dover, N. H. USA: Boydell Press. Hilton, R. H. (1975a). The English Peasantry in the Later Middle Ages: the Ford Lectures for 1973 and Related Studies. Oxford: Clarendon Press. Hilton, R. H. (1975b). The Small Town as Part of Peasant Society. In: R. H. Hilton (Ed.), The English Peasantry in the Later Middle Ages: the Ford Lectures for 1973 and Related Studies (pp. 256–289). Oxford: Clarendon Press. Hilton, R. H. (1982). Lords, Burgesses and Hucksters. Past & Present, 97, 3–15. Hilton, R. H. (1983). A Medieval Society: the West Midlands at the End of the Thirteenth Century. Cambridge: Cambridge University Press. Hoff, K., & Stiglitz, J. E. (1997). Moneylenders and Bankers: Price-Increasing Subsidies in a Monopolistically Competitive Market. Journal of Development Economics, 52, 429–462. Homans, G. C. (1941). English Villagers of the Thirteenth Century. Cambridge: Harvard University Press. Jones, W. R. (1974). English Religious Brotherhoods and Medieval Lay Piety: The Inquiry of 1388–1389. The Historian, 36, 646–659. Jordan, W. K. (1959). Philanthropy in England, 1480–1660; a Study of the Changing Pattern of English Social Aspirations. New York: Russell Sage Foundation. Jordan, W. K. (1960). The Forming of the Charitable Institutions of the West of England; a Study of the Changing Pattern of Social Aspirations in Bristol and Somerset, 1480–1660. Philadelphia: American Philosophical Society. Jordan, W. K. (1961). The Charities of Rural England, 1480–1660. London: Allen & Unwin. Karakacili, E. (2000). Placing Productivity In the Social Context: Average Agrarian Labour Productivity Rates in the Pre-1350 Period. Paper Presented at the UC Davis Agricultural History Mini-Conference. Kerridge, E. (1969). Agrarian Problems in the Sixteenth Century and After. London: Allen & Unwin. Kershaw, I. (1973). The Great Famine and Agrarian Crisis in England 1315–1322. Past & Present, 59, 3–50. Kimball, M. S. (1988). Farmers’ Cooperatives as Behavior Toward Risk. American Economic Review, 78, 224–257. Kosminski, E. I. A. (1956). Studies in the Agrarian History of England in the Thirteenth Century. Oxford: Blackwell. Kussmaul, A. (1981). Servants in Husbandry in Early Modern England. Cambridge: Cambridge University Press. Langdon, J. (1986). Horses, Oxen, and Technological Innovation: the Use of Draught Animals in English Farming from 1066 to 1500. Cambridge: Cambridge University Press. Lee, R. (1973). Population in Preindustrial England: An Econometric Analysis. Quarterly Journal of Economics, 87, 581–607.
4259 Ch07 4/12/02 10:20 am Page 353
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
353
Lucas, H. (1930). The Great European Famine of 1315, 1316, and 1317. Speculum, 5, 343–377. Ma, C. H., & Smith, B. D. (1996). Credit Market Imperfections and Economic Development: Theory and Evidence. Journal of Development Economics, 48, 351–387. Macfarlane, A. (1979). The Origins of English Individualism: the Family, Property, and Social Transition. Cambridge: Cambridge University Press. Mason, E. (1976). The Role of the English Parishioner 1000–1500. Journal of Economic History, 27, 17–29. Masschaele, J. (1993). Transport Costs in Medieval England. Economic History Review, 46, 266–279. Mattingley, H. (1989). The Medieval Parish Guilds of Cornwall. Journal of the Royal Institution of Cornwall, 10, 290–329. McCloskey, D. N. (1972). The Enclosure of Open Fields: Preface to a Study of Its Impact on the Efficiency of English Agriculture in the Eighteenth Century. Journal of Economic History, 32, 15–35. McCloskey, D. N. (1975a). The Economics of Enclosure: A Market Analysis. In: W. N. Parker & E. L. Jones (Eds), European Peasants and Their Markets: Essays in Agrarian Economic History (pp. 123–160). Princeton: Princeton University Press. McCloskey, D. N. (1975b). Persistence of English Common Fields. In: W. N. Parker & E. L. Jones (Eds), European Peasants and Their Markets: Essays in Agrarian Economic History (pp. 73–119). Princeton: Princeton University Press. McCloskey, D. N. (1976). English Open Fields as Behavior Towards Risk. Research in Economic History, 1, 124–170. McCloskey, D. N. (1977). Fenoaltea on Open Fields: A Comment. Explorations in Economic History, 14, 402–404. McCloskey, D. N. (1979a). Another Way of Observing Open Fields: A Reply to A. R. H. Baker. Journal of Historical Geography, 5, 426–429. McCloskey, D. N. (1979b). Explaining Open Fields: An Exchange of Letters with Charles Wilson. Journal of European Economic History, 8, 203–207. McCloskey, D. N. (1980). Scattering and Open Fields: A Comment on Michael Mazur’s Article. Journal of European Economic History, 9, 209–214. McCloskey, D. N. (1983). Theses on Enclosure. Agricultural History: Papers Presented to the Economic History Society Conference at Canterbury, 56–72. McCloskey, D. N. (1989). The Open Fields of England: Rent, Risk, and the Rate of Interest, 1300–1815. In: D. W. Galenson (Ed.), Markets in History: Economic Studies of the Past (pp. 5–51). Cambridge: Cambridge University Press. McCloskey, D. N. (1991). The Prudent Peasant: New Findings on Open Fields. Journal of Economic History, 51, 343–355. McCloskey, D. N., & Nash, J. (1984). Corn at Interest: The Extent and Cost of Grain Storage in Medieval England. The American Economic Review, 74, 174–187. Mcintosh, M. K. (1980). Land, Tenure, and Population in the Royal Manor of Havering, Essex, 1251–1352/1353. Economic History Review, 33, 17–31. Mcintosh, M. K. (1986). Autonomy and Community: the Royal Manor of Havering, 1200–1500. Cambridge: Cambridge University Press. Mcintosh, M. K. (1998). Controlling Misbehavior in England, 1370–1600. Cambridge: Cambridge University Press. Mcree, B. R. (1987). Religious Gilds and Regulation of Behavior in Late Medieval Towns. In: J. T. Rosenthal & C. Richmond (Eds), People, Politics, and Community in the Later Middle Ages (pp. 108–122). New York: St. Martin’s Press.
353
4259 Ch07 4/12/02 10:20 am Page 354
354 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
GARY RICHARDSON
Mcree, B. R. (1993). Charity and Gild Solidarity in Late Medieval England. Journal of British Studies, 32, 195–225. Moorman, J. R. H. (1945). Church Life in England in the Thirteenth Century. Cambridge: Cambridge University Press. Morduch, J. (1995). Income Smoothing and Consumption Smoothing. The Journal of Economic Perspectives, 9, 103–114. Nightingale, P. (1990). Monetary Contraction and Mercantile Credit in Later Medieval England. Economic History Review, 43, 560–575. Page, F. M. (1930). The Customary Poor-Law of Three Cambridgeshire Manors. Cambridge Historical Journal, 3, 125–133. Palmer, W. M. (1902). Village Gilds of Cambridgeshire. Cambridgeshire and Huntingdonshire Archaeological Society Review, 330–402. Phelps Brown, E. H., & Hopkins, S. V. (1955). Seven Centuries of Building Wages. Economica, 22, 195–206. Phelps Brown, E. H., & Hopkins, S. V. (1956). Seven Centuries of the Prices of Consumables, Compared with Builders’ Wage Rates. Economica, 23, 296–314. Platteau, J. P. (1991). Traditional Systems of Social Security and Hunger Insurance: Achievements and Modern Challenges. In: E. Ahmad, J. Dreze, J. Hills & A. K. Sen (Eds), Social Security in Developing Countries (pp. 112–170). Oxford: Oxford University Press. Platteau, J. P. (1997). Mutual Insurance as an Elusive Concept in Traditional Rural Communities. Journal of Development Studies, 33, 764–796. Popkin, S. L. (1979). The Rational Peasant: the Political Economy of Rural Society in Vietnam. Berkeley: University of California Press. Posner, R. A. (1980). Anthropology and Economics. The Journal of Political Economy, 88, 608–616. Postan, M. M. (1954). The Famulus; the Estate Labourer in the XIIth and XIIIth Centuries. Cambridge: Cambridge University Press. Postan, M. M. (1973). Essays on Medieval Agriculture and General Problems of the Medieval Economy. Cambridge: Cambridge University Press. Postan, M. M. (1975). The Medieval Economy and Society: an Economic History of Britain in the Middle Ages. Harmondsworth: Penguin. Postan, M. M., & Titow, J. Z. (1973). Heriots and Prices on Winchester Manors With Statistical Notes on Winchester Heriots by J. Longden. In: M. M. Postan (Ed.), Essays on Medieval Agriculture and General Problems of the Medieval Economy (pp. 107–149). Cambridge: Cambridge University Press. Power, E. E. (1941). The Wool Trade in English Medieval History. Oxford: Oxford University Press. Raftis, J. A. (1964). Tenure and Mobility; Studies in the Social History of the Mediaeval English Village. Toronto: Pontifical Institute of Mediaeval Studies. Raftis, J. A. (1965). Social Structures in Five East Midland Villages: A Study in the Use of Court Roll Data. Economic History Review, 18, 83–100. Razi, Z. (1981). Family, Land, and Village Community in Later Medieval England. Past and Present, 93, 3–36. Richardson, G. (1999). The Prudent Village: A Corroboration of Kimball’s Conjecture. University of California at Irvine, Department Working Paper, No. 101. Rogers, J. E. T. (1884). Six Centuries of Work and Wages. The History of English Labour. New York: G. P. Putnam’s Sons. Rogers, J. E. T. (1888). The Economic Interpretation of History. New York: G. P. Putnam’s Sons.
4259 Ch07 4/12/02 10:20 am Page 355
What Protected Peasants Best? 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
355
Rogers, J. E. T., & Rogers, A. G. L. (1866). A History of Agriculture and Prices in England, From the Year after the Oxford Parliament (1259) to the Commencement of the Continental War (1793). Oxford: Clarendon Press. Rosenthal, J. T. (1972). The Purchase of Paradise; Gift Giving and the Aristocracy, 1307–1485. London: Routledge & K. Paul. Rosser, A. G. (1984). The Essence of Medieval Urban Communities: The Vill of Westminister 1200–1540. Transactions of the Royal Historical Society, 5, 91–112. Rosser, A. G. (1988a). The Anglo-Saxon Gilds. In: J. Blair (Ed.), Minsters and Parish Churches: the Local Church in Transition, 950–1200 (pp. 31–33). Oxford: Oxford University Press. Rosser, A. G. (1988b). Communities of Parish and Guild in the Late Middle Ages. In: S. J. Wright (Ed.), Parish, Church and People: Local Studies in Lay Religion, 1350–1750 (pp. 29–55). London: Hutchinson. Rubin, M. (1987). Charity and Community in Medieval Cambridge. Cambridge: Cambridge University Press. Sayles, G. O. (1936). Select Cases in the Court of King’s Bench. London: B. Quaritch. Schumpeter, J. A., & Schumpeter, E. B. (1986). History of Economic Analysis. New York: Oxford University Press. Smith, J. T., Smith, L. T., & Brentano, L. (1870). English Gilds. London: Early English Text Society by N. Trübner & Co. Smith, R. M. (1984). Land, Kinship, and Life-cycle. Cambridge: Cambridge University Press. Snooks, G. D. (1994). Was the Industrial Revolution Necessary? London and New York: Routledge. Stewart-Brown, R. (1925). Calendar of County Court, City Court and Eyre Rolls of Chester, 1259–1297, With an Inquest of Military Service, 1288. Manchester: Chetham Society. Swanson, R. N. (1989). Church and Society in Late Medieval England. Oxford: Blackwell. Swanton, M. J. (1998). The Anglo-Saxon Chronicle. New York: Routledge. Thompson, J. A. F. (1965). Piety and Charity in Late Medieval London. Journal of Economic History, 16, 178–195. Tierney, B. (1959). Medieval Poor Law; a Sketch of Canonical Theory and its Application in England. Berkeley: University of California Press. Titow, J. Z. (1969). English Rural Society, 1200–1350. London: Allen and Unwin. Titow, J. Z. (1972). Winchester Yields; a Study in Medieval Agricultural Productivity. Cambridge: Cambridge University Press. Townsend, R. M. (1993). The Medieval Village Economy: a Study of the Pareto Mapping in General Equilibrium Models. Princeton: Princeton University Press. Townsend, R. M. (1995). Consumption Insurance: an Evaluation of Risk-Bearing Systems in LowIncome Economies. Journal of Economic Perspectives, 9, 83–102. Vale, M. G. A. (1976). Piety, Charity, and Literacy Among the Yorkshire gentry, 1370–1480. York: St Anthony’s Press. Walker, S. S. (1982). The Court Rolls of the Manor of Wakefield. Leeds: Yorkshire Archaeological Society. Walmsley, J. F. R. (1972). The Estate of Burton Abbey from the Eleventh to the Fourteenth Centuries. Birmingham: University of Birmingham Press. Waugh, S. L. (1991). England in the Reign of Edward III. Cambridge: Cambridge University Press. Westlake, H. F. (1919). The Parish Gilds of Mediæval England. London, Society for Promoting Christian Knowledge. Wood-Legh, K. L. (1965). Perpetual Chantries in Britain. Cambridge: Cambridge University Press. Woodward, G. W. O. (1966). The Dissolution of the Monasteries. London: Blandford Publishing.
355
4259 Ch07 4/12/02 10:20 am Page 356
356 1 2 3 4
GARY RICHARDSON
Wright, S. J. (1988). Parish, Church and People: Local Studies in Lay Religion, 1350–1750. London: Hutchinson. Wrightson, K., & Levine, D. (1979). Poverty and Piety in an English Village: Terling, 1525–1700. New York: Academic Press.
5 6 7 8
APPENDIX
9 10 11 12 13
The relationship between v, the standard deviation of average output, and s, the standard deviation of a scattered holding, can be derived as follows. By definition, Xv =
15
18 19 20 21 22 23
26 27 28 29 30 31 32
var(Xv) =
1 n2
35 36 37 38 39 40
i=1
Xs, i n
n
n
–W
(A1)
n
var(Xs, i) +
i=1
cov(Xs, i, Xs, j)]
(A2)
i=1 j=1
Since in general, cov(Y, Z) = yzRy, z where y = SD(y), z = SD(Z), and Ry,z is the correlation coefficient between Y and Z and since by definition, var(Xv) = 2v and var(Xs, i) = var(Xs, j) = 2s i and j, it is the case that cov(Xs,i, Xs, j) = ssRi, j = 2sRi, j, and after summing, simplifying and taking the square root of both sides, Eq. (2) becomes
33 34
n
n
=
where Xv is the random variable designating village average output, Xc, i is a random variable designating the ith individuals output in a village where residents cultivate consolidated fields, Xs, i is a random variable designating the ith individuals output in a village where residents farm sets of scattered strips, and W is a constant indicating the output lost due to the inefficiency of scattered strips. Since var(aZ + b) = a2var(Z) and var(Y + Z) = var(Y) + var(Z) + 2cov(Y, Z) where a, b R and Y and Z are random variable, Eq. (1) implies
24 25
Xc, i
i=1
16 17
n
14
v = s
1 + (n – 1)R n
(A3)
4259 Ch08 4/12/02 10:24 am Page 357
1 2 3 4 5 6 7 8 9 10
THE HESSE-CASSEL EMIGRANTS: A NEW SAMPLE OF TRANSATLANTIC EMIGRANTS LINKED TO THEIR ORIGINS
11 12 13 14
Simone A. Wegge
15 16 17 18
ABSTRACT
19 20 21 22 23 24 25 26 27
Over fifty million people emigrated from Europe between 1815 and 1930. Only a small fraction of these emigrants have been studied using data collected in their homelands. This paper presents a new sample of European emigrants who migrated from Germany in the mid-nineteenth century. The emigrants are linked to data describing the socioeconomic characteristics of their home villages. These data are promising for studying many issues in the migration literature, particularly migration behavior and the causes of emigration.
28 29 30
I. USING MIGRATION DATA TO STUDY EUROPEAN EMIGRATION IN THE NINETEENTH-CENTURY
31 32
Introduction
33 34 35
The years between 1815 and 1930 represent a period of large-scale European emigration that coincided with tremendous economic and demographic changes
36 37 38 39 40
Research in Economic History, Volume 21, pages 357–405. Copyright © 2003 by Elsevier Science Ltd. All rights of reproduction in any form reserved. ISBN: 0-7623-0993-8
357
4259 Ch08 4/12/02 10:24 am Page 358
358 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
SIMONE A. WEGGE
in both Europe and the New World.1 Social scientists continue to study this era of mass migration to seek an improved understanding of the nature of the migration decisions millions of Europeans made, their subsequent wealth and social mobility, and the effects of mass migration on the countries they left and those they settled in. This research is hampered by a lack of certain types of data sources. It matters since the issues we discuss and the hypotheses we test depend crucially on the kinds of data we have at our disposal. Many studies on international migration are based on aggregate data, in particular the various migration flow series Ferenczi and Willcox published in 1929. While these data exist in some form for many countries and are useful for addressing certain questions in economic history, they are inadequate for answering questions about how individual migrants behaved. Detailed information on the individual migrants and their origins can better elucidate issues in the microeconomics of migration behavior. Recent research also suggests that these data may seriously underreport population movements (Swierenga, 2000). To study migration with micro data many researchers have turned to the U.S. passenger ship records. These data are helpful in the context of immigration to the United States. As with any historical source, these records have their shortcomings, related primarily to data quality and under-representation of the actual immigrant flows.2 The existence of the American passenger records further encourages scholars to focus on assimilation issues. We know less about the origins of emigrants, their lives prior to moving, or how their home economies may have been affected by such emigrations. Thistlewaite’s 1960 critique of the state of migration scholarship as concentrating too much on the consequences and not the causes of migration remains relevant today.3 Micro sources that could facilitate an analysis of the causes of migration must come from the migrants’ homelands. Studies on emigration records are, however, rare, partly because such data do not always exist. Still, the emphasis on migrant origins is crucial to the story of nineteenth-century migration because migration decisions were most often made at home and in the context of a migrant’s life and his or her standard of living at home. Data collected at the origin by those who spoke the migrant’s language and were familiar with the local economy, for example, can facilitate a proper depiction of such important migrant characteristics as occupation. This paper presents a sample of migrants collected from the emigration lists of the German principality of Hesse-Cassel.4 Collected in the Hessisches Staatsarchiv in Marburg, Germany, the Hesse-Cassel data are highly unusual for what they can illustrate about migration behavior. These emigrant data are distinguished by the type of information they provide about Hessians who moved in the mid-nineteenth century. From them we obtain a more in-depth
4259 Ch08 4/12/02 10:24 am Page 359
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11
359
perspective on emigrants’ occupations, family structure, and wealth. I have further linked the emigrant data to census data describing their origins, which in turn allows one to examine the emigrants’ socioeconomic environment. Together, these sources of data allow for testing hypotheses about the importance of local and individual factors in migration decisions.5 The project discussed in this paper represents one of the few attempts in economics and history to use a large sample to study emigrants in the context of their homelands. More importantly, these various data, from emigrant and census records, shed light on the complex motivations of emigrants. They accomplish this by isolating individual and origin characteristics that may or may not have influenced migration decisions.
12 13
Sources of Micro Data on Migrants
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
The significance of micro data for social science research has been duly emphasized, especially in regard to demographic issues.6 Interesting micro data, particularly those at the individual level, are valuable for migration studies because they can illustrate better the complexities of migration behavior. Various sources of micro data exist on nineteenth-century European emigrants. European emigrants typically encountered data-collecting authorities at different points of their journey: in their country of origin before they departed, at the European harbor they departed from, in the harbors in which they arrived, and later in census records at the destination. These efforts have produced a variety of sources available for analysis, including emigration, port, immigration and finally census statistics. Each source that remains has its own advantages but also disadvantages, although most sources seem to undercount migrants.7 I address emigration data sources first. Unfortunately migration statistics were not always collected with future historians in mind. They often lack pertinent details about the migrants themselves. Many sources of emigration data, for instance, are basically aggregate figures on the emigration flows that occurred. Often what have survived are these summary statistics but not much data at the level of the individual, as is the case for England, Ireland, Scotland, and Wales (Baines, 1991, p. 12; Swierenga, 2000, p. 305). Tracing emigrants back to their origins and the site of their decision to leave is typically impossible for many European emigrants because we also do not know the identity of their home village. Knowing the home communities can be especially worthwhile if emigration was geographically concentrated, as Baines (1985, pp. 13, 23–25) has argued in the case of Austria, Finland, Germany, Italy, and Sweden. With a measure of the distribution of local emigration rates, a fuller understanding of emigration behavior can ensue, 359
4259 Ch08 4/12/02 10:24 am Page 360
360 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
SIMONE A. WEGGE
connecting emigration rates possibly to other issues such as urbanization and economic development. Knowledge of the home community can also facilitate a comparison of emigrants with non-emigrants and thus promote a deeper understanding of the migration decision. Three national groups stand out in the study of European population movements during the nineteenth century. According to U.S. historical statistics, the principal national groups who left during this period were the British, Irish, and Germans, with the latter comprising the largest group arriving in the United States for four decades, from 1854 until 1895. After the early 1890s, other national groups, especially Italians, Russians, and other eastern and central Europeans besides Germans were the largest ethnic groups leaving Europe. The scarcity of rich emigration statistics is particularly acute for the case of Britain. Statistics were collected that detail the numbers of persons departing from British ports, broken down by age and destination. These lists, for the most part, do not exist anymore. Among other deficiencies, no extensive amounts of data were ever collected which list British emigrants by their home community.8 It is a real shame that British authorities did not exhibit more interest in collecting information about their emigrants because this nation has such a long and sustained emigration history, the largest in absolute numbers of any European nation between 1815 and 1930 (Baines, 1991, p. 3). To get around the lack of high-quality emigration data, studies on British emigration have mostly used sources in the destination countries, such as the American passenger ship records.9 For very practical reasons, research on Irish emigration has also relied heavily on these American lists.10 Other studies on British emigration have used creative methodologies to skirt this major gap in data sources, such as Baines (1985), which imputed emigration rates for English and Welsh counties by backing them out from annual numbers on population, births, deaths, and indirect knowledge of internal migrations. To this date, gaps in the knowledge of British emigration remain. Little is known, for instance, about the specific local origins of British emigrants. Similarly, emigration documentation of many other European countries tends to be modest at best. In contrast to Britain and Ireland, the Dutch and Scandinavian governments compiled more interesting data during the nineteenth-century waves of emigration. Scholars tackling Dutch and Scandinavian emigration have been better equipped to rise to the Thistlewaite challenge. By linking three different micro data sources, for example, Robert Swierenga has described individual Dutch emigrants as they lived just prior to leaving, as ship passengers en route to the U.S., and finally over time in their new lives in the U.S. Such a study is only possible with rich sources of micro data, including emigration, U.S. passenger and U.S. census data.
4259 Ch08 4/12/02 10:24 am Page 361
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
361
Existing micro data on Swedish and Norwegian emigrants and their origins are also especially valuable.11 As a result, over the last few decades Scandinavian scholars have added a new level of sophistication in terms of the issues they have tackled in the migration literature. Their accomplishments include some testing of the importance of local factors in the migration decision, providing evidence of the relevance of chain migration networks in generating further migration, and emphasizing the link between internal and external migration as related to the process of industrialization and urbanization (Ostergren, 1986; Nugent, 1995). Certain regions in Germany also have data of potential interest to migration scholars. Before unification in 1871, emigration and immigration were both handled by the various German duchies and principalities. Thus, today some of the regional German archives contain useful sources of emigration data for the mid-nineteenth century. Valuable studies on German emigration, based largely on primary sources culled from regional German archives, include Kamphoefner’s study of Westphalians, Von Hippel’s study of Württemberg emigrants, and Auerbach’s study of Hesse-Cassel emigrants. As more German archives are making their emigration records accessible, other studies using local emigration records are emerging.12 Auerbach’s work is based on the same emigration records studied in this paper.13 Most research on German emigration, especially Auerbach’s work, has been very qualitative in nature. In contrast, my own work based on the Hesse-Cassel emigration records has used econometric analysis to test hypotheses about migration behavior.14 For those countries without emigration data at the individual level, selective port data in other European countries can often fill in some gaps. In Germany, for instance, the two ports of Bremen and Hamburg, used by Germans and many other central and eastern Europeans as a point of departure, kept detailed statistics on the emigrants who left from these seaports. While Bremen was the chief port of departure for German emigrants, its records on passengers no longer exist. The Hamburg lists, however, still exist today and can be used to study a subset of many large emigrations from Europe. They begin in the year 1850 and provide various items of information on individual emigrants. An example of these data can be found in Assion (1991).15 Work by Markus Günther compares the Hamburg lists with those on the American side for the same ships and can be used to understand the biases or quirks of each data source. He found that for many ships, the two different lists of passengers registered the migrants in the same basic order and were close to equivalent; the lists of other ships, however, were nowhere near as comparable and looked as if they had been constructed in very different ways (Günther, 1992). 361
4259 Ch08 4/12/02 10:24 am Page 362
362 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
SIMONE A. WEGGE
As discussed previously, many studies on nineteenth-century European emigration have made use of data in the destination countries, in particular the American ship passenger records. The American immigration statistics offer a convenient source of information, given that about two-thirds of the European migrants who made continental moves settled in the United States. These data are a continuous source of information, starting in September of 1819, and comprise the lists ship captains compiled and presented to customs agents upon arrival. Each ship captain arriving at an American port was required to provide a list of migrants on board, along with information regarding the migrants’ age, occupation, accompanying family members, and his or her homeland.16 Summary statistics exist based on these lists, but scholars and others who use them should be careful about their limitations.17 The U.S. passenger lists are often the only source of information on migrants who settled in the U.S. and have something to say about immigrants from all sorts of European backgrounds. In spite of the advantages of providing valuable demographic variables and wide coverage, the American passenger statistics are not a completely representative sample of who arrived in the U.S. For instance, the statistics only cover those who disembarked at an American port and miss a large portion of those who arrived from Canada before 1894 (Davis, 1931, p. 652). Thousands of Canadian-born individuals crossed the landborder to settle in the U.S. and multitudes of European migrants disembarked at Canadian or Mexican ports to then cross by land to the U.S. (Davis, 1931, p. 652). This is not a matter of small concern, as recent work on Canadian migration emphasizes that Canada was largely a country of emigration in the latter part of the nineteenth century (McInnis, 1994, p. 140). Measurement error and unnecessarily vague recording of information are two problems that exist for the U.S. passenger records. Captains, for instance, were required to record the occupation of immigrants. Some lists indicate that captains often did not take much care in distinguishing the immigrants’ various occupations, and instead wrote “laborer” or “farmer” for vast numbers of them. Cohn (1992), Cohn (1995), Ferrie (1999) and Swierenga (1986) address this measurement error in detail. Very often an immigrant’s country of origin was transcribed when it would have been easily possible to record the city or village of origin through a lengthier consultation with the emigrant.18 It is unclear how seriously captains took the task of preparing the passenger records. Many lists look to have been filled out by professional clerks in the port of embarkation before the captains departed, since lists from specific time periods from a given port appear to be compiled in the same handwriting (Swierenga, 2000, p. 318). For U.S. passenger lists between 1820 and 1850, Olsson (1967) also emphasizes the quality of lists from Dutch and German ports.
4259 Ch08 4/12/02 10:24 am Page 363
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
363
Not all lists, however, were so carefully transcribed. An interesting excerpt from Love Across Color Lines describes German immigrant Ottilie Assing’s experience as a volunteer secretary to the ship captain, illustrating how the compilation of a ship list could proceed quite haphazardly.19 Some of the comments are in Assing’s words, others in Maria Diedrich’s, the author: When the voyage was coming to an end, new passenger lists had to be prepared, and as one of the few people aboard who spoke German, English, and French, Assing volunteered as the captain’s secretary and translator. The description she gave of her work attributes an almost symbolic quality to the metamorphosis of every passenger. “It was on the high seas that all these Saxons, Hessians, Hannoverians, and Mecklenburgians were given a common fatherland, and even the Prussians and Austrians, who usually look down haughtily on the rest of Germany, arrived as “Germans,” she wrote . . . “Social rank, too, was treated rather lightly, and all those farmhands, day laborers, and working people were raised to the level of farmers, for as the captain said, it looks better.” She could not have agreed more. The Hessian and Bavarian, the Saxonian Serb and Jew from Bremen – America reinvented them as Germans; soon they would be German-Americans, and their children would call themselves Americans. The farmhand and the factory worker reemerged as farmers, the common soldier as captain.
While this account describes the transcription of immigrant occupations on only one ship out of thousands that make up the U.S. passenger ship lists, it casts a suspicious light on the quality of the occupational information from these records. Clearly, other methods were used to record information on immigrants, as the U.S. passenger ship records show that multitudes of laborers arrived in U.S. ports. Major differences on occupational identity between emigration sources and U.S. passenger records may not exist. This is at least the case for nineteenthcentury Dutch emigrants to the U.S., as Swierenga (2000) explains, based on his work in linking Dutch emigration lists to the U.S. passenger ship manifests. Dutch emigration authorities were, however, less successful in recording all the men leaving by themselves. Only further research that seeks to connect emigration sources and immigration sources for other European emigrant populations will shed more light on this matter.
32 33 34
II. THE PRINCIPALITY OF HESSE-CASSEL: A EUROPEAN SOCIETY BEFORE INDUSTRIALIZATION
35 36 37 38 39 40
The principality of Hesse-Cassel occupied the northern and eastern regions of the present-day German state of Hesse, bordering Bavaria, Hanover, Prussia, Saxony, Thuringia, Waldeck, Westphalia, the city of Frankfurt, and the other Hessian principalities. It was the largest in area of the three Hessian principalities. With proximity to the Rhine, Main, and Weser rivers, emigrants had 363
4259 Ch08 4/12/02 10:24 am Page 364
364 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
SIMONE A. WEGGE
various possibilities for getting to the northern shipping ports. The entire population of Hesse-Cassel did not even reach one million in the first half of the nineteenth century. In 1832 the census of population registered 677,849 as the total population of Hesse-Cassel and 726,739 by the December 1858 census.20 Over these three and a half decades, the additions to the population size that high birth rates made were dampened by migration, with 55,000 individuals, at the very least, leaving the principality.21 This area of Germany was exceptionally rural in the nineteenth-century, with only a handful of towns with over 7,000 residents. In many respects the principality of Hesse-Cassel mirrored the social and economic challenges confronted across large swaths of Europe in the nineteenth-century. During the early decades of the century the principality experienced substantial population growth but was very slow to industrialize. The main economic sector, agriculture, was characterized by diminishing returns to labor, given its fixed resources of land and the slow adoption of new farming technology. Thus, with 75% of the population still engaged in some way in agricultural production in the middle of the nineteenth century, the outcome of such population growth was an increase in poverty over the 1840s (Kukowski, 1995). Within the context of all of Germany, historians have described the Hesse-Cassel principality as relatively backward and poor. More generally, German agriculture in the middle of the nineteenth century was not at all impressive, according to recent estimates of European agricultural productivity (Clark, 1993; Perkins, 1981).22 Half a century later, however, German agriculture was more productive than much of the rest of continental Europe and on a par with Great Britain (Perkins, 1981). For Hessians seeking economic opportunity elsewhere, artisan trades could rarely serve as their salvation because opportunities were limited in the crowded artisan job markets. The archaic guild laws of Hesse-Cassel permitted many specialized artisan trades to be practiced only in officially designated urban communities, where only a fraction of the population lived.23 And as was the case for craftspeople all over Europe, factories that could manufacture goods more productively were beginning to compete with artisans. Thus, guild ordinances doomed the rest of the towns and villages to a local economy based on agriculture, some mining, and petty artisan trades.24 Hesse-Cassel also possessed no recent history of an enlightened ruler who may have seen ways to promote economic growth or the interests of the Hessian people, not a very unusual state of affairs for a nineteenth-century European state. With no particular absolute economic advantage in the way of agriculture or industry to promote the political standing of Hesse-Cassel, its rulers typically maximized their power within the Holy Roman Empire by building
4259 Ch08 4/12/02 10:24 am Page 365
Transatlantic Emigrants 1 2 3 4 5 6 7 8
365
up large armies. The Hessian landgraves paid for such extravagances by leasing the services of their soldiers out to other political entities (Pedlow, 1988, pp. 11–12). In 1775, for instance, Frederick II was able to cushion his government’s coffers by renting out the services of thousands of his soldiers to the British for subsequent battles in the American colonies. A more encouraging event was the adoption of Hesse-Cassel’s first constitution in 1831, through which citizens gained the right to leave the principality for any foreign country (Auerbach, 1993, pp. 134–136).25
9 10 11 12
III. THE HESSE-CASSEL EMIGRANTS: AN ILLUSTRATION OF EUROPEAN MIGRANT BEHAVIOR
13 14 15 16 17 18 19 20 21 22 23 24 25 26
The first great influx of German migrants to the United States occurred during the years 1816–1817, years of great crop failures. Outgoing German population flows subsided until the 1830s, when, yet again, people began to seek a life elsewhere. The first Hessians left soon after the passage of the Hesse-Cassel constitution in 1831. Hessian archival records document these early departures from the principality for the year 1832. During the early to mid-nineteenth century, German migrants came primarily from Westphalia, Württemberg, Baden as well as from the Hessian principalities. Migration rates from HesseCassel were some of the highest in Germany and in Europe during the 1850s.26 The percent of migrants from the north of Germany increased steadily over the century.27 As previously discussed, from the mid-1850s and until the mid-1890s, Germans were the largest group arriving annually in the U.S., even larger than Irish inflows.
27 28
The Hesse-Cassel Emigrant Permit Lists
29 30 31 32 33 34 35 36 37 38 39 40
Successful emigration from Hesse-Cassel required official approval. The interior ministry of Hesse-Cassel and authorities at the Kreis (district) and provincial levels maintained records on the individuals and their families who secured approval to emigrate, hence the existence today of the emigrant permit lists at the Hessian State Archive in Marburg, Germany. Such permission was no small matter, as obtaining it required the emigrant to relinquish his rights and responsibilities as a subject of the principality.28 Potential emigrants applied through the town mayor, who communicated the information to the sub-Kreis or Kreis authorities. For the government, the emigration-permit system served a number of roles, chief of which was control over who could leave. Emigration permits helped 365
4259 Ch08 4/12/02 10:24 am Page 366
366 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
SIMONE A. WEGGE
the government to preserve the size of its military force, to make sure that people paid debts before leaving, and to ensure that those with community responsibilities like an ailing parent stayed home. Thus, young men who had not completed their military service and individuals with outstanding debts or family responsibilities in the home village found it very difficult to leave with permission. Many ended up leaving illegally.29 Besides control over departures, authorities also wanted to keep track of the cash assets emigrants were taking out of the country. These various priorities affected the kind of data that were collected on the emigrants. The Hessian emigration permit lists concern over 55,000 emigrants, and the project and index are formally referred to by the acronym HESAUS.30 This data set is the fruit of a more than decade-long effort by European historian Inge Auerbach and the Hessian State Archive to catalog and index the extensive records on the individuals who emigrated from the principality of Hesse-Cassel.31 This undertaking represents one of the largest archival efforts in Germany to systematically record existing regional emigrant records, and has been described by one migration historian as “the most recent and ambitious project.”32 Academic researchers have, of course, compiled smaller data sets with specific purposes in mind (Assion, 1991). The data I analyze in this paper cover emigrants who left in the years from 1832 to 1857. In 1852, the Hesse-Cassel authorities started keeping track of illegal emigrants as well, and my data contain information on both legal and illegal emigrants for the years 1852 to 1857. The following information is fairly complete (where it applies) for each emigrant in the data set: (1) Village and Kreis (district) of origin. (2) Family name. (3) First name, or family relation to the head of household leaving (e.g. “son of John”). (4) Country (and occasionally city) of destination. (5) Year of emigration (and sometimes month). (6) Archival document code. (7) Whether emigrant is traveling with family or by himself or herself. (8) Gender. (9) Legal or illegal classification in the years 1852–1857 (with or without permission). While the first seven variables come directly from the HESAUS data set, I have incorporated the last two characteristics on gender and legality of the emigration through extensive programming and other archival documents respectively.33 The data set contains the most interesting information on the
4259 Ch08 4/12/02 10:24 am Page 367
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
367
heads of households, mostly men, as only men served in the military and most of the taxpayers were men. These variables are as follows: (10) (11) (12) (13)
Age. Money being exported. Occupation. Number of family members accompanying the head of household.
Except for the data on ages, information on exported cash, occupation and the number of traveling companions were generally recorded only for heads of households or single persons traveling alone. Table 1 shows examples of the emigrant records. The first group is the Becker family, who left the village of Wehrda in the Kreis of Marburg (MR) in the year 1852. Johann Peter is the male head of the household and is listed twice because he shows up in two different government documents. He worked as a carpenter (TI), was 44 years old, and left with 40 Thaler of currency, worth about $28 U.S. Johann Peter was also married with two children, one daughter and one son. Both of these children are listed twice. I inferred that there are two, not four children, because each separate set of archival records, coded 4/593,197 and 1A/23,213 respectively, lists two children. “99” refers to the fact that an individual is traveling within a group or family, and “AN” denotes that he or she is traveling to North America. Records on an individual who traveled alone are displayed for Mr. Oskar Maus in Table 1. Mr. Maus is listed twice in the records.34 As previously discussed, in general the first nine variables are very complete and the remaining four are not, especially for the last five years of the data set.35 The data in its original form contain some discrepancies, mostly because several levels of government collected or stored them, and the HESAUS index contains all existing governmental records on any emigrant. Many emigrants are thus listed several times in the HESAUS index. An important task of preparing these data for analysis involved combining the information from all governmental records. In addition, I reformatted many of the variables. These changes are described in the appendix.
33 34
The Advantages of the Hesse-Cassel Data
35 36 37 38 39 40
The Hesse-Cassel records have various advantages over other sources of migration behavior. As a large and rich data set, concerning over 55,000 emigrants, it can more easily yield testable hypotheses. Of significant value are the details about individual emigrants that are not available in other sources. The Hesse-Cassel data, for instance, list the cash 367
16
17
18
19
20
21
22
23
24
25
26
27
28
31
32
33
35
36
37
38
39
40 Christine Frau, Joh. Pet. Johann Peter Johann Peter Sohn, Joh. Pet. Sohn, Joh. Pet. Toch., Joh. Pet. Toch., Joh. Pet. Oskar Oskar
1852 1852 1852 1852 1852 1852 1852 1852 1851DE 1851
AN AN AN AN AN AN AN AN AN AN 2 20 20
1
44
44
KF KF
TI TI
T 60 T 55
T 40 T 40
99 99 99 99 99 99 99 99
Family
4/593,197 1A/23,213 1A/23,213 4/593,197 1A/23,213 4/593,197 1A/23,213 4/593,197 4/593,120 1A/23,108
Archival Code
368
11
29
30
34
Source: HESAUS data, see Auerbach (1987–1988). Note: For this table the names and characteristics have been changed to protect the identities of actual emigrants.
Becker Becker Becker Becker Becker Becker Becker Becker Maus Maus
13
Wehrda/MR Wehrda/MR Wehrda/MR Wehrda/MR Wehrda/MR Wehrda/MR Wehrda/MR Wehrda/MR Cappel/MR Cappel/MR
9
Cash
10
1 1 1 1 1 1 1 1 2 2
12
Job
7
Age
6
Goal
5
First Name
4
Surname
3
Village
2
Date Year, Month
1
Group
15
Examples of Emigrant Records.
14
Table 1.
4259 Ch08 4/12/02 10:24 am Page 368
SIMONE A. WEGGE
8
4259 Ch08 4/12/02 10:24 am Page 369
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
369
assets of emigrants. Although information on exported cash is incomplete in this data set, such data cannot be found at all in the oft-used American passenger ship lists or in the Hamburg port data. Such figures help to determine the financial and social background of migrants. Exported cash is not always available in the existing sources of other European emigration lists either, but some sources offer substitutes for such information. The Dutch emigration lists provide particulars on the social class of emigrants, and the Westphalian emigration lists provide occupational information for some emigrants couched in a class interpretation (Swierenga, 2000; Kamphoefner, 1987).36 In general, emigration data provide better information about occupations than immigration or port data do. High-quality emigration data gathered from the localities of emigrants, like the Hesse-Cassel lists, can make a real difference in measuring the skill composition of the emigrant population. Emigration data are recorded in the language of the migrant’s native tongue. Local officials have better information than ship captains or their designated transcribers on the lives of departing migrants, which can affect the manner in which migrant characteristics, especially occupation are described. The Hesse-Cassel data, for instance, contain more accurate information on migrants because the local authorities recording the data were familiar with their local economy and typically with the actual persons leaving. This is most significant in terms of comparing the Hesse-Cassel occupational data with that of the U.S. passenger lists. Familiarity with the local dialect and rural economy helped local officials to leave behind richer descriptions of emigrant occupational information than the Hamburg authorities. The occupational descriptions in the Hamburg data reflect a “Hanseatic city-perspective” and are too vague in regards to rural emigrants, especially in terms of the types of farmers and laborers (Kamphoefner, 1995, p. 26). Richter (1980, p. 146) has also criticized the Hamburg passenger records for the amount of data collected “by ear.” These variations are reflected in the number of different occupations of various migration sources. Ferrie’s U.S. passenger sample yielded 135 different occupations, Assion’s project generated 161 for the Hamburg Hessian sample, and the Hesse-Cassel data produced 332 different occupational identifications (Ferrie, 1999, p. 209; Assion, 1991, p. 173; Auerbach, 1987–1988). Another great advantage of the Hesse-Cassel data is the information provided on the emigrant’s home village. Every emigrant’s home village and Kreis are known. Such information cannot be found to a similar degree of precision in the ship lists or the census records of the United States. The U.S. passenger lists do provide the name of the hometown for some individuals, but only for a minority of the immigrants recorded. Such information is crucial for studying the importance of local factors in their migration decisions or the effect of their 369
4259 Ch08 4/12/02 10:24 am Page 370
370 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
SIMONE A. WEGGE
departure on the local home economy. It is also relevant for understanding the variation in local emigration rates and for accurate testing of chain migration effects (Wegge, 1998). While the Hamburg data also list village and Kreis, these data commence only in 1855 and concern only a fraction of the Germans leaving Europe during this period. Port authorities (as well as ship captains) also relied more on recording emigrants’ names and places of residence by ear, which created more chances for measurement error. Scholars suspect that a lot of Europeans left under suspicious circumstances. Emigration data sources are less helpful in describing such individuals, since most lists do not contain any information on surreptitious movers (Kamphoefner, 1995, p. 27). The Hesse-Cassel data are exceptional in this regard, as they provide a lot of information on individuals who left illegally between 1852 and 1857. Beginning in 1852 Hessian authorities began to keep track of those leaving without formal permission. The Hesse-Cassel authorities, however, managed to count almost as many illegal emigrants as legal ones, with 15,126 individuals emigrating legally and 13,584 emigrating illegally (as seen later in Table 3B). As a consequence of unknown clandestine emigration flows, European nations’ emigration statistics may grossly understate aggregate flows (Swierenga, 2000, p. 305). In linking Dutch emigrants from the Dutch emigration lists to the U.S. passenger lists, Swierenga (2000) provides important lessons. In his study both sources, but especially the American lists, understated the true number of Dutch people who permanently left the Netherlands. The U.S. passenger lists and the Dutch emigration lists both undercounted, but in different ways. While the U.S. passenger lists tended to undercount all types of individuals, the emigration lists had a much higher probability of missing young men, who are notorious in the migration literature for clandestine departures. These differences must be kept in mind when using either type of list for characterizing the background of a migrant population.
29 30
The Hesse-Cassel Census Data: Sources for Studying the Homeland Economy
31 32 33 34 35 36 37 38 39 40
The Hessian State Archive in Marburg contains various documents on the economic and social characteristics of Hesse-Cassel. Many of these sources are at the village level and provide insights into the agricultural and manufacturing sectors of the Hessian economy. These data have been linked to information on the emigrants through the village variable, so that one knows both details about individual emigrants and the socioeconomic background of the emigrant’s village. By linking such data to the emigrant lists one can test the importance of village characteristics in the decision to migrate. Aside from their use for studying migration decisions, such data can also be used to explore various
4259 Ch08 4/12/02 10:24 am Page 371
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
371
debates in European economic history with regards to property institutions and to processes of industrialization and urbanization. These debates and the use of these data for studying them are beyond the scope of this paper. Some of the most important information on the homeland comes from a survey completed of Hesse-Cassel villages in the mid-1850s, authorized by the Hesssiche Historische Kommission.37 Historians have previously used these data, but no study I am aware of has exploited the records for more than a circumscribed area of the principality. Although all 1376 (or so) Hesse-Cassel villages were surveyed in the 1850s, my project covers those villages in the contiguous area of the principality, comprising 19 different Kreise and encompassing slightly more than one thousand villages.38 The questionnaire is composed of 186 individual questions divided among seventeen different topics.39 Foremost are the information on land prices and on the land distribution within a village, showing specifically how many large, medium and small farm families as well as landless families lived in the village. Information was also provided on many other socioeconomic factors, such as the occupational distribution, the religious background, the number of schools, population growth, the number of houses, wages of laborers, and inheritance customs of the village. The occupational distribution of Hessians is well described in two other sources, namely an artisan census and a factory census, available for 1847, 1852 and 1862.40 The artisan data provide details on approximately one hundred different artisan occupations, broken down between master and apprentice. The factory census concerns more than eighty different types of factories or large commercial establishments and provides counts for workers in these industries.
25 26 27
IV. THE HESSE-CASSEL EMIGRANTS: VARIATION IN MIGRATION EXPERIENCE ACROSS TIME AND SPACE
28 29 30 31 32 33 34 35
The heyday of Hessian emigration took place during the 1850s, with 1852, 1854, and 1857 the years of greatest emigration activity. As one can see from Table 2A, there were fluctuations in the total number of emigrants leaving Hesse-Cassel. The first peak occurred in the year 1846, when over 2,000 people emigrated. The second peak was much larger, and occurred in 1852. Numbers for 1848 through 1850 must be considered tentative, as archivists suspect that these are understated due to missing information on several Kreise.41
36 37
Seasonal Strategies
38 39 40
Springtime was the typical season for travelling, with 63% leaving the principality between the beginning of February and the end of May (Table 2B).42 371
4259 Ch08 4/12/02 10:24 am Page 372
372
SIMONE A. WEGGE
Table 2A.
1
Emigrant Numbers.
2 3
Total Emigrants
Emigrants per 1,000 Population, by selected Kreis
Year
Number
Melsungen
Rotenburg
Frankenbergg
Hanau
22
1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857
1900 767 793 703 1019 686 498 1088 1544 2497 3323 1544 414 346 2067 7328 6272 8847 3570 4211 6099
1.01 0.03 0.46 0.66 5.37 1.49 1.43 5.32 3.16 2.6 3.3 1.9 3.1 2.3 8.2 13.8 12.1 14.4 3.6 7.0 11.9
2.51 1.23 7.55 0.83 3.56 2.38 0.90 1.02 2.14 9.9 5.6 4.6 1.2 0.5 1.0 14.1 16.1 22.0 10.0 14.9 11.8
0.47 0.0 2.27 0.62 1.04 1.05 0.62 1.21 1.97 1.4 3.2 1.2 0.0 0.1 1.8 4.9 13.2 21.0 9.6 11.8 12.8
3.14 3.14 0.96 1.48 0.54 0.39 0.59 0.47 0.84 2.4 6.5 2.5 0.0 0.0 2.3 6.6 8.3 12.8 2.7 3.1 3.1
23
Total
55516
N/A
N/A
N/A
N/A
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
24 25
Source: HESAUS data, see Auerbach (1987–1988).
26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
This pattern did not change much over time either, as the patterns for the 1840s are roughly the same as the ones for the 1850s. Travel through the interior of the United States was not possible during the winter months, when many of the waterways became impassable. In addition, migrants may have also avoided traveling during the winter months when mortality rates were highest (Cohn, 1984). A person leaving in mid-February could not expect to get to America before mid-April, given the time necessary to get to the harbor, acquire ship passage, and finally sail to North America. This suggests that the typical emigrant could have planned his or her trip already months prior. Accordingly, the large numbers for February through May represent more than pent-up demand from the winter months for migration opportunities.43 Work by Hubrich et al. (1991, p. 77) shows similar heavy traffic for the spring months, but they also show that many emigrants left in September and October, which may indicate the increasingly widening set of opportunities to travel by rail (Hubrich,
19
20
21
22
23
24
25
26
28
29
30
31
34
35
38
39
40
373 100% 613
27
32
33
36
37
Source: HESAUS data; see Auerbach (1987–1988).
18
0.8 15.8 28.6 19.6 13.5 2.3 4.2 4.4 7.2 0.7 2.1 0.8 100% 493
2.2 5.3 11.4 14.2 20.7 11.6 6.4 8.5 7.1 7.1 3.7 1.8 100% 1152
0.9 13.4 22.4 15.9 12.4 7.4 4.3 6.1 6.7 6.2 3.1 1.2 100% 461
1.7 6.9 13.7 19.3 13.4 11.7 5.9 7.6 6.3 10.2 0.9 2.4 100% 652
0.6 5.5 11.8 18.1 17.4 4.4 5.1 14.1 7.7 8.6 5.2 1.4
2.8 13.0 25.2 18.2 9.1 6.3 7.4 5.9 5.2 3.6 1.9 1.4
1857
100% 638
7
100% 1421
17
100% 278
16
Total No. Obs.
15
0.9 6.9 27.2 28.1 3.3 2.9 4.3 1.8 7.3 0 0.4 16.8
14
5.8 16.2 8.6 12.9 12.2 10.1 12.9 8.6 3.2 3.9 3.9 1.4
13
January February March April May June July August September October November December
9
1856
10
1855
6
1854
5
1853
100% 14,771
3.2 10.9 21.6 19.6 10.8 6.3 5.3 5.4 5.5 3.6 3.2 4.6
1832–1857
4
1852
3
1851
2
1850
1
Month
12
Month of Departure, 1850–1857 (in percent).
11
Table 2B.
4259 Ch08 4/12/02 10:24 am Page 373
Transatlantic Emigrants 373
8
4259 Ch08 4/12/02 10:24 am Page 374
374 1 2
SIMONE A. WEGGE
Meesenburg & Schriewer, 1991). In contrast, the Hessian data do not show large outflows for the autumn season, except for a few aberrations.44
3 4
Variation Across Kreise
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
Substantial variation in the rate of emigration existed between the different Kreise, suggesting that local factors may have been important in convincing people to move or to stay. Emigration rates also varied across time. No single Kreis exhibited a steady rate of migration. Instead, most Kreise show increasing emigration rates over time. Several Kreise stand out for their remarkable patterns of migration rates, some of which are displayed in Table 2A. Poverty seems to be a relevant characteristic of the Kreise with the highest emigration rates. The two Kreise that experienced the highest migration rates on a consistent basis were Melsungen and Rotenburg. These Kreise were centers of linen weaving, mostly home production, with steady business prior to the fall of Napoleon, but in ruin by the 1840s (Möker, 1977, pp. 86, 172). Persistently high emigration rates may have been caused by the demise of this industry. The emigration rates of the Kreis of Frankenberg, on the other hand, remained at a moderate level compared to other Kreis rates until 1852, when they reached some of the highest levels for the whole principality. The villages of Frankenberg are located in the hinterlands of HesseCassel and at very high elevations, relative to other Kreise. Here, the land was not especially productive for agriculture, nor did the area possess any commercial or industrial center. Several authors have commented on the economic crises that persisted throughout the 1840s and 1850s in this particular Kreis, even when other areas of the principality began to industrialize.45 Cities or regions with more industrial activity, especially a variety thereof, had consistently lower emigration rates. This is particularly evident for the Kreise of Cassel, Hanau, and Schmalkalden; emigration rates for Hanau are displayed in Table 2A. The capital city of Cassel and the city of Hanau had the most specialized set of artisans in the entire principality. Hanau benefited from close proximity to the city of Frankfurt and the Main River. Diversity of their local economies may have been a factor in the relatively low emigration rates of these Kreise.46
34 35
Destinations of the Emigrants
36 37 38 39 40
Overwhelmingly Hesse-Cassel emigrants left for the United States, over 80% before the 1850s and 90% thereafter. Table 3A describes the countries or continents migrants moved to. In addition, very few came back to resettle in Hesse-Cassel.47
18
19
20
21
22
23
24
25
26
27
28
29
30
32
33
34
35
36
37
38
39
40
375
31 14,454
11,964 0 120 16 1 5 3 172 1,933 240
Source: HESAUS data, see Auerbach (1987–1988). * What is primarily meant here by ‘America’ is the United States.
100
17
4,217
16
Total
15
84.7 – – – – – – 8.1 6.8 0.3
14
3,572 0 0 0 1 4 0 342 287 11
13
America* Canada Central America South America Africa Asia Australia Europe, non-German German states Not stated 100
82.8 – 0.8 0.1 – – – 1.2 13.4 1.7 36,332
33,193 0 2 434 16 6 428 342 1,603 252 100
91.4 – – 1.2 – – 1.2 0.9 4.4 0.7
9 55,003
48,729 0 122 450 18 15 431 912 3,823 503
Total Number
8
1852–1857 Percent
7
1852–1857 Number
6
1841–1851 Percent
100
88.6 – 0.2 0.8 – – 0.8 1.7 7.0 0.9
Total %
4
1832–1840 1841–1851 Percent Number
3
1832–1840 Number
2
Destination
12
Destinations of Emigrants, by Time Period.
11
Table 3A.
4259 Ch08 4/12/02 10:24 am Page 375
Transatlantic Emigrants 375
1
5
10
4259 Ch08 4/12/02 10:24 am Page 376
376
SIMONE A. WEGGE
Table 3B.
1
Destinations of Legal/Illegal & 1848 Emigrants.
2 3 4
Period
Number of Illegal 1852–1857
Percent of Illegal 1852–1857
Number Percent of Legal of Legal 1852–1857 1852–1857
Number of Percent of 1848ers 1848ers
5
13
America* Canada Central America South America Africa Asia Australia Europe, non-German German states Not stated
13049 0 0 90 5 3 184 118 67 68
14
Total
13,584
6 7 8 9 10 11 12
96.1 – – 0.7 – – 1.4 0.9 0.5 0.5 100
13399 0 1 214 4 2 128 187 1128 63 15,126
88.6 – – 1.4 – – 0.9 1.2 7.4 0.4 100
1274 0 0 1 1 0 2 8 168 5 1,459
87.3 – – – – – 0.1 0.6 11.5 0.3 100
15 16 17
Source: HESAUS data; see Auerbach (1987–1988). * ‘America’ generally means the United States.
18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
While the numbers of people who left for non-U.S. destinations were small, certain areas were preferred over others. The second favorite set of destinations, for instance, consisted of other nearby German states, with Hessian emigrants moving in large numbers to nearby areas, including Bavaria, the city of Frankfurt, Hanover, other Hessian principalities, the Rhine provinces of Prussia, and what is currently Thuringia (Saxe-Weimar-Eisenach). Still, the western part of Prussia (Rhine provinces) attracted well over 1,200 Hessians compared to the 233 who moved to the bordering country of Bavaria. This part of Prussia offered many job opportunities in the Ruhr area in places like Cologne and Düsseldorf and even experienced a net immigration in the 1840s, as Walker (1964, pp. 55–56) observes. Other German states had strict admission criteria for the most part, but they also could not always provide migrants all that they were looking for, which helps explain why movements to other German regions were not greater.48 Further afield but within Europe, a number of Hessians also settled in the Austrian-Hungarian Empire, East Prussia, Great Britain, the Kingdom of Poland, and Russia. One can understand some of these movements for language reasons, since German was spoken in many parts of Central Europe. Only a few Hessians in the nineteenth century found in Russia an attractive possibility, even though quite a few Hessians had moved there in the 18th century, posing a counterexample to a simple version of the chain migration model.49 Outside of Europe and the United States, Hessians moved to all sorts of places in the world, but in very small numbers. Interestingly, of this small group, an equal percentage moved to Australia and to South America.
4259 Ch08 4/12/02 10:24 am Page 377
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14
377
For the vast majority of emigrants, or over 47,000 persons, we know only that they left for the United States. A number of the “America-bound” may have ended up in Canada, but this is not quantifiable from the emigration lists. A few of these U.S.-bound emigrants, approximately 5% of them, expressed to authorities the exact city they planned to move to. Surely there were many more who had specific plans. Most likely, this figure is grossly understated and may reflect unfamiliarity with American place names on the part of local authorities, who found it easier simply to record North America. For those moving to other European but non-German nations, the probability that they knew what their plans were (and told authorities) was much higher at 33%. For those leaving for other German places, particularly in the East, it was even higher, at 60%. Many German states restricted immigration and required formal permission in order to move there, which would have forced emigrants to know ahead of time where they were going.
15 16 17
V. THE HESSE-CASSEL EMIGRANTS: WHO WERE THESE EMIGRANTS AND WHY DID THEY LEAVE?
18 19 20 21 22
The Hesse-Cassel data are rich in detail and illustrate the economic and demographic background of the emigrants and how they differed from those who stayed home. Linked with census data in the homeland, they suggest the complex socioeconomic motives of the emigrants.
23 24
Occupational and Wealth Characteristics
25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
One of the more interesting clues that emigrants left behind about their lives is their occupational classification. Occupation as well as lack thereof is interesting for what it may suggest about the emigrants’ motives for leaving. More individuals left in the 1850s than in the 1840s, so the bulk of the occupational information on Hesse-Cassel emigrants is for the latter period. A quite remarkable feature about the data is the incredible variety of occupations listed, with over three hundred separate ones identified. A gross summary is provided in Table 4A. For the entire period, 47.3% of the emigrants classified themselves as artisans, while 29.5% were manual laborers. The overwhelming difference between the 1840s and the 1850s is among the laborer group. Laborers migrated in much higher percentages in the later years, which in absolute numbers is a difference between roughly 700 laborers in the 1840s and 4,500 laborers leaving in the 1850s. The increase in the number of laborers emigrating overwhelmed these summary occupational classifications. So, over time, as a percentage of all 377
4259 Ch08 4/12/02 10:24 am Page 378
378
SIMONE A. WEGGE
Table 4A.
1
Occupational Distribution.
2 3 4 5 6 7 8
Period
Number
1841–1851
Number
1852–1857
Number
1832–1857
Professionals Farmers Artisans Laborers Linen Weavers Shoemakers
122 632 2375 669 161 329
2.8% 14.7% 55.4% 15.6% 3.8% 7.7%
250 1542 5888 4551 272 736
1.9% 11.6% 44.5% 34.4% 2.1% 5.6%
390 2189 8402 5236 438 1092
2.2% 12.3% 47.3% 29.5% 2.5% 6.2%
No. Observations
4,288
100%
13,239
100%
17,747
100%
9 10 11
Source: HESAUS data, see Auerbach (1987–1988).
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
emigrants with occupational information, most artisan groups decline. A few groups, however, increase over time as a percentage of the total: bakers and butchers for one, and merchants for another.50 Some information exists for women, mostly for the 1850s. These records primarily cover women who left on their own, including those who had never married, were widowed, or were just leaving at a separate time from their husbands.51 The vast majority of women with occupational identification, over 88%, was classified as laborers, and a smattering worked in the textile sector, mostly as weavers (Wegge, 2001a). With limited information on female job classifications before 1852, little can be said about changes over time. Married women who worked alongside their spouses on farms or in artisan workshops, or in other capacities, received no occupational classification. In the case that adult women traveled with male family members, what was important to record keepers, when they indeed made a note of it, was a proper clarification of the relationship of women to the head of household. Thus, most women were classified as wives, daughters, mothers-in-law, or sisters. Very distinct differences exist in the occupational distributions displayed in Table 4B. More than half of the legal emigrants with stated occupations were artisans, while almost half of the illegal emigrants classified themselves as laborers. Of the legal emigrants, there were twice as many farmers as among the illegal population. Illegal emigrants almost all went to America. The amount of cash emigrants took with them provides a proxy for wealth. One must be careful, however, in using such information.52 I computed the cash statistics in several ways. A few, very wealthy emigrants took very large amounts of cash, and such figures obscure the picture for the majority. So, the most important numbers to look at are those that exclude cash figures over 1,000 Thaler (equivalent to $0.71 U.S.). Looking at such figures in Table 5A,
4259 Ch08 4/12/02 10:24 am Page 379
Transatlantic Emigrants
Table 4B.
1
379
Occupations of Legal/Illegal & 1848 Emigrants.
2 3 4 5 6 7 8 9 10 11 12
Period Professionals Farmers Artisans Laborers Linen Weavers Shoemakers
Illegal 1852–1857
Legal 1852–1857
1848ers
1848–1851
1841–1847
1.3% 8.3% 36.9% 47.6% 1.6% 4.2%
1.9% 16.1% 51.0% 22.2% 2.1% 6.8%
2.2% 12.9% 65.3% 12.9% 2.4% 4.3%
3.0% 11.8% 61.0% 13.9% 3.2% 7.0%
2.8% 15.9% 53.0% 16.4% 4.0% 7.9%
4901
5519
582
582
4288
No. Observations
Source: HESAUS data; see Auerbach (1987-1988).
13 14 15 16 17 18 19 20 21 22 23 24
emigrants took on average 153 Thaler with them, approximately $108 U.S. This is less than the average real wealth of native-born American men above the age of 15 in the year 1850, approximately $1200 U.S., but more than their median real wealth of zero (Ferrie, 1999, p. 105). Not surprisingly, medians for all of these same samples are considerably lower. Corresponding to the average of 153, the median was almost 50% less, at 80 Thaler. Half of the emigrants were taking less than $60 U.S.! Breaking this up over time, emigrants in the 1830s and the 1840s took sufficiently more money with them than the emigrants of the 1850s did. Emigrants leaving in the 1840s were more likely to leave as families and took more in assets, while emigrants in the 1850s were younger.53
25 26
Table 5A.
Exported Cash.
27 28 Period
29 30 31 32
Cash Figures, with no zeros
33 34 35 36 37 38 39 40
Cash Figures, with no zeros, with no values above 1000
No. Obs.
Average Thaler
Median Standard Thaler Deviation
Min.
Max.
1832–1840 1841–1851 1852–1857
1,298 4,352 12,902
339 344 230
171 142 63
638 772 1,224
2 1 2
10,285 20,000 99,999
1832–1857
18,552
264
86
1,101
1
99,999
1832–1840 1841–1851 1852–1857
1,213 4,091 12,546
226 207 129
143 114 60
219 213 160
2 1 2
1,000 1,000 1,000
1832–1857
17,850
153
80
182
1
1,00
Source: HESAUS data; see Auerbach (1987–1988). Notes: One Thaler was worth $0.71 U.S.
379
19
20
21
22
23
24
25
26
27
28
29
30
34
35
36
37
38
39
40
33 1,229 3,967 12,323 17,519
1832–1857
17,893
1832–1857 1832–1840 1841–1851 1852–1857
1,247 4,107 12,539
1832–1840 1841–1851 1852–1857
31,677
31
32 107
128 136 96
176
163 229 159
99
133
140 161 120
697
431 685 721
531
365 526 543
1
2 1 1
1
2 1 1
0
0 0 0
1,000
1,000 1,000 1,000
30,000
10,285 20,000 30,000
30,000
10,285 20,000 30,000
Maximum
380
Source: HESAUS data; see Auerbach (1987–1988). One Thaler was worth $0.71 U.S.
18
Cash Figures, with no zeros and no values above 1000
17
Cash Figures, with no zeros
16
1832–1857
13
112 129 89
12
1812 7306 22,559
9
1832–1840 1841–1851 1852–1857
10
Cash Figures, with zeros
8
Minimum
7
Standard Deviation
5
Average Thaler
4
No. Groups
3
Period
15
Exported Cash Per Capita.
14
Table 5B.
4259 Ch08 4/12/02 10:24 am Page 380
SIMONE A. WEGGE
1
2
6
11
4259 Ch08 4/12/02 10:24 am Page 381
Transatlantic Emigrants 1
Table 5C.
381
Exported Cash In Thaler ($0.71 Equiv.) Legal/Illegal & 1848ers.
2 3
Period
4 5 6
9 10 11 12
Average in Thaler
S. D.
Min.
Max.
Cash Figures With no zeros
1852–1857, Illegal 1852–1857, Legal 1848
5,127 5,017 469
132 310 299
433 1,760 574
2 4 2
16,000 99,999 5,714
Cash Figures With no zeros, No cash > 1000
1852–1857, Illegal 1852–1857, Legal 1848
5,062 4,812 448
99 152 201
124 181 211
2 4 2
1000 1000 1000
7 8
No. Obs.
Source: HESAUS data; see Auerbach (1987–1988).
13 14 15 16 17 18 19 20 21 22 23 24 25 26
Legal emigrants had a lot more cash on hand. For the cash figures in Table 5C with no zeros and no values above 1000 Thaler, the legal average is over 50% higher than the illegal average. For cash per capita figures, the same method garners only a 20% difference. So, in spite of the fact that legal emigrants more likely settled debts prior to departure, they were much richer than the illegal emigrants were. Still, the legal emigrants of the 1850s were not as wealthy as the emigrants of the 1840s were.54 Wegge (2002) shows that wealth differed significantly by occupation, with laborers taking the least and farmers and merchants the most. Among almost all occupational groups, legal emigrants took more cash with them than illegal emigrants did. Compared to those who stayed home, emigrants were not the poorest and definitely not the richest of their villages, rather among the lower middle class (Wegge, 2001b; 2002).55
27 28
Lifecycle Aspects of Emigration: Families, Solo Strategies, Networks
29 30 31 32 33 34 35 36 37 38 39 40
The ages of emigrants can allow us to infer in what ways emigration represents a life-cycle choice, one made at specific ages. This is only possible with detailed micro data. Age distributions are shown in Tables 6A-6D. Emigrants are separated between persons who traveled alone and those who traveled in groups. Those not traveling with family are more likely than not to be unmarried. For the data on families, I included only those with complete age information. If even one person listed in a family did not have an age, I excluded the entire group. There are many families with partial age information. The age of the wife or a child is often not known, for instance, and this biases the age distributions. In many cases wives were not listed at all, which means that under-reporting of wives understates the age distributions in Table 6B of people over 20 a bit.56 381
4259 Ch08 4/12/02 10:24 am Page 382
382
SIMONE A. WEGGE
Table 6A.
1
Age Distributions, Persons Traveling Alone.
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Age Bracket 0–1 2–3 4–5 6–7 8–9 10–11 12–13 14–15 16 17 18 19 20 21–23 24–27 28–30 31–35 36–40 41–45 46–50 51–55 56–60 61 + Total No. Obs.
1841–1851
1852–1857
1832–1857
0.1% 0.1 – 0.1 0.1 0.2 0.2 2.2 2.4 4.6 9.7 14.4 11.8 6.3 15.9 12.2 11.8 3.9 2.4 1.2 0.2 0.1 0.1
0.5% 0.5 0.4 0.4 0.3 0.6 1.0 4.6 5.4 9.4 11.0 14.7 17.5 9.0 9.7 5.7 5.8 1.4 0.5 0.3 0.3 0.5 0.5
0.2% 0.2 0.1 0.2 0.2 0.2 0.5 2.8 3.0 5.7 9.8 14.2 13.1 6.8 14.4 10.8 10.6 3.5 1.9 1.1 0.2 0.3 0.2
100.0% 3,474
100.0% 1354
100.0% 5138
Source: HESAUS data; see Auerbach (1987–1988).
27 28 29 30 31 32 33 34 35 36 37 38 39 40
Not surprisingly, the vast majority of persons traveling alone are between the ages of seventeen and thirty-five. In the 1850s a real change occurred so that the bulk of individuals traveling alone were individuals in their mid- to late-teens. Unfortunately, the records are less complete in the 1850s. One cannot say whether there may have been as great an outflow of people in their twenties and thirties in the 1850s, but they made up a much smaller percentage of all of the emigrants leaving alone.57 The lack of individuals in their early twenties in Table 6A for the years 1841 to 1851 is suspicious. These data concern almost entirely men. At such an age, however, men were eligible for military service and were not allowed to emigrate under most circumstances. Men of this age may not have emigrated at all, but they also emigrated secretly without formal approval, which we know
4259 Ch08 4/12/02 10:24 am Page 383
Transatlantic Emigrants 1
Table 6B.
383
Age Distributions, Persons Traveling in Groups of two or more.
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Age Bracket 0–1 2–3 4–5 6–7 8–9 10–11 12–13 14–15 16 17 18 19 20 21–23 24–27 28–30 31–35 36–40 41–45 46–50 51–55 56–60 61 + Total No. Obs.
1841–1851
1852–1857
1832–1857
8.3% 8.0 6.3 6.0 5.9 5.9 5.1 4.8 1.9 1.8 2.1 1.7 1.8 3.0 3.8 3.3 5.8 7.4 6.3 5.3 3.2 1.3 1.0
3.6% 3.7 3.6 3.0 4.3 7.3 4.8 9.1 3.9 7.3 7.9 5.4 2.8 4.8 2.4 1.9 3.9 3.9 5.5 5.7 4.0 0.9 0.3
7.6% 7.6 6.0 5.6 5.5 5.7 5.0 4.9 2.1 2.1 2.4 2.0 1.8 3.0 3.9 3.5 6.1 7.8 6.3 5.3 3.4 1.4 1.0
100.0% 3378
100.0% 330
100.0% 3979
Source: HESAUS data; see Auerbach (1987–1988).
27 28 29 30 31 32 33 34 35 36 37 38 39 40
for 1852 and thereafter. Of the people traveling unaccompanied, the legal emigrants were on average older by a number of years, as reported in Table 6D. The high frequency years were twenty-one to thirty, while for illegal emigrants the high frequency years were seventeen to twenty three.58 Table 6C displays the age distribution for women. Most women leaving without family members are between the ages of nineteen and thirty. Interestingly, there are no holes in the female age distribution, as there are in Table 6A for the 21–23 age group and in Table 6D for the 20-year-old legal persons.59 Age of the migrant was also related to inheritance practice. In the first thirty years after Napoleon’s defeat, migration from the southwest of Germany was strong in the areas where partible inheritance was practiced. Social historians, 383
4259 Ch08 4/12/02 10:24 am Page 384
384
SIMONE A. WEGGE
Table 6C.
1
Age Distributions of Women.
2 3 4
Age Bracket
Traveling Alone, 1852–1857
Traveling Alone, 1832–1857
Groups of two or more, 1832–1857
Groups of two or more, 1852–1857
0.8% 2.6 0.8 0.9 0.8 0.9 0 2.5 6.0 3.3 6.0 9.3 7.6 23.8 16.9 7.6 5.1 1.7 0 0 0 0.9 2.5
1.5% 3.6 0.7 1.5 0.7 0.8 0.7 2.2 5.8 4.4 5.8 8.8 7.3 22.6 16.1 7.3 5.1 1.5 0.0 0.0 0.0 0.7 2.2
5.0% 5.0 6.3 1.2 5.5 5.0 6.2 2.5 3.8 6.2 2.5 2.5 1.8 6.2 7.5 3.8 5.0 5.0 6.2 2.5 7.5 1.3 1.2
9.9% 10.3 9.1 7.2 7.7 8.8 7.8 5.5 1.8 1.8 2.7 1.1 1.8 4.4 3.7 2.0 3.1 3.4 2.5 2.6 1.8 0.2 0.8
100.0% 118
100.0% 137
100.0% 80
100.0% 1196
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
0–1 2–3 4–5 6–7 8–9 10–11 12–13 14–15 16 17 18 19 20 21–23 24–27 28–30 31–35 36–40 41–45 46–50 51–55 56–60 61 + Total No. Obs.
27 28
Source: HESAUS data; see Auerbach (1987–1988).
29 30 31 32 33 34 35 36 37 38 39 40
thus, have naturally concentrated on this feature, observing that the constant division of the land created many small farms, a Zwergwirtschaft, on which families eked out an existence.60 Some areas of unigeniture like Westphalia, however, experienced large outpourings to America (Kamphoefner, 1987). Wegge (1999) also demonstrates that migration rates were higher for Hessian villages using unigeniture or non-partible inheritance traditions, higher than those villages using partible institutions to complete intergenerational property transfers. Emigrants leaving non-partible villages in Hesse-Cassel tended to be younger, partly because of the local economic differences among such villages and partly because many knew at an early age whether or not they would inherit the farm.
4259 Ch08 4/12/02 10:24 am Page 385
Transatlantic Emigrants 1
Table 6D.
385
Age Distributions, Legal/Illegal Emigrants, Traveling Alone.
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Age Bracket 0–1 2–3 4–5 6–7 8–9 10–11 12–13 14–15 16 17 18 19 20 21–23 24–27 28–30 31–35 36–40 41–45 46–50 51–55 56–60 61 + Total No. Obs.
Legal Persons, 1852–1857
Illegal Persons, 1852–1857
– – – – 0.7% 0.7 0.7 2.8 5.7 4.2 4.2 8.5 5.6 21.8 20.5 10.5 6.4 1.4 – 1.4 – 1.4 3.5
1.6% 1.0 1.0 1.3 0.7 0.7 0.3 3.6 4.6 9.2 10.9 14.8 13.1 9.5 7.6 3.3 3.3 0.9 – – – 0.4 0.3
100.0% 142
100.0% 304
25 26
Source: HESAUS data; see Auerbach (1987–1988).
27 28 29 30 31 32 33 34 35 36 37 38 39 40
Family composition of the groups that left also affects age distributions. Over a span of three decades emigration seems to have evolved from a family activity into a pursuit of the young and single. In the 1830s, 69% of emigrants left in a group of three or more. This figure drops in the 1840s to 58%, and tumbles to 43% in the 1850s. In absolute terms, more families left in the six short years from 1852 to the end of 1857 than in the eleven years from 1841 to 1851; these families just make up a smaller proportion of the total than before. Hence, an increase in the propensity of both groups and individuals traveling without family members exists, but the increase is much larger for the people traveling solo. An interesting pattern also exists among the legal and illegal emigrants: unlike the legal emigrants, the majority of illegal emigrants traveled alone, driving up the probability overall that an emigrant was unmarried, without children and less wealthy (Table 7). 385
4259 Ch08 4/12/02 10:24 am Page 386
386 1
Table 7.
2
SIMONE A. WEGGE
Size of Emigrant Groups, Percent of Individuals in Each Group, Legal/Illegal & 1848ers.
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Group Size, No. Persons 1 2 3 4 5 6 7 8 9 10 11 12+ Total No. Obs.
Illegal 1852–1857
Legal 1852–1857
1848ers
60.7% 9.3% 7.3% 6.8% 5.2% 4.3% 2.4% 2.0% 0.8% 0.8% 0.2% 0.3%
42.2% 4.0% 7.2% 10.2% 10.9% 8.9% 7.1% 4.4% 2.2% 1.3% 0.7% 1.0%
49.5% 7.0% 9.3% 11.8% 6.5% 5.8% 5.3% 1.6% 1.9% 0.7% 0.8% –
100.0% 13,563
100.0% 15,126
100.0% 913
18 19
Source: HESAUS data; see Auerbach (1987-1988).
20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
A very large proportion of the families emigrated with very small children, but the percentages of small children decreased in the 1850s, although this conclusion is based on a small sample. For the 1840s the number of accompanying teenagers is remarkably small compared to the percentage of infants, but this may indicate that young families emigrated at higher rates than families with older children. Parents could easily avoid emigration themselves and send off their older children. In the 1850s, families with older children and older parents constitute a much higher percentage than before. Why this should be, is not immediately apparent. Some families may have delayed emigration to find out about the experiences of others. More realistically, families may have sent off their older children in groups: a comparison of the 1840s and 1850s shows that the percentage of teens increases quite a bit, while the percentage of adults over the age of 35 decreases. In addition, many Europeans, including Hessians, had high expectations that the 1848 revolution would bring economic and social change. During this time, emigration outflows slowed for the lower and middle classes and did not rise until 1852. Some families may have delayed emigration plans by a few years and subsequently left at more advanced ages than was typical for the average emigrant family of the 1840s. With each passing year, it became easier to send a family member to join another family member or friends in the United States. Wegge (1998) shows
4259 Ch08 4/12/02 10:24 am Page 387
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
387
that over time more and more migrants could be traced to a migrant with the same last name and from the same village, suggesting that more migrants were using family networks as a way of accomplishing their migration goals. Those leaving later and within a network took less cash with them as well compared to those who did not have any family member to join up with in the states. Great differences exist between those who left in the 1850s and those who left earlier, based on several factors. Those leaving in the 1850s were less skilled, which can explain in great part their comparative poverty. Demographics may have been an additional underlying reason. We know from above that in the 1850s young people and people traveling without family members (presumably more likely to be single, unmarried, childless, and illegal) made up a more significant portion of the emigrant population. From the cash per capita information, Table 5B, one can see that cash per capita figures decreased significantly over time as well. In an analysis of cash data by group size, (not shown), it was significant that more individuals were traveling without family members, but that they were also poorer than the single travelers in the 1840s. So, average cash taken decreased with time, not only because more young (presumably unmarried people) were leaving, but also because they were poorer than the single travelers of the 1840s. Finally those leaving later were more likely to participate in a migrant network, which decreased the necessity of cash for those joining up with family members already at the destination.
22 23
The Forty-Eighters: Political Refugees?
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
In March of 1848, political demonstrations in Paris touched off a wave of revolutionary activity across continental Europe. In Frankfurt, which bordered Hesse-Cassel, revolutionaries met repeatedly in an attempt to pass a controversial German-wide constitution. Hesse-Cassel experienced its own turmoil, during which the many government departments ceased to perform properly many of their usual functions. Many emigrants probably left without being noted in the official records, as authorities may have had great difficulties in keeping track of all the migrants during the political unrest.61 Emigrant records are less complete for 1848 than for other years. They may also be less complete for 1850 and 1851. These were years during which overall German immigration to the U.S. well surpassed the recent previous high in 1847. The same is not true, however, for Hesse-Cassel: the emigrant totals for 1849 through 1851 are all well below that of 1847 (Table 2A). By 1852, Hessian authorities had figured out that counting illegal emigrants was important for the accuracy of emigration numbers. That Hessians were willing participants in the 1848 movement is not surprising, because 1847 agricultural prices for Hesse-Cassel were the highest 387
4259 Ch08 4/12/02 10:24 am Page 388
388 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
SIMONE A. WEGGE
they had been in many years. The year 1848 is associated in the popular literature with emigration, but it is not the case that more Germans actually left Germany in 1848. Rather it may be the case that more intellectuals and political activists left in 1848 and in subsequent years than normally observed earlier in the 1840s. Some of these people were well known and thus, emigrant destinations such as Milwaukee, Wisconsin, are tied with the ‘Forty Eighters.’ What are the characteristics of the Forty Eighters from Hesse-Cassel? Mostly, they are not that different from those who left in earlier years. Records for 1,459 individuals exist, specifically for the year 1848, which may be on the low side, but nothing stands out to indicate under-recording in a particular Kreis or month.62 One cannot assert that 1848ers were wealthier, as the differences in the cash data are negligible (Table 5C). The proportion of professionals is not higher for 1848 but it is ever so slightly higher for the period of 1848 to 1851; percentage-wise more farmers left in earlier years. Instead, there are many more artisans who left, as shown in Table 4B. Those leaving by themselves were also younger and more likely to be teenagers, with families making up a smaller proportion of the emigrants than for the rest of the 1840s. Was this a rebellious set? Although over 11% still left for other German states, a higher proportion than in previous years went to America. Nevertheless, the emigrant population changed very much in character after 1847 and especially in the year 1852 and thereafter. The emigrants’ expectations for the future had undoubtedly been affected by the failure of the 1848 revolution.
23 24 25
VI. CONCLUSION: EMIGRATION DATA AND THE MICROECONOMICS OF MIGRATION
26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
European emigration had a great impact on economic growth in the Americas. The millions who left the European continent, mostly for North or South America, participated in an era of New World economic development that proceeded at a pace much faster than it otherwise could have. Between 1815 and 1930, a massive transfer of physical and human capital to the New World occurred. To understand this immense population movement fully, historians of all stripes must continue to study historical migration from Europe. By relying exclusively on an analysis of the U.S. passenger records, they will miss important aspects of this grand epic. The nature of migrant self-selection is of particular prominence here because who came and who did not may have had an effect on the pace of economic development in both the destination countries as well as in the origin countries. The Hesse-Cassel emigrant records show a complicated profile of an emigrant population, in terms of its age, occupation, wealth, family, departure, and
4259 Ch08 4/12/02 10:24 am Page 389
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
389
geographical characteristics. The typical profile changed dramatically over time. Most striking was the rise in illegal emigration after 1851. In comparison to the legal emigrants, the illegal emigrants were very young, relatively strapped for cash, more prone to choose America as their destination, and they had little occupational training. 1850s emigrants in general, both legal and illegal, were relatively poorer than those who departed in the 1840s and increasingly more likely to settle in the U.S.. As the Hessian migration tradition entered its third decade, families were a smaller proportion of the total, while individuals migrating on their own became a much larger proportion. The Hesse-Cassel data have various advantages over other sources of data on overseas migration from nineteenth-century Europe. Unlike most sources, they were recorded in the native language of the emigrants and are more precise because they limit measurement errors that can arise through translation. More importantly, they provide the home village of origin, figures on exported cash, and distinguish between legal and illegal emigrants. This information allows one to address better the issue of who these emigrants were and where they came from. By linking the Hessian emigration lists to census data, the importance of certain local factors in the migration decision, like inheritance traditions or local-based migrant networks, has been explored. This work and continued study of what lies behind the large variation in village migration rates will lead to an improved understanding of the causes of migration. Outstanding issues remain to be addressed with these data. In the future, I will examine the effect of migration on the subsequent development of the HesseCassel economy as well as investigate further the sustaining power of migration networks. Continued exploration of such issues with the rich Hessian data will expand our knowledge of migration behavior in nineteenth-century Europe.
28 29 30
NOTES
31 32 33 34 35 36 37 38 39 40
1. In the early part of the nineteenth century, emigration restrictions eased for many Europeans, partly as a result of the 1815 Vienna Congress proceedings. By 1930 most of the destination countries had managed to put into place restrictions on immigrant flows. In absolute numbers, emigration from Europe was greatest in the first ten years of the twentieth century. It dropped off considerably in the next decade, by approximately 30%, but the number of emigrants dropped off even more between the 1920s and the 1930s, by approximately 80%. These estimates are derived from statistics on European emigration in Mitchell (1975), Table B8, page 135. 2. By linking the U.S. passenger records to emigration lists in the migrants’ homelands, more can be learned about the limitations of both types of records. With few exceptions this has rarely been done. Exceptions include Swierenga (2000), an account 389
4259 Ch08 4/12/02 10:24 am Page 390
390 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
SIMONE A. WEGGE
of the linkage of thousands of Dutch migrants from Dutch emigration lists to the U.S. passenger lists. Olsson (1967) has linked many Swedish immigrants in the U.S. passenger records back to Swedish sources, and Günther (1992) has compared for the same ships Hamburg passenger lists with U.S. passenger lists. 3. On this point, Thistlewaite stated in 1960, “Settlement, assimilation, and, more recently, acculturation: these have been the central, urgent issues for American historians and sociologists alike. As a result, it has been the consequences and not the causes of migration which have received most attention; and, moreover, the consequences in the receiving, not the sending country” (Thistlewaite, 1991). 4. These data are collectively known as the HESAUS data. Portions of what are analyzed in this paper have been published in Auerbach (1987–1988). 5. These data have been analyzed in other work, including Wegge (1997), (1998), (1999), (2001a), (2001b) and (2002). 6. See Brown, Guinnane et al. (1993) for a discussion of the relevance of micro data and what they mean for the study of fertility. See Ferrie (1996) for a discussion of the importance of migrant data at the micro and longitudinal levels and their relevance for economic, geographic and social mobility studies. 7. Swierenga (2000) warns that many sources, both emigration and immigration may seriously undercount migrants, as is the case for international Dutch migrants. 8. See Baines (1985, p. 3), on missing lists. The way they were collected poses additional problems. They do not distinguish, for instance, between British and foreign citizens before 1853, or between cabin and steerage passengers until 1863. The foreign/British classification and the cabin/steerage designation are discussed in several places, namely Ferenczi and Willcox (1929, p. 622) and (Davis, 1931, p. 650). See Ferenczi and Willcox (1929, pp. 621–624), and Baines (1985, p. 3), on the lack of data by migrant origins. On the low-quality occupational information in British statistics, see Baines (1985, p. 51). 9. The U.S. data are known as the customs passenger lists before the 1890s and as the immigration passenger lists thereafter. I refer to them collectively as the U.S. passenger ship records (lists). See Tepper (1993) for details. Scholars who have used these statistics to study British emigrants include Cohn (1992, 1995), Erickson (1981, 1990), Ferrie (1999), and Van Vugt (1999). 10. See Mokyr (1983), Mokyr and Ó Gráda (1982), and Ó Gráda (1986) for studies of Irish emigration, based on both Irish and American data sources, but in large part on American passenger ship records. Some Irish emigration statistics do indeed exist which contain detailed information about the individuals leaving, such as age, gender, conjugal status, home county, and destination. Although this source does not contain information on an emigrant’s occupation or the home village, other Irish sources do not contain equally specific data (Ó Gráda, 1975). Most studies of Irish emigration patterns have thus relied on American immigration passenger lists. 11. Scandinavian governments typically required persons departing to have a passport that had been issued by the police. Such data can make for more complete emigration statistics at the individual level. The Swedish passport journals, for instance, provide names, occupations, dates of departure, home districts, and destinations of the individuals traveling to foreign lands (Tedebrand, 1976, p. 82). With respect to Danish emigration, Hvidt’s study of Danish emigrants used data gathered from contracts between emigration agents and prospective emigrants to analyze over 90% of the Danish emigrants leaving between 1868 and 1900. These data, covering over 170,000 emigrants,
4259 Ch08 4/12/02 10:24 am Page 391
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
391
provide many interesting details, including time of departure, gender, occupation, place of residence, and age (Hvidt, 1975). 12. Kamphoefner (2001b), for example, uses new records from the Staatsarchiv of Düsseldorf on overseas emigrants leaving the Regierungsbezirke of Düsseldorf, Cologne, and Aachen. 13. Kamphoefner (1987) includes a statistical analysis of the amounts of cash emigrants took with them. It is one of the few, besides my own study, to do so. New studies of internal migration, based on archival documents, are Jackson (1997) and Hochstadt (1999). The project discussed in Brown, Guinnane et al. (1993) will also study internal migration. For more information on regional German emigration, see Adams (1980), a collection of papers from a conference on the state of migration history scholarship in Germany. 14. Work of this nature includes Wegge (1998) and Wegge (1999). 15. Assion and his research team studied the Hamburg passenger ship data on emigrants from all the three Hessian principalities of Hesse-Cassel, Hesse-Darmstadt and Nassau (Assion, 1991). This project examined the Hamburg ship lists for Hessians who left between 1855 and 1866. Hamburg shipping laws required ship captains to complete lists of emigrants upon embarkation. These Hamburg data were transcribed in German, and are a second but different source of information on emigrants from the same region and other regions close to the homeland of the Hesse-Cassel emigrants. They are fairly complete in terms of the information provided on age and occupation, which Assion and his research team have analyzed. Specifically, the records provide name, occupation, age, gender, number of accompanying family members, home village and country, year of departure, ship number and type of ship and accommodation. Exact identification of home village and country, however, is only possible with the records in 1855 and thereafter. 16. Scholars have gone to great lengths to study various international migrant populations using these records. For most, this has usually involved a lot of arduous work with raw data. Some of the more interesting studies have been those that link names in the passenger records to subsequent U.S. census data, including Ferrie (1999) for immigrants who arrived in the 1840s, Swierenga (1991) and Swierenga (2000) for Dutch immigrants between the 1830s and 1880. Kamphoefner (1987) follows migrants from emigration records in Westfalia to U.S. census records for Missouri. Such methodologies allow scholars to ask questions about migrant progress at the destination in terms of wealth and occupation, or about the affect of immigration on the economic outcomes of native-born Americans. 17. The summary statistics for certain time periods have excluded the better-off emigrants who traveled in cabin class. The cabin/steerage classification problem is discussed by Hutchinson (1958, p. 980), and by (Davis, 1931, p. 650). 18. For passengers arriving from Bremen during the 1847–1854 period, 79% are described as from “Germany” and only 21% from a specific place of origin (Zimmerman & Wolfert, 1988, p. xi). 19. Love Across Color Lines is an account by Maria Diedrich of the decades-long relationship between the abolitionist Frederick Douglass and the German immigrant journalist Ottilie Assing, based on, among other sources, Assing’s correspondence with her sister. 20. The 1832 figure is from Hildebrand (1853, p. 43). 1858 figures are from a report on the 1861 census, in the Landwirtschaftliche Zeitschrift of 1863, pp. 209–234. 391
4259 Ch08 4/12/02 10:24 am Page 392
392 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
SIMONE A. WEGGE
21. Most of the inhabitants were Protestant. An 1827 census characterized the population as 83% Protestant, 15% Catholic, and 2% Jewish, with the majority of Catholics and Jews living in rural areas (Hildebrand, 1853, p. 63). 22. Clark’s estimates of German agricultural productivity show that Germany was less productive than much of Northern and Western Europe, and about half as productive as England (Clark, 1993, p. 228). Perkins (1981, p. 114) describes German agriculture in the 1840s as “amongst the most backward in Europe.” In detailing the many strides German farmers made towards improving the yields of their arable lands during the latter half of the nineteenth century, Perkins (1981, pp. 112–116) argues that Germany went through an agricultural revolution during this period. 23. Craftsmen could, however, apply for exemptions from this law. 24. About 75% of the citizenry lived in small towns and villages and earned their living directly or indirectly through the agricultural sector (Hildebrand, 1853, pp. 53, 99). 25. A few Hessians emigrated to other countries prior to 1831, but under very select circumstances. In the eighteenth century, various groups acquired special permission to settle in Russia and Hungary (Auerbach, 1993, pp. 45–63, 95–113). After the American Revolutionary War, some Hessian soldiers settled in the U.S. This was one of the largest eighteenth-century migrations for the principality, albeit an involuntary one (at least initially) on the soldiers’ part. There was also a sizable amount of emigration from the Kreis of Hanau in the eighteenth-century, at that point in time not a part of Hesse-Cassel. 26. Between 1851 and 1857 the average annual migration flow from Hesse-Cassel to the U.S. was 7.6 individuals per 1,000 Hessian-Cassel persons, while the figures for the British Isles and Norway during the 1850s were 5.8 and 2.4 per 1,000 people respectively. Ireland, however, had higher overseas emigration rates, 14 persons per 1,000 (Baines, 1991, p. 4). Population figures for the source country are from Mitchell (1992). Numbers of migrants are from U.S. Historical Statistics, and the Hesse-Cassel emigrant data. 27. After 1870, some of the highest emigration rates were in the northern regions of Germany (Nugent, 1995, p. 67). 28. Formal permission was referred to as “Entlassung aus dem Unterthanenverband” or in translation, release and/or discharge from the community of subjects. 29. Local village or town authorities recorded information on illegal emigrants and sent their reports on to higher-level government offices. Such records are available first in the year 1852. Most communities had a population of fewer than 2,000 individuals. Thus, it was not difficult for a local person to file a fairly informative report about neighbors who had left surreptitiously. 30. See Auerbach (1987–1988) for details on the HESAUS data. Auerbach (1993) is an extensive account of migration from Hesse-Cassel over several centuries. 31. The decision of authorities to relocate the state archive in the capital city of Cassel to the smaller and less central town of Marburg in the 1870s turned out to be most propitious. Essentially, it ensured the safety of the rich collection of archival documents for the principality of Hesse-Cassel during World War II, in particular the data set on nineteenth-century emigrants. Allied bombers leveled the city of Cassel; had the archive still resided in Cassel, fire would have destroyed most of the documents. Only two bombs fell on the city of Marburg, and these bombs caused only minor damage. 32. As Kamphoefner et al. (1991, p. 38), explains, “Several archives (German ones), too, have started index files of emigrants – the most recent and ambitious project of its kind is . . . from Hesse-Cassel in the Marburg State Archives.”
4259 Ch08 4/12/02 10:24 am Page 393
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
393
33. Using the first names of emigrants, a SAS program I wrote classified the names as either men’s or women’s names. This procedure provided a gender classification for 97% of the emigrants. Specific archival codes identified the illegal emigrants. The appendix contains further details. 34. These records are only examples of what is typically found in the HESAUS data. Names, codes and other variables have been changed to protect the dead but innocent. 35. Age information is the most incomplete variable, with ages existing for approximately 25% of the 55,000 different emigrants. 36. For example, the Westfalian data do not merely list that an emigrant was a farmer but offer some idea as to his holdings (Kamphoefner, 1987). 37. The Historical Commission of the Hesse-Cassel principality developed the survey and implemented the collection of the data. The village survey data are found in Bestand H3, Volumes 1–80, at the Hessisches Staatsarchiv, Marburg. A village resident, typically the mayor or a school teacher, collected and completed the survey. 38. Documents for a number of villages are missing, especially in the Fulda Kreis and the Kassel Kreis, meaning that over a hundred villages cannot be included in the study. The two Kreise excluded are Schaumburg and Rinteln, making up together 132 villages. Of the 1376 different Hesse-Cassel villages, archival documents still exist for 1,233 of them. 39. An English translation of the questions is available in Munter (1983). 40. The artisan and factory data are in Bestand 30, Repositur II, Klasse 10, Nr. 9–10, at the Hessisches Staatsarchiv, Marburg. The factory census also provides information on looms, mills and other power sources like steam engines. 41. Evidence exists that the records undercount young men eligible for military service and illegal emigrants before 1852. In spite of such sources of incompleteness, the data represent across many dimensions a very large proportion of the actual population of Hesse-Cassel emigrants. 42. The figure of 63% is based on a sample of 14,771. 43. If month of departure were uniformly distributed, 8.3% of the migrants for each year would leave every month. If one makes a conservative estimate and considers the months of October through January as the source of pent-up demand, on average these months deviate from 8% by 4.7%. By adding 4.7% to the 8.3% of each of the four months of February through May; each would show 13% of the total yearly migrants, for a combined total of 52% of the total migrants for the spring season. But, on average, 63% of all migrants left in these months for the years 1832 through 1857. Lastly, emigrants may have wanted to avoid the death rates on ships, which were highest in the winter months. 44. A great number left in November of 1848, but deviations during this year of political unrest are no great surprise. Two months later, a great number left in January, either related to the political events of 1848 or perhaps to gold rush fever in California. Steady increases in agricultural prices over 1846 and early 1847 may have inspired many to leave in the first few months of 1847: by April of 1847, 80% of the migrants leaving that year had left. In 1843, on the other hand, 27% of the migrants departed HesseCassel in the months of September and October. 45. See Kukowski (1992, p. 205), and Hahn (1986, p. 417). 46. As the capital city and the city with the largest population, the Cassel economy actually consisted of a number of sectors, including a sizable professional class and a lot of specialized and highly-skilled artisans. Schmalkalden boasted comparatively 393
4259 Ch08 4/12/02 10:24 am Page 394
394 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
SIMONE A. WEGGE
advanced industrial development and was known for its large exports of iron and iron products. Within Hesse-Cassel, many buildings, or farm and household goods containing iron were produced by Schmalkalden artisans (Bovensiepen, 1909, p. 102). 47. This is according to the document Bestand 16, Repositur II, Klasse 14, Nr. 23, Vols I–III in the Hessisches Staatsarchiv. Population censuses provide a slightly higher estimate, at 117 to 244 persons annually between the years 1843 and 1858. See the Landwirtschaftliche Zeitschrift (1850s or 1860s), pp. 365–375. If this had actually been the case, it represents a 10–20% return rate in the years prior to 1852, which fits in with return migration rates of 13–22% for Europeans estimated between 1868 and 1873 (Gabaccia, 1996, p. 94). 48. Many emigrants wanted land at a good price or an increase in social status (Kamphoefner et al., 1991, pp. 51–61). Other German principalities could not provide the land at the desired prices, and those which provided opportunities of factory jobs could not satisfy the desire for a higher station in life either. Perhaps many emigrants did not believe that another German state was an acceptable option. Kamphoefner et al. (1991) further discusses how potential land ownership was very important to German immigrants. The HESAUS data may also have possibly missed some stage migrants, who started out first as seasonal migrants in other German states but later became permanent migrants. This would bias emigration rates to other German states downward. 49. This is significant because many Hessians from Hesse-Cassel had established new homes in Russian provinces in the more restrictive times of the eighteenth-century. During this period, emigration entailed obtaining special permission from both the homeland and the host country. From such figures, one must be wary of any explanations of migration patterns that rest to a great degree on chain migration. Previous emigrants to Russia had few reasons to encourage more Hessians to emigrate to Russia. In the aftermath of Napoleon’s Russian campaign, out of more than 7,000 Hessian soldiers, the whereabouts of 5,540 of them were still unknown in 1818 (Auerbach, 1993, p. 95). For various reasons, Russia was no longer an attractive destination. 50. Bakers and butchers make up 2.9%, and this rises to 3.6% of the total. Similarly, the broadly defined group of merchants (which included shopkeepers, salesmen, pawnbrokers, and others) increases from 3.1% to 3.6%. More detailed tables are in Wegge (1997), Tables 4.5.2a–c. 51. Cases of divorced or separated women exist, but they are not immediately apparent from the emigrant records. 52. For various reasons, I surmise that the cash figures are biased downward, mostly because there seem to have been few good reasons for a departing Hessian to be exceptionally open with the authorities about personal financial matters. The Hesse-Cassel government required payment of a small administrative fee that increased with the amount of cash taken (Auerbach, 1993, p. 134). Many may have understated their true net worth upon applying for permission to emigrate. Of course, when an emigrant had sold off all his or her physical assets just prior to emigration, land especially, it was somewhat difficult for such knowledge to remain private. Unless the civil service employees who transcribed the information at the time of emigration had reason to increase the figures emigrants reported to them, I see no justification for believing that they are otherwise upwardly biased. The records show, in addition, absolutely nothing in terms of cash for many individuals. One does not know how many of these concern individuals with no assets beyond travel tickets or persons who managed to keep their personal financial business private.
4259 Ch08 4/12/02 10:24 am Page 395
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
395
53. There is no evidence to believe that through deflation the Thaler of the 1850s was worth more than in the 1840s. Emigrants may have also lied more to the authorities about their assets in the later period. 54. The average cash figure for 1841 to 1851 is 207 Thaler (no zeros and no values above 1000), while the average cash figure for legal emigrants between 1852 and 1857 is 152 Thaler. 55. Wegge (2001b) compares the wealth of emigrants with that of those who stayed home, proxying for the stayer wealth with land values. 56. The under-reporting seems to be worse for 1853 than for other years. Other sorts of errors exist as well. There are some individuals not emigrating with family and under the age of ten. Some of these are data errors. And sometimes a very young individual is traveling alone and is listed with an artisan classification. Whether the age of the individual has been properly stated is questionable. It is quite possible that many emigrants who did not travel with family made journeys, nonetheless, with distant relatives or friends, which might explain unaccompanied migrating infants in the data. 57. Whether this change is as prevalent among the women is unclear because sample sizes for women emigrating alone (or without family members) are incredibly small for the 1840s. 58. I did not calculate statistics for people traveling with family because the sample sizes were less than twenty-five. 59. Women leaving with their families are mostly under sixteen. For some years, there is a plethora of families inclusive of several children, a male head of household, but no wife or mother. One cannot be sure whether wives stayed at home, with younger children perhaps, or whether this constitutes sloppy (old-fashioned) recording by the authorities. 60. This literature can be traced back to Friedrich List’s writings in 1842, as described by Walker (1964, p. 48). 61. Inge Auerbach, archivist at the Hessisches Staatsarchiv, suspects that authorities had great difficulties in accounting for all those who left. 62. Johnson (1950, p. 45), argues that the so-called 48ers from all of Germany numbered in the few thousands and left anywhere between 1846 and 1856. 63. Mr. Maus is listed twice in the records; the two cash values differ by 5 Thaler, and only one of the records lists the month he left in. Likewise, Johann Peter’s age is listed in only one of the two records. 64. No pattern exists in the data that determines for which emigrants information is incomplete. I thus assume that statistics computed from the cash, occupation, and age variables are based on a random sample of the entire emigrant population. 65. Johann Anton and Joh. Anton is a typical example for a first name. Some of these are mistakes, while others arise due to the use of abbreviations. Unfortunately, a computer program recognizes each use of these as separate names. 66. Errors confined to particular Kreise or a certain province may exist too. These discrepancies would most likely be due to differences in Kreis-level administrative procedures and could create biases affecting the figures at the Kreis or province level. 67. These references included a village lexicon (Reimer, 1926), contemporary and historical maps (Hesse, Der Deutsche Planungsatlas, Hessen, Vol. IV; Hesse. Generalkarte von dem Kurfürstenthum Hessens 1859) and the state handbook for listings of communities (Hesse-Cassel. Staatskalendar, 1843, 1859). 395
4259 Ch08 4/12/02 10:24 am Page 396
396 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
SIMONE A. WEGGE
68. A second problem is the existence of several records for “Son of Georg” (anonymous child), each with a different age. If multiple ages for an anonymous child existed under the same archival record code, I assumed these were different children, even if two ages were only one year apart. If a child with a specific name, Maria for example, had two ages, I assumed this was the same person only if the ages were one year apart.
ACKNOWLEDGMENTS I wish to thank Dudley Baines, Greg Clark, Joseph Ferrie, Christiane Harzig, Walter D. Kamphoefner, Robert P. Swierenga, and anonymous referees for valuable comments and other contributions to this paper. A part of the work on this paper took place during the 1999–2000 academic year while I was a visiting scholar at the Hoover Institution of Stanford University, the Agricultural History Center at UC Davis and at the Institute of European Studies at UC Berkeley. I gratefully acknowledge the generous financial support of a grant for advanced research from the German Marshall Fund, an Arthur H. Cole Grant-in-Aid from the Economic History Association, and a summer research award in 2001 from the College of Staten Island of the City University of New York.
19 20
REFERENCES
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Adams, W. P. (Ed.) (1980). Die deutschsprachige Auswanderung in die Vereinigten Staaten: Berichte über Forschungstand und Quellenbestände. Berlin: John F. Kennedy Institut für Nordamerikastudien, Freie Universität Berlin. Assion, P. (Ed.) (1991). Über Hamburg nach Amerika. Hessische Auswandernde in den Hamburger Schiffslisten 1855–1866. Marburg: Institut für Europäische Ethnologie und Kulturforschung der Universität Marburg. Auerbach, I. (1987–1988). Hessische Auswanderer (HESAUS). Index nach Familiennamen (Vol. II). Marburg: Institut für Archivwissenschaft. Auerbach, I. (1993). Auswanderung aus Kurhessen: nach Osten oder Westen? Marburg: Hessisches Staatsarchiv Marburg. Baines, D. (1985). Migration in a Mature Economy: Emigration and Internal Migration in England and Wales 1861–1900. Cambridge: Cambridge University Press. Baines, D. (1991). Emigration from Europe 1815–1830. London: MacMillan. Bödiker, T. (1873). Die Auswanderung und die Einwanderung des preussischen Staates. Zeitschrift des Königlich Preussischen Statistischen Bureaus, 8, 1–42. Bovensiepen, R. (1909). Die kurhessische Gewerbepolitik und die wirtschaftliche Lage des zünftigen Handwerks in Kurhessen von 1816–1867. Marburg: Elwert. Brown, J. C., Guinnane, T. W., & Lupprian, M. (1993). The Munich Polizeimeldebögen as a Source for Quantitative History. Historical Methods, 26(3), 101–118. Clark, G. (1993). Agriculture and the Industrial Revolution 1700–1850. In: J. Mokyr (Ed.), The British Industrial Revolution: An Economic Perspective. Boulder, Colo.: Westview Press. Cohn, R. L. (1984). Mortality on Immigrant Voyages to New York, 1836–1853. Journal of Economic History, 50(2), 289–300.
4259 Ch08 4/12/02 10:24 am Page 397
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
397
Cohn, R. L. (1992). The Occupations of English Immigrants to the U.S., 1836–1853. Journal of Economic History, 52, 377–387. Cohn, R. L. (1995). A Comparative Analysis of European Immigrant Streams to the United States during the Early Mass Migration. Social Science History, 19, 63–89. Davis, M. R. (1931). Critique of Official United States Immigration Statistics. In: W. Willcox (Ed.), International Migrations (Vol. 2). New York: National Bureau for Economic Research. Diedrich, M. (1999). Love Across Color Lines: Ottilie Assing and Frederick Douglass. New York: Hill and Wang. Erickson, C. (1981). Emigration from the British Isles to the USA in 1831. Population Studies, 35, 175–197. Erickson, C. (1990). Invisible Immigrants: The Adaptation of English and Scottish Immigrants in Nineteenth Century America. Ithaca: Cornell University Press. (Originally published in 1972, Coral Gables: University of Miami Press.) Ferenczi, I., & Willcox, W. (1929). International Migrations (Vol. I). New York: National Bureau of Economic Research. Ferrie J. P. (1996). A New Sample of Males Linked from the Public Use Microdata Sample of the 1850 U.S. Federal Census of Population to the 1860 U.S. Federal Manuscript Schedules. Historical Methods, 29(4), 141–156. Ferrie J. P. (1999). ‘Yankees Now’: Immigrants in the Antebellum U.S. 1840–1860. New York: Oxford University Press. Gabaccia, D. (1996). Women of the Mass Migrations: From Minority to Majority, 1820–1930. In: D. Hoerder & L. Moch (Eds), European Migrants: Global and Local Perspectives. Boston: Northeastern University Press. Glazier, I. A. (1984). Ships and Passengers in Emigration from Italy to the U.S., 1880–1900. In: W. Miodunka & A. Brozek (Eds), Emigration from Northern, Central and Southern Europe: Theoretical and Methodological Principles of Research. International Symposium, Kraków, November 9–11, 1981. Kraków: Nakladem Uniwersytetu Jagiellonskiego. Glazier, I. A., & Kleiner, R. J. (1996). Comparative Analysis of Emigrants from Southern and Eastern Europe from U.S. Ship Passenger Lists: 1910. In: A. Newman & S. W. Massil (Eds), Patterns of Migration, 1850–1914. Proceedings of the International Academic Conference of the Jewish Historical Society of England and the Institute of Jewish Studies, University College London. London: Jewish Historical Society of England and the Institute of Jewish Studies, University College London. Günther, M. (1992). Auswandererlisten und Passenger Lists, 1855–1864: Eine vergleichende Quellenkritik als methodischer Beitrag zur Auswanderungsforschung, M. A. Thesis (History), Ruhr Universität Bochum. Guinnane, T. (1997). The Vanishing Irish. Princeton: Princeton University Press. Hahn, H.-W. (1986). Der hessische Wirtschaftsraum im 19. Jahrhundert. In: W. Heinemeyer (Ed.), Das Werden Hessens. Marburg: Veröffentlichungen der Historischen Kommission für Hessen. Hansen, M. L. (1961). The Atlantic Migration 1607–1860. New York: Harper Torchbooks. (Originally published by Cambridge: Harvard University Press, 1940.) Harris, J. F. (1989). The Arrival of the Europamüde: Germans to America after 1848. In: C. L. Brancaforte (Ed.), The German Forty-Eighters in the United States. New York: Peter Lang. Hatton, T., & Williamson, J. (1998). The Age of Mass Migration: Causes and Economic Impact. New York: Oxford University Press. Hesse. Der Deutsche Planungsatlas, Hessen (Vol. IV). Hannover: Akademie für Raumforschung und Landesplanung.
397
4259 Ch08 4/12/02 10:24 am Page 398
398 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
SIMONE A. WEGGE
Hesse. Generalkarte von dem Kurfürstenthum Hessens 1859. (originally from Koenigl. Kart. Inst. Berlin) Wiesbaden, Germany: Hessische Landesvermessungsamt. Hesse-Cassel. Landwirtschaftliche Zeitschrift für Kurhessen, Landwirtschaftliche Anzeiger, 1855–1868. Hesse-Cassel. Staatskalendar, 1843, 1859. Hessisches Staatsarchiv, Marburg. Bestand H3, Vols 1–80 (Village survey). Hessisches Staatsarchiv, Marburg. Bestand 16, Repositur II, Klasse 14, Nr. 23, Vols I–III (Interior Ministry of Hesse-Cassel). Hessisches Staatsarchiv, Marburg. Bestand 30, Repositur II, Klasse 10, Nr. 9 & Nr. 10 (Statistical Commission). Hildebrand, B. (1853). Die Statistische Mittheilungen über die Volkswirtschaftlichen Zustände Kurhessens. Berlin: Trautweinschen. Hippel, W. von. (1984). Auswanderung aus Südwestdeutschland: Studien zur wurttembergischen Auswanderung und Auswanderungspolitik im 18. und 19. Jarhundert. Stuttgart: Klett-Cotta. Hochstadt, S. (1999). Mobility and Modernity: Migration in Germany, 1820–1989. Ann Arbor: University of Michigan Press. Hubrich, M., Meesenburg, J., & Schriewer, K. (1991). Strukturmerkmale der hessischen Amerikaauswanderung – Eine Analyse der Schiffslisten. In: Assion (Ed.), Über Hamburg nach Amerika. Hessische Auswandernde in den Hamburger Schiffslisten 1855–1866. Marburg: Institut für Europäische Ethnologie und Kulturforschung der Universität Marburg. Hutchinson, E. P. (1958). Notes on Immigration Statistics of the United States. Journal of the American Statistical Association, 53, 963–1025. Hvidt, K. (1975). Flight to America. The Social Background of 300,000 Danish Emigrants. New York: Academic Press. Jackson, J. H., Jr. (1997). Migration and Urbanization in the Ruhr Valley, 1820–1914. Atlantic Highlands, NJ: Humanities Press. Johnson, H. B. (1950). Adjustment to the United States. In: A. E. Zucker (Ed.), The Forty-Eighters, Political Refugees of the German Revolution of 1848. New York: Columbia University Press. Kamphoefner, W. D. (1986). At the Crossroads of Economic Development: Background Factors Affecting Emigration from Nineteenth-Century Germany. In: I. A. Glazier & L. De Rosa (Eds), Migration Across Time and Nations: Population Mobility in Historical Contexts. New York: Holmes & Meier. Kamphoefner, W. D. (1987). The Westfalians: From Germany to Missouri. Princeton: Princeton University Press. Kamphoefner, W. D. (1995). German Emigration Research, North, South, and East: Findings, Methods, and Open Questions. In: D. Hoerder & J. Nagler (Eds), People in Transit: German Migrations in Comparative Perspective. New York: Cambridge University, German Historical Institute. Kamphoefner, W. D. (2001a). Ansätze und Ergebnisse der modernen Migrationsforschung. Kölner Vorträge zur Sozial- und Wirtschaftsgeschichte 42. Cologne: Fördererkreis des Forschungsinst. für Sozial- und Wirtschaftsgeschichte, Universität zu Köln. Kamphoefner, W. D. (2001b). Rheinlanders and Rheinlands in the USA: Factors Influencing Immigrant Destinations. In: Schöne Neue Welt: Rheinländer erobern Amerika, Ed. Landschaftsverband Rheinland, conference proceedings (pp. 296–306). Kommern: Mechernich. Kamphoefner, W. D., Helbich, W., & Sommer, U. (Eds) (1991). News From the Land of Freedom. German Immigrants Write Home. Ithaca: Cornell University Press.
4259 Ch08 4/12/02 10:24 am Page 399
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
399
Karl, T. (1983). Die Waldeckische Auswanderung zwischen 1829 und 1872. Cologne: M. Thomas. Köllmann, W., & Marschalck, P. (1973). German Emigration to the United States. In: D. Fleming & B. Bailyn (Eds), Perspectives in American History, 7, 499–554. Kukowski, M. (1992). Vom ‘Pauperismus’ zum ‘Proletariat’: Stufen und Phasen der Pauperisierung in Kurhessen in der ersten Hälfte des 19. Jahrhunderts. Hessisches Jahresbuch für Landesgeschichte, 42, 185–217. Kukowski, M. (1995). Pauperismus in Kurhessen: Ein Beitrag zur Entstehung und Entwicklung der Massenarmut in Deutschland 1815–1855. Darmstadt und Marburg: Hessische Historische Kommision Darmstadt, Historische Kommision für Hessen. List, F. (1926). Friedrich Lists kleinere Schriften. In: F. Lenz (Ed.), Die Herdflamme, Sammlung der gesellschaftswissenschaftlichen Grundwerke aller Zeiten und Völker (Vol. 10). Jena, Germany: Gustav Fischer. Lubinski, A. (1995). Overseas Emigration from Mecklenburg-Strelitz: The Geographic and Social Contexts. In: D. Hoerder & J. Nagler (Eds), People in Transit: German Migrations in Comparative Perspective. New York: Cambridge University, German Historical Institute. Marschalck, P. (1973). Deutsche Überseewanderung im 19. Jahrhundert. Stuttgart: Ernst Klett. McInnis, M. (1994). Immigration and Emigration: Canada in the Late Nineteenth-Century. In: T. J. Hatton & J. G. Williamson (Eds), Migration and the International Labor Market, 1850–1939. London: Routledge. Mitchell B. R. (1975). European Historical Statistics 1750–1970. New York: Columbia University Press. Mitchell B. R. (1992). International Historical Statistics: Europe 1750–1988. New York: Stockton Press. Moch, L. P. (1992). Moving Europeans. Migration in Western Europe since 1650. Bloomington: Indiana University Press. Moe, T. (1970). Demographic Developments and Economic Growth in Norway 1740–1940: An Econometric Study. Unpublished dissertation, Stanford University. Möker, U. (1977). Nordhessen im Zeitalter der Industriellen Revolution. Cologne: Böhlau. Mokyr, J. (1983). Why Ireland Starved: A Quantitative and Analytical History of the Irish Economy, 1800–1850. London: Allen & Unwin. Mokyr, J., & Ó Gráda, C. (1982). Emigration and Poverty in Prefamine Ireland. Explorations in Economic History, 19, 360–384. Munter, C. (1983). Das Land der Armen Leute: Poverty in the Marburg Region of Hesse in the Nineteenth Century. Unpublished dissertation, Johns Hopkins University. Noback, C., & Noback, F. (1858). Münz-, Mass- u. Gewichtsbuch. Das geld-, maass- und wechselwesen, die Kurse, Staatspapiere, Banken, Handelsanstalten und Usanzen aller Staaten und wichtiger Orte. Leipzig: Brockhaus. Nugent, W. (1995). Crossings: The Great Transatlantic Migrations, 1870–1914. Bloomington: Indiana University Press (reprint of 1992 edition). Ó Gráda, C. (1975). A Note on Nineteenth-Century Irish Emigration Statistics. Population Studies, 29(1), 143–149. Ó Gráda, C. (1986). Across the Briny Ocean: Some Thoughts on Irish Emigration to America, 1800–1850. In: I. Glazier & L. DeRosa, Migration Across Time and Nations: Population Mobility in Historical Contexts. New York: Holmes & Meier. Olsson, N. (1967). Swedish passenger arrivals in New York, 1820–1850. Chicago: Swedish Pioneer Historical Society. Ostergren, R. C. (1986). Swedish Migration to North America in Transatlantic Perspective. In: I. Glazier & L. DeRosa, Migration Across Time and Nations: Population Mobility in Historical Contexts. New York: Holmes & Meier.
399
4259 Ch08 4/12/02 10:24 am Page 400
400 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
SIMONE A. WEGGE
Pedlow, G. (1988). The Survival of the Hessian Nobility 1770–1870. Princeton: Princeton University Press. Perkins, J. A. (1981). The Agricultural Revolution in Germany 1850–1914. Journal of European Economic History, 10(1), 71–118. Perlmann, J. (1996). Selective Migration as a Basis for Upward Mobility?: The Occupations of the Jewish Immigrants to the United States. Mimeo, Levy Economics Institute. Reimer, H. (1926). Historisches Ortslexikon für Kurhessen. Marburg: Elwert. Richter, K. (1980). Die hansestädtischen, vor allem Hamburger Quellen zur Nordamerika – Auswanderung. In: W. P. Adams (Ed.), Die deutschsprachige Auswanderung in die Vereinigten Staaten: Berichte über Forschungsstand und Quellenbestände. Berlin: John F. Kennedy Institut für Nordamerikastudien, Freie Universität Berlin. Swierenga, R. P. (1986). Dutch International Migration and Occupational Change: A Structural Analysis of Multinational Linked Files. In: I. A. Glazier & L. DeRosa (Eds), Migration Across Time and Nations: Population Mobility in Historical Contexts. New York: Holmes & Meier. Swierenga, R. P. (1991). Local Patterns of Dutch Migration to the United States in the Mid-Nineteenth Century. In: R. J. Vecoli & S. M. Sinke (Eds), A Century of European Migrations, 1830–1930. Urbana: University of Illinois Press. Swierenga, R. P. (2000). Faith and Family: Dutch Immigration and Settlement in the United States, 1820–1920. New York: Holmes & Meier, Ellis Island Series. Tedebrand, L. G. (1976). Sources for the History of Swedish Emigration. In: H. Runblom & H. Norman (Eds) From Sweden to America: A History of the Migration. Minneapolis: University of Minnesota Press. Tepper, M. (1993). American Passenger Arrival Records: A Guide To The Records Of Immigrants Arriving At American Ports By Sail And Steam. Baltimore: Genealogical Pub. Co. Thistlewaite, F. (1991). Migration from Europe Overseas in the Nineteenth and Twentieth Centuries. Reprint of essay presented at XI International Congress of Historical Sciences, Stockholm, 1960. In: R. Vecoli & S. Sinke (Eds), A Century of European Migrations, 1830–1930. Urbana: University of Illinois Press. United States Census Bureau (1975). Historical Statistics of the United States: Colonial Times to 1970. Washington: Government Printing Office. Van Vugt, W. E. (1999). Britain to America: Mid-Nineteenth-Century Immigrants to the United States. Urbana: University of Illinois Press. Walker, M. (1964). Germany and the Migration 1816–1885. Cambridge: Harvard University Press. Wegge, S. A. (1997). Migration Decisions in Mid-Nineteenth Century Germany. Unpublished dissertation, Northwestern University. Wegge, S. A. (1998). Chain Migration and Information Networks: Evidence From NineteenthCentury Hesse-Cassel. The Journal of Economic History, 58(4), 957–986. Wegge, S. A. (1999). To Part or Not to Part: Emigration and Inheritance Institutions in Mid-19th Century Germany, Explorations in Economic History, 36(1), 30–55. Wegge, S. A. (2001a). A Historical Perspective on Female Migrants. In: P. Sharpe (Ed.), Women, Gender and Labour Migration: Historical and Global Perspectives. London: Routledge. Wegge, S. A. (2001b). Do migrant origins matter? Migrants vs. Non-Migrants: Self-Selection of Migrants Out of the Home Population. Mimeo. Wegge, S. A. (2002). Occupational Self-Selection of European Emigrants: Evidence from Nineteenth-Century Hesse-Cassel. Forthcoming in European Review of Economic History. Zimmerman, G. J., & Wolfert, M. (1988). German Immigrants: Lists Of Passengers Bound From Bremen To New York, 1863–1867 With Places Of Origin (Vol. 3). Baltimore: Genealogical Publishing Co.
4259 Ch08 4/12/02 10:24 am Page 401
Transatlantic Emigrants 1
401
APPENDIX
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
The emigrant data of Hesse-Cassel, referred to at the Hessian State Archive as the HESAUS data are explained in Auerbach (1987–1988). Many adjustments were made to the data set before any analysis commenced. This appendix describes the procedures used. Various levels of government kept information on emigrants, which subsequently created multiple observations for many individual emigrants. Different records for the same person do not always provide consistent information, as evident from the records on Mr. Oskar Maus in Table 1.63 Government authorities recorded these data with a particular focus on the heads of households, primarily men. They wanted to make sure that emigrants did not leave behind family or financial responsibilities and that young men were not shirking their duty of military service. A final concern was to keep track of the cash being taken out of the country. As explained in the main text, the data consist of eleven main variables, to which I augmented it by two more. Below I discuss some of the problems I encountered with several of the variables.
19 20
Preparing the Hesse-Cassel Data for Analysis
21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
A person was defined in the computer algorithms by unique village, surname, first name, and year of migration. To minimize the double counting of the same person, and to ensure an accurate emigration rate, I prepared the data for analysis by removing as many errors as possible in these four variables. These are described below. Although the information on most variables is fairly complete, the information on age, occupation, and exported cash is often missing. As a result, some of the statistics on emigrant characteristics are based on smaller sample sizes.64 An important piece of my study involves an analysis of village emigration rates (defined as the number of emigrants divided by the village population). In calculating these rates I assumed that the probability that two families from the same village with identical surnames emigrating in the same year was very small, and close to zero. I defined a family or a person emigrating alone by a unique combination of surname, village, and year. To define an individual within a family, I included first name to this list. The family assumption biases calculations of the number of emigrant families downward, if during one year, two unrelated individuals from the same village with the same surname emigrate. Other problems upwardly bias the emigration rate. Often two different records exist for the same person, but each 401
4259 Ch08 4/12/02 10:24 am Page 402
402 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
SIMONE A. WEGGE
contains a slightly different spelling for the surname or more commonly, the first name.65 Sources of this measurement error exist primarily because many different levels of government kept information on the same emigrant, so that several different offices and thus bureaucrats handled the same set of cases, inducing multiple spellings of the same name. Different levels of government naturally varied in the information they thought was important to collect. Mistakes by archive employees in transferring the data from the original documents to computer files have occurred as well. The data are most accurate for 1832–50, for which years the archive has reviewed all the original documents and published the emigrant lists. For the years 1851 through 1857, however, the problem of missing data is more severe because the archive has transcribed only the emigrant records for one administrative level, the interior ministry. In the 1850s authorities may have also been less careful about collecting detailed information on each emigrant because of the great numbers of emigrants leaving during these years.66 Each of the issues described above hinders an accurate examination of the data. To prepare the data for analysis I inspected each variable, correcting the names of many individual observations and using computer algorithms to properly summarize the information on each person. Minor problems still exist: some of them can only be corrected by further manual corrections, and others require investigation of church and Kreis records. The work I have so far completed provides, however, emigration rates and other statistics that approach the true values fairly closely. Manual inspection of the records for one Kreis (Kirchhain) provides further reassurance.
25 26
Correcting Village Name, Surname, and First Name
27 28 29 30 31 32 33 34 35 36 37 38 39 40
The emigrant data set listed 2,400 unique villages, more than one thousand too many. Most of the “extra” villages were multiple spellings of the same village. Many others were areas or subdivisions of one of the 1,376 recognized villages. Residents sometimes identified their neighborhoods within their village and not by the proper name for their village. I reclassified records under the appropriate village names. To assign these locales to their proper village, I relied on various reference materials, including a village lexicon, nineteenth- and twentiethcentury maps, and various years of the State Handbook for Hesse-Cassel.67 For most of these locales I was able to cipher their location and find which village they belonged to. A handful that listed a foreign city as their home were deleted from the analysis. Kreis misclassifications were also present, stemming mostly from post-1857 border alterations. The remaining ones that I could not trace I deleted, which amounted to the deletion of a couple of hundred migrants. Each
4259 Ch08 4/12/02 10:24 am Page 403
Transatlantic Emigrants 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
403
of these corrections involved one line of code per “village problem” in a computer algorithm, thus over 1,000 manual changes. These efforts have allowed a precise matching of emigrant data set with the census and village survey data that are both broken down by village. To calculate accurate village emigration rates, and to properly study emigration patterns in terms of the numbers of families or unaccompanied persons leaving, a proper accounting of the emigrants’ surnames was necessary. Various errors in the surname variable were possible. Multiple spellings of surnames, such as Anton Jakob listed separately under Jacob and Jakob, or Caroline Noeding listed in addition as Caroline Doeding, were present. Double names were also prevalent, such as Johann Schmidt listed under two surnames, Schmidt and Schmitt. A double name such as Johann Mueller/Schmidt indicated that Johann was the wife’s child from her first marriage or an illegitimate child, and meant that he probably traveled with the Mueller family. To eliminate multiple spellings I examined all surnames that showed up more than once either with the same first two letters, or the same letters but a different first letter. For each I made a judgement as to whether two records represented the same person; if most of the other information for the record agreed, I could be fairly sure that the two records concerned the same individual. Approximately 1,500 manual changes were handled. To take care of the double names I corrected most of these by writing a matching program that assigned a person with two names to a family, if one existed, with the same name and departure dates. Remaining problems were treated individually, with corrections hardcoded in the appropriate computer algorithms. Likewise, remaining minor problems were also hard-coded individually. I investigated each stepchild and illegitimate child individually, and similarly hard-coded the necessary changes into the programs. The government’s interest in the heads of households is particularly evident in how first names were described. The data, for instance, refer to many wives and children merely as “Wife of Johann” or “Son (or Daughter) of Johann.” Again, the accuracy of emigration rates is improved if one corrects for inaccuracies in the first name variable or recognizes the types of multiple records in the emigration rate calculation procedure. Two types of complications existed. First, multiple spellings of a first name are possible, with Isaak Kaufmann listed also as Isk. Kaufmann. I pinpointed the most common misspellings and wrote a program that abbreviated most of the men’s and women’s names to make the names consistent within a single family. For instance, Wilhelm cropped up in all of the following ways: Wihelm, Wih., Wilh., Wh., Wl., etc. All of these were rewritten as “W.” Using these first names, a computer algorithm determined the gender of over 97% of the emigrants. 403
4259 Ch08 4/12/02 10:24 am Page 404
404 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
SIMONE A. WEGGE
Secondly, wives and children are listed multiple times under their own specific first name, and/or respectively as “Wife of Johann” and “Son of Johann.” Multiple observations often exist for the same individuals, but the archival codes can eliminate the unnecessary duplications. A glance at the Becker family in Table 1 illustrates this situation. These observations come from only two archival records, coded as 4/593,197 and 1A/23,213. By comparing the two sets of records, one can conclude that they contain only one head of household, Johann Peter, that Christine, age 44, is most probably “Wife of Johann Peter,” and that two children exist, a son of age one and a daughter of age two. This comparison precludes the possibility that there are four children, two adult aged women, and two adult aged men. Multiple listings of the heads of household were easiest to eliminate because more information exists for them. In contrast, wives are often listed both under their first names and as “Wife of . . .,” which makes it impossible sometimes to discern whether a second adult woman, aside from the wife, was emigrating as well. Although one cannot be absolutely sure whether in this case Christine and “wife of Johann Peter” represent the same person, most likely they do. Wives listed both under their own first name and as “Wife of . . .” is a problem that arises primarily in the year of 1852.
20 21 22
Correcting Information on Age, Occupation, Destination, Cash, and Accompanying Family
23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Age information for an emigrant is very often missing, to the extent that one cannot use this variable in any extensive manner to identify individuals. Nevertheless, for studying the age distribution of the emigrants, accuracy is important. Several records, for example, can exist for Johann Schmidt, each with a different age, but leaving the same village at the same time. If several records existed for Johann with identical name, village, emigration date, and with varying ages, I assumed that the records concerned a single person if the ages were all greater than twenty and the ages were within five years of each other.68 For some emigrants with multiple records, conflicting information existed on exported cash, occupation, destination, or the number of family members accompanying heads of household. For a specific last name, village, year of emigration, first name, and sex, I compressed the records into one observation and carried the extra information in extra variables. For the cash information I constructed a lower and upper bound. I assumed that the more credible number for family size was the higher one, whether from adding up the observations on each family or using the higher number for accompanying family members.
4259 Ch08 4/12/02 10:24 am Page 405
Transatlantic Emigrants 1
405
Determining Illegal/Legal Status
2 3 4 5 6 7 8 9 10
After the 1848 revolution many citizens began to leave without applying for permission, hence departing on an illegal basis. In 1852 authorities began keeping track of those who left illegally in addition to those who successfully applied for permission before going. Certain archival codes in the HESAUS data set identified illegal emigrants for the years 1852 to 1857, as described in the archival document, Bestand 16, Repositur II, Klasse 14, Number 23, Vols I–III. From this information, I was able to assign illegal or legal status for most of the emigrants from the years 1852 to 1857.
11 12
Counting the Number of Families and Individuals
13 14 15 16 17
Multiple listings of the same person make it difficult to figure out whether one or more persons are concerned. It is particularly difficult when trying to ascertain how many members are in a family. I calculated the number of family members in three ways and then took the maximum:
18 19 20 21 22 23 24 25 26 27
(1) If heads of households were listed with the number of accompanying family members, I added one to this number to come up with a family size. (2) Sometimes children were listed anonymously as Son of Johann, etc. I recognized the sons per unique archival code as separate individuals and added these up to arrive at a total for the number of children. (3) Many families were displayed twice, once with the children anonymously listed as in the second method, and a second time each by their own first names. For each possibility I added the family members up and took the higher number.
28 29 30 31 32
To test the accuracy of the final tabulations my programming methods produced, I calculated by hand the emigration figures for the Kreis of Kirchhain. These tests show that my hand calculations differ from the computer calculations by less than 10%.
33 34 35 36 37 38 39 40
405
