Aggregate Behaviour of Investment in China, 1953-96
Also published in association with the Institute of Social Studie...
35 downloads
506 Views
14MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Aggregate Behaviour of Investment in China, 1953-96
Also published in association with the Institute of Social Studies, The Hague Selected titles: An]an Kumar Datta LAND AND LABOUR RELATIONS IN SOUTH-WEST BANGLADESH Resources, Power and Conflict Luis Carlos Jemio DEBT, CRISIS AND REFORM IN BOLIVIA Biting the Bullet Joke Luttik ACCOUNTING FOR THE GLOBAL ECONOMY Measuring World Trade and Investment Linkages S. Parasuraman THE DEVELOPMENT DILEMMA Displacement in India Jan Nederveen Pieterse (editor) WORLD ORDERS IN THE MAKING Humanitarian Intervention and Beyond Laixiang Sun AGGREGATE BEHAVIOUR OF INVESTMENT IN CHINA, 1953-96 An Analysis of Investment Hunger and Fluctuation Howard White (editor) AID AND MACROECONOMIC PERFORMANCE Theory, Empirical Evidence and Four Country Cases
Institute of Social Studies, The Hague Series Standing Order ISBN 0-333-71477-6 (outside North America only)
You can receive future titles in this series as they are published by placing a standing order. Please contact your bookseller or, in case of difficulty, write to us at the address below with your name and address, the title of the series and the ISBN quoted above. Customer Services Department, Macmillan Distribution Ltd, Houndmills, Basingstoke, Hampshire RG21 6XS, England
Aggregate Behaviour of Investment in China, 1953-96 An Analysis of Investment Hunger and Fluctuation Laixiang Sun Economist International Institute for Applied Systems Analysis Austria
in association with Institute of Social Studies
© Institute of Social Studies 2001 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, 90 Tottenham Court Road, London W1P0LP. Any person who does any unauthorised act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The author has asserted his right to be identified as the author of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2001 by PALGRAVE Houndmills, Basingstoke, Hampshire RG21 6XS and 175 Fifth Avenue, New York, N. Y. 10010 Companies and representatives throughout the world PALGRAVE is the new global academic imprint of St. Martin's Press LLC Scholarly and Reference Division and Palgrave Publishers Ltd (formerly Macmillan Press Ltd). ISBN 0-333-94809-2 This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Sun, Laixiang, 1957Aggregate behaviour of investment in China, 1953-96: an analysis of investment hunger and fluctuation / Laixiang Sun. p. cm. Includes bibliographical references and index. ISBN 0-333-94809-2 1. Investments—China. I. Title. HG5782 .S87 2001 332.6'0951—dc21
2001021626 10 10
09
9 08
8 7 6 5 4 3 2 07 06 05 04 03 02
Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire
1 01
Contents
List of Tables and Figures Preface
IX
xii
1
Introduction
1.1 1.2 1.3 1.4 1.5 1.6 1.7
Key Questions Growth Cycle Model Background Cycles or Fluctuation? Methodological Issues Scope and Notes on the Database Organization of this Book
1 3 8 15 20 24 27
2
Integrating Selected Theories Based on China's Experiences
29
2.1 Introduction 2.2 Incorporating the Investment Cycle Theory of the 'Hungarian School' with China's Experiences 2.2.1 Soft budget constraint, expansion drive and investment hunger 2.2.2 Bauer's four-phase theory and China's investment cycles 2.2.3 Single control equation model and the relevant problems 2.3 Kaleckian Economic Growth Theory and China's Capital Accumulation Mechanism 2.3.1 The causality line: From growth rate to investment to saving to bottleneck constraints 2.3.2 Generalizing the Kaleckian labour constraint equation to a key factor constraint equation 2.3.3 China's capital accumulation mechanism and Kaleckian agricultural-determining growth theory
29 31 31 35 38 45 45 48 51
vi
Contents
23 A Equilibrium growth path and fluctuations 2.4 Existing Researches on China's Investment Cycles 2.4.1 The efforts to link investment cycles to agricultural fluctuations 2.4.2 Reform cycle theory and the persistence of substitution between growth and bottleneck 2.5 Insights from Western Business Cycle Theories 2.6 Summary: The Implications for Modelling Investment Cycles in China 3
The State Investment System and its Response to Reform
3.1 Introduction 3.2 A Historical Overview of the State Investment System 3.3 The Project Approval System and Project Approval Norms 3.3.1 The project approval system: Formal procedure versus real practice 3.3.2 Project approval norms and locality's evasion by collusion 3.4 Credit Plan and Government Control over Financial Resources 3.4.1 Investment plan, credit plan and mandatory loans 3.4.2 Central government investment hunger and key state projects 3.4.3 Lending outside the credit plan 3.4.4 Adjustments of nominal interest rates and patterns of real interest rates 3.4.5 Recurring cycle of inflation and retrenchment during the reform period 3.5 The Material Supply System 3.5.1 Material supply system: Function and characteristics 3.5.2 Declining importance and its special focus since reform 3.6 Soft Budget Constraint and Investment Hunger in State-owned Enterprises 3.7 The Development Drive and Investment Hunger of Local Governments 3.8 Summary: Insatiable Investment Demand Exists at All Levels 4
Agricultural Constraint to the Insatiable Investment Demand
4.1 Introduction
55 57 57 59 63 69
71 71 72 79 79 82 84 84 86 89 90 95 97 97 100 102 106 109
112 112
Contents 4.2 Contribution of the Agriculture to the National Economy 4.3 The Change of Factor Proportions in China's Agriculture 4.4 The Specific Institutional Setting to Help Minimize Agricultural Fluctuation 4.5 Selection of Indicator System 4.6 Agricultural Fluctuations and Macroeconomic Adjustment: Empirical Evidence 4.7 Agricultural Fluctuations and Macroeconomic Adjustment: Stylized Facts 4.8 Summary Data Appendix 5
Energy as the Representative of Producer Goods Constraints
5.1 5.2 5.3 5.4 5.5
Introduction Energy Situation in China: An Overview Widespread and Chronic Shortage of Energy in China Transport Bottleneck and Effective Energy Supply Energy Constraint to Investment Demand: Some Primary Econometric Evidence 5.6 Summary
6
Estimating Investment Functions Based on Cointegration
vii 114 118 122 124 130 149 156 158
162 162 166 174 179 184 186
188
6.1 Introduction 6.2 Unit Roots, Equilibrium Relationship and Error Correction Mechanism 6.3 Modelling Strategy and Steps: A General Framework 6.4 Estimate of Cointegration and Investment Level Equation 6.5 Estimate of Conditional Investment Growth Rate Equation 6.6 Theoretical and Empirical Implications: A Summary Appendices Al Data A2 Cointegration analysis of the vector system A3 Exogeneity
189 192 195 202 205 208 208 210 217
7
221
Conclusions
7.1 Introduction
188
221
viii
Contents
7.2 Major Theoretical Contributions of the Research 7.3 Aggregate Investment Behaviour in China: Stylized Facts 7.3.1 System-generated insatiable investment demand exists at all levels 7.3.2 Supply and distributive barriers to investment expansion and retrenchment campaigns 7.4 Inefficiency as a Consequence of Investment Hunger and Bureaucratic Coordination 7.5 The Difficulties and Possible Selections of Reforming the State Investment System 7.6 Limitations of the Research 7.7 Summary
222 225
Notes Bibliography Index
246 260 282
225 229 232 235 241 242
List of Tables and Figures
Tables 1.1 a 1.1b 1.1c 1.2 1.3 1.4 3.1 3.2 3.3 3.4 3.5 3.6
4.1 4.2 4.3
Subsectoral shares of investment in state-owned industry, 1981-95 Subsectoral shares of employment in state-owned industry, 1965-92 Subsectoral shares of output in state-owned industry, 1965-95 Subsectoral shares of output and employment in rural industry, 1985-94 A chronology of investment cycles in China, showing the rough correlation between cycles and political campaigns Position of foreign investment in the state section fixed investment, 1977-96 Approval norms for investment projects in the 1970s and 1980s Sources of finance for fixed investment in the state sector, 1953-95 National priority projects, 1982-96 Main interest rates on state bank loans, 1979-96 Percentage of key materials subject to central allocation, 1965-92 Reported financial losses of industrial SOEs with independent accounting systems, fiscal subsidies to loss-making SOEs, and policy loans of the People's Bank of China, 1985-96 The output value of light industry using agricultural products as raw materials The values and shares of the exports of unprocessed and processed agricultural products in China's total exports The changes of factor proportions in China's agriculture, 1949-96
IX
12 13 14 16 19 24 83 84 88 92 100
103 115 117 119
x
AA 4.5 4.6 4.7 4.8 4.9 4.10 4.11 A4.1 A4.2 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 5.10 6.1 6.2 6.3 6.4 6.5 7.1
List of Tables and Figures
The average annual growth rates of key indicators of agriculture, 1949-96 1954-62: The Great Leap Forward and the agricultural crisis it induced 1962-69: The readjustment after the crisis and construction of inland industrial bases 1969-72: Post-cultural Revolution Advance 1972-77: Initial Attempt at the 'Four Modernizations' 1977-80: Post-Mao import of technology and the subsequent economic readjustment 1980-89: Economic reform and the readjustment induced by stagnation of grain production in 1985-89 1989-96: Three-year retrenchment, 'Deng Whirlwind' and 'soft landing' Basic data on output, labour, and population The source of Tables 4.5-4.11 Primary commercial energy production and consumption, 1953-95 Percentage shares of final energy consumption, 1980-95 Energy intensity by sector, 1980-94 Indexes of energy intensity in relation to physical unit of output in energy intensive industries Plan versus market prices for steam coal (1989) and crude oil (1993) Incremental capacity of power-consuming equipment induced by fixed investment Stock and increment of power-generating and consuming equipment (G We) Average railway traffic density in selected countries, 1979 and 1990 Raw coal reserves and production by region, 1985-2000 Granger causality between real investment and effective energy supply Tests for unit roots Residual misspecification tests of the UVAR model (6.4) The cointegration analysis of model (6.4) Tests for linear restriction on (3 and weak exogeneity The estimation of investment growth rate equation in China Industrial concentration by major industrial sectors in China, USA and Japan, 1985
120 132 135 136 138 13 9 142 146 158 160 165 169 169 170 172 176 176 180 182 186 196 196 199 200 205 233
List of Tables and Figures
xi
Figures 1.1 4.1 4.2 4.3 6.1 6.2
Investment ratio and real growth rate of state sector investment The fluctuations of per capita grain output and real value added per peasant Accumulation ratio and investment ratio The correlations between agricultural fluctuation and investment adjustment The graphic tests for cointegration relation P'Z, One-step residuals, ±2 standard errors, and one-step-ahead Chow tests for the investment growth equation
17 126 129 151 200 203
Preface Business cycle theory has been a major branch of macroeconomics in the west. Industrial capital accumulation and investment cycle are recurring subjects in the literature about socialist economies. Agricultural issues in general and the interaction between agricultural development and industrial capital accumulation in particular has been central to macroeconomic analysis in a developing economy. While both the general investment cycle theory of socialist economies and the distributive barrier-constrained growth theory of developing economies can help reveal certain features of pronounced investment cycles in China, an integration of them with the methodological and technical progresses emerging in the western business cycle theory and econometrics would bring us a new advantage. The integration makes it possible to model both investment level (growth) and change (cycle) without artificially imposed separation between them, thus leading to an econometrically advanced growth-cycle framework. This is the major theoretical intention of this book. I believe that this integration will have a sound academic implication, and will be of interest to scholars who work on transition economics and development economics, and to general readers who are concerned about economic transition and development. In addition to the theoretical intention, this book also documents and stylizes the evolutionary dynamics of China's state investment system, the policy trade-off between industrial expansion and agricultural development, and the persistent transportation and energy constraints to the economy, from the perspectives of both political economy and historical viewpoint. This will be of interest to China specialists and other general readers who are interested in China's development and reform. This book is based on my doctoral dissertation, which was submitted to the Institute of Social Studies (ISS) in the Hague in 1997, and on research continued and extended after that date. I am very grateful to ISS for providing financial and institutional supports to my PhD study. I am also grateful to the International Institute for Applied Systems Analysis (IIASA) and the
xn
Preface
xiii
World Institute for Development Economics Research (UNU/WIDER) for their support to the continued research in this field. This book has greatly benefited from the support and advice of many different persons in different areas. I am very grateful to my three supervisors, Professor Ashwani Saith, Professor Marc Wuyts and Dr Karel Jansen, whose complementary expertise covers political economy, monetary and financial economics, applied econometrics and development theory. Their common professional interest in behaviour analysis clearly interacted in a manner that broadened and deepened my understanding of the issues in investigating, stylizing and modelling aggregate investment behaviour in China. I am indebted to many of my former colleagues in Peking University in Beijing, particularly Professors Justin Yifu Lin, Yining Li, Shali Zhu, Liangkun Chen, Liyan Wang and Qiwen Wang, for their unreserved support and help in my field work. During my field work in China, many scholars, officials, friends and relatives granted me interviews and gave me very useful assistance. I would like to express my sincerest gratitude to them all and regret that they are too many to name. On a more personal note I want to thank Hao Lu, Fumin Mo, Kexiang Sun, An Wu and Jie Zhang. I benefited considerably from the comments by the four anonymous referees and the editors and from discussions with John Bonin, Valpy FitzGerald, Bill Wansing Hung, S0ren Johansen, Michiel Keyzer, Peter Nolan, Haris Psaradakis, Servaas Storm, and Rob Vos. I owe gratitude to the ISS Publications Committee and Publications Office for their encouragement and prompt assistance, particularly to George Irvin, Cris Kay, Linda McPhee, Joy Misa, Gary Debus, Sharmini Bissessar and Iris Qureshi. I thank all of my former colleagues at ISS, particularly Ank v.d. Berg and Dita Walenkamp, for their profound concern and constant support. For her enduring love, support and commitment, I thank my wife, Sylvia Ya Xu. Without her continuous assistance and contributions, this book would not have been completed.
Laixiang Sun
1
Introduction
1.1 Key Questions China has maintained fairly high capital accumulation levels and aggregate investment ratios,1 which contribute to its impressive economic growth. At the same time, state-sector fixed investment, which constitutes the main share of gross fixed investment,2 has shown conspicuous cyclical patterns in its annual growth rate since 1953, when a consistent and reliable accounting system for the state sector came into operation. The relevant amplitudes are impressive in comparison with those of other socialist countries, such as the former Soviet Union and Eastern Europe.3 An interesting question arises: which mechanisms produce such persistent high investment ratios and what forces shape the remarkable cyclical patterns of these investment growth rates? The basic purpose of this study is to reveal the nature of these fundamental mechanisms and determinant forces and the way together they generate chronic investment hunger and shape the pronounced investment cycles. The determination of a 'rational scale of investment' has constantly perplexed Chinese economic authorities and scholars. Many of the problems in the development process, such as shortages of consumer goods, energy and raw materials, fiscal deficit, over-expanding credit and inflation are linked to the required yet impossibly high level of investment. Correspondingly, any type of'adjustment' or 're-adjustment' always starts by reducing the scale of investment. This process is reflected repeatedly in the policy documents of the Chinese government and much of the literature written by Chinese economists. The determinants of the investment cycle embody essential characteristics of the Chinese economy, which reflect both the general features of a centrally planned economy and those of a developing econo-
Chapter 1
my. It is impossible to adequately describe and simulate the Chinese investment fluctuation process by analysing the general characteristics of the centrally planned economy only or by analysing those of a developing economy on their own. In other words, investment cycles in China synthesize the characteristics of both without being reducible to either. The former characteristics shape our understanding of the institutional pressure and incentives that drive the mechanism of maximizing growth through investment. The latter draw our attention to the interaction between industrialization and supply constraints in general, and the agrarian question in particular, which are central to the dynamics of a typical developing economy (cf. e.g. Patnaik 1995, Rakshit 1989). Analysing and identifying these characteristics from both the perspectives of political economy and of history also form an important subject of this research. According to the general theory of socialist economies, the system constantly initiates pressures and incentives to maximize investment and cannot generate internally self-imposed restraint to resist expansionary drives (Kornai 1992: Chapter 9). As a result, the realized real investment level has, on the one hand, taken quite a high proportion of the national income, and on the other hand, been constrained by the supply limits of bottleneck sectors. In the case of China, investment is limited by the tolerable adverse distributive impact it has on food supply. Against the background of significant policy change, a more interesting question is: has reform fundamentally altered the investment coordination mechanism? Initial data analysis shows a surprising lack of structural change in state-owned industry prior to 1993, in terms of the subsectoral shares of output, employment and investment when compared with townshipand village-run industry. This structural rigidity may imply that statesector investment was not yet based mainly on market criteria and that the state investment system did not respond actively to the changes in demand occurring in a rapidly growing and transforming economy until 1993. This finding shows the need for more accurate econometric modelling and testing based on recently developed cointegration and error correction approaches. There has been a sustained effort to set up formal models of aggregate investment behaviour and investment cycles in socialist and/or developing socialist economies. Two research strands have been influential. The first is distinguished by Kalecki's 'bottleneck constraint
Introduction
type of growth theory'. Kalecki's model focuses on a stable growth path, which, in a developing socialist economy, typically represents a trade-off balance between the desired high rates of investment growth and the toleration limits to the adverse distributive consequences that a high investment rate puts on agriculture. Understanding this trade-off balance and its functioning mechanism is important in the case of a typical developing economy like China. However, this strand does not pay much attention to demand analysis and cycles. The second one is known as the 'Hungarian School'. This strand bases investment demand and fluctuation analysis on notions of soft budget constraint, expansion drive and investment hunger. These notions are particularly helpful in analysing China's case. In terms of modelling, this school focuses on short-run disequilibrium adjustment behaviour of the planners, which is modelled in a response function of the representative planner. The model does not explain the co-movement between real investment level and its constraints. In addition, the approach based on a representative planner faces the problem of intertemporal inconsistency of the planner's rationality (Mihalyi 1992). The attempt to advance both these approaches and to integrate them into a new growth cycle model is of theoretical interest and importance.
1.2 Growth Cycle Model As mentioned above, Kalecki's growth theory emphasizes an equilibrium steady growth path, rather than demand analysis and cycle. Following Kalecki's theory, the starting point is the constant growth pressure produced by a socialist system. Such pressure determines in turn a desired investment growth rate and a corresponding savings rate. Thanks to the effective wage and price control by the socialist state, the desired savings rate can be insured by the state distributive policy on national income. It is this distribution which causes investment to create its corresponding saving. Thus, the main barriers to growth reside in bottleneck constraints such as shortage of agricultural products, energy and raw materials, problems in balancing foreign trade, and the tolerable limit to the adverse distributive consequences of a rising investment rate on agriculture and urban real wage level. Kalecki's analysis provides insight into the relation between the micro phenomenon of the soft budget constraint faced by state sector and the macro context of the investment/saving relation. Its first prem-
4
Chapter 1
ise is that 'investment finances itself from the perspective of income distribution (Kalecki 1976: 43). According to Kalecki's convenient simplification, workers spend what they earn and capitalists earn what they spend. In the context of China, it is the consolidated state sector that earns what it spends. In a capitalist economy, the expansion of individual capitalist investment can be disastrous if corresponding demand is not forthcoming. Reckless investment spending will lead to bankruptcy, or no earning at all, although this spending still contributes to the income of the capitalist class as a whole. This type of restraint on investment puts a hard budget constraint on individual capitalists. However, the soft budget constraint in a socialist economy takes away the capitalist form of individual sanction on investment decision-making because the volume of investment will generate the necessary income to cross-subsidize investment activities. As a result, the real limit to investment expansion is, besides bottleneck constraints, the distributive barrier of a rising investment rate to the food balance. This is reflected in the form of a shortage barrier during the pre-reform period and as both shortage and inflationary barriers in the post-reform years. The Hungarian School's way of dealing with investment cycles in socialist economies focuses on modelling planners' response functions to one or several key shortage signals.4 To advance this most influential approach, we need to first incorporate the diverse actors into cycle analysis and to introduce the behaviour assumption of their intertemporal rationality. This is because in China, investment decisions are taken at different levels of the central, provincial, municipal, prefectural and county authorities, and in enterprises, partly interactively, partly independently. Different brakes operate at different levels in a complex interactive fashion. The decision-makers at each level have their own specific interests and incentives. These indicate that there may be different response mechanisms operating interactively at the different levels within the system. Moreover, these response mechanisms may not be invariant over time, but instead change significantly with the recurring cycles of decentralization and re-centralization, particularly in the reform period. There is no evidence that there is a specific response mechanism that constitutes the constant characteristics of China's investment behaviour. Hence, it is necessary to step back from the immediate mechanisms so as to be able to observe the deeper constancy prevailing in the determination of the investment cycles.
Introduction
In this connection, the cointegration approach can serve the required advancement. As a statistical expression of a long-run equilibrium relationship between two or more non-stationary time series, cointegration can help to establish cycle analysis based on the behavioural assumption of intertemporal rationality. In other words, based on a cointegration approach, we can assume that the investment decisionmakers at different levels and in different sectors are all rational economic agents and that, because of their own specific interests and incentives, they cannot be considered as representative planners. As a consequence, coordination mechanisms that prevail among economic agents are of decisive importance and cannot be ignored. Full information and rationality of individual choice are not sufficient to preclude the business cycle in a well-functioning market economy. The bureaucratic coordination mechanism that characterizes the state investment system of China has also failed to prevent an investment cycle. Secondly, modelling planners' response functions inevitably involves the definition of a norm (cf. among others, Kornai 1982, Simonovits 1992) which itself is not fully explained (Hare 1982, 1989). The notion of norm is intended to represent determinants behind what could be called planners' perceptions of what is feasible. In this direction, the cointegration approach can support the advancement as well. The cointegration relation presents a long-run equilibrium co-movement between the real investment level, supply and distributive barriers to investment expansion. It shows clearly the long-run equilibrium determined by the fundamental tension between system-generated investment ambition and the supply and distributive barriers to that ambition. Therefore, it can replace the traditional univariate norm concept, which in practice, takes the form of a univariate moving average and may produce spurious deviations as pointed out by Slutsky (1927) and Frisch (1928) (see Morgan 1990). In addition, such a co-movement path also serves as an attractor for the disequilibrium adjustment of the investment decision-makers toward a dynamic equilibrium as shown by the error-correction model. Thirdly, a single response function cannot be used to model interaction between shortage and investment tension. This interaction works in two ways. First, the above-normal shortage intensity makes the planner restrain new investment starts and conversely, if the difficulties caused by shortage have diminished, this will stimulate and support further investment expansion. Second, shortage generates in-
6
Chapter 1
vestment tension and at the same time investment tension is one of the major causes of shortage (Kornai 1980: 201). This two-way interaction is present in both the short and long run, although in the long run the positive correlation may be dominant. In the long run, a strong investment expansion initially increases shortage and inflationary pressures, but it eventually leads to an increase in capacity and potential output, which tends to dampen shortage and inflationary pressures and allows the economy to expand investment again. Here, the demand for modelling such short-run two-way interaction and long-run complementarity between investment expansion and bottleneck constraints can be fully met by applying cointegration and error-correction modelling. The long-run equilibrium co-movement can well represent this long-run complementarity and the short-run error-correction equation(s) highlights the different interaction mechanisms in an effective and intuitive way. Following the above analysis, we may conclude that the strong points of Kalecki's growth theory and the Hungarian School's cycle theory can be combined into a new framework of growth cycle theory. In this new framework, the modelling of investment cycle is based on an understanding of what determines the equilibrium level around which the investment fluctuates. The behavioural logic behind the new framework can be seen in a standard way: almost all investment decision-makers, at all levels, have tried to maximize the investment scale within the supply constraints and the inflationary barrier under their jurisdiction. This type of maximization behaviour, subject to certain constraints, can be modelled and tested by employing recently developed econometric approaches such as cointegration and error-correction. In modelling both growth and cycle, the most desirable advantage of the novel econometric approaches is that they allow researchers to single out the underlying equilibrium relation (attractor) which characterizes the supply constraints on investment demand level while investigating its dynamic structure (change) through error-correction. Thus one can work directly with readily observed data rather than having to guess at the shortage indicators that planners at various disjointed levels may use. Adoption of the new framework of growth cycle is also supported by data analysis. By comparing the input-output relations in China's agriculture with several important distributive relations in the uses of the national income, there are evident two-way causalities to be found
Introduction
between the cycles of the capital accumulation ratio (investment ratio also) and the fluctuations of agricultural value-added per labourer, and especially of grain output per capita. This result is clearly different from those obtained from common linear regressions based on some forms of constructed shortage indicators. This type of learning from data has led to employing the recently developed modelling strategy and approaches, and to modelling the economic agents' investment maximization behaviour subject to the representative bottleneck constraints. The behaviour logic of the growth cycle theory can be intuitively interpreted as follows. Investment hunger, which is initiated by expansion drive and becomes feasible under the soft budget constraint (Kornai 1980, 1992), is the driving force which causes investment behaviour to be constrained by shortage and inflationary barriers in the bottleneck sectors rather than by demand. Investment hunger is not about fine-tuning the level of investment; rather it relentlessly pushes the economy towards overheating, thereby producing tensions which lead to subsequent error corrections. The retrenchments characterized by ad hoc administrative measures can be likened to 'pushing a basketball under water': just as this mechanism rarely seems to involve letting some of the air escape, investment hunger is suppressed but still present; the refused proposals are not swept away but only postponed. On the supply side, food as the typical necessary consumer good and energy as the representative basic producer good together set the limits on the feasible rate of investment growth. Foreign exchange does not form an independent constraint since the food and energy constraints together encompass the dynamics of foreign exchange earnings.5 Together they indicate that probably only by combining what drives the system (investment hunger) with the supply constraints encountered (represented by agriculture and energy) can we come to grips with the nature of investment growth rate cycles in China. The cycles cannot be understood independently from the forces that determine the level of investment. While investment hunger is a constant feature of the economic process, planners' responses to tensions are endogenous and have changed over time depending on the specific economic mechanisms through which they are effected. These specific mechanisms may affect the features of the cycles, but not what causes them to happen in the first place. This is the reason why the cycles may be understood and
8
Chapter 1
modelled in relation to the fundamental tensions between investment hunger and the bottleneck constraints. In addition to the general logic presented in the previous paragraph, two points are worth noting. Firstly, there are different conducting channels from investment demand to the bottlenecks between the prereform and reform periods. In the pre-reform period, over-investment resulted in supply shortages of energy, raw materials and agricultural products through material conduction. During the reform period, in contrast, over-investment usually induces over-expansion of credits first, followed by shortage in the planned component and inflation in the market component of the economy through both material and value conduction. In other words, in the transitional period, looser (or more loosely enforced) price controls on transactions, which take the forms of an administratively negotiated market (white market) and a free market, provide a mechanism for a partial translation of cumulated shortage into higher prices. Secondly, in a bureaucratically coordinated economy, though the increase of price levels may be used as a reason to reduce investment scale, it can also be cited, possibly more significantly and frequently, as a good reason to request extra investment, at least in the short run (Kornai 1992). The research below suggests that with regards to shaping real investment cycles, the effect of the inflation rate as a special short-run shortage signal seems to be outweighed by its ability to offer bargaining power for requesting additional investment (for details, see section 6.5).
1.3 Background China's development model and experience, particularly its economic reforms and the resultant rapid economic growth and structural changes, have been of major interest to economists. At a practical level, its development has a significant impact on the well-being of over one fifth of the world's people, on international trade, on the evolution of the Asian-Pacific economy and on the international balance of economic and political power. On a theoretical level, a better understanding of its development and transition would provide useful lessons for other transitional and/or developing countries. It would also contribute insights into the processes involved in the transformation of economic systems, the relationship between economic development and institutional changes, the interaction between industrialization and the agrar-
Introduction
ian question and the nature of 'Asian' models of development. Moreover, the fact that China's impressive development and reform occurred despite many sharp policy oscillations, economic fluctuations, internal political conflicts and a lack of coherent reform strategy has compounded the challenges for economic analysis (Chow 1997, Lin et al. 1995 & 1996, Watson 1994, Zhang & Yi 1995). For the pre-reform period, as identified by World Bank (1983),6 China's development strategy comprises two basic but often conflicting objectives. The first is poverty reduction or securing basic needs. This has been realized in the initial stages through land reform and collectivization in agriculture and nationalization in the non-agricultural sectors, and later through a programme of comprehensive rural development and provision of basic social services utilizing local resources and initiatives. The second objective is industrialization, particularly the development of a heavy industrial basis, mainly based on a massive infusion of centrally mobilized resources with less concern about cost effectiveness, and relying on technology largely descended from Soviet designs of the 1950s. Persistent tension between these two objectives has contributed to sharp policy oscillations. The dominant policy trend, however, has been in favour of industrialization. During this period the overall development of China was impressive. With adjustments for international comparability, GNP per capital appears to have grown at 2.0-2.5 per cent per annum in the 1957-77 period and 2.5-3.0 per cent per annum in 1957-79. The former rate is significantly above the average of 1.6 per cent for other low-income countries, though the latter is well below the average of 3.7 per cent for middle-income developing countries. China's most remarkable achievement is regarded as making its low-income groups far better off in terms of basic needs compared with their counterparts in most other poor countries. This was accomplished despite a relatively high population growth rate (2.0 per cent) and an unprecedented degree of international isolation, which means that development has been almost entirely self-financed. At the same time, China's impressive economic growth has been characterized by profound structural unbalances and gross inefficiencies. Driven by persistent industrialization bias, the net output of industry grows at an annual average of 10.2 per cent (1957— 79) in real terms, far above the average for other low-income countries (5.4 per cent) and well above the average for middle-income developing countries (7.5 per cent). Industry accounts for about 40 per cent of
10
Chapter 1
GDP, similar to the average for middle-income developing countries, although China's per capita GDP (US$279 in 1980) remains one of the lowest in the world. Its rapid industrialization provides China with a very comprehensive and basically self-sufficient industrial system. On the other hand, the cost of rapid industrialization has been very high. The two most important issues are as follows. First, apart from the agricultural crisis and the consequent great famine (1959-61) induced by the Great Leap Forward,7 the inadequate agricultural output growth contrasts sharply with the rapid industrial growth. Agricultural gross output has increased by only 2.1 per cent (1957-77) despite significant progress in developing sources of intensive growth (e.g. multiple cropping, irrigation, flood control, new high-yielding seed varieties and chemical fertilizer application). Growth of foodgrain output in particular has been slower than that of gross agricultural output. In 1980, after a remarkable increase in output following the agricultural reform introduced in late 1978, per capita grain output is only 7 per cent higher than in 1957. Rural per capita incomes have hardly risen, with net output per capita dropping by 12 per cent between 1957 and 1977. Secondly, from 1957 to 1979, industrial labour productivity grew slowly while capital productivity declined, and the total factor productivity in industry either stagnated or declined. These facts may indicate that industrial expansion in the pre-reform period was achieved mainly by increasing the quantity of factor inputs (i.e. extensive growth) and not by increased efficiency (intensive growth). Another startling fact is that about 70 per cent of total commercial energy use is accounted for by industry, which on a per capita basis is nearly four times the average for other low-income countries. Energy consumption of per dollar of GDP is also about 2.5 times the average for other developing countries or for industrialized market economies, and about 1.5 times the average for other centrally planned economies. The stylized facts of the pre-reform period can be spelt out from the literature. China's economy achieves a rapid industrial growth, but this is impaired by sharp fluctuations in growth rates, extraordinarily high and oscillating investment rates, declining total factor productivity and inadequate agricultural output growth. There has been a remarkable stagnation in per capita levels of consumption of key items (e.g. grain, vegetable oils, cotton cloth) and in rural living standards over a twodecade period (1957-78) (see, among others, Ishikawa 1983, Lin 1988, Lin et al. 1996, World Bank 1983, Yeh 1984). This necessitates a re-
Introduction
11
examination of the fundamental capital accumulation mechanism and the process which generates the impressive industrial growth, an analysis of the incentive sources of the industrialization drive and relevant supply constraints, and modelling the co-movement between real investment and the supply frontier of representative bottleneck sectors and investment cycles. This will provide a deeper understanding of China's development model and experience. An analysis based on historical and institutional perspectives would also be desirable. Since the economic reforms and open-door policies began in the late 1970s, China's economic performance has been much better than in the pre-reform period. The average annual growth rate of GDP from 1978 to 1997 was 9.8 per cent. This is arguably the fastest in the world and rivals the record achieved by the four Small Dragons - the creators of the East-Asian Miracle in their fast-growing period. On a per capita basis, the annual growth rate of GDP is more than double that of the prereform period {People's Daily, 25 September 1998). While the industrial net output has grown at a rate of above 12 per cent, there have been important changes in the efficiency of the resource use. In stateowned industry there has been an evident reversal of the long-term decline in total factor productivity. The typical Stalinist relationship between the growth rates of heavy and light industry has been reversed, with explosive growth of the rural light industry in township and village enterprises (TVEs). The agricultural growth rate accelerated far ahead of that achieved during the 1957-78 period, and with much more efficiency in resource use. The growth of commerce, transport and communication is also much more impressive than in the pre-reform years. In addition, China remains relatively unburdened by foreign debt and has achieved fast growth with relatively low inflation (see, e.g. Chang & Nolan 1996, Jefferson & Rawski 1994, Lin et al. 1995, 1996, McKinnon 1994, World Bank 1994). Such phenomenal growth, however, has also been impaired by the recurring grain problems, the increasingly sharp energy shortage and crises, and in particular by the pronounced cycles of reform and growth (cf. e.g. Ash 1992, Kambara 1992, Lin et al. 1996, Oppers 1997, Sicular 1989 & 1993, Zhou & Chu 1992). A typical reform and growth cycle is usually stylized as follows (cf. Lin et al. 1996, Oppers 1997, Watson 1994, World Bank 1994, Yusuf 1994, among others). The scenario commences with a series of reform experiments and relaxation of control. As their scope is gradually
12
Chapter 1
broadened, investment and growth accelerate and are accommodated by credit expansion. Soon the economy is expanding by double-digit rates. The decentralization of administrative and economic authority decreases the power of the central government to enforce macroeconomic control. Inflation follows the overheating, and the increasing rent-seeking by those with the opportunity to exploit the institutional rents between the plan and market components of the economy increases corruption. Once the fear of inflation and corruption is widely acknowledged, the government is empowered to contain reform initiatives and to curb growth by severe austerity measures. There is still a lack of the sophisticated mechanisms required to direct an increasingly complicated economy. After a period of 'restoring order and rectifying the economic environment', the call for a resumption of reforms and faster growth becomes irresistible. This is because the benefits of reform and faster growth have been widely recognized, and also a large number of state enterprises cannot exist without subsidies and face real difficulties in circumstances of deflation. Thus new reform experiments are initiated, policies are relaxed and a new cycle begins.
Table 1.1a
Subsectoral shares of investment in state-owned industry, 1981-95 Shares of Investment
Metallurgy Power Coal & coke Petroleum Chemical Machinery Building materials Food Forest Textile & clothing Paper & cultural articles Others Total
1981
1984
1987
1989
1992
1995
11.4 12.5 10.4 14.2 8.8 11.8 4.5 5.3 2.2 10.3 1.6 1.2
12.3 13.5 13.0 16.0 9.7 11.6 5.7 5.0 1.8 7.1 1.4 1.3 100.0
12.7 16.7 8.2 13.7 11.3 12.5 5.5 7.4 1.4 7.3 1.9 1.5
12.0 18.5 8.9 16.8 11.9 11.0 3.9 5.5 1.1 7.3 1.9 1.7
11.4 19.4 8.3 14.4 11.3 13.6 4.5 5.6 0.7 6.4 2.4 2.0
14.1 23.0 7.1 14.6 14.1 9.4 4.1 5.2 0.7 3.5 1.5 2.5
100.0
100.0
100.0
100.0
100.0
Introduction Table 1.1b
13
Subsectoral shares of employment in state-owned industry, 1965-92 Shares of Employment
Metallurgy Power Coal & coke Petroleum Chemical Machinery Building materials Food Forest Textile & clothing Paper & cultural articles Others Total
1965
1978
1985
1990
1992
8.6 2.5 12.4 1.4 6.3 26.0 5.3 8.3 5.8 12.3a 1.3b n.a.
10.2 2.5 12.4 1.6 9.5 30.9 6.0 6.7 3.9 8.9a 1.3b n.a.
9.3 3.0 11.7 2.1 9.4 27.8 6.6 8.1 3.6 11. T 1.5b 1.3
9.3 3.6 11.4 2.7 10.2 26.1 6.2 8.1 3.2 12.3a 1.7b 1.5
9.2 3.8 11.3 2.9 10.4 25.7 6.4 8.4 3.0 11.9a 1.7b 1.5
100.0
100.0
100.0
100.0
100.0
Compared with the conspicuous reform and growth cycle, the structural adjustment of state-owned industry seems to be relatively insignificant, although the state sector monopolies represent 60-70 per cent of total fixed investment and national industry underwent a profound structural change between the 1970s and the early 1990s. Tables 1.1 a-c and 1.2 present the subsectoral shares of output, employment and investment for the state-owned industry, and output and employment for township and village-run industry (TVI). Different patterns of structural change between state industry and TVI are evident in these two tables, particularly, during the period before 1992 (including 1992). Within seven years (1985-92) - quite a short period - the TVI sector underwent a significant structural adjustment. The output shares of textiles and clothing, papermaking and cultural articles, chemicals, metallurgy and, especially, machinery increased by 3-7 percentage points; correspondingly, the output shares of coal and building materials decrease remarkably. The same applies to the employment structure, with a profound change in shares. However, for a fairly long period before 1992, the subsectoral structure of state-owned industry had remained stable. For example, machinery led with 27.7 per cent of gross output value of state-owned industry in 1975 and remained the lead sector
Chapter 1
14
with 26.6 per cent of gross output in 1992, 17 years later. The output shares of textiles, chemicals, food and metallurgy, the next four most important subsectors, have also changed little, since 1975.
Table 1.1c
Subsectoral shares of output in state-owned industry, 1965-95 Shares of Output Value
Metallurgy Power Coal & coke Petroleum Chemical Machinery Building materials Food Forest Textile & clothing Paper & cultural articles Others Total Notes:
1965
1975
1980
1985
1990
1992
1995
10.7 3.1 2.6 3.2 12.9 22.3 2.8 2.9 12.6 19.2 3.9 3.8
9.0 3.9 2.8 5.6 11.3 27.7 3.1 1.9 12.0 15.4 3.2 4.1
8.6 3.8 2.3 5.1 12.5 25.5 3.6 1.7 11.4 18.4 3.5 3.4
8.0 3.3 2.5 4.5 11.2 27.1 4.2 1.6 11.5 18.7 3.9 3.3
7.2 3.0 2.1 3.8 12.7 28.9 4.1 1.3 10.5 18.0 3.7 3.8
10.7 3.7 2.5 5.5 12.4 26.6 5.6 1.4 10.6 15.4 4.3 n.a.
13.6 7.3 3.7 12.2 11.7 20.0 4.2 1.0 13.7 8.8 2.3 n.a.
100.0
100.0
100.0
100.0
100.0
100.0
100.0
(a) Textile only, (b) Papermaking only. For translation between 'new1 and 'old' industrial classifications, see Statistics on Labour and Wages of China, 1949-1985 (1987: 275-82). Some rounding errors are larger than normal because those shares in some small subsectors are missing, as indicated by (a), (b) and n.a.
Sources: For investment: The data for 1981-91 are from Statistics on Fixed Investment of China (1987: 23 for 1981-85; 1989: 31-32 for 1986 & 1987; 1991: 30-31 for 1988 & 1989; and 1993: 39-40 for 1990 & 1991), and for 1992-95 are taken from China Statistical Yearbook on Investment in Fixed Assets, 1950-1995 (1997: 54-55). For employment: The data for 1965 & 1978 are taken from Statistics on Labour and Wages of China, 1949-1985 (1987: 37), and others are from Statistical Yearbook of China (1991: 392; 1993: 410). For output: The figures for 1965-90 are taken from Almanac of China Industry (1991: 969), the data for 1965 are in 1957 constant prices, for 1975 and 1980 are in 1970 constant prices, and for 1985 and 1990 are in 1980 constant prices. The numbers for 1992 and 1995 are inferred from Statistical Yearbook of China (1993: 417; 1996: 418), and in current prices. Columns may not tally due to rounding.
Introduction
15
Other data on changes in subsectoral shares of employment and investment in state-owned industry reveal a similar stability over the period of 1978-92. Taking into consideration that the proportion of state industry dropped from 78.5 per cent in 1978 to 48.1 per cent in 1992 {Statistical Yearbook of China 1993: 414) and that the economic reforms had been carried out for one-and-a-half decades, the lack of change in the structure of state-owned industry during 1978-92 is surprising. This remarkable structural rigidity may imply that before the most radical reform in the state sector was initiated in mid-1992, the state sector investment was not yet based mainly on market criteria, and that the state investment system did not actively respond to changes in demand occurring in a rapidly growing and transforming economy. In other words, the investment approval process in the state sector might still have been dominated by bureaucratic negotiation and coordination based on vested sectoral interests and structural inertia, although the concrete mechanisms had changed. This rigidity, combined with the similarity between the reform/growth cycle in the reform years and the policy/growth cycle in the pre-reform period may justify the persistence of the fundamental tension between investment hunger and the supply possibility for both pre- and post-reform periods. Although there have been numerous changes in miscellaneous policy details and concrete mechanism. The absolute increase of supply possibility in key bottleneck sectors will certainly induce higher rates of economic growth, but the tension between growth drive and supply possibility may remain. This fundamental tension may determine the co-movement between the real investment level and the supply frontier of the representative bottleneck sectors and shape the investment cycles.
1.4 Cycles or Fluctuation? Economists define 'cycle' using the triplex of recurrence, reinforcement and regularity. Recurrence means that the phases of expansion and contraction of certain time series data follow each other. Reinforcement and regularity indicate that each phase 'produces the conditions which usher in the next phase of the cycle' and 'successive cycles ought to resemble each other' (Ickes 1986: 43, Simonovits 1991: 466).
16 Table 1.2
Chapter 1 Subsectoral shares of output and employment in rural industry, 1985-94 1985
1989
1992
1994
2.63 0.30 3.90 n.a. 7.44 20.00 18.64 2.67 7.80 17.02 4.12 14.76
5.14 0.31 3.05 0.23 10.25 23.93 17.43 2.46 8.45 18.50 6.20 4.05
6.34 0.27 2.24 0.27 11.11 27.64 14.49 2.30 7.75 19.99 6.36 1.24
100.00
100.00
100.00
8.23 2.33 1.66 1.08 9.52 24.02 13.82 3.42 8.93 18.86 3.87 n.a. 100.00
2.18 0.45 5.45 n.a. 5.68 14.62 30.64 3.08 7.81 13.81 5.21 10.72
2.91 0.44 5.10 0.09 7.06 18.56 28.86 2.84 6.97 14.60 8.56 4.00
3.03 0.41 4.89 0.11 7.96 19.67 24.41 2.73 6.77 17.07 8.65 4.29
6.87 3.15 4.14 1.85 6.28 19.42 19.57 5.44 6.99 17.36 4.80 n.a.
100.00
100.00
100.00
100.00
Shares of Output Value Metallurgy Power Coal & coke Petroleum Chemical Machinery Building materials Forest Food, beverages & tobacco Textile, clothing & leather Papermaking & cultural articles Others Total Shares of Employment Metallurgy Power Coal & coke Petroleum Chemical Machinery Building materials Forest Food, beverages & tobacco Textile, clothing & leather Papermaking & cultural articles Others Total Note:
For translation between 'new' and 'old* industrial classifications, see Statistics on Labour and Wages of China, 1949-1985 (1987: 275-82). Sources: Statistical Yearbook of China, (English version) 1987: 226-7; 1991:419-20; 1993: 441-2; and 1995: 399.
But the periodicity, or replication, should not be explained in terms of abstract mathematics. Instead, it should be understood from the perspective of the behaviour of investment decision-makers. In other words, one should attempt to understand why investment decisionmakers and, especially, planners in a socialist economy, replicate their
Introduction
17
expansion and retrenchment behaviours, and what forces drive them to do so repeatedly. Figure 1.1 provides a clear picture of investment fluctuations in China by employing both time series of investment ratio (cf. note 1) and real investment growth rate. In Figure 1.1 the investment ratio series presents the cycle of investment levels in terms of national income distribution, and the real investment growth rate exhibits its own cycle. It is clear that the cycles show recurrence, reinforcement and regularity; the relevant amplitudes are also very impressive. What should be emphasized here is that the recurrence, reinforcement, and regularity of planners' expansion and entrenchment behaviours can be clearly spelt out as follows.
Figure 1.1
Investment ratio and real growth rate of state sector investment
Investment ratio
-- 80 - 60
- Growth rate of real Investment
40 -- 20
g E 1/5 >
_20
--40 --60 --80
*&
2
2
o
-100 ( 0 ( D O ( M * ( D 0 0 O C M V 1 0 I O ( 0 ( 0 ( 0 ( 0 ( O N N . K 0 ) 0 ) 0 ) 0 ) 0 ) 0 ) 0 ) 0 ) 0 ) 0 )
Note:
The traditional National Income statistics has been abandoned since 1994.
Source:
Investment ratio series are taken from Table A4.2; real investment growth rate is derived from Table A6.1.
18
Chapter 1
Once we measure the cyclical pattern of investment ratio from trough to trough, it can be seen that there is a rough correlation between investment fluctuations and political/economic campaigns, as shown in Table 1.3. The political economy logic behind this preliminary correlation is as follows. Expansion in each cycle begins with the economy in good condition and is accelerated by the corresponding political/economic campaign. Once danger signals appear the expansion is interrupted, ameliorative measures are initiated, and a retrenchment campaign, which is usually called 'readjustment' or 'rectification' follows. The political/economic campaigns are essentially endogenous within the system (Kornai 1992), which will be analysed further in Chapters 2 and 3 and in section 4.6. In order to identify more precisely the behavioural characteristics of investment decision-makers in each phase of each cycle, it is helpful to employ Bauer's (1978) four-phase description. According to this description, each investment cycle begins with a 'run up' phase, in which a new round of investment expansion is initiated and driven by growth pressure and institutional incentives. In the accelerating or 'rush' phase investment expansion is increased by the high outlay commitments made in the 'run up' phase as well as by the political/economic campaigns. The upper turning point in each cycle, the 'halt', is accompanied by the sharpening of shortages and other danger signals. Finally, the retrenchment phase, or 'slowdown', is usually brought about by administrative coercion in retrenchment campaigns. A chronology for each investment cycle and its corresponding phases is presented in Table 1.3. The relevant detailed description and analysis will be presented in sections 2.2 and 4.6. The behavioural characteristics of planners and the demand-supply forces behind planners' activities, as one of central concerns of the research, will be examined step by step in the following chapters. This introductory section aims to show that China's investment ratio and real investment growth rate are indeed characterized by a pronounced cyclical pattern, with a periodicity of about five years. In mathematical terms, this cyclical pattern is not very regular. However, in the sense of economic behaviour, the pattern represents a typical 'business cycle'.
The 'Socialist High Tide' The Great Leap Forward Construction of inland industrial bases Advance after Cultural Revolution chaos Initial attempt at the 'Four Modernizations' The'Foreign Leap Forward'and 3-year readjustment Rural and urban economic reforms 3-year retrenchment and'Deng Whirlwind'
1 2 3 4
1955-57 1957-62 1962-68 1968-73 1973-77 1977-81 1981-89 1989-96
Notes:
Initial expansion Accelerating expansion Upper turning point Retrenchment
Political/economic campaign 1955 1957-58 1962-63 1968-69 1973 1977 1981-82 1989-90
Run up
1
Phases
1956 1958-59 1964-65 1969-70 1974-75 1978 1983-86 1991-92
Rush
2
1956 1960 1966 1971 1975 1978 1987 1993
Halt3
1957 1961-62 1967-68 1972-73 1976-77 1979-81 1988-89 1994-96
Slowdown4
A chronology of investment cycles in China, showing the rough correlation between cycles and political campaigns
Period
Table 1.3
8
20
Chapter 1
1.5 Methodological Issues8 The growth cycle framework established in this book is rooted in the application of the recently developed methodological framework as summarised in Spanos (1990, 1995). It has been symbolically labelled the LSE (London School of Economics) methodology in the literature though many eminent contributors are not directly linked with the LSE. Within the LSE framework, both theoretical analysis and the probabilistic structure of observed (nonexperimental) data have a common denominator in the form of a data-generating process (DGP) which demarcates the framework of analysis. Theoretical analysis proceeds not for the sake of theorizing but in order to understand certain observable economic phenomena. A reliable summary of the probabilistic information in the data enable us to relate the empirical econometric model to the actual mechanism underlying the observable phenomena of interest and not just to the existing theory. Grounded in this framework, the aim of econometric modelling is no longer the quantification of the existing theoretical model, irrespective of the observed data chosen. The probabilistic structural model of the data plays an important role in interpreting the actual mechanisms that give rise to the observed data, while theoretical analysis serves to clarify the behaviour rules of the economic agents, and suggests relevant observable variables and theoretical parameterization of interest. The reformulation of the wellspecified estimated statistical model in view of the relevant theoretical analysis generates the economically well-interpreted econometric model. Based on this framework, political economy analysis is used to reveal the institutional pressure and incentives confronting the decisionmakers at each level of the central and local governments and the enterprises for maximizing investment growth. It also shows why there is a lack of internal, self-imposed restraint capable of resisting the investment drive, and how this has allowed real investment to move along the supply possibilities frontier of the economy. The Sims's type of unrestricted vector autoregressive (UVAR) representation is then employed to parameterize the joint dynamic structure of the basic data system suggested by the political economy analysis and some primary data exploration. Following the 'probabilistic reduction' (Spanos 1995) type of 'general to specific' procedure and based on cointegration and error correction approaches, we finally obtain a well-specified, well-
Introduction
21
fitted, and economically meaningful long-run equilibrium investment level equation and short-run investment growth rate adjustment equation. Just as the methodological framework predicts, these two equations are adjustment equations rather than traditional demand/supplytype functions. More concretely, based on this framework and relevant techniques we can seek direct empirical proof for the existence of equilibrium relationships without having to tackle such difficult problems as how to set up separate demand and supply functions first. The shortrun adjustment behaviour can then be modelled as an error correction process with the equilibrium relationship as an attractor. The major advantage of modelling the adjustment equations is that it enables us to avoid tackling certain thorny issues around 'bridging' the gap between theory and data which arise when the usual equilibrium assumption is prior-imposed. In fact, such an ex ante assumption implies the imposition of a behaviour rule for at least one of the original 'control' (exogenous) variables in the corresponding demand and supply functions (e.g. price in the common demand and supply schedules), suggesting that the situation envisaged by theory differs greatly from the mechanism underlying the DGP.9 By contrast, based on the LSE framework, the political economy analysis for state investment system, primary data exploration and cointegration modelling can be consistently synthesized within the process of specifying the dynamic adjustment functions without imposing any ex ante behavioural assumption. The expression 'political economy analysis' is used here to mean that besides discussing economic issues about investment decisionmaking, financing, implementation and so on, this book will also examine many other problems beyond the boundary of economics. These include the orientation, interests and decisions of China's top and regional authorities, the central ministries and local economic bureaucracies; conflicts of interest between the centre and the locality, and between different bureaucracies; the correlation between the recurring cycles of decentralizing and recentralizing and the investment cycles led by the local and/or the central; the collusive behaviour between local governments and local enterprises, between enterprise and its immediate supervisory agency, and among enterprises, local governments and the local branches of state banks. As a result, the analysis extends into the fields of political science, history, sociology, and social psy-
22
Chapter 1
chology. This extension is what the term 'political economy' is intended to encompass. The choice of a concrete modelling strategy is taken based on the following two observations. First, as mentioned in section 1.2, in the literature on modelling planners' reaction function to shortage, scholars formerly employed univariate moving averages to represent the normal growth path (norm) which served as an attractor similar to the equilibrium path in Walras' economics (Kornai 1982). The econometric problems linked with this approach are: (a) univariate moving averages may produce spurious cycles and (b) the normal path should be a structural relationship when one deals with investment control function and should be determined by the interaction between relevant demand and supply forces rather than by such univariate techniques as moving averages. In order to overcome these two problems, it is necessary to introduce the concept of an equilibrium relationship or a structural equilibrium so that we can talk about equilibrium against the background of supply shortage (Banerjee et al. 1993).10 Secondly, to regress one nonstationary process on another may result in a spurious (even nonsensical) regression, as discussed in detail by Yule (1926), Granger & Newbold (1974), Granger (1981), Nelson & Kang (1981), and Nelson & Plosser (1982), among others. The idea behind this discussion is that if no bounded combination of the levels of several series exists, then the error term in the regression must be non-stationary under the null hypothesis so that known distributive results do not apply. To obtain an economically sound normal path (norm) or long-term equilibrium relationship between real investment and the representative bottleneck supplies, it helps to employ the recently developed cointegration approaches. This approach enables us to detect and test long-term stable relationships among non-stationary variables based on the interaction between theoretical exploration and statistical testing. The cointegration analysis based on Johansen's (1988, 1992a, 1992b, 1992c) procedure can also help to check weak exogeneity of relevant variables. Once this type of long-term norm relation is found, it is possible to model economic agents' dynamic adjustment function toward the longterm steady-state target in the form of an error correction mechanism. A practical modelling strategy based on dynamic cointegration is recently proposed in Hendry & Mizon (1993). This strategy can be characterized as a 'probabilistic reduction' type of 'general to specific' approach (Spanos 1995), and allows us to simultaneously model the
Introduction
23
dynamics of both short-term (changes) and long-term (levels) adjustment processes. In concrete terms, the modelling starts from a wellspecified, log-linear, and unrestricted vector autoregressive (UVAR) representation of the data. The cointegration relation is detected based on the UVAR and introduced into a reduction process. The modelling then gradually reduces and reparameterizes the UVAR into a most concise conditional error correction model and other marginal models. The work in this book shows that this strategy works well in the context of modelling aggregate behaviour of investment in China. Therefore, based on the LSE methodological framework and the corresponding modelling strategy and approaches, we can model the economic agents' dynamic investment-maximization behaviour subject to the representative bottleneck constraints directly without looking for and constructing complex shortage indicators and artificial normal paths. In this model set-up, the cyclical adjustment processes of real investment can also be better understood. The cointegration relation represents the long-run equilibrium relationship or 'norm' path (hyperplane) between real investment and relevant bottleneck constraints. On the supply side, so long as the supply of energy and agricultural products (as the representative bottleneck sectors) increases along this longrun equilibrium path, the basic demand of real investment for them can be satisfied, and real investment thus proceeds regularly under the restriction of the normal bottleneck constraints or normal shortage states. If, say, the supply of agricultural products exceeds this norm path, it means that the shortage of agricultural products is relaxed, and the 'bottleneck' is widened; thus with the help of those long-run and shortrun adjustment coefficients (bottleneck multipliers) real investment can be accelerated. If the supply falls lower than the norm path, it means that the shortage is intensified, and the 'bottleneck' is narrowed; thus again with the help of those adjustment coefficients, the growth rate of real investment will decline. On the demand side, if an industrialization drive is initiated by a political/economic campaign and real investment exceeds this norm path, in the short-run it is possible to support such an investment rush by drawing on stock and overloading the supply fertilities of bottleneck sectors. However, eventually shortages in bottleneck sectors will re-emerge and be rapidly sharpened by the exhaustion of stock and the overloading of supply capacity. As a result, the 'brake may suddenly be pulled' by the central authorities and the administrative hierarchy in their struggle to cope with emerging danger signals,
24
Chapter 1
and the growth rates of both planned and actual investment outlays will fall, in some case even becoming negative. Thus, by means of the cointegration norm (attractor) and the error correction mechanism we can incorporate Bauer's four-phase description of investment cycles into a more comprehensive and formalised framework.
1.6 Scope and Notes on the Database In consideration of relevant data limitation and the emphasis of the research, investment efficiency and sectoral breakdown of investment will not be discussed in detail.11 Because of data limitations (cf. note 2), this research can only deal with fixed investment in the state sector, for the period of 1953-96. For simplicity, 'investment' hereafter represents the state-sector fixed investment, i.e. 'investment in fixed assets of state-owned units' as described by the Statistical Yearbook of China. Although the over-investment of township and village enterprises (TVEs) is also reported frequently in China's media, both the scale and significance are secondary. In addition, TVE investment hunger may be fuelled mainly by very low real interest rates of credits, by their using the collective lands and facilities with less costs or even without costs, and by local protectionism (Chen 1993, Saith 1995). Dealing with TVE investment drive goes beyond the scope of this book. However, so long as TVEs over-investment exists, the supply constraints on state sector investment expansion are bound to become more binding.
Table 1.4
Position of foreign investment in the state sector fixed investment, 1977-96 1977
1981
1990
1992
1993
1996
n.a. n.a.
3.64 3.80
28.46 6.30
46.87 5.80
95.43 7.30
274.74 11.96
1.31 2.40 n.a.
3.61 5.40 99.26
27.18 9.10 95.50
43.99 8.00 93.86
48.35 6.10 50.67
81.11 6.73 29.52
Total foreign investment Y billions % of national total fixed investment Foreign investment in the state sector Y billions % of state sector fixed investment % of total foreign investment
Sources: China Statistical Yearbook on Investment in Fixed Assets, 1950-1995 (1997: 19-24); Statistical Yearbook of China (1997:151-54).
Introduction
25
Foreign investment has been counted as a component of fixed investment statistics. Table 1.4 presents the relative position of foreign investment in the national total of fixed investment and in the statesector fixed investment. It can be seen from the table that: (a) Before 1993, foreign investment was concentrated in the state sector, which mainly consisted of official loans from OECD countries and international organizations such as the World Bank and Asian Development Bank. It financed the purchase of capital goods from the lending countries (Lees 1997, Chapter 10; Lardy 1994, 1995). (b) Starting from 1993, foreign direct investment and portfolio investment have appeared to constitute the bulk of foreign investment. Because state enterprise joint ventures and shareholding companies are counted as 'other ownership enterprises' in China's statistics, foreign direct investment and portfolio investment have statistically entered the non-state sector. (c) Foreign investment has accounted for a small proportion of China's fixed investment in both the state and non-state sectors, though its contribution to the transfer of advanced technologies and management experiences, to the expansion of China' trade, and to the alleviation of foreign exchange shortage should not be underestimated. In this book, foreign investment is also treated as a component of fixed investment. While foreign investment may not bring direct demand pressure upon domestic investment goods, in the short run it will certainly put pressures on energy supply and on the consumer goods market. The intention of this book is to identify and model the critical forces that determine the equilibrium investment level and investment cycle. Therefore, some important macroeconomic components, such as international trade in general, and export and import of energy and agricultural products in particular, play only implied roles in the modelling process. The economic reason behind this is that China has, in most years, depended on net export of energy and agricultural products for gaining foreign exchange, which, together with capital inflows, has in turn served as a important means for importing advanced techniques and equipment. This implies that the foreign exchange constraint to investment expansion has been transferred to the representative bottleneck constraints of energy and agriculture, which is, to a large extent, encompassed by the modelling process. This transfer process and its
26
Chapter 1
quantitative magnitude will be analysed in some detail in Chapters 4 and 5. The technical reason for this simplification is that many econometric experiments with several relevant international trade indicators and net foreign savings reveal that none of them have statistical significance.12 Due to an almost complete statistical blackout from approximately 1960 until 1980, the reliability of complete sets of hundreds of time series, going back to the early 1950s and which later appeared in the early 1980s, is questioned by most Western scholars. It is unclear whether the data before 1980 come from recent estimations or are actually based on professional statistics. In fact, many people believe that China simply stopped collecting data during the more chaotic years of the Cultural Revolution. However, many confidential and/or unpublished statistical books and documents that were based on professional statistics were written prior to 1980. Recent publications are mainly based on these then-confidential or unpublished data. Apparently exaggerated output data as a result of political campaigns were usually corrected in the following year and/or in the consequent rectification period, rather than in 1980s. For instance, on 14 April 1959 the State Statistical Bureau published 1958 grain output figures double the 1957 total, amounting to 380 million tons {People's Daily, 14 April 1959, p. 1). However, in the same year, the State Statistical Bureau published the 'Revision of Agricultural Statistics of 1958', in which grain output was aggregated to 250 million tons {People's Daily, 26 August 1959). After 1960, the official figure for grain output in 1958 was 200 million tons. During the Cultural Revolution, most data was collected within functional systems, as a part of the day-to-day management of the economy. Li Chengrui (1984), then Director of the State Statistical Bureau, explained that statistics were collected even during the Cultural Revolution and declared that they are 'basically reliable' and 'basically conform to the political and economic changes of the times' (Li Chengrui 1984: 23). The problems related to China's statistical data are mainly caused by the following two factors. First, China's statistical system has undergone several reorganizations over the years, and these reorganizations may have affected the quality, coverage, and definition of individual data series. Secondly, many changes in coverage and definition of data series, particularly structural ones induced by institutional changes and reform and/or by structural shifts of production and con-
Introduction
27
sumption, have not been, and may continue not to be, corrected. For example, in terms of fixed investment, consistent statistical distinctions between in-budget and out-budget financing, between central and local financing, and between fiscal appropriation and bank's policy lending, are hard to obtain. This is due to recurring cycles of decentralization and recentralization, considerable overlap of fiscal and monetary operations, and adjustments of budget coverage. Therefore, before any data can be used in econometric exercises, the relevant institutional context and the definitions used need to be clearly understood. The data series employed in this book in general, and in the modelling exercise in particular, are those considered to meet minimum criteria of reliability and consistency. The modelling process uses aggregate data such as total fixed investment of the state sector, grain output, energy consumption, total population, and a deflator of accumulation in fixed assets by the state sector. Those are all relatively independent of the changes of coverage and definition. Data on state sector fixed investment have always been closely monitored by planners. A single and most powerful institution - the State Planning Commission - has always borne responsibility for collecting these data. Because of the specific position of the State Planning Commission in comprehensive planning over the years, it has maintained the capacity to collect and aggregate investment data on the state sector in a meaningful fashion. The other data used in the primary data analyses are also mainly derived from aggregate figures such as national income, total output of agriculture and agricultural national income, total cultivated land and labour force engaged in agriculture. Where problems in coverage and definition arise, explanations and corrections are provided.
1.7 Organization of this Book Chapter 2 attempts to select the most relevant theories from the literature and to incorporate them into a discussion of China's experiences such as its heavy industry-oriented development strategy and the rigidity of this strategy, the fundamental mechanism of capital accumulation and the reform cycle. Chapter 3 presents the basic characteristics of China's state investment system and its response to reform. It shows that at all levels of central and local government and in enterprises, there is a ubiquitous growth drive and investment hunger. Both of them are endogenously generated within the political/economic system and
28
Chapter 1
structure, and have created pressure and incentives to maximize the likelihood of expanding investment. Due to a lack of internal, selfimposed restraint capable of resisting the expansion drive, real investment will expand until it overshoots the supply possibility frontier of bottleneck sectors. Chapters 4 and 5 present primary evidence of key bottleneck constraints to investment expansion. Chapter 4 aims to explore the adverse distributive consequences that a rising investment rate puts upon agriculture and to examine the short-run two-way interaction between agricultural fluctuation and macroeconomic adjustment in general and investment modification in particular. Chapter 5 shows how energy shortage has checked China's economic growth and played a representative role among producer goods constraints. The short-run two-way dependence between the growth rate of real investment and effective energy supply is also tested in this chapter by using the Granger causality test. In Chapter 6, a comprehensive econometric exercise based on dynamic cointegration approach is carried out. The integration and cointegration properties of the data are analysed. A long-run investment-level function and a conditional error-correction model of investment growth rate are presented. The possibility of the existence of a structural break caused by reform or other policy shifts is examined by the one-step-ahead Chow test (standard structural break test), and by one-step residuals and the corresponding standard errors in a recursive estimation. Finally, Chapter 7 summarizes the book, explores further the theoretical and policy implications of the research, reviews the major limitation of the analysis in this book and discusses the most recent development of China's state sector reform beyond 1995 and 1996.
Integrating Selected Theories Based on China's Experiences
2.1 Introduction The purpose of this chapter is to select complementary theories based on China's specific experiences and to reveal the possibility of integrating these selected theories into a new growth cycle framework. Because this book intends to undertake a theoretical exercise grounded on empirical and historical analyses, the interaction between empirical examination and theoretical searching has played a very important role in the process of understanding problems and difficulties, in looking for the right modelling strategy and approaches, and in exploring the theoretical and policy implications of the findings. Among the numerous publications dealing with investment cycles in socialist economies such as the former Soviet Union and Eastern Europe, the works of the 'Hungarian School', particularly, Kornai and Bauer, are most influential. These works present a brilliant demand analysis and vividly descriptive explorations of the investment behaviour of planners and the formation mechanism of investment cycles. Such new terms as 'soft budget constraint', 'expansion drive', and 'investment hunger', coined by Kornai and his colleagues, have become key concepts in the discussion of the socialist economy. In particular, the 'soft budget constraint' concept has gone beyond the boundary of the socialist economy and become one of central metaphors in the debates about the dilemmas of 'limited liability', 'the central bank', and 'strategic bankruptcy' as well (Cui 1993). The next section introduces the demand analysis of the 'Hungarian School' based on the concepts of soft budget constraint, expansion drive and investment hunger first, and then uses Bauer's (1978) description of four-phase investment cycles in
29
30
Chapter 2
a socialist economy to sketch out the cyclical patterns of investment in China. As noted in section 1.2, in terms of modelling, a key tradition in the literature is to analyse and estimate the planners' reaction function to several shortage signals, which represent the economy's supply capacity constraints on planners who are seeking to maximize the growth rate. This tradition also dominates the relevant research on China (Imai 1990, 1994a; Naughton 1986, 1987). However, as will be shown below, this modelling tradition suffers from certain theoretical and statistical difficulties. These include the absence of intertemporal rationality of the planner, statistical possibility and macro-indivisibility assumptions of investment commitments, two-way dependency between investment tension and shortage strength (and thus the weak exogeneity assumption of shortage), the effectiveness of planners' control, and the definition and measurement of shortage in consumption and investment goods sectors. In order to overcome these contradictions and limitations and to advance the standard investment cycle theory, we need to find complementary theories and advanced econometric methodology and approaches. A parallel theoretical strand, which attempts to build up the 'bottleneck constraint type of growth theory' for a socialist and/or developing economy, is distinguished in Kalecki's (1972, 1976) seminal works. Kalecki's theory is insightful for directly examining bottleneck constraints on a fairly high rate of investment growth, as in the case in China. He focuses on the trade-off balance between the desired high rates of investment growth and the limits to tolerate the adverse distributive consequences that a high investment rate puts on agriculture. He interprets the problem of financing development as more than an adjustment of planned aggregate output to the available supply of necessities. By incorporating the comparative advantages of both the Hungarian School and Kalecki's theory it is possible to set up a new 'growth cycle' framework (Goodwin 1967 & 1982, Dore 1993). Business-cycle theory has been a major branch of the Western economics. Many approaches have been established to model business cycles in the west. Among them, the modern approach employed by the proponents of real business cycle theory has followed the tradition of stochastic analysis of Frisch (1933) and Slutsky (1937). It distinguishes shocks from a steady growth path of an economy, and more importantly it distinguishes between the shocks that cause variables to differ
Integrating Selected Theories Based on China's Experiences
31
from their steady state values and the propagation mechanisms that convert the shocks into longer-lived divergences from steady state values (Fischer, 1988). The modelling strategy and approach used in this book is, to a great extent, rooted in the real business cycle approach. This chapter is organized as follows. Section 2.2 gives a brief introduction of the investment demand analysis of the Hungarian School, and then sketches the cyclical patterns of investment in China based on Bauer's four-phase theory. It also reviews the tradition of establishing single control equation models, and shows the relevant problems in that tradition. Section 2.3 discusses and extends Kaleckian 'bottleneck constraints' growth theory in a socialist developing economy, and then explores the mechanism of fluctuation around the Kaleckian equilibrium steady growth path. Section 2.4 traces the existing researches on China's investment cycles and reviews Justin Lin's (1995, 1996) reform cycle theory. Section 2.5 outlines the relevance of the Western business cycle theories and highlights the methodological significance of real business cycle approach and Goodwin's growth cycle model to the research of this book. Section 2.6 summarizes this chapter.
2.2
Incorporating the Investment Cycle Theory of the 'Hungarian School' with China's Experiences 2.2.1 Soft budget constraint, expansion drive and investment hunger
The term 'budget constraint' is familiar from microeconomic theory of the household, in which the sum available to a decision-maker places a constraint on consumer spending, with own expenses covered by income generated by selling household output and/or by earning a return on assets. Clearly, the budget constraint presents a behaviourial characteristic of the decision-maker. It is a constraint on ex ante variables and primarily on demand. The 'softening' of the budget constraint appears when the decision-maker expects that excess expenditure over earnings will be paid by some other institution. In other words, external assistance is highly probable and this probability is firmly built into decision-making behaviour (Kornai 1986).
32
Chapter 2
In socialist economies, such as Hungary and China, state-owned enterprises (SOEs) used to receive regular external assistance. Four main forms of this assistance can be distinguished: (a) Soft appropriation and subsidies granted by national or local governments. These are soft because their amount is negotiable, subject to bargaining, lobbying, etc. (b) Soft taxation, where the taxation rate is not low, but the amount of tax payment is subject to prior and/or subsequent bargaining, where the fulfilment of tax obligations is not strictly enforced, and where there are leaks, ad hoc exemptions, postponements, etc. (c) Soft credit, where again the fulfilment of credit contracts becomes the subject of bargaining, contract fulfilment is not enforced, unreliable debt service is tolerated, and postponement and rescheduling are in order. And (d) Soft administrative pricing, in which administrative pricing, which dominated the price-setting processes before reform and still plays a significant role during transition, can be 'softened' by vertical bargaining with the price authorities according to some permissive 'cost plus' principle. Such external assistance is not granted automatically and some effort is needed to obtain it. Influence costs (Milgrom & Roberts 1988, Milgrom 1988), rent-seeking costs (Krueger 1974) and bargaining costs should be take into account. Therefore, even a softened budget constraint exerts some influence on the behaviour of the enterprise (Kornai 1986). The concept of soft budget constraint illustrates the collective experience of a large group of enterprises and, in China, the sum of the SOEs. It reflects in financial form a deeper, socio-economic phenomenon, which in Marxian terms would be indicative of a certain social relationship between the state and the economic micro-units, particularly the SOEs. One of the most impressive consequences of the soft budget constraint syndrome may be the formation of investment hunger. For example, consider an economy where hard budget constraints dominate. With hard budget constraints a firm will start a project only if it seriously believes that the flow of revenues from the sale of output generated by the new project will cover the flow of expenditures needed to
Integrating Selected Theories Based on China's Experiences
33
accomplish it. It is true that in a world of uncertainty, different decision-makers will have different degrees of risk-aversion. Nevertheless, given the distribution of risk-aversion over all investment decisionmakers, the total demand for investment resources will be constrained. There will be self-restraint in capital formation decisions because of genuine fear of financial failure. Therefore there will be a relatively symmetrical relationship between demand for investment resources and the supply generated by the same investment resources.1 In contrast, in an economy where a sufficiently large number of decision-makers enjoy soft budget constraints, this kind of symmetry breaks down because external financial support can be obtained with little cost. An enterprise might start a project even though it may suspect that the cost might exceed planned levels by a significant margin or that the revenue might be much less than estimated. In case of financial failure it can expect to be bailed out by the state. In such a situation there is no internal, selfimposed restraint in investment intentions and the demand is not counterbalanced by a 'dead serious' consideration of revenues or of supply. Therefore, as long as an expansion drive exists, investment hunger is inevitable. Expansion drive is found at all levels of the economic hierarchy in a socialist country. The major motivations behind it can be divided into two categories. The first relates to the internal and external pressure to provide evidence of socialist superiority, which should be demonstrated by the fact that a socialist economy can quickly catch up with the developed countries. This belief is a major aspect of official ideology, and has been constantly repeated since the victory of the socialist revolution. As a result, the revolutionaries themselves are not only impatient, but they also feel pressure from large masses of the public. On the other hand, this drive to catch up is reinforced by military and defence considerations. Modernization and economic strength are needed to create a powerful army (Kornai 1992). The second category relates to the motives of the bureaucracy. These motives create such a strong inner expansion drive that the top-level leadership does not need to impose a forced growth policy on the mid-level and lower-level leaders. Among the seven motivations listed by Kornai (1992, section 7.4), five provide an explanation of bureaucracy's powerful inner expansion drive: political and moral conviction, identification with the job, power, prestige and material benefit. The members of the bureaucracy hold the same political convictions as the top-level leaders, who perceive the need to
34
Chapter 2
catch up with increasing speed. Identification with one's job also inclines one toward expansion due to the conviction that the activity of the unit under one's charge is important and that it has to be extended. Someone burdened with the internal problems of a unit also often believes that these could be solved by investment. Leaders at all levels believe that their power and prestige grow along with the expansion of the unit under their charge, and in many cases there are material rewards involved. Finally, expansion drive is further stimulated by shortage. There is a waiting list for the firm's products; buyers increasingly demand more, and investment is needed to increase output. The difficulties the firm suffers in obtaining its own inputs may prompt it to internalize the production of these inputs, which again requires investment (Kornai 1980, 1992). Expansion drive is not, in fact, system-specific. Capitalist entrepreneurs also expect expansion to bring greater profits, and so greater power, prestige and financial rewards. The main system-specific distinction lies not in the actual effort to expand but in the internally generated self-restraints that run counter to the expansion drive. For a capitalist firm even with limited liability, any loss caused by a faulty investment decision hits both owners and managers. Expansion is an attraction but also a high risk because although they expect to make good investments, the risk of bad investment limits unbridled expansion. By contrast, in a socialist economy subject to a soft budget constraint, an SOE can suppose that liquidation will not follow from any defective investment decision, however high the costs and financial losses may be. For an investment project, many people at different levels of the hierarchy play a role in the decision-making (i.e. project approval) process so that it is almost impossible to lay the blame on anyone in particular. The loss incurred will not affect the income or assets of any of the bureaucrats. Since there is an almost total absence of internally generated self-restraint that could check expanding, expansion drive becomes a 'natural instinct' for the bureaucracy and investment hunger is ubiquitous. The above distinction may be alternatively explained by incorporating some thoughts of Kalecki and Kaldor. For each class, from the perspective of income distribution, 'capitalists earn what they spend, and workers spend what they earn' (Kalecki, cited in Kaldor 1955). For individuals, however, the expansion of capitalist investment can be disastrous if corresponding demand is not forthcoming. By contrast, in a
Integrating Selected Theories Based on China's Experiences
35
socialist economy, the consolidated state sector (state enterprises and government bodies) earns what it spends. In other words, the volume of investment spending will generate the necessary income to cross-subsidize reckless investment activities. Such a soft budget constraint means that the abovementioned individual reluctance is groundless, and the fear of failure is largely eliminated. As a result, the limits to investment expansion are bound to be resource constraints and distributive barriers between investment and consumption in general, and industrial expansion and requisite incentives to food production in particular.
2.2.2 Bauer's four-phase theory and China's investment cycles Section 1.4 presents a preliminary indication of the existence of investment cycles in China. This section shows that the cyclical patterns of both investment ratio and investment growth rate as well as the linkage between them can be adequately described by Bauer's four-phase model of investment cycles in a socialist' economy (Bauer 1978: 24360 & 1988). Bauer's model is instructive for an intuitive understanding of the fundamental mechanism generating investment cycles in China.2 Bauer's initial or run-up phase (cf. Table 1.3) corresponds to the years 1955, 1957-58, 1962-63, 1968-69, 1973, 1977, 1981-82, and 1989-90. During these years, a large number of projects were approved. According to Bauer's model, newly started projects initially involve only a relatively moderate increase in investment outlays. This does not lead to tensions, although clusters of new projects may cause a rapid extension in the stock of investment projects in progress. This leads to a significant increase in the 'investment engagement' or the unspent part of the investment budget allotted to these projects. In the second or rush phase, namely the years 1956, 1958-59, 196465, 1969-70, 1974-75, 1978, 1983-86, and 1991-92 (cf. Table 1.3), new investment projects were continuously being approved and the high outlays committed in the run-up phase were beginning to be felt. Since the investment outlays were systematically understated in the previous plan bargaining process, actual investment outlays exceeded planned volume. The growth rate of investment reached its maximum in this cycle, and exceeded the planned growth rate even if the latter was excessive due to political or economic campaigns (for example, the 'High Tide of Socialist Construction' in 1955-57, the Great Leap For-
36
Chapter 2
ward in 1958-60, Construction of Inland Industrial Bases in 1962-68, the 'Foreign Leap Forward' i.e. 'Yang Yuejirf in 1978, and the 'Deng Whirlwind' in 1992-93). Excess demand in the investment sector increased, the shortage of investment goods and services re-emerged or became sharper, and the shares of fixed investment and of capital accumulation in the national income rose. Other sectors sacrificed their growth rates; in the case of China these were agriculture and consumption sectors, which are particularly vulnerable. According to Bauer's model, the upper turning point in the cycle occurs in the third or halt phase. This occurred in China in the years 1956, 1960, 1966, 1971, 1975, 1978, 1987 and 1993 (cf. Table 1.3), during which there was a sharpening of the shortages of investment goods and services, and planners, especially central planners, became more vigilant and began to turn down a large proportion of investment requests. In the four-phase model, investment requests do not decrease during this phase, but a relatively smaller proportion of them is approved. That is, the 'approved coefficient' falls. At the same time, work on existing projects is hastened even at the sacrifice of a shift between the uses of national income, since 'the planners intend to work off the investment engagement through forcing a rapid growth of investment outlays and of construction' (Bauer 1978: 250). Owing to increased spending on existing projects the shortages may be exacerbated and the investment ratio may rise further, though compared to the previous phase, the growth rate may fall. Once such danger signals as bottlenecks become more frequent in the investment sector, as the supply of investment goods and services for the implementation of ongoing projects becomes more irregular and uncertain, the authorities begin to fear the development of social disturbances as a consequence of growing dissatisfaction with stagnating agricultural production and increased shortages in the consumption sector. As a result, the 'brake may suddenly be pulled'. The collective action of the authorities initiates the fourth or 'slowdown' phase, seen in China in 1957, 1961-62, 1966-67, 1972-73, 1976-77, 1979-81, 1988-89, and 1994-96 (cf. Table 1.3). Not only does the 'approval coefficient' continue to fall, but the annual limits of investment outlays are also cut. In extreme cases, planners may even retroactively cancel investment projects approved during the 'rush' and 'halt' phases, in their struggle to cope with emerging danger signals (which appears to be especially significant in China):
Integrating Selected Theories Based on China's Experiences
37
The most important feature of this phase is the fall in the planned and actual growth rate of investment outlays (in some cases even negative growth rates). Due to this the rate of investment and of accumulation falls, and an opposite shift takes place among the uses of national income. (Bauer 1978: 252) This situation continues until the tensions in the utilization of national income and the shortages of investment goods and services are alleviated. As the shortages of investment goods and services are relieved, the investment ratio falls and the situation improves with respect to those uses of national income that were suppressed earlier. As a result, restrictions on starting new investment projects begin to weigh more heavily on supervisory planners who are authorized to approve them. As pointed out by Ickes (1986: 47), 'in a resource-constrained economy all projects will appear socially beneficial'. At the micro-level the factory and shop managers say they cannot produce more under the given conditions, and the desire to increase production comes up against one bottleneck after another. Therefore, it seems logical to open the bottlenecks through expansion of capacity. Each investment project, viewed in isolation, appears worthy of investment funds which are needed to alleviate shortages. No measure can be used to compare the opportunity costs of different projects, which means that while each project by itself seems deserving of funds, not all projects can be funded. This means that there are only 'important' and 'more important' investment projects in bureaucratic-coordinated economies. In this context, if selection becomes more rigorous and the approval coefficient falls (in the halt and slowdown phases), then the refused proposals are not denied outright, but are only postponed. Hence, the pressure of these postponed claims increases in the slowdown phase. At the same time, a certain alleviation of tensions and shortages also stimulates the growth-related goals of planners, who become less sensitive to efforts endangering macroeconomic equilibrium. Thus, the selection of investment projects is less strict, the approval coefficient rises again and a new cycle begins. The end point of slowdown marks the lower turning point in the cycle. It is worth noting that the run-up phase usually provides suitable conditions for initiating political/economic expansion campaigns and such campaigns typically accelerate the run-up phase to rush. However, the concrete radical extent of each political/economic campaign, which cannot be folly explained by the general process examined above and is
38
Chapter 2
to a great extent exogenous, seems to play a leading role in determining the amplitude of each cycle. Most significantly, the most radical 'Great Leap Forward' (1958-60) and the resultant economic collapse (196162) and recovery (1963-65) resulted in the widest cycle (1957-62) and the recovery type of high growth rates in 1963-65. The extremely radical 'Cultural Revolution' brought in the second biggest dip (1967-68) and a subsequent recovery rush (1969-70) (cf. Figure 1.1). It may be thanks to such tragic lessons that China has avoided similarly radical political/economic campaigns since 1970. However, less radical economic/political campaigns have been generated within the system (cf. Table 1.3) and may also occur cyclically in the near future. While the radical extent of these political/economic campaigns determines the amplitude of each corresponding cycle, it also determines the nature of interaction between investment expansion and bottleneck constraints. Radical campaigns induce extraordinarily high investment demands, creating the unsustainable exhaustion of stock and an overloading of supply capacity of bottleneck sectors, leading finally to the collapse of some bottleneck sectors or even of the entire economy. In brief, Bauer's investment cycle model is an important component of the Hungarian School's work. The conclusion is that investment fluctuations are endogenous to the centrally planned economic system and are caused largely by planners' response to shortage signals and to internal systematic tensions. The shortage signals represent resource constraints on planners who are seeking to maximize the investment growth. The internal tensions epitomize investment hunger and the expansion drive which stem from incentives of local and departmental institutions and enterprises to attain the objectives of the output plan and to expand their economic and administrative power. On the other hand, it is clear that Bauer's theory focuses on short-run disequilibrium only and does not tackle the question of what determines the level of investment.
2.2.3 Single control equation model and the relevant problems The work of the 'Hungarian School' implies that our understanding of investment cycles in socialist countries should be improved by trying to model planners' behaviour. In fact, analysing and estimating the planners' reaction function have occupied central positions in the literature
Integrating Selected Theories Based on China's Experiences
39
(Mihalyi 1992, Chapter 2), in which the works of Kornai (1982), Roland (1987), and Christin & Short (1991) are most influential and therefore of greatest interest to us. Nevertheless, the successful elaboration of a testable model would not be an easy task, particularly when the necessary econometric approach were not yet available. Kornai (1982: 38-53) establishes a control equation governing the investment process by the planners in which the dependent variable, namely the volume of what he calls investment vintages (new projects), is determined by the normal volume of investment vintages, and the deviations of consumption, investment commitment, and shortage intensity from their normal values or averages (e.g. five-year moving average). This model can be written as:
where M(t) represents the volume of the investment vintage and symbolizes the value of machines and buildings put into operation in future as a result of the projects started in year t, H(t-\) represents consumption lagged by one year, K(t) represents the investment commitment, that is, the total scale of investment required for completing all projects under construction in the year /, and Z(t) indicates the shortage intensity. The (*) designation presents normal levels of M(t), H(t-\), K(t) and Z(t), respectively. The \i's (>0) represent the feedback strength coefficients. Kornai affirms that as a result of the influence of the three non-price feedback signals, decision-makers cause M(t) to deviate from its normal value M (t) - new investment starts - so as to drive the system back to the normal paths of consumption, investment and shortage. Clearly, this model also focuses on short-run disequilibrium behaviour without tackling the question of what determines the investment level, because in equilibrium M(t) = M*(t). Apart from this, the theoretical and statistical confusion inherent in equation (2.1) makes it impossible to test this single equation. First of all, there is an absence of intertemporal rationality of the representative planner in the model setup. Secondly, as pointed out by Mihalyi (1992, section 2.2.2), both K(t) and M(t) are stock variables; the measurement of K(t) is both theoretically and statistically confusing because it represents a mixture of actual (i.e. past expense at mixed price level) and planned (i.e. future expenditure estimated) costs; and
40
Chapter 2
M(t) is completely an ex ante estimate of the expenditure needed to complete the given investment projects started in the year /, based only on engineers' calculations. In fact, the relevant actual costs are always much higher than the pre-estimates (Kornai 1980: 195-8). In addition, as in other socialist countries, data on new investment projects are not available in China. Thirdly, perhaps more importantly, there exists a type of 'fallacy of composition' within the theoretical derivation related to investment commitment. The concept of investment commitment has played a central role in quantifying effective demand for investment goods. 'The investment decision-maker is - to a large extent - the prisoner of his own previous decisions' (Kornai 1982: 54). The basic assumption here is that current decisions concerning the volume of investment in the next period are determined by past decisions on investment starts. The intuitive rationale behind the derivation is based on numerous observations, which indicate that once an investment project has been started, it will sooner or later be completed regardless of profitability considerations, as a half-ready factory is a guaranteed loss from the point of view of the planners. As a theoretical abstraction, this can be expressed as an assumption of the indivisibility of investment decisions. Although at microeconomic level an investment decision usually remains indivisible in a socialist economy, on the macroeconomic level, projects can and do substitute and complement one another. If the plans for opening a new coal mine are shelved, there will be no need for a new coal-fired power station. In this sense, the idea of investment commitment provides a theory of demand for investment goods only at the individual firm level rather than at the macroeconomic level (Mihalyi 1992). Fourthly, using a univariate moving average as a proxy of the 'normal path' may, itself, produce spurious cycles as discussed in details by Slutsky (1927) and Frich (1928). In addition, if the 'normal path' does represent the equilibrium state of planners' behaviour, there should be a structural relationship determined by the interaction among the relevant demand (investment, here) and supply (e.g. supply of investment goods and consumer goods) variables. In fact, this type of structural equilibrium relationship presents the determining function of investment level. Fifthly, the notion of shortage is theoretically powerful but statistically difficult to handle. In a market economic framework ex ante demand would be simulated and modelled through structural behaviour relationships, e.g. consumption should express the 'revealed preferen-
Integrating Selected Theories Based on China's Experiences
41
ces' of consumers given their income and prices. In case of insufficient supply (shortage) price adjustment would lead to an adjustment of both demand and supply to equilibrium direction. Here one may define 'shortage' as excess demand for goods or factors and trust that price and/or quantity adjustment will clear the shortage. But even in a market economy we are not always able to be certain that we have modelled ex ante behaviour properly. In a centrally planned economy, we may have some certainties, but also some uncertainties. For example, in the area of consumption it is difficult to know the 'revealed preferences' of the agents if consumer goods have been rationed. Certainty used to be linked to the central planners' 'revealed preferences' as expressed in economic plans, but these plans are also argued to be endogenous to economic performance. Such uncertainties imply that it is almost impossible to estimate ex ante demand functions in the traditional manner. In the literature, the term 'shortage' has been linked to several different situations: (a) bureaucratic inefficiencies, (b) real bottlenecks such as insufficient supply of investment goods and consumer goods, (c) financial bottlenecks such as insufficient foreign exchange or domestic finance to meet investment requirements, (d) sectoral imbalances, and (e) gestation lags or duration of the 'construction period'. There are various measuring exercises focused on (b), showing diverse results. For example, in China's case, there are about ten competitive indicators measuring shortage in consumption markets (Portes & Santorum 1987, Table 6; Peebles 1991, Table III.l & III.2; Peebles 1992; Imai 1994a & b). Finally, due to the simultaneous interaction between shortage and investment tension, the sign before |i z is less certain than might be wished. On one hand, more intense shortage reflects the fact that the system meets its own resource constraints more and more frequently and suffers increasing losses. Abovenormal shortage intensity therefore induces the decision-makers to restrain new investment starts. Conversely, if the difficulties caused by shortage have diminished and complaints about under-utilization are beginning to be voiced, this will provide a stimulus to expand investment activity. (Kornai 1982: 49-50) Thus, the sign before \iz would be negative. But proceeding from the perspective of correlation analysis for time series, we will obtain an opposite conclusion. Shortages and investment tension
42
Chapter 2 are in close interaction: they form a special 'vicious circle'. Awareness of shortage is one of the main motives for expansion drive and the associated investment hunger. Shortage signals play an important role in the selection of investments. Thus, shortage generates investment tension... At the same time investment tension is one of the main causes of general shortage. Since investment hunger is insatiable, it creates an almost-insatiable demand. This extends as far as the resource constraints on investment activities, and even goes beyond them... The stronger the investment tension, the more it is felt that investment demand tries to draw resources away from other fields of utilization thus amplifying general shortages. (Kornai 1980: 201)
This circle suggests that there is a positive correlation between investment tension and shortage intensity in general, and the shortage of investment goods in particular, because investment tension results in shortages of investment goods almost without lags. These two opposite forces make the sign before \iz uncertain. Roland (1987) develops a more practical model based on equation (2.1) to display the former Soviet planners' behaviour with regard to macroeconomic investment growth. In his model the growth rate of investment is a function of structural pressure for investment and the measures of shortages in the investment goods, foreign trade, and consumption sectors. The structural pressure is measured in the first specification by the constant term and, in the second one, by the five-year moving average. He estimates the model using Soviet data for the period 1960-80 and concludes that with the exception of a positive coefficient for consumption tension, his results are consistent with the Hungarian School's explanation of investment cycles. Christin & Short (1991) improve Roland's model by removing trend and average indicators and employing more appropriate and theoretically well-established measures of disequilibrium in foreign trade and the consumer goods market. However, both works use 'unfinished construction' as a proxy for shortage in the investment goods sector which may be misleading. 'Unfinished construction' is the stock of unfinished capacity at historical cost, and there is no way to build a meaningful price index to deflate this stock series (Mihalyi 1992: 81).3 Following Bauer's (1978) line of analysis, the ratio of unfinished investment stock in current investment is much more sensitive to the 'approval coefficient' and 'the planners' intention to work off the investment engagement'. In other words, planners are more responsive to shortage signals and investment engagement at the beginning of the cause-effect chain; they then decide whether or not to change the approval coefficient and to hasten the
Integrating Selected Theories Based on China's Experiences
43
work on existing projects even at the sacrifice of growth in agricultural and consumption sectors. As a result, the ratio of unfinished investment rises or decreases passively and cannot be considered causative. Finally, as Roland (1987) argued, 'increased shortage in this sector leads to a lengthening of construction periods and thus to a rise in the rate of unfinished construction' (Rolands 1987: 197). As a theoretical concept, the construction period should be determined by technical characteristics of the investment process and the shortage intensity of investment goods. However, in the light of statistics, it is almost fully determined by investment starts and current investment outlays. In fact, the relevant statistical formulas of the construction periods, which have been used in China and in the other socialist countries (Statistics on Fixed Investment in China, 1989-1990, 1991: 347; Chen & Niu 1990: 916), are as follows: Construction period = [Number of projects under construction in this year] [Number of projects completed in this year]
(2.2)
or Construction period = [Plan target of investment in all projects under construction this year] [Investment outlays in this year]
(2.3)
Relating formulae (2.2) and (2.3) to Bauer's (1978) description of the investment cycle (section 2.2.2), it can be found that there is a strong negative correlation between the construction period cycle and that of the investment growth rate. In the 'run-up' phase of an investment cycle, a large number of projects are approved, and the number of projects under construction and the planned investment in all projects under construction rise rapidly. However, because 'the newly started projects in the first period of construction involve only a relatively moderate increase in investment outlays' (Bauer 1978: 249), the construction period reaches its maximum in the cycle while both growth rates of investment outlays and the number of completed projects are still very low. In the 'rush' phase, actual investment outlays exceed planned volume by a large margin and the growth rate of investment gradually reaches its maximum in the cycle. However, there is a slowdown both
44
Chapter 2
in the growth of the number of projects under construction and of planned investment. Therefore, the construction period gradually shortens. In the 'halt' phase, planners begin to refuse a larger proportion of investment requests, and at the same time, work on existing projects is hastened even at the sacrifice of growth of other sectors, so that the construction period reaches its minimum. In the 'slowdown' phase, the approved coefficient continues to fall, both the number of projects under construction and planned investment in all projects stagnate, and the annual limits of investment outlays are cut. Thus the construction period begins to stretch out. The above analyses suggest that the empirical correlation between investment growth and the uncompleted construction rate presented in Roland (1987) and Christin & Short (1991) may represent this type of negative correlation stemming from the statistical formulae (2.2) and (2.3) rather than the expected relationship between the investment growth and the shortage intensity of investment goods. Grosfeld (1987) models a planner's reaction function based on the notion that a planner's reaction to shortage signals is not smooth and continuous. She incorporates thresholds into her model, above and below which the planner has different intensities of reactions to shortages. In this way she effectively models Kornai's idea of shortage indicators (1980, 1982) as departures from normal levels. Furthermore, by allowing the reaction threshold to vary, Grosfeld tries to eliminate the problem of attempting to specify the subjective normal level of shortage. Nevertheless, there have been criticisms of (a) her proxies for the shortages of consumer goods and foreign trade sectors, which face the same contradictions as in the Roland's model (Christin & Short 1991), and (b) the impossibility to test planner's synthetic reactions to more than two shortage signals using her model. The latter would involve some very complicated econometric models switching among more than four regimes. Two other implied assumptions in the modelling tradition should also be considered. The first is that all shortage variables are assumed to be at least weakly exogenous. Taking into consideration the close interaction between investment tension and shortage intensity, this assumption has to be tested at the outset, because without weak exogeneity, the single reaction function is bound to be misspecified (Engle et al. 1983, Engle & Hendry 1993). The other assumption is that the planners' control is effectual, at least in the 'halt' and 'slowdown' phases of
Integrating Selected Theories Based on China fs Experiences
45
the investment cycle. However, as shown in their empirical estimates by Brada & King (1992), in a number of markets in former Soviet Union and Poland there was an appreciable, long-lasting, and increasing excess demand, most notably for grain, although the planners had made efforts to achieve market equilibration. The separation of motives and effects suggests that empirical testing based only on the single reaction equation might become secondary or insignificant in statistical tests.
2.3
Kaleckian Economic Growth Theory and China's Capital Accumulation Mechanism
2.3.1
The causality line: From growth rate to investment to saving to bottleneck constraints
By generalizing his experience of perspective planning, Kalecki (1972) established an outline of an economic growth model for a socialist economy. His basic equations are as followings. r<5 + n
(2.4)
AY 1 / r— = a+u Y mY (2.5) Equation (2.4) indicates that the growth rate of national income, r, is determined by the growth of labour productivity, 5, and the growth of labour supply, n. Beyond this constraint further increase in investment will have no effect in a labour-scarce economy. In equation (2.5), Y represents national income and AY is the increment of Y\ I indicates investment level; m denotes incremental capital-output ratio; a and u denote parameter of depreciation and coefficient of improvement in efficiency respectively, both of which can be treated as constant in a short-run dynamic analysis. Equation (2.5) is mainly used to analyse how the time path of national income and consumption are affected by raising the economic growth rate using various alternative strategies. Given the parameters m, w, and a, the rate of growth r in equation (2.5) determines the constant share of investment, I/Y, in national income, which is necessary in order to sustain r (Kalecki 1972: 24-5). In Financing Development, Kalecki (1976: Chapters 5 & 7) treated foodstuffs supply - which he called 'necessities' - as central to macroeconomic analysis in a developing economy rather than as an additional
46
Chapter 2
welfare issue. He interpreted the 'problem of financing development' as the adjustment of planned aggregate output to the available supply of necessities. FitzGerald (1993) reconstructs Kalecki's implicit model of a semi-industrialized macroeconomy with algebraic clarification and full consistency, so that the maximal industrial growth is as follows (FitzGerald 1993: Chapter 6):4 r,=rfl+(l-e)5
(2.6)
where r, and ra represent the growth rates of industrial output and food supply; e indicates the income demand elasticity of food; and 5 is productivity growth in the agricultural sector which is assumed to be equal to that in the industrial sector. A direct implication of equation (2.6) is as follows. When growth is accelerated under the condition that the internal terms of trade is effectively contained by coercive state procurement and rationing (as the means to sustain real wages), the food supply growth rate will decline with a certain lag as will the real wage. As a consequence, macroeconomic fluctuation becomes inevitable (FitzGerald 1993: 126, Sun 1996). Kalecki's outlines are indeed simple and fundamental5 and the corresponding theoretical analyses can provide us with valuable insights when considering the investment cycle in a socialist developing economy like China.6 Firstly, in a socialist economy, the causation among saving, investment and growth can be spelt out from growth drive to investment to saving. Initially, the system determines growth pressure, then the desired investment growth rate and saving rate. The desired saving rate is guaranteed by effective wage and price control by the state. In other words, investment generates its own desired saving. Finally, the feasible growth rate is determined by bottleneck constraints, such as shortages of agricultural products, energy and raw materials, problems in balancing foreign trade, and the adverse distributive impact of a rising investment rate on real wages and food supplies, which form barriers to growth (cf. also Osiatynski 1988: Chapter 5, Kalecki 1976: Chapters 5 & 7). This is the reason why Kalecki's growth theory is known as a 'bottleneck constraint type of growth theory'. Secondly, although labour constraint has not been binding in China, derivation of the labour constraint equation (2.4) can easily be generalized, as shown in section 2.3.2. In other words, it is not difficult to build a key scarce factor constraint equation similar to equation (2.4)
Integrating Selected Theories Based on China's Experiences
47
following Kalecki's derivation procedure (Kalecki 1972: 15-24). In China, energy, rather than labour, has acted as this type of key scarce factor in economic development, particularly in industrial and agricultural inputs productions (as will be demonstrated in Chapter 5). Therefore, we can consider equation (2.4) as an energy constraint equation. It means that the growth rate of China's national income is determined by the growth rates of both energy productivity and effective energy supply. Thirdly, equation (2.6) not only represents a direct food supply constraint to industrial expansion, it also implies a distributive barrier between raising industrial investment and maintaining necessary agricultural growth, as manifested in the derivation process.7 This distributive barrier can be further substantiated by considering China's specific capital accumulation mechanism. This is because its main characteristics are coercive extraction from agriculture and forced saving through depressing real wage and consumption,8 as analysed at some length in section 2.3.3 below. The logical progress of the capital accumulation mechanism is as follows: High capital accumulation policy -» state compulsory procurement system + collectivization of agriculture -> low prices for agricultural products -» low wages and low prices of inputs for the non-agricultural sector + low income of the peasantry -> low consumption of both peasantry and non-agricultural workers + high profit in the non-agricultural sector -» high capital accumulation. This indicates that China's capital accumulation mechanism is not characterized by agriculture's unfavourable terms of trade alone. It is also distinguished by the extra profits of the industrial sector stemming from depressing the wage rate and coercively procuring low-priced agricultural products as inputs and basic wage goods.9 When a mechanism for extracting resources from the agricultural sector through undervaluing agricultural produce and coercion is put into effect, the immediate result is a reduction in surplus products and/or a decrease in labour income. Consequently, the growth rate of agricultural production and/or of peasants' consumption declines. When agriculture is exploited to the extent that it loses its comparative profitability, agriculture shrinks, disinvestment occurs, and real consumption levels of both the rural and the urban population fall. The continuous drop in con-
48
Chapter 2
sumption levels may induce economic, social and political instability. Once this kind of situation occurs, the only way to solve the problem or crisis is to re-adjust the distributive pattern of national income so as to transfer some resources into agriculture. It is this mechanism that links China's macroeconomic adjustment and related investment cycles much more closely to agricultural fluctuation than is the case in other countries.10 In comparison to the Hungarian School's work, which shows comparative advantage in terms of vivid investment demand and fluctuation analyses, Kalecki's works on bottleneck constraints adds insight into understanding of the representative bottlenecks and the determination of the equilibrium steady growth path in China's economy. The two theories are complementary and it is possible to incorporate the comparative advantages of both theories into a new framework. In this way we can directly examine decision-makers' investment maximization behaviour subject to certain bottleneck constraints, and thus be relieved of the task of finding and constructing shortage indicators. Since Kalecki's theory does not show how the bottleneck constraints lead to a cycle, this first requires an elaboration of cycles, which will be done in section 2.3.4 by theoretical description and in Chapter 6 based on an error correction mechanism.
2.3.2
Generalizing the Kaleckian labour constraint equation to a key factor constraint equation
This section generalizes the Kaleckian labour constraint equation (2.4) to a key factor constraint equation. The key factor (KF) can be labour in a labour shortage economy like Poland before 1989. It may also be energy in an energy-shortage labour-surplus economy like China. It should be pointed out here that this generalization is only a theoretical exercise. In making this generalization, it should perhaps be noted that both the urban and rural sectors in China have been confronted with significant labour surplus. In urban China, the official unemployment rate fluctuated between 1.8 and 5.3 per cent during 1978-97 {Statistical Yearbook of China, hereafter Yearbook, 1995: 106, 1998: 127). This official rate is popularly believed to underreport the real figures on unemployment. In addition to open unemployment, it is estimated that about 6-20 per cent of employees in the state sector are redundant
Integrating Selected Theories Based on China's Experiences
49
(Naughton 1996: 211-12, Bell 1993). In 1996 and 1997, the job-off (xiagang) rates in the state sector was more than 6 per cent (Statistical Survey of China 1997: 31; People's Daily, 29 May 1997 & 9 Feb. 1998).11 In China's agricultural sector, it is well known that farmland is scarce, capital is fairly scarce, while labour is relatively abundant. Although the percentage of the labour force engaged in agriculture fell from 93.5 per cent in 1952 to 70.5 per cent in 1978, and further to 56.4 per cent in 1993, the number of agricultural labourers doubled in the same period, from 173 million in 1952 to 283 million in 1978 to 374 million in 1993. The rapid expansion of the rural industrial sector has created employment for more than 120 million rural labourers since 1992. However, the trend of growth in the absolute number of farming labourers was not reversed until 1984, and this trend for labourers in the broad agricultural sector was not reversed until 1993 (Lin 1992, Table 4; Yearbook 1997: 94 & 400). In 1990, the average size of family farms was only 0.42 hectare but the average number of agricultural labourers employed per farm was 1.73 (Ministry of Agriculture 1991). A recent econometric estimation based on agricultural and natural conditions data at county level (Sun et al. 1998) shows that even in the northwest region, where the ratio of land to labourers employed is much greater than that in the coastal regions, the farming output elasticity of labour still indicates that the marginal output of labour in 1990 (323 kg of grain) is lower than the subsistence wage level. Based on the national average dependent rate of 1.9 and taking roughly 250 kg of grain as the subsistence requirement per person,12 this indicates an overall labour surplus in rural China. As this research into modelling investment cycles in socialist economies introduces energy constraint for the first time, it is necessary to point out the difference between labour and energy constraint. Labour may be an absolute constraint because its supply is determined by both population growth and human life-cycle, both of which change very slowly and are thus not so responsive to demand-supply pressure. In contrast, energy supply can be increased by new investments, cutting back export and increasing import, and promoting energy saving technology, etc. In addition, the energy situation in China has certain distinguishing features. The most significant one is that demand seems to be insatiable and therefore is always greater than supply, no matter how fast the supply grows. In addition, the irrational price structure, com-
50
Chapter 2
bined with the relatively longer construction period and large-scale investment requirement, has made investment in the energy sector much less attractive to growth-minded investors under both local 'block' and ministerial 'line' authorities. As a result, energy development has lagged far behind macroeconomic development, and acted as an almost absolute constraint on China's economy (for details, see Chapters 3 & 5). In his Introduction to the Theory of Growth in a Socialist Economy (Kalecki 1972), Kalecki analysed an abstract uniform growth process based on the following assumptions: (a) the growth rate of the national income, r, is constant; (b) the parameters m, a and u in equation (2.5) remain unaltered, as does k- 'capital-output ratio for total capital'; (c) the productivity of labour in a new plant, which is brought into operation in successive years, increases at a constant rate, 8, owing to technical progress (inclusive of organizational progress) (Kalecki 1972: 17). All three assumptions can be maintained here by substituting the KF for the labour factor. Thus, (c) indicates that KF productivity in plants brought into operation in successive years increases at a constant rate, 8, meaning that in any given year it is higher than it was a year before in the proportion (1+8). Since investment increases at an annual rate r, as proved by Kalecki (1972: 17-8), and the capital-output ratio m is constant, the output produced by a new representative plant in any given year must be (I +r) times greater than that produced by an old representative plant in the preceding year. When both output and KF productivity increase at constant rates in the new plants brought into operation each year, the same must be true of KF employment. If denoting the rate of increase in KF employment in new plant by 8 we may write: 1+8 By repeating Kalecki's procedure with the unique substitution of KF for labour (1972: 21-4), we can then prove that equation (2.7) also holds for total KF employment and overall KF productivity. This means that total KF employment and the overall KF productivity increase pari passu with KF employment and KF productivity in new plants. We then need to make an additional assumption that full employment of KF prevails in the economy. In fact, this is a natural conse-
Integrating Selected Theories Based on China's Experiences
51
quence of the key scarce factor concept because of KF's chronic shortage in a socialist developing country like China. Let n denote the growth rate of effective KF supply. For full KF employment to be maintained, the growth rate of KF employment must be equal to that of effective KF supply. Hence we get: s=n
(2.8)
and 1 + r = (1 + 5)(1 +rc)= l + 8 +rt+ 5-rc
(2.9)
Since the annual growth rates 5 and n are rather small fractions, we may disregard their product 5 • n in equation (2.9) and write the latter in an approximate form: r =5+
rc
(2.9a)
in which the maximum growth rate of national income r is determined jointly by 8, which depends upon technical progress, and n, which depends on the growth rate of KF supply.
2.3.3 China's capital accumulation mechanism and Kaleckian agricultural-determining growth theory China's capital accumulation mechanism has been shaped by the adoption of a heavy industry-oriented development strategy in a capitalscarce agrarian economy. The relevant political, economic, and institutional logic is well explained in Lin et al. (1995, 1996).13 In 1949 when the People's Republic was founded, China inherited a war-torn agrarian economy, in which 89.4 per cent of the population resided in rural areas and industry accounted for only 12.7 per cent of national income (Yearbook 1991: 79; Comprehensive Statistics on China's Rural Economy, hereafter Rural Statistics 1989: 64). At that time, the general belief was that a developed heavy-industry sector was the clearest symbol of a nation's power and economic achievement. Like the leaders in India and many other newly independent developing countries, Chinese leaders intended to accelerate the development of heavy industry. After China's involvement in the Korean War in the early 1950s and the consequent embargo and isolation from the Western camp, catching up with the industrialized powers also became a necessity for national security. In addition, the Soviet Union's outstand-
52
Chapter 2
ing record of industrialization in the 1930s and 1940s provided the Chinese leadership with both inspiration and experience for adopting a heavy industry-oriented development strategy. Therefore in 1953, after recovering from wartime destruction, the Chinese government set heavy industry as the priority sector. The purpose was to build up the nation's capacity to produce capital goods and military materials as rapidly as possible. This strategy was developed and implemented through a series of five-year plans.14 As a capital-intensive sector, the construction of heavy industry has three specific features: (1) each project requires a considerable period of time, maybe five to ten years or more, to be completed; (2) most equipment, at least in the initial stage, is imported from more developed economies; (3) heavy-industry projects require larger initial investments than do projects in other industries. In the early 1950s, the basic condition of the Chinese economy was obviously mismatched to these three features. At that time available capital was limited and, as a consequence, the market interest rate was high (2-3 per cent per month, Lin et al. 1996: 30). Foreign exchange was scarce and expensive because exportable goods were scarce and consisted mainly of low-priced agricultural products. Most significant was that economic surplus was small and dispersed across the country due to the agrarian nature of the economy. These facts imply that a spontaneous accelerated development of the capital-intensive industry in the economy would be impossible. In order to bring this about, a set of distorted macro policies such as low interest rates, overvalued exchange rates, low input prices and low wage rates were put into effect. The basic assumption behind the policy choices was that the low prices of input factors would enable industrial enterprises to create large enough profits for state-fixed investment and capital accumulation. If the enterprises were privatelyowned, the state could not be sure that the private entrepreneurs would invest the policy-created profits in the intended projects. Thus, private enterprises were soon nationalized and new key enterprises were stateowned to secure the state's control over profits for investment in heavy industry. Meanwhile, to make the low-wages policy feasible, the state had to provide urban residents with cheap food and other necessities, including housing and clothing. As such, the distorted macro policies would create overall supplydemand imbalance for credit, foreign exchange, raw materials and other living necessities. Non-priority sectors would compete with the priority
Integrating Selected Theories Based on China's Experiences
53
sectors for low-priced resources. Thus, in order to avoid these consequences, plans and administrative controls were employed to replace markets as the basic mechanism for allocating credit, foreign exchange, labour, raw materials and living necessities. In this way, scarce resources could be used for the planned projects. This development strategy, together with the resultant policy environment and allocation system, reshaped the nation's farming institutions. In order to secure cheap supplies of grain and other agricultural products for industrial inputs and low-priced urban rationing, a compulsory procurement policy was imposed on the agricultural sector in 1953. This policy forced peasants to sell certain quantities of their products to the state at government-set prices. In addition to the provision of cheap food and input for industrialization, agriculture was also the main foreign exchange earner. In the 1950s, unprocessed agricultural products made up more than 40 per cent of all exports. While processed agricultural products are also taken into account, agriculture generated more than 70 per cent of the foreign exchange earnings in the 1950s and more than 60 per cent up to the 1970s (Table 4.2). Foreign exchange was just as important as capital as a constraint for the heavyindustry development. For a long time, China's capacity to import capital goods for industrialization clearly depended on agriculture's performance. Agricultural development requires as much resources and investment as industrial development does. In order to keep agriculture from competing with industrial expansion for resources, the government adopted a specific agricultural development strategy. The core of this strategy involved the mass mobilisation of rural labour to work on labour-intensive projects, such as irrigation, flood control, and land reclamation. Unit yields were raised through intensive and meticulous farming, such as closer planting, more careful weeding, and the use of more organic fertilizer. The state believed that the collectivization of agriculture would perform these functions. It also viewed collectivization as an institutional guarantee for its low-priced procurement programme involving grain and other agricultural products (cf. Mao 1953, published 1977: 116-20; Chen Yun 1952 & 1953, published 1984: 159-60 & 202-10; Chen, X. 1993: 25-35). This distorted macro-policy environment, planned allocation system and induced institutional arrangements enabled the maximum mobilization of resources for the development of heavy industry in a capital-
54
Chapter 2
scarce agrarian economy. As state-sector investment dominated gross investment after 1953 (cf. note 2 of Chapter 1), the pattern of state investment is the best indicator of bias in the official development strategy. Despite the fact that more than three quarters of the population lived on agriculture, agriculture received less than 10 per cent of state investment in the period of 1953-85, whereas 45 per cent went to heavy industry {Statistics on Fixed Investment in China: 1950-1985, 1989: 97). This analysis indicates that the fundamental mechanism of China's capital accumulation was not characterized by lowering agriculture's terms of trade with industry alone, and thus will not be apparent in comparisons of trade balance and financial accounts of these two sectors, which is customary in the literature.15 From these kinds of accounts alone, it is hard to calculate the extra profit of the industrial sector stemming from depressing wage rates and from using low-priced agricultural products as inputs. Between 1952 and 1978, wage rates in China were kept fairly constant, increasing by only 12.7 per cent in real terms, while the real national income per capita nearly tripled (Yearbook 1993: 132, 33-4 & 81). Even in 1992, the output value of those parts of light industry that mainly used agricultural products as raw materials still accounted for 68 per cent of the total (cf. Table 4.1). Directly or indirectly this extra profit may have provided the most significant proportion of finance for industrialization. This capital accumulation mechanism may be summarized as follows: The heavy industry-oriented development strategy, distorted macro-policy environment, planned allocation system and induced institutional arrangements all together guarantee low wages and low prices for inputs in the non-agricultural sector. The resultant low income and consumption of both peasantry and non-agricultural workers plus the high profit in the non-agricultural sector contribute to high capital accumulation. The essentials of this mechanism appeared to have changed little until 1994, when a powerful system entitled 'provincial governor's grain bag responsibility system' (midaizi shengzhang fuzezhi) was enforced. Although agriculture's terms of trade improved significantly between 1979 and 1984, the compulsory procurement system in rural areas and a food rationing system in urban areas have functioned effectively. The government tried to stop low-priced rationing in 1993, but revived it in most cities in 1994 and 1995 (Crook 1998). Most significant is that the urban and industrial biases of local governments have
Integrating Selected Theories Based on China's Experiences
55
been strengthened by the intense expenditure pressure they face as well as by a tax system that depends on industry for the generation of over two-thirds of total revenues (cf. section 3.7; also, Guan & Jiang 1990, Naughton 1986, Wong 1987 & 1991). As a result, agricultural policy has been guided by budgetary considerations and local industrialization drives, rather than by the desire to promote efficient growth in agriculture. The combination of budgetary and inflation pressures encouraged the government to lower agriculture's terms of trade, to reduce other material rewards whenever possible and to increase procurement prices and other incentive rewards only when absolutely necessary. As a consequence, by 1989 the ratios of market to state procurement prices were 3.2:1 for rice, 2.1:1 for wheat, and 2.3:1 for corn (Sicular 1989: Table 5). IOUs (dabaitiao) were commonplace in the process of procuring agricultural products. The tied sales of fertilizer and diesel fuel for grain contracts were often 20 per cent below the levels promised in the central directives (Central Rural Policy Research Office 1988; China Information Daily, 18 August 1993, pp. 1-2; China Youth Daily, 8 February, 12 April 1993; Sicular 1993). It is not possible for a huge developing economy to enjoy rapid dvelopment and industrialization while agriculture is depressed and shrinking. As will be demonstrated in Chapter 4, swings in agricultural production and marketing were followed by dramatic swings in industrial capital accumulation and investment. Agricultural surpluses were followed by major shifts in intersectoral distributive policy and by significant reductions in incentive rewards to the agricultural sector. At the same time, agricultural shortfalls caused a passive reversion to increases in incentives and prices, and also caused cutbacks in industrial capital accumulation and investment. This linkage is consistent with Kaleckian agricultural determining growth theory, and demonstrates that the mark-up rate in Kalecki's profit formation equation (cf. FitzGerald 1993), derived from the degree of industrial monopoly, is endogenous.
2.3.4 Equilibrium growth path and fluctuations Integrating the Hungarian School's fluctuation analysis with Kalecki's bottleneck constraint type of growth theory requires a reconciling of their main difference, which lies in their understanding of 'normal
56
Chapter 2
growth path' or 'equilibrium steady growth path'. In the Hungarian School's works, the 'normal growth path' is a univariate concept, and usually takes the form of a moving average. This makes it difficult to deal with the relationship between the 'normal growth paths' of two variables. Statistically, the univariate moving average may produce spurious cycles, as pointed out long ago by Slutsky (1927) and Frisch (1928) (see Morgan 1990). In contrast, Kalecki's equilibrium steady growth path is a kind of equilibrium relation among growth-related variables, similar to the recently defined equilibrium relation - cointegration - which presents the long-run relationships among these variables.16 Based on this comparison, the Hungarian School's fluctuation analysis can be reinterpreted in a more rigorous way, as follows.17 The united 'norm' path among real investment and representative bottleneck constraints can be thought of as the equilibrium steady relation (hyper-plane) among these variables. Let the new interpretation start from the supply side. If the supply of the representative bottleneck sectors - in China's case energy and agriculture - increases along the equilibrium steady growth path, the normal demand of real investment for them can be satisfied, and real investment thus progresses smoothly under normal bottleneck constraints or normal shortage states. Once the supply of some bottleneck sectors, say, agricultural products, exceeds this equilibrium growth path, the shortage of agricultural products is alleviated, the 'bottleneck' is widened, and thus real investment can be accelerated. If the supply falls below the steady growth path, the shortage is strengthened, the 'bottleneck' is narrowed, and thus the growth rate of real investment has to be cut back. Starting from the demand side, an investment drive initiated by a political/economic campaign may induce real investment to exceed this equilibrium relation hyperplane. In the short term it is possible to support such an investment surge by drawing on stock and overloading the supply capacity of the bottleneck sectors. However, sooner or later, shortages in the bottleneck sectors will be sharpened by the exhaustion of stock and the overloading of supply capability. As a consequence, the 'brake may suddenly be applied' by the central authorities and the administrative hierarchy in their struggle to cope with emerging danger signals. Both growth rates of planned and actual investment outlays will then drop, and sometimes even become negative (cf. section 2.2.2).
Integrating Selected Theories Based on China's Experiences
2.4
57
Existing Researches on China's Investment Cycles
2.4.1 The efforts to link investment cycles to agricultural fluctuations There have been continuous efforts to detect and empirically prove the interaction between agricultural fluctuation and macroeconomic adjustment in general and investment modification in particular, in the context of China. However, the picture given by the literature is ambiguous. Eckstein (1968) undertook the pioneering work on linking China's economic cycles with agricultural fluctuation. Based on a survey of economic data from 1949 to 1966, Eckstein found a cyclical pattern of economic fluctuations which was 'generated by the interaction of a harvest cycle and a policy cycle'. According to his theory, economic cycles were produced by the tension between the communist regime's economic objective of rapid industrialization and the predominantly agricultural economy, which is subject to harvest fluctuations. The concrete process of the interaction is as follows. A good harvest creates new, mobilizable resources, which planners can appropriate and allocate for investment expansion as resources under their control increase. This appropriation produces strong disincentive effects in the agricultural sector by reducing individual rewards, eventually creating a crisis in the agricultural sector (see also Skinner & Winkler 1969). Planners are forced to reduce their resource extractions from agriculture, which results in the reduction of investment. In order to detect the causes of agricultural fluctuation, Tang & Huang (1982) estimate the total factor productivity of China's agriculture using both the Solow method and the ratio of gross output value to the aggregate input index. They apply weather dummies to the former method and the standard time-series decomposition technique to the latter in order to distinguish between influences of weather and policy. Agricultural productivity cycles are found to coincide precisely with the known policy or political cycles from 1952 to 1979. This finding suggests that peasants have historically responded in a rather predictable and speedy manner to Beijing's policy gyrations that impinged upon them economically.
58
Chapter 2
Naughton (1986) attempts to retest Eckstein's theory based on the data from 1953-83. He concludes that 'while there is some evidence of this in the 1950s, attempts to attribute a causal role to agricultural production or procurement trends in the generation of investment fluctuations after 1958 were completely unsuccessful' (Naughton 1986: 157). Naughton (1987) argues that since the tests of agricultural production or procurement fail to provide meaningful predictive power to the changes in investment, the hypothetical investment cycles are not linked with harvest fluctuation, but respond to the condition of the consumer goods market. Investment expansion follows a period in which the supply of consumer goods is adequate to meet the demand generated by household money income, and an investment contraction sets in after the emergence of a significant shortage of consumer goods. The mechanism that links investment cycles to the shortage intensity of consumer goods means that investment expansion (contraction) leads to higher (lower) employment and larger (smaller) household money income, and these developments augment (restrain) demand for consumer goods. His tension indicator, as a measurement of repressed inflation, had lost its relevance by the late 1970s, and thus Naughton estimates the planner's reaction function for the 1956-78 period with annual data. His estimated equation is as follows (usual standard errors in parentheses): Igr=
52.76 - 0.14/ - 0.21 X M (16.49) (0.89) (0.05)
(2.10)
R 2 = 0.468 This means that the percentage growth of state real fixed investment, Igr, is regressed on a time trend, /, and an index of the relative degree of shortage of consumer goods in the preceding year, X M . 'The shortage index was the only time series found which predicted changes in investment; tests of agricultural production or procurement, energy supply, and foreign trade balance failed to provide any predictive power' (Naughton 1987: 342). Imai (1990, 1994a) develops and extends Naughton's work by elaborating a macroeconomic model to derive the adjusted inflation rate, which is the market-clearing price of consumer goods in the model and thus a combination of both open and repressed inflation. His estimated planner's reaction function is (usual /-statistics in parentheses):
Integrating Selected Theories Based on China's Experiences Igr= 1 7 . 5 9 2 - 1.906 OP g r ) M + 67.760 z (5.11) (-3.78) (5.01)
59
(2.10a)
Adjusted R2 = 0.567, D.W. = 1.888, Period = 1955-91 Thus the percentage growth of state real fixed investment is regressed on the previous year's adjusted rate of inflation, (P gr)t-u and z, which is a dummy variable associated with the Great Leap Forward (1958 = 1, 1961 = -1) (Imai 1994a: 207). While a significant improvement is made, some statistical weaknesses remain in the work. For example, the relatively higher R2 can be attributed to the dummy variable because of the robust outlier property of the data in 1958 and 1961. By removing dummy variables associated with 1958 and 1961, the adjusted R2 in equation (2.10a) will become 0.246. The graph of actual and fitted values in the estimate shows that actual and fitted values both move in the same direction in the periods of 1957-67 and 1987-91 only, and that the errors are significant (see also Imai 1994a: 207).18 In order to overcome the statistical obstacles appearing in the existing works and to make advances in these fields, we may need to go beyond the common linear regression and to detect and examine the main proportional relations in the economy which the decision-makers are most sensitive to. We also need to construct a relatively formalized framework to highlight historical traces or dynamics of these main proportional relations which have been recognized by Chinese leaders (e.g. Mao's famous ten major relationships; see Mao 1977). These are considered in Chapters 4 and 6.
2.4.2
Reform cycle theory and the persistence of substitution between growth and bottleneck
There is a vivid Chinese expression commonly used in official documents, newspaper articles, and academic papers to describe the reform cycles in China from 1978 onwards, namely, 'yifangjiu huo, jihuo jiu luan, yiluanjiu shou, yishoujiu si\ It means that relaxing restrictions and reducing state intervention stimulate economic dynamism and bring a boom. However, this is accompanied by disturbances, even disorders, which lead to intervention and control. These in turn lead to bust. This boom-and-bust cycle is the subject of intense discussion and research among Chinese economists, and has also interested Western researchers.19 While most researchers focus on describing the dilemma
60
Chapter 2
between decentralization and centralization, the conflicts of interest between central and local governments, and the soft budgetary constraint of the state enterprises, Lin et al. (1995, 1996) establish a comprehensive and theoretically consistent framework to analyse the formation mechanism of the boom-bust cycles. China's traditional economic system is characterized by a trinity consisting of the distorted macro policy environment, the planned resource allocation system and the puppet-like micro-management institution (as analysed in section 2.3.3). This was adopted to facilitate the priority development of heavy industry when China was a capitalscarce economy. To lower the costs for this development, the state artificially suppressed the prices of credit, foreign exchange, energy, raw materials, labour and living necessities. This distorted macro policy environment resulted in chronic shortages. The resource allocation system adopted to secure scarce resources for the development of heavy industry justified both state ownership and the creation of the people's commune, which were seen as the institutionalised guarantee to ensure that the surplus would be used according to the state's strategic directions. This meant that the micro unit was deprived of managerial autonomy. Although China's reform is officially based on a philosophy of experimentalism, which was vividly expressed by Deng as 'go a step and look for the next' and 'feeling for stones to cross the river' {mo zhe shitao guohe), the overall direction of reform has been fairly clear ex post facto. The main problems in the traditional trinity appear to be low efficiency arising from structural imbalance and poor incentive. The government had sensed this long before the late 1970s and had made several attempts to address the structural problems by decentralizing the allocation mechanism from central to local governments. However, the administrative nature of the allocation mechanism was not changed and the policy environment and micro management institution were not altered. Therefore these attempts failed to rectify the structural imbalance or improve economic incentive. The reforms in the late 1970s, also intended to rectify structural imbalance and improve incentives, differed from previous attempts in that they included enterprise reforms. The primitive intention of enterprise reform was to improve work incentives by allowing managers to provide workers with a more direct link between their efforts and rewards, by delegating administrative power, and by allowing a certain portion of profits to be retained in
Integrating Selected Theories Based on China's Experiences
61
state enterprises. This was a small concession, but it eventually led to the dismantlement of the traditional system. Reform proceeded in a logical manner. It started from the micro management institution. The improvement in the micro incentive mechanism greatly increased workers' and managers' work enthusiasm and thus labour's contribution to economic growth. Prompted by competition and profit motives, enterprises paid more attention to market conditions in their choices of products and technologies, and the change reversed the trend of negative total factor productivity and the old pattern of growth which depended mainly on increasing inputs (Lin et al. 1996: section 5.2, Jefferson & Rawski 1994, Yang 1991). The benefits of the micro reform and the increase in enterprise autonomy put pressure on governments to reform the resource allocation system. The rigidity of the planned allocation system was relaxed and later reformed to allow the newly generated resources to be allocated to sectors which had been suppressed under the traditional economic system. As a result, the economic structure improved and the growth rate accelerated (Lin et al. 1996: section 5.3). Finally, when conflicts arose between the distorted macro policy environment and the reformed micro management institution and resource allocation system, reforms were extended to macro policies as well. However, the state has not yet abandoned the heavy industryoriented development strategy, and still attempts to maintain low input prices in order to protect the large and medium-sized state enterprises that embodied the goal of the traditional development strategy. This means that reforms in macro policies, especially those regarding the interest rate, have lagged behind the reforms in the allocation system and micro management institutions. Constrained by such a rigid macro policy environment, the reform in the allocation system has to follow a 'dual track' plan/market system where there are both planned and market prices as well as transactions. Thus, the micro management units look to reap benefits both from bargaining within the bureaucratic hierarchy as well as from the market. At the same time, government coordination becomes 'keeping the planned partition under strict control and letting the rest go free' {guansi yikuai, fanghuo yikuai). In this context, relaxing administrative control, or even slackening credit control, will result in a rapid expansion of credits and investment, and lead to an overheated economy.
62
Chapter 2
Although the generating mechanism of the boom-and-bust cycle has departed from the one functioning in the pre-reform period, the most pronounced feature of this cycle continues to be the persistence of investment hunger and recurring bottleneck constraint to investment growth. Ever since local and enterprise interests have been strengthened by reform, their incentive to expand both output and profit have become stronger. Confronted with low-priced input factors, all enterprises (including local governments) were motivated and stimulated to obtain more credit to expand investment and production until the shortage of basic supplies (such as energy, transport, and raw materials) became widespread and inflation appeared. Due to political constraints, the government was reluctant to make the macro policy environment consistent with the liberalized micro management and resource allocation systems by, for example, increasing the interest rate to check investment. It resorted instead to centralized rationing of credits and raw materials and to direct control of investment projects which meant a return to the planned allocation system. The rationing and controls allowed the state sector a priority position, shortage and inflation were reduced, and a bust followed. The products and services of the energy, transport, raw material and other basic sectors have relatively smaller price elasticities of both demand and supply. If the heavy industry-oriented development strategy remains largely unchanged, there will be no effective price stimulation to remove the bottleneck constraints from the basic sectors. Yet the state has continued to protect large and medium-sized state enterprises by supplying the low-priced input factors. On the other hand, the basic industries stand in the upper reaches of the industrial chain in the national economy, so increases in their prices will have a significant impact on all other industries. The state has therefore lacked incentive to risk releasing price control in these basic sectors. When the state had to partially adjust prices in the basic sectors, it seemingly had to allow the lower reach industries to modify their prices by a lag mainly because it lacked effective control measures. The result has been the restoration of the original relative prices which meant that the relative prices of basic industries remained located at the bottom of the industrial price structure. This has led to continuing shortages of energy and other basic products, and the growth rate of the national economy has been forced to accommodate that lower level determined by bottleneck constraints (Lin et al. 1996: section 7.1).
Integrating Selected Theories Based on China's Experiences
63
One consequence of the inconsistencies between the distorted policy environment, the liberalized allocation system and the micro management institution is rampant rent-seeking. During the reform period a market price existed, legally or illegally, along with a planned price for almost every kind of input or commodity under state control. In 1992, the difference between the market price and the planned price (economic rent) for bank loans alone reached 220 billion yuan (Hu 1994, Lin et al. 1995: 12), about 10 per cent of the national income. Both non-state and autonomous state enterprises had incentives to seek rents through bribes and other measures from the state allocation agencies. Due to rent-seeking, state enterprises were often unable to obtain the credits and materials promised in the plans. Rent-seeking also caused widespread public resentment and became a source of social instability. To guarantee the survival of the state enterprises and to check social resentment, the government often attempted to reinstitute tight controls on the allocation system, but again, these controls caused growth to slow down. It needs to be recognized that Lin et al.'s analysis may be oversimplified and that there is a development strategy bias and a market bias. They did not pay much attention to factors such as interest rigidity in the state sector, political risks linked to a radical reform of the state sector, the territorial barrier and industrial monopoly and other political economy issues. As a result they came up with a rather abstract and optimistic policy suggestion - 'the core of economic reform is the change of development strategy' from the heavy industry-oriented strategy to the comparative advantages strategy. Although their perspective is ad hoc, they did supply a clear and theoretically consistent view of the logical process of the reform cycles in China and they also highlight the persistence of investment hunger and bottleneck constraints.
2.5 Insights from Western Business Cycle Theories The business cycle has been one of the central subjects of economic theories dealing with Western market economies. Countless works have appeared in the literature and some of the ablest minds in the history of economics have written about business cycles in the West. Following the comparative surveys conducted by Fischer (1988) and Dore (1993), the main contending approaches can be defined as the equilib-
64
Chapter 2
rium business cycle approach, the real business cycle approach, the New Keynesian approach and the endogenous cycle approach. Their common aim is to explain the stylized facts about the integral relationship of output, inflation, unemployment and wages over time in Western economies. The equilibrium business cycle approach is represented by the Lucas (1973) Philips curve model. The distinguishing features of this model are that it builds the Philips curve on rational expectations and asymmetric information, and that it emphasizes a monetary cause of economic fluctuations and the trade-off between output and inflation. Later on, the Lucas monetary misperception model (1981, 1987) shows that the governing mechanism of the cycle is misperceptions about prices caused by money supply growth. As pointed out by Fischer (1988), this approach seems to be caught between the theoretical difficulty of finding a route through which monetary policy can affect output and that fact that changes in aggregate demand for money do affect output. In contrast, the real business cycle models (cf. e.g. McCallum 1989, Prescott 1986) attempt to produce a business cycle theory in a purely non-monetary framework. The most distinguishing feature of the real business cycle approach is that it has followed the stochastic approach of Frisch (1933) and Slutsky (1937) and has been closely associated with the advancement of modern econometrics. This feature facilitates econometric implementation, and makes it possible to address and clarify issues that were simply too difficult to tackle before. The real business cycle models distinguish the shocks that strike the economy from the propagation mechanisms which convert the shocks into longerlasting divergences of economic variable from their steady state values. This theory emphasizes that virtually all business cycle phenomena are the result of productivity shocks impinging on an economy where markets are continuously in equilibrium. Productivity shocks may be treated as a combination of permanent changes that come from improvements in knowledge and transitory changes such as those brought about by natural disasters. The New Keynesian models are fairly diverse but focus their attention on obtaining Keynesian results (e.g. an underemployment equilibrium with involuntary unemployment) from microeconomic, representative agent models. The fundamental assumptions are the short-run non-neutrality of money, and imperfect competition in the goods, la-
Integrating Selected Theories Based on China's Experiences
65
bour and credit markets. The governing mechanism of the cycle is the oligopolistic nature of the economy (cf. Dore 1993, Mankiw 1985 and Mankiw & Romer 1991, among others). The endogenous cycle approach attempts to prove that business cycles are inherent to a free enterprise economy, a notion already stressed by Schumpter. This is a non-Walrasian analysis in that it rejects the mythical auctioneer. In a non-Walrasian model without an auctioneer, trade takes place in sequence and agents will often be quantity-constrained so that both goods and labour markets may fail to clear. The approach is thus open to a different market structure, and the governing mechanism of the cycle is the interaction between the goods and labour markets (cf. e.g. Benassy 1986, Goodwin 1967, Goodwin et al. 1984, Di Matteo 1984, Semmler 1989). Undoubtedly each of the abovementioned models is helpful to our understanding of specific characteristics of a typical capitalist economy. In fact, each of them (despite certain shortcomings) adequately explains a good number of the stylized facts about the integral relationship between output, inflation, unemployment and wages over time in Western economies.20 However, because of the significant differences in the major stylized facts between developed market economies and a socialist developing economy like China, we may need to avoid applying the Western business cycle theories automaticallly to China. Instead, we need to discover thought-provoking insights from these theories and to integrate them into our methodological framework. There are three major differences in stylized facts between developed market economies and a socialist developing economy like China. First of all, in China the agrarian question has played a central role in the developing economy. There is an agricultural constraint upon industrial growth in both the short and long run. If peasants are subject to disincentives due to institutional arrangements, there is a ceiling to the growth rate of real wage goods from the agricultural sector. Once the economy hits its ceiling, the agricultural constraint becomes dynamically binding unless the real wage rate can be pushed down infinitely. While the agricultural growth rate can be raised to any desired level, a price has to be paid in terms of the capital-output ratio because of limited land supply, as will be analysed in Chapter 4. Furthermore, not only would a higher growth rate require a higher capital-output ratio, but also sustaining this growth rate over time would necessitate a rising capital-output ratio. This indicates that an increasingly intense scramble
66
Chapter 2
for scarce modern resources between agriculture and industry has shaped and may continue to shape the overall process of China's industrialization in general and its business cycles in particular. All these agrarian questions are clearly beyond the stylized facts of business cycles in the West. Secondly, it has been established in the literature of comparative economics that a socialist and/or developing economy is typically resource-constrained (RC) whereas a typical Western capitalist economy is demand-constrained (DC) (Kalecki 1970, 1972, 1976; Kornai 1979, 1980). A typical socialist firm is RC because it would produce more output and/or invest more if additional quantities of certain material inputs were available. The physical limits implied by the input-output relationship are nearly always effective in that they serve as the binding constraint on the firm. However, the demand constraint is seldom effective as an increase in a firm's demand rarely leads to an increase in its output. On the other hand, a typical capitalist firm is DC as it would sell more by utilizing its productive capacity more fully if its demand curve were to shift outward. Excess capacity, availability of inputs and sensitivity to prices are inherent characteristics of the capitalist firm. In comparison with the general interpretation of Kalecki and others, Kornai's presentation of the concepts of RC versus DC economies is well established. He strongly emphasizes the nature of a firm's budget constraint. An 'almost-hard' budget constraint forces a firm to worry about survival, to accept responsibility for its own investment behaviour and to adjust to price changes. On the other hand, a soft budget constraint implies that a firm's survival and growth depend critically on firm-specific decisions made by governments through various forms of vertical bargaining. Verifiable conditions for an 'almost-hard' budget constraint are price-setting within narrow limits; scarcity of government subsidization viafirm-specifictax, credit, and other redistribution policies; loans based on orthodox creditworthiness indicators; no automatic external financial assistance. The corresponding verifiable conditions for a soft budget constraint are substantial price-setting behaviour by firms; the availability of tax exemptions, government subsidies and external financial assistance to offset losses; loans that are not based on orthodox creditworthiness indicators (cf. section 2.2.1, Marrese & Mitchell 1984). The basic behaviour of major economic actors in RC socialist economies are bound to be quite different from those in typical capitalist
Integrating Selected Theories Based on China's Experiences
61
economies. For instance, employees in China's state sector had enjoyed the advantages of life-long job security until very recently. The employees in Eastern European socialist economies have enjoyed the privileges brought about by chronic serious labour shortage. This also contrasts sharply with the situation of persistent unemployment in the West. Furthermore, firms in a RC economy do not pay much attention to minimizing cost because they operate with budget constraints softened by government assistance. As a consequence, their input costs do not necessarily dictate forced substitution of inputs or outputs. This also undermines a direct application of the Western firm theory. The most relevant example is the 'investment hunger' which explains longrun behaviour of the RC firm. As mentioned in sections 1.1 and 2.2, investment hunger constantly pushes up real investment to overshoot the binding resource constraints and can only be suppressed by retrenchment campaigns. Such a governing mechanism of investment cycles is clearly socialist-specific. The third difference lies in the fundamental mechanisms coordinating investment and other economic activities in socialist and capitalist economies (Kornai 1992: Chapter 6). In a typical capitalist economy, the fundamental coordination mechanism is the market even though it may represent far from perfect competition and its structures may differ among the different markets for goods, labour, credits and so on. For instance, firms buy labour services from the labour market, raise investment funds in financial markets, and supply products to the goods market, which are all mainly based on profit-maximization considerations. Households purchase goods from the goods market and supply labour services to the labour market based on utility-maximization considerations. In a socialist economy, bureaucratic coordination dominates economic activities. During the pre-reform period, the relations between firms had been bureaucratically coordinated as there was no financial market and the goods market was insignificant. Although reform has introduced significant market relations between firms, China's investment resources are still mainly allocated to the state sectors and through bureaucratic coordination.21 In terms of employment, there is no free contract between employees and state firms. Employees enjoy the advantages of life-long job security but at the cost of giving up 'exit rights'. The governing mechanisms of the cycle propagated in the major Western schools of thought are all based on the market coordination mechanism, and theories differ only in interpretations of the market and
68
Chapter 2
relations between markets. None of them consider the bureaucratic coordination mechanism. While recognizing the inadvisability of directly transplanting the Western business cycle models into the East, their methodological advantages in general and their advanced econometric implementation in particular supply powerful means for modelling investment cycles in the East. In fact, the cointegration and the error correction model approach has developed under the strong stimulation provided by Lucas' famous econometric policy evaluation critique and by its interaction with the development of real business cycle models (see, e.g. Engle & Granger 1991, Spanos 1995). In addition to learning from the real business cycle modelling, the idea of integrating investment 'growth' with 'cycle' through singling out the underlying equilibrium growth path and modelling the error correction is also partly thanks to Goodwin's growth cycle model. Goodwin's model was the first to incorporate both growth and cycle and has received wide recognition as perhaps the most important business cycle model since Schumpter (Velupillai 1990, Dore 1993). In his model the cycle is endogenously generated, and the cycle and growth trend are integrated. Neither the growth trend nor the cycle can be independently analysed. The model reflects all the major stylized facts of the Western business cycles listed in Dore (1993: Chapter 2). It strongly emphasizes the fact that each subsequent cyclical trough of those macro indices such as real output, real disposable income, real wage, etc. does not cancel out the 'gains' made by the previous cyclical peak. It explains the movements of output, employment, inflation and money in one consistent framework. It shows that growth and cycle are inextricably intertwined and that the cycle is a necessary condition for growth and vice versa. These insights are fully consistent with the recent findings of econometrics. These indicate that non-stationary time series, as are most macroeconomic time series with growth over time, cannot be separated into its trend and cycle components except in a purely arbitrary manner (Nelson & Plosser 1982, Stock & Watson 1988). The economic logic behind this has been intuitively demonstrated by Goodwin as early as 1953. There is no such thing as a trend factor that continues to rise right through the depression. Conversely, the cycle would be quite different were it not for these special relations which give rise to the trend. Therefore we cannot separate out trend and cycle and discuss each independently. (Goodwin 1982: 117)
Integrating Selected Theories Based on China's Experiences
69
2.6 Summary: The Implications for Modelling Investment Cycles in China This chapter contains an attempt to extract the most relevant theories from the literature and to apply them to the specific experience of China in an integrated manner. These theories include first the investment demand and fluctuation analyses of the 'Hungarian School', which are based on notions of soft budget constraint, expansion drive and investment hunger. In particular, Bauer's four-phase description of investment cycles in a socialist economy also appears to fit the patterns of and the linkage between investment ratio and growth rate cycles in China. The theoretical and statistical difficulties inherent in the traditional modelling approach based on 'single control equation' or 'single response equation' calls for a major improvement on this approach based on an integrated framework and advanced econometrics. Some difficulties confronted in applying the single response equation approach are the following: (a) There is a lack of intertemporal rationality of the representative planner. (b) The investment commitment is a mixture of actual and planned expenditures at mixed price level. (c) Investment vintage is based on engineers' calculations and is always underestimated. (d) The indivisibility assumption of a macro-investment decision may involve the 'fallacy of composition'. (e)The simultaneous interactions between investment tension and shortage intensity make the signs of shortage feedback coefficients uncertain ex ante. (f) Using 'unfinished construction' as a proxy for shortage in the investment goods sector may be misleading and may only reflect that type of negative correlation coming from the statistical formulas of construction periods. (g) The weakly exogenous assumption of shortage intensity variables has to be tested at the beginning. (h) The 'normal path' based on a univariate concept and technique may produce spurious cycles and still faces the difficulty of dealing with
70
Chapter 2
the relationship between the 'normal paths' of two or more variables. It would be very difficult to overcome the majority of these problems before significant technical progress in both business cycle theory and econometrics had been made. Kaleckian growth theory and Lin et al.'s reform cycle theory highlight the persistence of the trade-off between growth rates and bottleneck constraints. An analysis of China's specific capital accumulation mechanism which incorporates Kaleckian and Lin et al.'s theories shows that China's capital accumulation mechanism is not characterized by a lowering of agriculture's terms of trade alone. A significant part of capital accumulation in China has gone beyond the common accounts of trade balance and financial flows between the agricultural and industrial sectors. The extra profit of the industrial sector stems mainly from depressing the wage rate and the prices of agricultural products. This extra profit has contributed greatly to China's high capital accumulation. Consequently, the conflict between a capital accumulation drive and the requisite agricultural development links China's macroeconomic adjustment and relevant investment cycles much more closely to agricultural fluctuation than elsewhere. The complementary comparative advantages of the Hungarian School, Kalecki's growth theory and Lin et al.'s reform cycle theory may all be integrated into a more comprehensive framework. The insight gained from Goodwin's growth cycle model and recent technical progress from both Western business cycle theories and econometrics also make such an integration feasible. This integration enables us to model the long-run comovement path and short-run two-way interaction between real investment expansion and the constraints from agricultural and energy supply. It also enables us to empirically test the possible structural breaks that may occur as a result of reform.
The State Investment System and its Response to Reform
3.1 Introduction This chapter presents a political economy analysis of the development of China's state investment system and its response to reform. Political economy analysis as used here means that the focus of the analysis is on the orientation, interests and interest conflicts among the different economic authorities and agents, and on showing how these have shaped the processes of concrete investment decision-making, financing and implementation. The main focus is on fundamental forces and interaction mechanisms shaping the investment system rather than on the miscellaneous policy details and related shifts. The conflicts of interest have existed not only between central and local governments, but also among different bureaucracies of the same rank (mainly planning, fiscal, banking and specific industrial bureaucracies). On the other hand, there has been widespread collusion among local government agents, locally run enterprises and local branches of state banks, as well as between enterprises and their immediate supervisory agencies. All aim to avoid project approval requirements and procedures and to obtain loans outside the credit plan. The complicated interaction between interest conflicts and strategic collusion has shaped the correlation between the recurring cycles of decentralization and re-centralization and the investment cycles led by local or central government. Hopefully, this political economy analysis will help to understand the reasons for ubiquitous growth drive and investment hunger, how growth drive is translated into investment boom, and what induces the central government to initiate retrenchment campaigns and to reinforce these campaigns by ad hoc administrative measures and Party discipline.
71
72
Chapter 3
The chapter is set up as follows. Section 3.2 overviews the history of the state investment system. Sections 3.3 to 3.5 present the basic characteristics of the project approval system and the relevant changes in approval power distribution. These sections also describe how governments at different levels control financing resources by formal credit plans as well as by informal measures. The fundamental function of the material supply system and its change during the reform period are also described. Section 3.6 analyses the worsening of the soft budget constraint in state-owned enterprises and the consequent investment hunger. Section 3.7 examines the development drive and investment hunger of local governments. The key conclusion, i.e. that insatiable investment demand exists at all levels and in both pre- and post-reform periods, is summarized in section 3.8.
3.2 A Historical Overview of the State Investment System In the standard model of the centrally planned economy, control of the investment process is far more strictly centralized than day-to-day production. Following the implementation line of the national economic plan, the national plan distributes investment funds among the various ministries in the course of the planning and decision-making process. The disaggregation of funds proceeds from the top downward through the hierarchy of national economic control to the state-enterprise level. The quotas are set in value terms and are usually broken down into a few main items of expenditure, such as construction, purchasing machinery and equipment, etc. For some priority projects, decisions are made separately. The centre determines what the installation will produce, what technology it will use, where it will operate, when it should be ready and how much it should cost. Following the formation line of the national economic plan, starting from the enterprises, the subordinate institution in the hierarchy of national economic control makes initial proposals and comments on drafts. If necessary it can ask for a target to be amended during the implementation of the plan (Kornai 1992: Chapters 7 and 9, Naughton 1986: Chapter 4, Eckstein 1977: Chapter 4). In spite of the enormous influence of central planners over the economy, planners' investment decisions are not automatically and exogenously imposed. Rather, planners must struggle with insufficient information and clumsy indirect measures of control to generate the
The State Investment System and its Response to Reform levels of investment they believe appropriate. Planners thus have difficulty controlling the economy and need to constantly reassess the appropriateness of investment levels due to the speed with which excessive levels of investment cause adverse effects on living standards. This puts pressure on as well as supplies incentives to the system and/or its authorities to initiate institutional adjustment and reform. This may also apply to all former European centrally planned economies, up until the mid-1960s, when some of them introduced significant economic reforms. In the 1950s, China adopted the classical investment system of the former Soviet Union. During the First Five-Year-Plan period (1953— 57), the centre directly controlled 82 per cent of capital construction {Statistics on Fixed Investment in China, 1950-85: 64). The other 18 per cent, managed by local government, was limited to agriculture, urban infrastructure, education, culture and so on. In addition, the concrete projects had to be approved by the central ministries if the project required investment of 10 million yuan or more (Guan & Jiang 1990: 199, Eckstein 1977: 153). However, China's size - combined with its remarkable regional differences, diversity and its backwardness - severely limited the ability of central planners operating from Beijing to control economic activity throughout the economy. It was recognized by the Chinese leadership in the second half of the 1950s that the economic system would need to provide incentives for local fiscal, investment and development responsibility. It would also need to allow flexibility to local governments, particularly to provinces, so that the local authorities could adapt central directives to highly varied local conditions (Mao 1956). From that time on, reconciling the conflicting needs for central control and local latitude has been one of the most striking features of China's economic institutional development. Although from one period to another, policy emphasis has swung between decentralizing and re-centralizing, planning has been made not only at central government level, but also at various local levels including province, city, county and even commune (township). It should be mentioned here that given the many levels of decision-making, decentralization is a complex process and may lead to a number of outcomes. Furthermore, a reduction in the scope of central planning does not necessarily mean a shift to market allocation. The shift between decentralization and re-centralization usually depends on whether policy goals are expansion or retrenchment, and also
73
74
Chapter 3
on whether their intention is to foster local initiatives or to impose national uniformity. When the central leadership or pressure within the system favours economic expansion, local authorities receive more financial resources and greater autonomy to allocate these resources. On the other hand, recentralization is accompanied by efforts to restrict investment growth and economic expansion. The first important decentralization took place during the 'Great Leap Forward' period (1958-60). In September 1958, the Central Committee of the Communist Party of China published The Stipulation for Improving the Planning Administration System, which described plans to establish such a planning system 'based on regional comprehensive balance and coordinated by both specialized functional departments and regional authorities.' The programme consisted of 'profit retention' for most enterprises and granted power to approve projects locally, subject to a ceiling on total investment. Due partly to the fervour and dislocation of the Great Leap Forward, the programme indeed contributed to the enormous expansion of investment in 1958-60 but also to the consequent economic crisis. Thus, in 1961, the profit-sharing ratios were scaled back considerably, and in 1962 profit retention was cancelled. At the same time, approval for investment projects was re-centralized. It was stipulated that all large and medium-sized projects had to be approved by the State Council or State Planning Commission. The ability of local governments to approve projects was cut drastically, although local governments could still approve non-industrial investment (Guan & Jiang 1990: Chapter 8).1 The second important strategic decentralization process was initiated in 1965. At first, due to increased administrative complexity, investment in the technical updating and transformation of existing enterprises was no longer determined as part of a technical updating plan, but was simply allocated as a 'block grant' to localities. Around the same time, non-industrial capital construction was also converted to 'block grant' status. Combining these two grants the local governments gained control over about 20 per cent of budgetary investment (He 1984: 64). Local governments were also provided with 'revolving funds', technically bank loans under the control of the Construction Bank of China. These funds were used to grant interest-free but repayable loans to locally sponsored projects (Luo 1979). In 1967 local governments and their industrial bureaux were authorized to retain the depreciation funds of locally controlled enterprises, while block grants for
The State Investment System and its Response to Reform technical updating were cut back. From this time forward, the predominant funding source for technical updating and transformation investment became locally retained depreciation funds, all of which were ' extra-budgetary'. In 1970, the central authorities formulated the Programme of the Fourth Five-Year Plan (Draft) to establish the so-called 'Ail-Around Contract System (dabaogan)\ between central and local authorities, on fiscal revenue and expenditure, material supply, and capital construction. The contract system on capital construction meant that the local mainly provincial - governments took responsibility for capital construction targets assigned by central government. Local governments could now make overall arrangements and coordinate the allocation of investment, materials and facilities. The surplus of these allocations could be retained by the local authority. Certain key projects beyond local ability were to be dually directed (centrally and locally). In 1974, an allocation-sharing programme was introduced with 40 per cent of investment to be arranged directly by the central ministries and commissions, 30 per cent to be allocated by local governments, and the remaining 30 per cent to be co-arranged in negotiation between the central and local authorities (Guan & Jiang 1990: Chapter 8). These changes in power distribution in investment decisions contributed to the decentralization process of the state investment system from the mid-1960s to the mid-1970s. Through the successive rounds of decentralization and rapid local industrialization, the local sector grew rapidly to rival the state sector under the central control. By the mid-1970s, the local sector had developed substantial growth momentum based on resources generated within the sector itself. As a result of rapid industrialization at different local levels, large numbers of small enterprises were added. The impossibility of incorporating them into the planning structure led to the creation of a multi-tiered, regionally based system, in which much of the responsibility for planning and coordination devolved to local governments. In this system, stateowned enterprises were divided according to their importance; that is, large-scale key enterprises remained in the central plan, while non-key enterprises were left to planning and coordination at the provincial, prefectural and county levels. During the pre-reform period, China's state-owned enterprise was neither an investor nor an autonomous saver, though it did produce surplus. This means that the central and local governments had been the
75
16
Chapter 3
real decision-makers concerning investments and savings. In fact, most important microeconomic decisions resided in government agencies. National, provincial, prefectural and county authorities were guided, at least in theory, by mandatory plans. Major decisions regarding output level, product mix, source and nature of input (including the level of employment and the selection or replacement of particular workers, prices, wages, investment, sources of financing, etc.) were taken outside the enterprise. This was a far from complete and integrated central planning system however, and authority was in fact fragmented among agencies at vertical administrative levels as well as among various levels of government. There was little horizontal communication among the vertically organized hierarchies. An individual enterprise might be answerable to several different agencies at different levels of government for various aspects of its operation. Within the constraints imposed by these external decisions, enterprise managers were responsible for meeting the production targets assigned to them. Prices, products, their uses and their users were all designated by government agencies, and producers faced no competition. They had only limited contact with their suppliers and customers. Prices were seldom changed and played a principally distributive rather than allocating function. The enterprise as a whole depended entirely on some form of government authority for its financing. Profits and depreciation allowances were largely remitted to the central or local government and losses were also covered by government. New investments, whose priorities were established by the plan, were directly funded by the governments' budget and credit plans. The economic reform initiated in the late 1970s has been characterized by decentralization and dominated by local government (Zhang & Yi 1995), although there were several cases of re-centralization and top-to-bottom reforms have also played an important role (Chen et al. 1992). Many reform programmes have been initiated at local government level and even at grassroots level and then adopted as national policies by the central authority.2 Much of the planning system was dismantled by local governments. Many local governments have been far ahead of central leaders in reforming the economy. This does not necessarily mean that locally-led reform is contrary to the will of the central leaders. In fact, decentralization of reform governance was an important feature of the central leaders' style of governing the reform.
The State Investment System and its Response to Reform
11
Along with the decentralization of investment decisions, there has been a sharp decline in the proportion of capital construction investment in the state sector financed directly by the budget: from more than 77 per cent in 1978 to 35.5 per cent in 1985, 22.3 per cent in 1992 and 6.8 per cent in 1995 {Yearbook 1997: 155), although part of this decline can be attributed to the programme of transforming budget grants for capital construction into planned bank loans {bokuan gai daikuan). Correspondingly, the share of capital construction investment from retained profits of the enterprises, domestic loans and foreign loans has risen substantially. With regard to the second component of fixed investment - technical updating and other investment - it can be found that the share of this component in the total fixed investments has grown considerably in contrast to that of capital construction investments. Starting at 25 per cent in 1978, the share of the former increased to 36 per cent in 1985 and to 43 per cent in 1992 (plus 'commercialized housing construction') {Yearbook 1993: 149). The importance of this shift is not in the changed nature of the investment itself, but rather in the manner in which it is determined and financed. The actual control of technical updating and other investments as well as commercialized housing construction has rested to a substantial degree not with the central government or with enterprises, but with local governments and their industrial bureaux, which traditionally have the authority to concentrate those type of funds (mainly, depreciation funds) to carry out larger projects.3 As a result of these two trends, local governments, especially provincial and municipal governments, have become major actors in investment decision-making. Their investment decisions increasingly influence the overall structure of investment, and thus the basis for sustained growth. This decentralization process was partly counterbalanced by the development of improved skills, information and control available to the central government. In the face of constant changes and increasing complexities, the central authorities have understandably used their newly available skills and information, and this has involved the central government in new activities, despite the general decentralization process. On the other hand, the experimental and evolutionary characteristics of China's reform have left a wide range of activities open to the centre's involvement. Thus, while the volume of resources directly under the control of the central government has declined, the centre's bargaining position has in other respects been enhanced by its greater
78
Chapter 3
access to information and the wide range of instruments at its disposal. A good example of this trend is the special programme of 'national priority projects', mainly related to the energy and transport sectors, which was created by the central government and has expanded since the early 1980s. Each project in the programme is planned with 'rational time-tables and a guaranteed supply of materials' at subsidized prices. The centre concentrates its own resources on those projects, harnesses the financial resources of the state banks to them, and draws on local financial resources through a version of 'matching funds' (peitao zijiri) for them.4 Besides this new development, the central government has used traditional methods to bring local government decisions in line with national priorities. The major indirect instruments include pricing policy, the tax system and interest rates. The impact of the last two has been limited and price reform had been constrained by inflation for many years. Therefore, the central government usually resorts to the following four direct methods to influence investment decisions by local governments: (1) setting project approval limits both for capital construction investments and for technical updating and transformation investments; (2) controlling financial resources through the mandatory plan and credit quota distribution; (3) allocating raw materials through the material supply system; and (4) selecting priority sectors for investment. During the reform period, it was still impossible for the enterprises to make investments without involving government, because many factors of production such as land, water, electricity, rail transport, bank credits, etc. remained under the control of government bureaucrats. For these factors markets are absent or rudimentary. Further, local governments allocate workers and appoint managers and this gives them substantial leverage for influencing enterprise behaviour (Qian 1995: 22830, Perotti et al. 1998). With the economy moving away from the use of physical allocations and in favour of financial indicators, enterprise dependence on the bureaucracy has similarly shifted from plan bargaining to the financial sphere, i.e. bargaining over profit and credit quotas, subsidies, investment, tax reduction, etc.
The State Investment System and its Response to Reform
3.3
79
The Project Approval System and Project Approval Norms
3.3.1 The project approval system: Formal procedure versus real practice The Project Approval System has been the nucleus of the state investment system. The approval system endows the State Council with the power to control investment scales and to impose the centre's preferences on investments. The 'above-norm' projects must be approved by the State Planning Commission (SPC), whether the investments are financed by the central government, local governments, or by the enterprises themselves. As the term suggests, the project approval process is passive. SPC initiates few new projects itself, but waits for project proposals submitted by provincial industrial bureaux, line ministries and/or other entities sponsoring investment projects. SPC's role is to veto those projects that do not conform to its project appraisal criteria. Local planning commissions are also required to submit 'above-norm' projects to the SPC for approval. The project 'cycle' of the state investment system, in terms of its documented formal procedures, consists of six phases: initiation, proposal, preliminary approval, feasibility study, approval and implementation {A Collection of Documents on Fixed Investment and Construction, Singh 1992, World Bank 1988a). Small and medium-sized projects are typically initiated by the enterprises concerned, which, after obtaining the support of the relevant industrial bureau, submit their initiating proposals to the local planning/economic commission for approval and project preparation. Large projects and some small and medium-sized projects are initiated by local bureaucracies or line ministries. Project proposal and pre-feasibility studies are organized by the local industrial bureaux, are normally entrusted to a local design institute and include a preliminary financial plan. The formalized project proposal is then sent to the line ministry, to the SPC and, if its financing is sought for the project, to a State Investment Corporation (SIC). This formal proposal is usually a short document, often only a few pages in length, which gives a general overview of the project with its benefits, costs and feasibility. Preliminary approval of this document by the SPC means that the project is authorized to go ahead with ^feasibility study and, in the case of technical updating and transformation
80
Chapter 3
projects, with an inclusion in the relevant rolling investment plan or pipeline. The feasibility study is a detailed, technical document on a project prepared by a local or ministerial engineering company. At this stage, bargaining also takes place between enterprises, industrial bureaux and planning commissions, with the enterprises attempting to secure as many inputs as possible at state-planned prices. Firm commitments for the supply of key factors of production, such as land, electricity, transport and water have to be obtained. Also, the financing package needs to be agreed on, and contributions from local government/enterprise, banks, budget and/or SIC finalized. Once the feasibility study is complete, it will be submitted to the line ministry, the SIC and the SPC. The line ministry usually invites a specialized institute to undertake a project review, although at this stage this will usually be rather superficial. The line ministry then submits the review of the feasibility study as well as a recommendation to the SPC, and another recommendation is submitted by the SIC. The SPC then asks the China International Engineering Consulting Company (CIECC, owned by the SPC) to do a final review, which consists of a detailed economic analysis of the project according to the SPC's manual, using its approximately 400 shadow prices. A 'social analysis' is also carried out to consider factors like regional income distribution. The CIECC produces an evaluation and a recommendation to the SPC. Though all procedures have been adhered to, once the SPC decides to go ahead with a project, all the rest become pro forma. Once the project is finally approved by the SPC, in the case of capital construction projects a plan responsibility document is prepared. The project enters the annual capital construction plan and work on the preliminary design and design can start. With technical updating and transformation projects there will be no plan responsibility document, but at a certain stage the project enters into the annual plan. The implementation stage then begins. In contrast to those procedures outlined above, project approval process is characterized by the bargaining that pervades economic decision-making. 'Nodding approval' by some leaders at high level and in key positions (shouzhang diantou) is the most important passport for a project to succeed in official approval procedures. As a result of successive rounds of bargaining, project design as well as funding may reflect bureaucratic compromises more than a careful evaluation of al-
The State Investment System and its Response to Reform ternatives. This often results in sub-optimal project scales and underestimated project costs. As a number of different localities as well as ministries compete to get their own projects approved, the outcome of subsequent bargaining with the central government (or authorities at higher levels) is to reduce the scale of each proposal and to underestimate costs rather than to turn down some projects and maintain others at optimal scale and reasonable cost levels. In the official procedures, the SPC's authority over large and medium-sized projects has not been diminished by the decentralization of investment financing. However, in practice, there are several commonly reported bureaucratic strategies used by the increasingly financially independent local governments to evade central government controls. These include (a) breaking projects down into separate subprojects that do not require approval; (b) underestimating project costs and using state-planned prices, even if inputs need to be purchased at market prices; and (c) 'stretching' projects and presenting the first phase alone as an investment project. In fact, all these bureaucratic strategies are common in centrally planned economies, and are known in East European business parlance as 'clambering into the plan' (Kornai 1992: 164). The separation of formal procedure and real practice reflects a common paradox faced by planners in a bureaucratically coordinated system. Even disregarding these tactical distortions and other problems such as insufficient incentive and lack of real owners pointed out by Kornai (1992: 127), the vast quantity of information required for bureaucratic coordination causes serious problems. Gathering and processing huge masses of information and the subsequent coordination are all too onerous to be undertaken efficiently through centralized planning and administration. Therefore, initial project proposals, even large ones, are usually cursory, and their preparation and evaluation are necessarily perfunctory. Moreover, at this initial stage, the SPC is not presented with alternatives, but with a single proposal that its creators attempt to defend as the sole feasible and therefore 'inevitable' design. Once this is approved, a considerable amount of costly preparatory work is undertaken, making it harder to reject the project at a later stage. In the succeeding stages, because it remains impossible for the SPC to do a thorough technical and economic evaluation of complex major projects, the SPC relies almost entirely on the line ministries and local industrial bureaux which propose the projects in the first place.
81
82
Chapter 3
3.3.2 Project approval norms and locality's evasion by collusion As noted in section 3.2, China's size as well as the strong regional diversity have led to local authorities being given certain powers to approve local projects, subject to a quota on total investment. Responsibility for project approval has depended on the size and nature of the project. Capital construction projects are generally categorized as large, medium or small, mainly depending on the size of the investment (see note 1 of this chapter). In the pre-reform period, projects involving investments of Y10-30 million or more were usually considered as the medium-to-large category and required SPC approval, while very large projects and designated 'key' projects were submitted to the State Council for approval (Eckstein 1977: 153). Throughout the 1980s, approval norms for capital construction projects remained the same (see Table 3.1). However, technical updating and transformation investment seemed to be less controlled and also less controlled than investment in capital construction. Provincial and sub-provincial governments had the authority to approve almost all small projects and some medium-sized projects. Most significantly, all projects needing investment of more than Y500 thousand needed some form of government approval. In the years of retrenchment (such as 1982, 1986 and 1989), the 'free norms' were lowered and in some cases approval rights shifted by more than a single level. Table 3.1 lists only the normal approval norms as they appeared in official documents. In fact, at the sub-provincial level, the approval norms vary greatly from location to location, from sector to sector, between state and collective owned enterprises, and, in some cases, even between similar enterprises. Such variations can be attributed to specific negotiations and, to further complicate matters, SPC approval was (and is) required for all projects that receive central government funds and material allocation. There have been strong motivations for subordinate projectapplication entities to avoid project approval requirements. Two of them are to promote local growth and to create employment for their local communities. This leads local governments to collude with banks and enterprises within their jurisdiction (World Bank 1994: 48 & 120). The most popular collusion in investment planning process is that the different subordinate entities cooperate to break down projects into
The State Investment System and its Response to Reform
83
smaller components in order to get around or simplify project review and approval procedures. Other reasons for the reduction in size of investment projects include the fact that larger investments also involve the problem of obtaining larger material allocations from the centre or from a higher-level government. Larger projects also find it exceedingly difficult to procure materials independently and underdeveloped capital markets make it difficult to finance larger-scale projects through self-collected funds and loans. Finally, retained earnings in the form of production funds and depreciation allowances accrue in relatively smaller amounts.
Table 3.1
Approval norms for Investment projects in the 1970s and 1980s* Approval norms (million Yuan)
Level of Government
State Council State Planning Commission Provinces & cities directed by central planning Cities directly under provinces Prefecture, municipalities & provincial bureaux Comprehensively reformed counties Counties Notes:
Capital construction1'
Technical updating0
>200 < 10-200 >50 < 30-50 <5-50 < 10-30 <3-10 < 1-5, some 10 < 5-10, some 30 < 0.5-2.0 <5-10 < 1-5, a few 10 < 0.5-1.0
(a) Refers to total costs of individual project. (b) For productive (vs. unproductive) capital investment the 'free norms' are usually higher, for example, Y10 million versus Y5 million for provincial level. (c) For energy, transport and raw material sectors, the 'free limits' are usually higher relative to the others, for instance, Y50 versus Y30 million for the provincial level. Sources: State Statistical Bureau 1992: 91-2,101-3,116-9 (A Collection of Documents on Fixed Investment and Construction); The Standard for Classifying Large, Medium and Small Investment Projects, 1979 version; China Investment and Construction 1991, (2): 32. People's Daily (overseas version), 18 August 1992: 2; & 6 January 1993: 2; Wang & Zhu 1987: 56-7; and interviews in the author's fieldwork.
84
Chapter 3
Table 3.2
Sources of finance for fixed investment in the state sector, 1953-95
Yoar Tcdl
Total (Y billions)
1953 1957 1962 1965 1970 1975 1979 1982 1985 1988 1990 1992 1995
9.16 15.12 8.73 21.69 36.81 54.49 69.94 84.53 168.05 276.28 291.86 527.36 1089.82
State budget (%) 83.7 88.6 83.4 85.8 75.3 64.4 63.0 31.4 24.0 14.6 13.2 6.3 4.9
Domestic loans (%)
Foreign loans (%)
Self-raised3 & others (%)
0.0 0.0 0.0 1.3 0.8 1.6 3.6 16.2 23.0 24.1 23.6 30.4 23.4
0.0 0.0 0.0 0.0 0.0 0.0 2.3 7.1 5.3 9.0 9.1 8.0 7.8
16.3 11.4 16.6 12.9 23.9 34.0 31.1 45.3 47.7 52.3 54.1 55.3 63.9
Note: (a) Self-raised investment refers to investment financed mainly by the retained depreciation fund and expanding production funds of local governments, industrial ministries and bureaux and enterprises, and by issuing construction bonds and stocks. Sources: Statistical Yearbook of China 1988: 560; 1991:148; 1993:150. China Statistical Yearbook on Investment in Fixed Assets, 1950-1995: 24.
3.4
Credit Plan and Government Control over Financial Resources
3.4.1 Investment plan, credit plan and mandatory loans Prior to 1979, more than 75 per cent of fixed investment was directly financed by state budgets in the form of grants from the central and/or local governments. The remainder was financed by funds retained by the industrial ministries and bureaux or by the state enterprises themselves (Table 3.2). Bank loans were primarily considered as a revolving fund for state enterprises to cover their extraordinary and temporary needs for working capital ('non-quota' working capital).5 A major departure from the prevailing practices was made when the State Council authorized the People's Bank in 1979 to extend Y2 billion in short- and medium-term equipment loans for the technical updating and transformation of state enterprises. Following this departure, and strengthened
The State Investment System and its Response to Reform by the programme of 'transferring budget grants into bank loans' (bo ge dai) initiated in early 1980s, investment credit rapidly became the most important means of financing fixed investment, as shown in Table 3.2. The general credit plan is compiled annually by three government agencies (Ministry of Finance, State Planning Commission, and People's Bank of China). The plan is based on the policy goals or directives set by the State Council for the major aggregate economic variables - the annual growth rates of output, investment and inflation - and on the aggregation of sectoral and local needs. The planning process is typically an iterative one initiated from the 'bottom up' (for a comprehensive introduction of China's general credit plan, see World Bank 1994: Annex 3.1). Investment credit plans have been likened by Chinese officials and scholars to '[T]he planning commissions order the dishes, and the banks pay the bills' (jiwei dian cai, yinhang tao qian). In other words, investment credit plans are passively linked with investment plans through mandatory loans. At central government level, the mandatory investment plan covers key national projects, projects approved by the SPC and projects partly financed by the central government. All other investments at this planning level are included in the guidance plan, which is a result of negotiations between the SPC and local planning commissions and provincial governments and can be renegotiated in the implementation process. The mandatory investment plan forms the basis for the mandatory credit plan. The State Council and the SPC request the state banks to finance a share of the investment through credit for the projects which appear in the investment plan. In fact, the banks have little choice. As noted in last section, the banks have already been consulted early in the project negotiation stage, and they are required to carry out financial and economic analyses for large- and medium-scale investment projects. Sometimes, the banks use these analyses to bargain with the SPC in an effort to minimise the costs of financing unprofitable projects. However, it is the SPC that finally decides which bank should finance which project as part of the mandatory credit plan. Projects that are undertaken by provincial governments under guidance planning are, in turn, included in the mandatory provincial investment plan, which, as in the case of the central mandatory plan, is directly translated into a mandatory provincial credit plan. The banks,
85
86
Chapter 3
again, have little to say concerning the way in which credit is extended for mandatory investments. They can bargain, but they can hardly refuse to extend the credit. In fact, the local branches of the state banks (especially People's Bank, Construction Bank, Industrial and Commercial Bank), have been particularly vulnerable to pressure from local governments, who not only control the allocation of housing, children's schooling and employment and other benefits to bank employees, but until 1988 even controlled the appointment of senior employees of the branches. Therefore, local branches are willing not only to follow the local mandatory credit plan, but also to collude with local governments to expand the scale of credit (cf. section 3.3.2). As a result, the ceilings of the credit plan do not usually function as a maximum allowable quota for credit, but as minimum targets to be achieved by investment hungry localities (World Bank 1994).
3.4.2 Central government investment hunger and key state projects The expansion drive and investment hunger of the central government can be mainly attributed to the internal and external pressures to provide evidence of socialist superiority, to catch up with the developed economies as fast as possible, and to create modernised armed forces for national security (Kornai 1992, Lin et al. 1994 & 1995). Before the reform, expansion drive and investment hunger were practised principally through political/economic campaigns and financed through budget channels, which accounts for more than 70 per cent of total investment of the state sector in most years (China Statistical Yearbook on Investment in Fixed Assets, 1950-1995: 24). During the reform period, the central government's investment hunger was mainly stimulated by the divergent priorities between the central and local governments, although it continues to struggle to provide evidence of socialist superiority. Such divergence, which used to be linked to China's remarkable regional difference and diversity, is growing for three reasons. First of all, an irrational price structure, together with relatively longer construction periods and large-scale investment requirements, make investments in national priority sectors like power and transport less attractive to provincial and local investors. Secondly, the rigidity of the state investment system, accompanied by underdeveloped capital markets, means that investable surpluses are usually rein-
The State Investment System and its Response to Reform
87
vested on the spot, regardless of higher rates of return elsewhere. Thirdly, the uncertainty of legal claims makes investment across administrative boundaries (regional and industrial) risky. The central government has been concerned about the investment pattern emerging as local entities are left to make investment decisions alone while these factors are in force. In fact, the central government has been evaluating various direct and indirect administrative measures to bring local investment decisions in line with national priorities (cf. section 3.2). The central government has also used its greater access to information and the wider range of tools at its disposal to create (in the early 1980s) and expand a special programme of national 'key' or 'priority' projects mainly in energy, transport and communication (Naughton 1992, Singh 1992). These national priority projects are planned with a 'rational timetable and guaranteed supply of materials', and thus have first claim to materials still under state control. In order to ensure the implementation of these priority projects the central government employs three strategies: concentrating its own resources on these projects, harnessing the financial resources of the state bank to them, and drawing on local financial resources through a form of 'matching funds' (peitao zijin) for them. While the first strategy is merely a rationalized version of the traditional planned system, the other two involve the central government in continuous, tug-of-war bargaining with local interests. The significance and development of this programme can be seen from Table 3.3. From 1982 to 1987, the number of projects included in the programme increased from 50 to 206. Over the same period, investment in the programme grew from 1.3 per cent to 3.3 per cent of GNP, and from 7.5 per cent to 15.7 per cent of total fixed investment in the state sector. Since then, the programme appears to have stabilised, and its nominal size, in relation to GNP has fallen. However, this does not necessarily indicate a decline in importance. Inflation rates since 1987 have been fairly high for material.6 As national priority projects are budgeted at state planned prices, the nominal size of this programme does not increase in line with inflation. In other words, since these projects have access to in-plan materials provided at subsidized prices, the nominal size tends to understate their importance by significant margins. Thus the programme of key national priority projects allows the centre to keep (and even increase) the investment and output
88
Chapter 3
shares of the energy sector in industry (see Table 1.1), to relieve the persistent energy shortages. The programme of national priority projects also differs from the old central plan in the quality of its design and planning activities, despite widely reported problems of inefficiency and huge waste (zhongdian gongcheng, zhongdian langfei). These priority projects are carried out with fairly high quality preparatory work, with feasibility studies, and with complete sets of design documents. A number of the projects utilize international funds, such as those from the World Bank, the Japanese Development Bank, and other international financial institutions, and therefore must comply with international standards for project appraisal and implementation, including competitive bidding for some parts of the work. The central plan has thus been strengthened in a technocratic sense. As each priority project is better designed and the programme as a whole is targeted more effectively to bottleneck sectors, and also because their priority status and access to materials ensure their completion, these projects usually have higher economic returns than those initiated during the 1970s, when many large projects were characterized by 'designing and constructing simultaneously' (biansheji, bianshegong).
Table 3.3
National priority projects, 1982-96
Number of projects Value (Y billions) % of SOE investment %ofGNP
Number of projects Value (Y billions) % of SOE investment %ofGNP
1982
1983
1984
1985
1986
1987
50 6.3 7.5 1.3
70 9.4 9.9 1.7
123 15.9 13.4 2.4
169 19.8 11.8 2.4
190 27.9 14.1 3.0
206 36.1 15.7 3.3
1988
1989
1990
1991
1993
1996
203 43.1 15.6 2.9
205 36.8 14.5 2.3
200 42.6 15.6 2.4
182 44.0 13.0 2.2
151 105.1 13.7 2.4
155.2 12.9 2.3
Sources: The figures for 1982-89 are taken from Singh (1992: Table 2.8), data for 1990 are from China Investment and Construction (Zhongguo Touziyu Jianshe, 1991(4): 28), those for 1991 and 1996 from Economic Situation and Prospect of China (Zhongguo Jingjixingshi yu Zhanwang, 1991-92: 224; 1996-97:128), and figures for 1993 are taken from People's Daily (2 March 1995: 2), and part of share calculation is based on Statistical Yearbook of China (1994:144; 1997: 150).
The State Investment System and its Response to Reform
89
3.4.3 Lending outside the credit plan Several factors account for the growth of lending outside the credit plan in the reform period. Firstly, non-bank financial institutions have grown substantially in recent years. In particular, the assets of Urban Credit Cooperatives and Trust and Investment Companies have increased rapidly. The ratio of their assets to the total assets of the state banks rose from 1.0 per cent and 5.5 per cent in 1987, to 4.0 per cent and 8.5 per cent, respectively, in 1992 (World Bank 1994: Table 3.3). While state banks are still far from being profit-oriented and have remained constrained by heavy government intervention at all levels, a growing number of investment entities have turned to financial intermediaries such as non-bank financial institutions in their search for greater flexibility. Moreover, since the lending of these non-bank institutions has been subject to less strict control than regular bank lending, setting up these institutions has provided a way for the state banks to earn returns on funds that might otherwise be tied up as excess reserves at the People's Bank (Central Bank). The banks can also use them as subsidiaries to engage in unorthodox lending operations. This dynamic has been emphasized by local governments, which have seen these non-bank financial institutions as an avenue for diversifying the sources of funding for local development. In fact, most non-bank financial institutions are owned either by local branches of the state banks or by local governments themselves. This ownership structure has created vested interests in the localities that support their non-bank financial institutions, and collusion in efforts to avoid the limits set under the credit plan. Secondly, China's capital markets have also developed rapidly. For example, the Shanghai and Shenzhen Stock Exchanges have seen the number of companies and the volume of transactions increase dramatically since their establishment in 1990. The total number of firms listed on the two stock exchanges increased from 183 in 1993 to 323 in 1995 and over 500 in 1996. Total volume of transactions reached US$42.1 billion as of December 1995, which is equivalent to about 6 per cent of China's GDP (Xu & Wang 1997). The shares were traded by almost 450 authorized brokers and securities companies connected through a national network of over 800 offices in 1994 (World Bank 1994). Thirdly, the rapid growth of the interbank market has also provided the bank branches with a channel for circumventing regional credit
90
Chapter 3
ceilings in search of higher rates of return elsewhere in the country. In principle, the maximum allowable maturity is seven days and the interest rates are allowed to fluctuate within a 10-30 per cent range around the People's Bank's 'reference rate'. In practice, however, the transactions with longer maturities and higher interest rates have taken place in the market. Other regulation-breaking operations have included the deviation of funds for unauthorized uses (e.g., acquisition of real estate assets or purchase of securities) and by unauthorized institutions (nonbank financial institutions and enterprises). According to World Bank (1994) estimates, the size of these leakages through the interbank market was about Y70 million in the first quarter of 1993, and declined to about Y40 million by the end of 1993 following the increased supervision under the 16-point programme announced in July 1993. Finally, collusion among local governments, enterprises and the banks has led to the growing incidence of diversion of funds from their intended use within the credit plan. For instance, as popularly reported in Chinese media, the funds allocated for working capital have been diverted for financing fixed investment. The ratio of this diversion to the total fixed investment loans of the state banks was about 17 per cent in 1992. Funds intended for the national priority investment projects have been diverted to projects more in line with the priorities of the localities (Almanac of China's Finance and Banking 1993: 269). The growing incidence of lending outside the credit plan reflects a breakdown in the State Planning Commission's investment approval system (see section 3.3) and in the supervision of local branches of the state banks. The monetary consequence of such lending is that it pressures the headquarters of the People's Bank of China (Central Bank) to Till the gap' through incremental reserve money creation, so that the activities originally intended under the credit plan can be financed.
3.4.4 Adjustments of nominal interest rates and patterns of real interest rates Prior to 1979, China practised a long-term stable interest rate policy. For example, from 1959 to 1970 the lending interest rate of industrial working capital were fixed at 7.2 per cent and from 1971 to 1979 the rate for both industrial and commercial working capital were fixed at 5.04 per cent (Almanac of China's Finance and Banking 1993: 385). These nominal interest rates might also be regarded as real interest
The State Investment System and its Response to Reform rates in view of the stable price levels during 1963-78. Although price control has made it difficult to compare these derived real interest rates with the rates that would have resulted from a market economy, it is popularly believed that these lending interest rates were far below market-clearing levels and were planned to facilitate the heavy-industry oriented development strategy (see, among others, Lin et al. 1995 & 1996, World Bank 1988). From 1979 to 1989, China gradually increased nominal interest rates in ten steps, mainly following the trends of inflation. During 1990-92, along with the decrease in inflation rates, China cut interest rates three times. In 1993, with rates lagging behind rapidly rising inflation, China was forced to increase them again in May and again in July. Table 3.4 gives the main interest rates on state bank loans for the period of 1979-96. From this table, we can see that real interest rates have been fluctuating dramatically, and it often happened that the real interest rates were negative. The average real interest rates during this 17-year period ranges between -0.62 per cent (standard deviation a = 5.17) and 0.61 per cent (a = 5.74) for the working capital; for the fixed capital, the means are between -0.96 per cent (a = 5.15) to 1.90 per cent (a = 5.47). Statistically, this means that all the averages of the real interest rates are not different from zero. This pattern is in sharp contrast to the significantly positive real loan interest rates in the period of 1963-78. For a rapidly growing economy like China, these levels of real interest rates are clearly below the market clearing levels, and this alone can cause a serious imbalance of supply and demand for credit. Moreover, as noted by Hussian & Stern (1991, 1992), because the tax system allows enterprises to deduct the principal as well as the interest payment from taxable profit, the 'effective' real interest rates have been further reduced by a large margin. As a result, the 'effective' real interest rates may be considerably lower than the real loan interest rates. This tax deductibility of loan repayment provides an incentive to enterprises to substitute bank loans for own-funds when financing investment. The prevalence of low real interest rates, particularly negative rates, has several adverse consequences, as documented in the literature (see, among others, e.g. Lin et al. 1995 & 1996, Singh 1992, World Bank 1994). These include that:
91
5.04 5.04 5.04-5.52 3.60-7.20 3.60-7.20 3.60-7.20 3.60-7.92 7.92 7.92 9.00 9.00 7.92-11.34 8.64-9.36 8.64 9.36-10.98 10.98 10.98 10.08-10.98
7.22-8.46 2.65, 2.07
2.0 6.0 2.4 1.9 1.5 2.8 8.8 6.0 7.3 18.5 17.8 2.1 2.9 5.4 13.2 21.7 14.8 6.1
7.8 6.5
1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
Mean
0.16-2.32 -3.84-(-1.68) -0.24-1.92 2.42-3.86 2.82-4.26 4.40-5.12 -0.88-2.00 1.92-4.80 0.62-3.50 -10.58-(-7.70) -9.88-(-7.00) 1.50-9.90 5.74-8.26 3.06-4.32 -4.02-0.84 -10.72-(-9.46) -3.82-(-0.04) 3.98-9.02
3.04 -0.96 2.64-3.12 1.70-5.30 2.10-5.70 0.80-4.40 -5.20-(-0.88) 1.92 0.62 -9.50 -8.80 5.82-9.24 5.74-6.46 3.24 -3.84-(-2.22) -10.72 -3.82 3.98-4.88 -0.62-0.61 5.17, 5.74
2.16^.32 2.16-4.32 2.16^.32 4.32-5.76 4.32-5.76 7.20-7.92 7.92-10.8 7.92-10.8 7.92-10.8 7.92-10.8 7.92-10.8 3.60-12.0 8.46-11.16 8.46-9.72 9.18-14.04 10.98-12.24 10.98-14.04 10.08-15.12 6.87-9.71 2.97, 3.50
-0.96-1.90 5.15, 5.47
Fixed capital
Working capital
Fixed capital
Real interest rates
Notes: The nominal interest rates only represent the basic level of main rates for each year. For instance, those for working capital are for general loans with maturities of 3, 6 and 12 months, and those for fixed investment are linked to general loans with maturities of 1 to 10 years (cf. tables of 'Regulated Interest Rates' in the Almanac of China's Finance and Banking, various years). Sources: Retail price change: Data for 1979-92 are taken from Statistical Yearbook of China (1993: 237; 1997: 267). Nominal interest rates: Data for 1979-92 are taken from Almanac of China's Finance and Banking (1993: 385), and for 1993-96 from Statistical Yearbook of China (1997: 626).
a
Working capital
Nominal interest rates
Main interest rates on state bank loans, 1979-96 (per cent per year)
Retail price change
Table 3.4
The State Investment System and its Response to Reform
(a) Lower-than-equilibrium costs for capital produces over-demand for investment funds, causing over-investment and inflation. (b) There have been very substantial differences in costs of capital between state and non-state sectors, and between formal and informal markets, which are determined not by market forces but by artificial barriers. These differences encourage socially wasteful and rampant rent-seeking activities, the latter often involving bribes and other corrupt measures. (c) Investment decisions still entail extensive administrative rationing with attendant interference by bureaucrats at all levels. This obstructs reforms of the investment system and state enterprises. (d) Highly capital-intensive investments with unwanted labour-displacing effects become feasible, which is opposite to the comparative advantage of China's factor structure. (e) The negative real interest rates correspond to significant implicit interest rate subsidies in the system. This means that capital use is being extensively subsidized in an otherwise capital-scarce economy and that over-invested sectors such as household durable goods and general machinery receive extensive subsidies, while the underinvested sectors such as energy, transport and raw materials receive far less. Recognition of these negative impacts frequently surfaces in China's official documents (e.g. the '16-point Programme' announced in July 1993 and in the State Council's Decisions on Financial System Reform in 1993). It would be interesting to know what prevents the Chinese government from raising real interest rates to overcome these adverse effects and to improve efficiency. The most reasonable current explanations are as follows. The first is mainly based on political considerations surrounding state-owned enterprises (SOEs). SOEs bear the social burden of supporting an extensive work force by keeping redundant workers on their payroll. They also provide a number of services (including housing, health care, child care and schooling) regarded as public goods in most other countries. SOEs are supported because the social consequences of their collapse would be too dire and the economic costs too high. As SOEs are called upon to incur expenditures that they might not otherwise undertake, they need to be entitled to the low interest rate support
93
94
Chapter 3
and other assistance from government (McKinnon 1994, World Bank 1994). It is also argued that keeping unemployment within the state sector can provide much more favourable political and social conditions for the non-state sector to expand and for economic reform to proceed smoothly. In comparison to a market economy where social insurance accounts for a significant proportion of the GNP, China's SOE losses may be less than many have assumed (Perotti et al. 1998; Zhang &Yi 1995). The second explanation of interest-rate policy mainly considers operational feasibility. It is argued that due to the rigidity of the traditional heavy-industry oriented development strategy, state banks have a dual function as both commercial and policy safeguards. This prevents most of them from becoming purely commercial banks. In turn, this makes it impossible in the short term, and fairly difficult in the medium term, to liberalize interest rates (Lin et al. 1996). Perkins (1994) mentions that in the context of soft-budget constraints, although raising interest rates may lead to a reduction in the demand for credit from some enterprises, the queue for loans is likely to remain long. This is because too many state-owned enterprises know they will not have to repay the loans if they make losses. Higher interest rates could also lead to further financial losses of the state enterprises and increasing arrears between firms. This argument has also been confirmed by the experience of Russia. In Russia the liberalization of interest rates has been accompanied by a dramatic accumulation of arrears. A rapid build-up of bad debt followed the high increases in nominal interest rates, which were running 8 to 10 percentage points above monthly inflation rates in April 1994 (Lavigne 1995: Chapter 7, Sachs 1995). There is no doubt that these two political and feasibility considerations provide some sound explanations for the obstructive forces. However, as pointed out in Zou & Sun (1996), there may also be purely incentive-based reasons, in terms of quality of credit and the moral hazards, i.e. the choices of risk and private efforts following state enterprise reactions to a hike in their borrowing costs. Very low interest rates, and particularly negative rates, normally make it unattractive for savers to put their money in banks. This disincentive, however, has not yet had a significant effect. Household accumulation of financial assets increased from only 3 per cent of total income on average before 1978 to over 15 per cent in most of the 1980s and the early 1990s (Naughton 1995a: 317-18). The total household
The State Investment System and its Response to Reform savings deposit balance of state banks has increased rapidly and steadily from Y40 billion in 1980 to Y162.3 billion in 1985, to Y703.4 billion in 1990, and to Y3852.1 billion in 1996. Household savings deposits have become the most important source of state banks' funds (Statistical Yearbook of China, 1995: 259 & 572; 1997: 622), and bank deposits remain the major part of the portfolio of an average household investor in China. The rapid increase in household savings has been an essential constituent of China's economic achievement. Increased savings allow continuing high levels of investment. A range of factors contributes to the remarkable increase of household savings. Four of them can easily be identified. Firstly, within the household responsibility system, individual rural households need to carry a large monetary balance throughout the agricultural production cycle. In addition, the continuous fervour for building new houses and improving living conditions needs accumulated monetary balances for the relevant lump-sum expenditures. Secondly, in the absence of consumer credit, the purchase of consumer durables and other lump-sum expenditures (such as increasingly expensive life-cycle ceremonies) require preparatory savings. Thirdly, the emerging individual, private, and small-scale cooperative firms also need to carry large monetary balances throughout their production cycles. Fourthly and perhaps most significantly, there are a lack of alternatives for household investors. Although bond and stock markets have emerged and expanded rapidly in China, their scale is still insignificant in comparison with that of available saving funds of Chinese households. In this context, the lower-than-equilibrium interest rate forms an implicit wealth reallocation mechanism from the depositors to the investors and producers (Zou& Sun 1996).
3.4.5 Recurring cycle of inflation and retrenchment during the reform period Because credit has begun to play an increasingly critical role for financing investment in the reform period, investment cycles have shown a new scenario. It means that in the up-phase of each cycle, the rising investment demand, dominated by local governments and their SOEs, is fuelled by low or negative real interest rates and the lack of any significant risk burden. This has been accommodated by an expanding credit plan and by the increased lending outside the expanding credit
95
96
Chapter 3
plan. The expanding credit plan and lending outside the plan have in turn resulted in sharp declines in the growth of excess reserves of the banking system. These have resulted in an accelerated growth of currency in circulation. Constrained by its assignments to provide policy loans to finance priority government expenditures, the People's Bank (China's central bank) has been unable to contain the growth of base money. The People's Bank has usually had to 'fill the gap' or accommodate excess demand for credit. This has been done through incremental reserve money creation when the geographical/sectoral targets set by the central government under the credit plan have not been met because of the diversion of funds and/or the disintermediation of bank deposits. As a consequence, the up-phase of each cycle has been characterized by sharp increases in the People's Bank's net lending to the financial system, leading to excessive reserve money creation, and thus to the buildup of inflationary pressure. Once the fear of overheating and record inflation is properly acknowledged, the retrenchment phase will be initiated by the central government to curb over-investment and inflation. However, because of the absence of other means, each investment boom and inflation rush has been reversed mainly through such ad hoc administrative measures as (1) re-centralization of approval power of fixed investments, (2) recentralization (at the headquarters of the People's Bank) of licensing authority over non-bank financial institutions and over central bank's lending to the specialized banks, (3) tightened supervision over the activities of the state banks and their branches, and (4) the dispatching of inspection teams to provinces and autonomous regions to examine investment and financial probity. The re-centralization and administrative controls do reduce the inflation pressure. However, they lead quickly back to state intervention in most areas of the economy and a bust follows. When this bust occurs, the administrative lid comes off, local governments and other investment entities begin to re-assert their roles and sooner or later there is excess demand. In the pre-reform period, over-investment directly tightened the supply bottlenecks of energy, transportation and agriculture through material shortage conduction. However, during the reform period overinvestment has generally caused overexpansion of credit followed by shortage in the planned component and inflation in the market component of the economy through both material and value conduction. This
The State Investment System and its Response to Reform inflation is, in fact, a monetary manifestation of supply shortages in bottleneck sectors.
3.5
The Material Supply System
3.5.1 Material supply system: Function and characteristics China's material supply system covers both investment and production. The distribution of consumption goods was mainly managed by the commodity distribution system before the reform and has now been left increasingly to market forces, although a large part has been carried out mainly by local government-owned enterprises. Township and village enterprises have played an increasingly important role in this marketization process,. However, the distribution and production of investment goods and other raw materials required for investment have been strongly influenced by planning, even during the reform period. For example, in 1992, 14 years after the beginning of reform, of the 10 major materials listed in the Statistical Yearbook of China (1993: 503), 40 per cent of coal, 21 per cent of rolled steel, 34 per cent of caustic soda and soda ash, 13.5 per cent of timber and 6 per cent of cement were still under state plan allocation within the material supply system. Due to the specific focus of this book, the general material supply system is of less interest here than the relationship between the material supply system and the investment plan.7 In brief, the material supply system is not used as an independent instrument to influence the investment plan, but rather to reinforce and add weight to the mandatory investment plan at central and local levels. All projects included in the investment plans of the State Planning Commission and/or the local planning commissions receive allocations of key materials through the material supply system at central and local levels. The system's basic function, i.e. to reinforce mandatory investment plans, can be clarified by the following observations. First of all, the highest quality materials are distributed through the material supply system. Before reform, it was almost impossible for projects to purchase inputs of sufficient quality outside the material supply system. During the reform period, it has also at times been impossible to procure inputs of the required quality outside the material supply system. For instance, high-quality rolled steel and cement for structural com-
97
98
Chapter 3
ponents are mainly produced by central government owned enterprises and distributed through the material supply system. Secondly, the materials are distributed at state plan prices, which are much lower than market prices. This serves as a major avenue for subsidizing the projects that central government (higher-level government) wishes to impose on provincial authorities (lower-level governments). This is especially true for the key national projects (cf. section 3.4.2), which receive priority allocations of investment goods at state plan prices. Thirdly, the combination of the first and second functions supplies bargaining leverage for the central government in its negotiation with local governments. This is done by adjusting the amount of materials supplied through the material supply system. In this way, the central government may entice local authorities to go along with its priorities (Naughton 1992, Singh 1992). As in the former Soviet Union, the distribution of materials is divided into three categories. The first, unified distribution (tongpei), is allocated by the State Planning Commission on behalf of the State Council, and its day-to-day operations are delegated to the Ministry of Material Supply (the former State Material Supply Bureau). The second category is central ministerial distribution (buguari) through specialized ministries and the Ministry of Commerce. The third is residual materials (diguan) distributed by local governments. In the former Soviet Union in the early 1980s, a total of over 67,000 commodities were centrally allocated (2,000 by the Central Planning Commission, about 15,000 by the Material Supply Bureau, and over 50,000 by the central ministries; see Naughton 1986: 126 and World Bank 1988: 67). However, in China, the number of commodities (including materials) subject to 'unified distribution' at the peak of central management of the economy (i.e. 1964 and 1965) was 370. The largest number of commodities subject to the second category in the history of the People's Republic was 837 (256 by the State Planning Commission and Material Supply Bureau, 581 by central ministries, which was in 1981 and 1982) (China Material Economic Society 1983: 92). This significant difference shows that a fairly limited number of commodities are centrally allocated in China. It implies a very high degree of aggregation within the material supply system. In other words, the classifications of commodities and materials involve much broader categories than those of the former Soviet Union. As pointed out by Naughton (1986), it is impossible to believe that control over a few
The State Investment System and its Response to Reform hundred commodity specifications could give central planners detailed control over resource allocation throughout the economy. In fact, central planners are primarily engaged in redistributing materials and commodities among regions and ministries, rather than in assigning specialized materials to specific uses (Li, Jingwen 1981). This makes the material supply system strikingly different from that of the former Soviet Union. The second significant difference between China's material supply system and that of the former Soviet Union is that control over materials is quite fragmented. Along with progress in local industrialization, the allocation rights to materials have become more or less attached to ownership of the production units, rather than to centrality or importance in production and investment. For example, coal produced in mines under central control comes under central allocation, whereas coal produced in provincial mines comes mainly under provincial allocation. The latter is subject to an obligation to turn over a certain aggregate amount of output to be redistributed by the central authority to other provinces. As a result, by 1978, on the eve of reform, proportions of output of categories 'unified' and 'ministerial' distribution under local allocation included 46 per cent of coal, 20 per cent of steel, 64 per cent of cement, and 65 per cent of machine tools (Yu 1980, Zhang 1979). The third characteristic of the material supply system is that planners are in the habit of 'leaving gaps'. This means that they have to underallocate supplies to investment projects and production enterprises, requiring them to 'mobilize initiatives' and locate additional inputs for completing their investment and production targets (Chen 1980, Li, Jingwen 1981, Naughton 1986). In fact, this is a natural consequence of the fragmentation of the material supply system mentioned above. 'Leaving gaps' has exacerbated the tendency - common to all centrally planned economies - for hoarding and duplicative local production (Wong 1985). However, this practice may also act as a screening device. This is because shortage of supply may discourage a manager or officer, who knows a project is bad, from applying for financing in the first place. This improves the average quality of the projects (Qian 1994).
99
100 Table 3.5
Chapter 3 Percentage of key materials subject to central allocation, 1965-92 Materials
Year
1965 1978 1981 1985 1986 1987 1988 1989 1990 1991 1992
Coal
Iteef
75 54 53 46 45 43 42 40 42 42 40
95 80 52 51 47 43 38 34 31 29 21
Soda
n.a. n.a. 78 67 62 45 47 37 37 32 33
Cement 71 36 27 17 15 13 n.a. 10 10 8 6
Timber 63 81 72 30 28 27 22 25 28 19 4
Sources: The figures for 1965 are taken from Li, Kaixin (1981(iv): 124,1983:16), data for 1978 are from Yu (1980: 4) and Zhang (1979:15-18); those for 1981-92 are from Statistical Yearbook of China (1991: 482, and 1993: 503).
3.5.2 Declining importance and its special focus since the reform Some scholars use the number of materials under unified distribution as an index of centralized control. This index may be quite useful for understanding the material supply system in the former Soviet Union and other centrally planned economies; however, it is less applicable to China. In China local governments also allocate a portion of the output of materials from both unified and ministerial distributions. This institutional arrangement has its roots in the 1950s, when local governments were given initial responsibility for the management of numerous small and medium-sized factories as they became nationalized or established. Local control during this period was primarily a transitional strategy. Direct local control was gradually being reduced by the central ministries in the first five-year-plan period as the economy was coming under central planning. A peak was reached in 1957, when materials under central allocation accounted for 70-90 per cent of the total (Li, Jingwen 1981). However, after the first round of decentralization in 1958 and 1959 and the resultant reassertion of central control in 1962
The State Investment System and its Response to Reform
101
and 1963, the central authorities had to (in 1964) relax their grip once again on material allocation and to give localities more flexibility. In 1966, central government transferred almost all output from local small-scale industries to local allocation. With the development of the local 'five small industries' and commune and brigade enterprises (the predecessors of present township and village enterprises) during the Cultural Revolution, local production included key materials such as iron and steel, cement, coal, chemicals and machinery (Liu & Wang 1982). In the reform years, the proportion of materials under central allocation has declined more substantially due to two new factors. First of all, state-owned enterprises have been allowed to market above-quota production on their own, which has stimulated high production growth above the goals of the mandatory plans. Secondly, the rapid development of township and village enterprises has induced an extraordinary increase in material production outside the mandatory plans. The decline in the central allocations of five key materials is shown in Table 3.5. It should be pointed out that although central control over materials has declined substantially in the post-1978 reform period, the drop began well before the initiation of the reforms (Wong 1985). The decline of central control over materials does not mean a corresponding expansion of market transactions. In fact, such control has been largely taken over by local governments, in particular (as in other areas of policy-making), by provincial governments. Markets have expanded more slowly than expected with declining central control. In 1987, according to a World Bank estimation (based on Statistical Yearbook of China, 1988), more than one-third of coal, steel and lumber, and about one-third of cement were distributed by local authorities (Singh 1992: 31). This means that local governments control a sizeable proportion of the production and distribution of materials. Nevertheless, market transactions have expanded steadily, especially since reform. The same World Bank estimation shows that in 1987 the market sales of state-owned enterprises accounted for about 53 per cent of their cement production, 33 per cent of their lumber production, and over 17 per cent of their steel and coal outputs (Singh 1992: 31). Along wth marketization and decentralization, the material supply system at present appears to be more a means of distributing rents and of supporting the central government investment programme, than a sign
102
Chapter 3
that the economy could be considered as becoming centrally planned (cf. section 3.4.2).
3,6 Soft Budget Constraint and Investment Hunger in State-owned Enterprises At the onset of reform, China's state-owned enterprises (SOEs) functioned as passive agents of the government economic bureaucracy. Managers focused on quantitative targets set by bureaucracy rather than on financial objectives, especially on physical output volume and total output value. Profitability influenced neither the incomes of executives and workers nor the growth prospects of firms. Almost two decades of reform, however, has brought about dramatic changes in the originally puppet-like micromanagement system of the SOEs. New incentive mechanisms have re-defined relations between effort, financial outcomes and individual reward. They have also enabled enterprise to gain control over some resources. The first reform effort, implemented between 1979 and 1983, consisted of tentative steps to expand the role of financial incentives and to improve performance within a framework dominated by mandatory output planning and administrative allocation of inputs and products. These measures included allowing SOEs to retain a modest share of total profits, establishing performance-related bonuses for workers and permitting the SOEs to produce outside the mandatory state plan. The second set of reforms, dating from 1984, centred on two innovations: the dual-track price system and the enterprise contract responsibility system. The dual price system partitioned SOEs products as well as inputs into plan and market components. This meant that most SOEs transacted marginal sales and purchases on markets where prices responded increasingly to the forces of supply and demand. At the same time, bank loans began to replace budgetary appropriations as the chief source of external funding for both working and fixed capital. Under the contract responsibility system, a firm's chief executive manager, a group of managers, or sometimes the firm's entire workforce, agree to fulfil a set of specific obligations. This involves targets for total profits, delivery of profit to the government and productivity increases. In return SOEs acquire more control over their business operations, including control over the depreciation funds and the use and allocation of after-tax profits for bonus, welfare payments and reinvestment.
2.8 2.4 5.2
II. Fiscal subsidies (% of GDP) 1. Explicit subsidies 2. Estim. implicit subsidies3 3. Total estimated subsidies 3.1 4.2
3.7 1.9 5.6
18.02 10.2 1.13
1989
2.5 3.3
2.5 1.9 4.4
36.70 22.1 1.82
1991
2.5 3.3
1.8 1.9 3.7
36.93 19.0 1.54
1992
5.1 6.8
1.3 1.3 2.6
45.26 18.4 1.31
1993
0.78 0.35 1.13
48.26 16.8 1.03
1994
0.57 0.42 0.99
63.56 22.3 1.11
1995
0.50 0.52 1.02
79.07 28.9 1.18
1996
(a) The implicit subsidies are estimated as the difference between what is reported in the state budget as capital construction expenditures and what is reported in the Statistical Yearbook of China as state appropriations for financing SOEs' fixed capital investment. Most of them are channelled through the People's Construction Bank of China as interest rate subsidies. In 1994, China changed the definition of capital construction in the state budget, which causes a problem for comparison of 11.2 between two periods before and after 1994. (b) & (c) 60 and 80 per cent of PBC lending to the financial system, respectively. The short-term policy lending is mainly for working capital loans to loss-making SOEs. All PBC loans are subsidized to a different degree below the base administered interest rate for bank loans, and it is estimated that more than half are not repaid (World Bank 1994: 27).
0.9 1.3
3.3 1.7 5.0
6.10 4.0 0.54
1987
Sources: I: Statistical Yearbook of China 1993: 31 & 430; 1997:42 & 439. II: China Fiscal Statistics: 1950-88, pp. 19 & 146; Statistical Yearbook of China 1993: 219, 31, 222 & 149; 1994:140 & 219; 1997:42, 155 & 233; and World Bank 1994: 36; III: World Bank 1994: 26.
Notes:
III. Policy loans of PBC (% of GDP) 1. Lower boundb 2. Upper bound0
3.24 2.4 0.38
1985
Reported financial losses of industrial SOEs with independent accounting systems, fiscal subsidies to loss-making SOEs, and policy loans of the People's Bank of China, 1985-96
Reported losses 1. Amount (Y billion) 2. Ratio to total profit (%) 3. Ratio to GDP (%)
I.
Table 3.6
o
I
f 1 1
I
I
1
104
Chapter 3
This increased control over retained earnings forms the nucleus of the new incentive mechanism in the SOE sector. In fact, managerial autonomy would be meaningless if the enterprise were required to remit all its profits to the state. Conversely, the larger the fraction of its profits the enterprise is allowed to retain and control, the stronger the managers' incentive to increase profits. The increase in the SOEs' marginal profit retention rate has brought the general shift of SOEs' emphasis from physical targets to financial ones. In particular, it leads to the significant positive correlation between profits and retained earnings and rewards to managers.8 With autonomy in output decisions and with higher marginal profit-retention rates, the managers of SOEs strengthen workers' incentive, pay more in bonuses and can hire more workers on fixed-term contracts. The new incentive mechanism was effective, as Groves et al. (1994) demonstrated empirically.9 Productivity increased with increases in bonus payments and in contract workers. The increase in autonomy raised worker's incomes and also investment in the SOEs, but tended not to raise remittances to the state. Unfortunately, greater SOE autonomy has not eliminated intrusive regulations. State agencies sometimes refuse to allow enterprises to exercise their new 'rights', especially with respect to financial and investment management, employment and foreign trade. This is because of the lack of adequate management accountability and because the increase in autonomy has been accompanied by managers' and workers' discretionary behaviour in a distorted macroeconomic environment. Managers are burdened with objectives only partially related to the market's requirement to make profits by cutting costs and increasing sales. If managers acted in accordance with government officials' wishes, their losses would be made up with bank loans that would not have to be repaid. This means that managers are only willing to take responsibility for profits and expect the state to cope with financial losses (Liu & Zou 1992). The SOEs' budget constraint remains fairly soft. Table 3.6 presents the reported financial losses of industrial SOEs with independent accounting systems. It also includes the reported and implicit fiscal subsidies to the loss-making SOEs (1985-96) and the estimated policy loans of the People's Bank of China to the financial system (1987-93). As can be seen, the industrial SOEs have made increasing losses in the period of industrial reform. The ratios of their losses to their total profits (before tax) increased dramatically from 2.4 per cent in 1985 to over 20 per cent in 1991, 1995 and 1996. The loss/
The State Investment System and its Response to Reform
105
GDP ratio also increased considerably from 0.38 in 1985 to 1.82 per cent in 1991 and over 1 per cent in the period of 1991-96. Some of these losses can be attributed to 'planned' or 'fixed quota' losses. However, such overall loss is not possible without 'soft' subsidies from the government budget and 'soft' credits from the state banks. Panels II and III in the table indicate, similarly, that fiscal subsidies to lossmaking SOEs had remained rather high, and that the decrease in explicit fiscal subsidies had been offset by implicit fiscal subsidies and state banks' policy loans. Besides the policy loans reported in the table, there is an unknown proportion of working capital loans of state banks from the banks' own resources. These are, in effect, non-performing because they have been extended as subsidies under the direction of the central and local governments. Such a soft budget constraint, combined with increasing autonomy, stimulates the SOEs' expansion drive and investment hunger (cf. section 2.2.1), and encourages the SOEs to take too much risk in their investments (Zou & Sun 1996). The logical link between soft budget constraint and investment hunger as well as willingness to run higher risks is understandable. A manager of a large SOE in China, interviewed by Shih (1992), gave a telling explanation about a project with a high and clearly visible risk in 1988 and 1989: Socialism means that everyone eatsfromthe same big bowl and no one pays. How could I come up with money to invest? I used the state's money. If I lose money, actually it is the state which loses money. If I have to use my own money to invest, I guarantee you that I would not have done what I did. So you see, sometimes socialism can be a good thing (Shih 1992: 50) [to encourage him running high-risk investment]. Theoretically, there is a formidable array of administrative controls between enterprise investment desire and real investment allocation, which are largely enforced through project approval procedures (section 3.3). Several overlapping controls augment the project approval process. These include annual fixed investment plans and ceilings, multiyear investment plans, fund allocation processes and administrative controls over the allocation of investment goods (cf. sections 3.3, 3.4 and 3.5). In practice, however, enterprises cooperate easily with their immediate supervisory agency on many issues, and find means to obtain approval or even to bypass the limitations altogether. For instance, collusion between enterprises and local governments induced by their compatible interests and incentives is well known. Local
106
Chapter 3
governments need their enterprises' investment enthusiasm to promote local growth and to create employment. Local governments have to rely on their SOEs to provide a crucial social safety net for the urban population in terms of housing, schooling, health care, retirement benefits, etc. The result is upward pressure on the demand for investment funds that can lead to overinvestment and run counter to the central government objective of macroeconomic stability.
3.7 The Development Drive and Investment Hunger of Local Governments In China there are six government levels: central, provincial, prefecture, county, township (previously, commune) and village (previously, brigade). In urban areas, there are three levels: municipality, district and neighbourhood. Each level of the governments owns a large number of enterprises. For example, according to the data in the 1985 Industrial Census of China, of the 70,342 state-owned industrial enterprises in 1985, the 3,825 that were owned by the central government produced only 19.57 per cent of the total industrial output. 31,254 were owned by provincial and municipal governments, accounting for 44.57 per cent of the total output. The other 35,263 SOEs were owned by county governments and produced the remaining 8.98 per cent of the total output. Besides these industrial SOEs, in 1984 township and village governments owned 901,000 industrial enterprises (TVEs) and urban districts and neighbourhood governments owned 134,900 industrial enterprises (the Urban Collective Enterprises) (World Bank 1988a: 60). Industrial enterprises only accounted for about 13 per cent of the total number of non-agricultural enterprises and in 1984 there were almost 12 million enterprises engaged in non-agricultural activities (World Bank 1988a: 55). The large number of mostly small enterprises, combined with China's size, regional differences, and poor communication and transportation facilities, mean that the economy has been organized mainly along regional lines. It is also consistent with China's historical experiences of managing multi-regional economic and political organizations for more than 2,000 years (Qian & Xu 1993a & b). Ideological and political reasons dictated that China's first Five-Year Plan (1953-57) should be formulated with the help of Soviet experts, by copying the classical Soviet model. However, towards the end of the First Five-Year Plan,
The State Investment System and its Response to Reform
107
Mao and other leaders increasingly recognized the unsuitability of the Soviet model. In his famous 1956 speech on the ten major relationships, Mao pointed out: Our territory is so vast, our population is so large and the conditions are so complex that it is far better to have the initiatives come from both the central and the local authorities than from one source alone. We must not follow the example of the Soviet Union in concentrating everything in the hands of the central authorities, shackling the local authorities and denying them the right to independent action. ... The central authorities want to develop industry, and so do the local authorities. ... The central authorities should take care to give scope to the initiative of provinces and municipalities, and the latter in their turn should do the same for the prefectures, counties, districts and townships; in neither case should the lower levels be put in a strait-jacket. (Mao 1977: 284-307). A general consensus on this point has led to attempts to deviate from the Soviet model and to move towards 'administrative decentralization' from, the late 1950s onwards. Further, throughout the mid1960s and 1970s, when rapid industrialization at local level added large numbers of small enterprises, the impossibility of incorporating them into the central planning led to the creation of a multi-layer, multiregional system. This led to much of the responsibility for promoting and coordinating economic and social development being devolved to local governments (cf. section 3.2). In this institutional set-up, apart from the need to provide evidence of socialist superiority, to catch up on more developed regions, and to expand the power, prestige and material benefits of the local bureaucrats (see section 2.2.1), a local government has to be responsible for its own region's development, employment, stability and welfare. In fact, local governments have been faced with much stronger pressure to develop than the central government. For example, they have to accept liability for solving social disturbance due to such factors as unemployment, housing shortages, infrastructure deficiency and dissatisfaction in consumer sector. They must struggle to meet expenditure obligations imposed by central policies and commands, which is likened to 'the centre throwing dinner parties and leaving the localities to pay the bills'. They also strive to promote more rapid local economic growth so as to improve their own negotiating position within the bureaucracy. In order to promote local economic development and extend employment, local governments have been eager to set up new factories
108
Chapters
and expand their owned enterprises. In order to ensure the production and supply of basic necessities, local governments take responsibility for food processing, clothing, housing, civil fuel processing, public transportation, urban infrastructure, etc. This has significantly corrected the bias of the heavy-industry-oriented development strategy. However, local governments have also been enthusiastic to develop industries and other sectors with low entry barriers and quick profits. This results in the convergence of local economic structures, undersized enterprises, proliferation of short-term projects, a continuous build-up of obstacles to economic structural adjustment, as well as low investment efficiency (Guan & Jiang 1990, Lou 1992, Wong 1992). Because local governments rely on their SOEs for a crucial social safety net for the urban population, they are willing to protect their enterprises from bankruptcy even in the event of chronic losses. The most popular way to avoid bankruptcy is to grant investment to a 'virtually bankrupt' enterprise by fiscal appropriation and/or soft credit so that the enterprise can update its equipment and try to produce new products. This measure is known as 'helping an enterprise to cast off the difficulty' or, in a very bad case, 'rescuing' (wanjiu). Like the central government, local governments have also been revenue-starved because they have to carry heavy economic and social responsibilities as well as to cope with growing local fiscal expenditures. On the supply side, the fiscal structure has depended overwhelmingly on industry for raising revenues. During both the pre-reform and post-reform periods, this 'industrial domination' characteristic has persisted, though many changes have been introduced to the tax system. From 1964 to 1984, over 70 per cent of total government budget revenues (at both central and local levels) came from industrial profits and taxes. In 1991, this share fell to 41 per cent but extra-budget revenue (mainly from depreciation funds and other fees of industrial enterprises) almost equalled the budget revenue in the early 1990s (Yearbook 1993: 215-17 & 230). Fiscal revenues, especially extra-budget revenues, are also apportioned mainly on the basis of enterprise ownership (Naughton 1986, Wong 1991 & 1992). Faced with intense expenditure pressure and a tax system that depends on industry for the generation of over two-thirds of total revenues (both in- and extra-budget), local governments have little choice but to engage in industrial expansion. Therefore, the expansion drive and investment hunger of local governments in industrial development are much greater. This has long
The State Investment System and its Response to Reform
109
been recognized by both Chinese officals and scholars (see, among others, Guan & Jiang 1990, Naughton 1986, Wong 1992, World Bank 1994).
3.8 Summary: Insatiable Investment Demand Exists at All Levels The discussion so far indicates that there has been insatiable investment demand at the central, local and enterprise level, and that this has been true in both the pre- and post-reform periods. The central government has been faced with internal and external pressures to provide evidence of socialist superiority, to catch up with the industrialized powers as fast as possible and to create a modernized military force for national security. This pressure led to the formation and implementation of a heavy-industry-oriented development strategy in the pre-reform period. The purpose of this strategy was to build the nation's capacity to produce capital goods and military materials as rapidly as possible. During the reform period, the investment hunger of the central government was further stimulated by the divergent priorities between the central and local governments. In the early 1980s, due to greater access to information and a wider range of tools at its disposal, the central government created and later expanded the special programme of national priority projects which focused on energy, transport and communication, as well as other basic infrastructure sectors. The centre concentrated its resources on these projects, harnessed the financial resources of the State Bank to them, and drew on local financial resources through a version of 'matching funds' for them. As a result, the central government has maintained and increased the investment and output shares of the energy sector in industry since 1981, and this has served to significantly relieve the long-lasting energy shortages. The expansion motives of the local governments may also be attributed to a desire to provide evidence of socialist superiority, to catch up with more developed regions and to expand the power, prestige and material benefits of the local bureaucrats. However, local governments also have to take more responsibility for solving social disturbances coming from unemployment, housing shortage, infrastructure deficiency and growing dissatisfaction in consumption sector. They must struggle to meet expenditure obligations imposed by the central policies and
110
Chapter 3
commands. They also need to promote more rapid local economic growth in order to raise their own bargaining position within the hierarchy. Therefore, local governments have been faced with much stronger pressure to develop than the central government. In order to meet the basic demand of local residents for necessities of life, local governments support the development of food processing, clothing, housing, civil fuel processing, public transportation and other urban infrastructure sectors. This effort has served to improve the industrial structure of the economy. However, local governments have also favoured the expansion of processing industries and other sectors with low entry barriers and quick profits. They have gambled on the fact that the bottlenecks in energy, transportation and communication would ultimately be taken care of by the central government, and that the regions with the most promising early development of profitable industries would be able to hold on to those assets. This local priority raises obstacles for economic structural adjustment and causes lower investment efficiency. The expansion drive and investment hunger of the state-owned enterprises, can partly be explained by managers' bureaucratic motives such as political and moral conviction, identification with the job, and extending their power, prestige, and material benefits. However, it should be attributed more to the SOEs' soft budget constraint. The soft budget constraint has stimulated the SOEs' expansion drive and investment hunger and also encouraged them to take too many risks in their investments. This is because they do not expect liquidation to follow. Looking at the formal set-up of the state investment system, a seemingly strict project approval system or sound distribution of project approval norms exist among the various levels of local governments. In practice, however, in order to promote their own regional growth and create employment, governments at lower levels used to collude with state bank branches and enterprises within their jurisdiction to circumvent project approval requirements set at higher levels. The most popular collusion in the investment planning process is for subordinate entities to cooperate to disaggregate projects into smaller components with underestimated costs, so as to evade or simplify project review and approval procedures. The outwardly formidable credit plan system has operated in a similar way. The local branches of the state banks have been vulnerable to pressure from local governments under the institutional arrangement of 'dual subordination' (shuangchong lingdad) along the lines of both vertical and regional accounta-
The State Investment System and its Response to Reform
111
bility. In addition, the local branches have depended on their local governments for basic supports such as housing, children's schooling and employment, and other social services for bank employees. This means that the local branches are inclined to follow the local credit plan and to collude with local authorities to expand credit scale in various ways, formally or informally, legally or 'illegally'. In brief, on one hand there is an ubiquitous expansion drive and investment hunger at the central, local, and enterprise level which is endogenously generated by the political/economic system. On the other hand, there is a lack of internal, self-generated restraint to resist this expansion drive. The credit system is far from independent, and financial obligation is usually not binding unless some strict ad hoc administrative measures are imposed by the central administration through the hierarchy. As a consequence, real investment may only be constrained by the supply possibilities frontier of bottleneck sectors. The following two chapters will focus on evidence of such supply constraints.
Agricultural Constraint to the Insatiable Investment Demand
4.1 Introduction As reviewed in section 2.4.1 there have been continuous efforts to link China's investment cycles and agricultural fluctuation. On the one hand, Eckstein (1968) combined agricultural fluctuation with policy cycles to explain the fluctuation of economic growth during the 1950s. In the same vein, Tang & Huang (1982) used estimates of total factor productivity from 1952-79 to suggest that the peasants had historically responded in a predictable and speedy manner to Beijing's policy gyrations which impinge upon them economically. On the other hand, Naughton (1986) retested Eckstein's theory based on the data from 1953-83 and suggested that although there is some evidence of this link between policy and production in the 1950s, attempts to attribute a causal role to agricultural production or procurement trends in the generation of investment fluctuations after 1958 were unsuccessful (Naughton 1986: 157). Naughton's argument is confirmed by Imai (1994a). It is clear that empirical evidence on this issue in the literature is ambiguous. This chapter presents an alternative way of dealing with the interaction between agricultural fluctuation and macroeconomic adjustment and in particular the investment modification. The intention of the alternative analysis is to set up an indicator system and to document empirically the input-output relations in agriculture and the distributive relations in the uses of national income. The indicator system can help reveal the causality between agricultural fluctuations and relevant macroeconomic adjustments. It comprises appropriate indexation of real value added per labourer in agriculture, per capita output of grain, purchase price levels of farm products, real material inputs into agricul-
112
Agricultural Constraint to the Insatiable Investment Demand
113
ture, fixed investment ratios and capital accumulation ratios. Employing such as a comprehensive indicator system enables us to reject the likelihood that the interaction between cycles of agriculture and investment (or capital accumulation) ratios may be an accidental phenomenon appearing in the statistical data. The comparison of the dynamics of these indicators are presented in seven tables and corresponding graphs (Tables 4.5^.11 and Figure 4.3 a-g), which are elaborated for each of the seven main cycles of both agriculture and investment from 1954 to 1996. These tables and graphs present the general, short-run two-way interaction between investment expansion and agricultural growth. In other words, they indicate that in each phase of each investment cycle, there is a scramble for scarce resources between agricultural development and industrial investment. It is worth noting that the short-run interaction between investment expansion and agricultural development cannot be simplified as substitution alone. Investment expansion may be at the cost of agricultural development and intensify shortage. However, the increase of agricultural production will stimulate and support investment expansion. The same also applies in the long run, and it is more likely that complementarity may be dominant in the long run (cf. section 1.2). The terms substitution and complementarity may oversimplify the issues and cause confusion. That is another reason why this research uses the notions of cointegration and error correction rather than those of substitution and complementarity. The long-run equilibrium comovement of real investment level and agricultural development will be analysed based on the cointegration equation in Chapter 6. This chapter focuses on the short-run two-way interaction. This two-way interaction represents the characteristic link between the cycles of investment growth and agricultural production. It also indicates the relevant patterns of policy trade-off by decision-makers in a typical dual economy. It should also be kept in mind that the magnitude of investment fluctuations is greater than that of agricultural fluctuations. This implies that other factors, particularly, energy and transport bottleneck constraints, have played an important role in determining the investment cycle. The methodological implications of this chapter may be of independent interest. If our purpose is to detect causality between two groups of economic events rather than to explain exclusively the fluctuation magnitude of one selected variable, the approach used in this chapter may have a clear advantage over common regression analysis. It is
114
Chapter 4
more intuitive and easily understandable, allowing one to obtain insights through a direct and interpretable analysis. The rest of this chapter is organized as follows. Section 4.2 briefly reviews the contribution of agriculture to the national economy. Section 4.3 reveals the change of factor endowments in agriculture and its consequences. Section 4.4 analyses how China's specific institutional setting has helped to minimize agricultural fluctuation. The selection of an indicator system is discussed in section 4.5. In sections 4.6 and 4.7 the empirical patterns of agricultural fluctuation and macroeconomic adjustment are analysed and stylized. Section 4.8 summarizes these issues.
4.2 Contribution of the Agriculture to the National Economy The importance of agriculture in China is in its provision of food and other basic necessities to 20 per cent of the world's population, using only seven per cent of the world's cultivated area, rather than its share in national income or in total national production. 'Agriculture is the basis of the national economy and grain is the centre of that basis'. 1 Moreover, as indicated in section 2.3.2, although the percentage of the workforce engaged in agriculture has been falling (from 93.5 per cent in 1952 to 70.5 per cent in 1978 and 56.4 per cent in 1993), the number of agricultural labourers nearly doubled in the same period (from 173 million in 1952 to 283 million in 1978 to 374 million in 1993), and the growing labour force in agriculture was not reversed until 1993. Agriculture had also long served as a shock absorber of urban unemployment by employing the surplus labour from urban sectors, and urban 'young intellectuals' sent down to the countryside (Lardy 1985). The idea that macroeconomic instability is inevitable if the nation lacks agricultural stability {wunong buwen) in terms of both food supply and employment generation is frequently put forward by Chinese leaders and economists. It is no simple matter to meet the demand for the main agricultural products of over a billion people. In the case of grain, in 1996 the production of grains was a record 504 million tonnes, of which cereals amounted to 451 million tonnes (Yearbook 1997: 383).2 In terms of international comparison, this was one of the highest levels of national grain production in the world, but the per capita production of cereals in China was still only 368 kg, only 12 kg more than the 1996 world average of 356 kg (Yearbook 1997: 831, 836). Among the various competi-
Agricultural Constraint to the Insatiable Investment Demand
115
tive uses of grain, its direct consumption as a foodstuff accounts for 70 per cent of the total production of grain, and this percentage is unlikely to change significantly in the near future. As a result, the supply of grain has been rather tight. At least 80 million people in 1994 still suffered from basic food shortage {People's Daily, 4 Feb. 1995: 2). At the same time there were millions for whom meeting of subsistence requirements was no longer a problem but who desired to purchase more processed food and animal products. This led to the increase of demand for grain from the food processing and animal feed manufacturing sectors.
Table 4.1
The output value of light industry using agricultural products as raw materials • " V billions
1952 1957 1965 1978 1985 1992
19.4 33.0 50.4 123.6 275.7 1196.4
A
^
87.5 81.6 71.7 68.4 67.0 68.4
Sources and Notes: The data for 1952-85 are taken from Comprehensive Statistics of China's Rural Economy (Rural Statistics) (1989: 60-61), in which the figures before 1957 are based on 1952 constant prices, before 1970 on 1957 constant prices, before 1980 on 1970 constant prices, and before 1986 on 1980 constant prices. The numbers for 1992 are from Statistical Yearbook of Rural China {Rural Yearbook) (1993:37) and based on current prices.
Apart from supplying basic food to the population, agriculture has been closely related to the industry. Raw materials for industry are largely provided by agriculture. Agriculture supplies the food processing industry, textile industry, breweries, tanneries, the pharmaceutical and other industries with grain, cotton, vegetable oil, sugar, cocoons, tea leaves, wool, hides, medicinal herbs and other products. Table 4.1 clearly depicts this linkage. In 1952 the output value of light industry using agricultural products as raw materials was 88 per cent of the total of light industry. After the 1960s the development of the petrochemical industry made chemical fibres, plastics and synthetic materials available for light
116
Chapter 4
industry. The use of these oil-based materials became increasingly widespread, and led to a revolutionary change in the composition of the supply of raw materials for industry. By 1985 the output value of light industry using agricultural products as raw materials had fallen to 67 per cent of total output, and this figure has remained stable. Despite this relative decrease, however, the total demand for agricultural products by industry has been steadily increasing due to the rapid expansion of light industry. The purchase of farm and sideline products by industry in the total state procurement increased from 12 per cent in 1978 to 20.4 per cent in 1992. The purchase of grain by the state rose from about 21 per cent at an average in the 1970s to 35 per cent during 1986-92 {Yearbook 1993:606,609). Agriculture supplies industry with a large quantity of raw material, and in turn, the agricultural sector and rural population provide an extensive market for the industrial products. The retail sales in rural areas accounted for about 50 per cent of total retail sales in the whole country during 1952-78, and after 1979, this share increased to over 55 per cent {Yearbook 1993: 611). At present about 60-70 per cent of the products produced by light industry are sold in the countryside. Products produced by heavy industry, such as chemical fertilizers, motor vehicles, tractors, large and medium-sized farm machines and farm implements, also find an expanding market in rural areas. The export of agricultural products makes up a substantial proportion of the export trade and thus contributes to releasing foreign exchange constraint. The items exported include grain (especially rice and soybeans), peanuts, fresh eggs, live pigs, poultry, fruit, aquatic products, cotton yarn, flue-cured tobacco and a wide range of other plant, animal products and by-products as well as processed items. Table 4.2 provides relevant data for selected years. Although the percentage contribution of agricultural exports, particularly that of unprocessed agricultural products, to the export trade has fallen, in volume terms the exports of both unprocessed and processed agricultural products have increased. In comparison with exports, unprocessed and processed agricultural products form only a small part of total imports, despite an increase in the related shares. For instance, the value of agricultural products imported in 1985 was US$6.39 billion, or 18.6 per cent of the total import {Comprehensive Statistics of China's Rural Economy, hereafter Rural Statistics, 1989: 499, 515).
Agricultural Constraint to the Insatiable Investment Demand Table 4.2
117
The values and shares of the exports of unprocessed and processed agricultural products in China's total exports Exports of agricultural products
Year
1953 1957 1965 1978 1986 1992
Tntfll pvnnrtQ — (US$ millions)
1,022 1,597 2,228 9,745 27,014 85,000
Unprocessed (US$ millions)
Processed (US$ millions)
569 640 737
265 503 802
2,691 5,202 n.a.
3,414 8,472 n.a.
As % of total exports 81.6 71.6 69.1 62.6 50.6 41.9
Sources: Data for 1953-86 are taken from Rural Statistics (1989: 516-19). Figures for 1992 are from Rural Yearbook (1993:37) and Yearbook (1993:633).
As indicated in Tables 1.4 and 3.2, more than 90 per cent of the funds needed for economic development nationally have been generated through domestic financing. Of the domestic savings, a substantial proportion has come from the agricultural sector. It is widely recognized that the main characteristic of China's capital accumulation mechanism is extraction from agriculture. As discussed in section 2.3.3, the political economy logic of the extraction mechanism is that a heavy industryoriented development strategy, a distorted macro-policy environment, planned allocation of resources and induced institutional arrangements together insure low wages and low prices for inputs in the non-agricultural sector. This causes low peasant incomes, low consumption by both peasantry and non-agricultural workers and high profit in the non-agricultural sector. This high profit and low consumption contribute to high capital accumulation. The real wage rate of non-agricultural workers increased by only 12.7 per cent in the 26 years from 1952 to 1978, when real national income per capita nearly tripled (Yearbook 1993: 132, 3 3 34, 81). During the reform period, especially, between 1984 and 1993, this basic mechanism seems to have changed very little (cf. section 2.3B}.brief, fluctuations in the supplies of agricultural raw materials significantly affect industrial planning and production, while the entire nonagricultural workforce is dependent upon food supplies from agriculture. Moreover, a large part of foreign exchange earnings had been generated from the export of farm and farm-related products until very recently. Thus, it is of great importance that the economic authorities should seek
118
Chapter 4
to minimize the degree of agricultural fluctuation. It is also imperative to raise farm capital accumulation for investment in improvements such as irrigation and drainage capacity and in modern material inputs, for a traditional agriculture like that of China to be improved and modernized, thereby enabling farming to enter a stage of sustainable and steady growth. In the mid-1950s the government became aware of the competition for resources between industrial expansion and agricultural development (Mao 1956). The government then adopted a specific agricultural development strategy characterized mainly by accumulating production capital through mass labour inputs. This is known as 'labour accumulation' in the literature. The intention of the strategy was (a) to minimize the competition for resources through the mass mobilization of rural labour to work on labour-intensive investment projects such as water conservation, irrigation, prevention of soil erosion and land reclamation, and (b) to raise unit yields through traditional labour-intensive methods and inputs, such as closer planting, more careful weeding and the use of more organic fertilizer (Lin et al. 1996, Saith 1985,1995). However, because the relation between labour accumulation and agriculture's demand for modern material inputs is far from a perfect substitution (as will be shown in the following sections) the policy effort to stabilize and increase agricultural output has led to diverting investment resources from the preferred industrial sectors into the agricultural sector. This is why the macroeconomic adjustment and investment cycle has been closely linked to price scissors between agricultural and industrial products, to income disparity between the urban and rural population, and most importantly, to how much agriculture can be exploited.
4.3 The Change of Factor Proportions in China's Agriculture It is well known that in China, cultivated land is scarce, capital factors fairly scarce and labour relatively abundant. The factor structure of China's agriculture has experienced two phases of evolution since the foundation of the Peoples' Republic. The relevant data are listed in Tables 4.3 and 4.4.
163.33 173.16 193.00 283.13 308.62 339.66 329.10
31.07 46.10 57.53 92.12 143.09 222.04 292.59
22.88 34.00 40.83 54.81 83.98 117.72 135.09
Gross output Value added (billions of (billions of 1952 yuan) 1952 yuan)
9.00 9.35 8.70 5.25 4.72 4.20 4.35
Cultivated land per labourer (mu)
140.1 196.4 211.6 193.6 272.1 346.6 410.5
Value added . per labourer5 (1952 yuan)
24.9 26.2 20.9 29.4 30.0 42.5 40.7
Current prices
26.4 26.2 29.0 40.5 41.3 47.0 53.8
1952 prices
Material input ratio0 (%)
Sources: Cultivated land: The data for 1949-84 are taken from Xu & Peel (1991: 58), for 1993 from Yearbook (1994: 329), and for 1996 from Agricultural Yearbook of China (1997). For other data, see the Data Appendix of this chapter.
(a) In millions of mi/; 1mu = 0.0667 hectare or 0.1647 acres. (b) Value added presents the gross output value minus the sum value of material inputs, namely, the net output in Chinese statistics. (c) The material input ratio of agriculture means that the material input as a percentage of agricultural gross output value.
1949 1952 1957 1978 1984 1993 1996
Notes:
1470 1619 1680 1490 1460 1427 1432
Year
Labour force (millions)
The changes of factor proportions in China's agriculture, 1949-96
Cultivated land3
Table 4.3
I
I !
8.01 2.27 7.62 5.00 9.60
-0.42 -2.38 -1.76 -1.29 1.18
Landper
5.29 -0.42 5.84 2.45 5.80
Real value added per labourer (%) labourer (%) 1952 prices
1.18 1.60 0.33 1.45 4.61
Current prices
-2.17 1.64 0.34 3.95 -1.43
Material input ratio (%)
13.03(57) 22.99 (77) 28.91 (84) 37.48 (92) 43.39 (96)
Organic fertilizer nutrients (million tons)
31.52(93) 38.28 (96)
0.37 (57) 8.84 (78) 17.40(84)
Chemical fertilizer appliedb (million tons)
The average annual growth rates of key indicators of agriculture, 1949-96
1.21 (57) 117.50(78) 194.97(84) 318.17(93) 385.47 (96)
Power of agricultural machinery (billion watts)
(a) GOVA indicates the gross output value of agriculture. (b) Chemical fertilizer applied is calculated on the basis of 100 per cent effectiveness. Sources: All growth rates are derived from Table 4.3. Figures of chemical fertilizer applied and total power of agricultural machinery are taken from Yearbook (1993: 349, 341; 1997: 373, 371). Data on organic fertilizer nutrients for 1957 and 1977 are directly cited from Tang & Stone (1980: Table 17), and the others are my estimation based on Tang's method (Tang & Stone 1980).
Notes:
1949-57 1957-78 1978-84 1984-93 1993-96
Periods
Table 4.4
to
I
o
Agricultural Constraint to the Insatiable Investment Demand
121
In the first phase, from 1949 to 1957, cultivated land area increased from 1.47 to 1.68 billion mu. The net increment was 210 million mu and the average annual net increment was 30 million mu. The agricultural labour force increased from about 163.34 million in 1949 to 193.09 million in 1957, with a net increase of 29.75 million. The ratio of incremental cultivated land to incremental agricultural labour force was 7.06 mu per labourer. It was slightly less than the 8.7 mu per labourer stock ratio of 1957. Therefore, the increased demand for agricultural products in both processes of population growth and industrialization could be satisfied on the basis of increased output coming from the combination of increased land and labour inputs. In fact, from 1949 to 1957 the annual growth rate of agricultural output in 1952 prices was 8.01 per cent, and between 1952 and 1957 the growth rate was still 4.53 per cent. This occurred in spite of the ratio of material input to output in current prices decreasing from 26.2 per cent in 1952 to 20.9 per cent in 1957, and the real material input ratio (in 1952 prices) rising slowly compared to the rapid growth of agricultural output (for details, see Table 4.3). This enabled the creation of the 'maximum speed and selective growth' development strategy - to support priority industrialization with resource transfers from agriculture in the form of food (the principal wage good), labour, raw materials and exportable farm products. In the second phase, namely after 1957, cultivated land area decreased. From 1957 to 1993, the net decrease in cultivated land was 253 million mu. At the same time the net increase in the agricultural labour force was 147 million, which means that cultivated land per labourer decreased from 8.7 mu in 1957 to 4.20 mu in 1993 (Table 4.3). Due to the decrease of the cultivated land area and the law of diminishing marginal output of labour, the ways to support the growth of agriculture must shift to raising the productivity of land through increasing both traditional and modern material inputs into agriculture. These land saving means mainly include the intensive application of organic and chemical fertilizers, increasing use of pesticides, heavy dependence on machine-powered irrigation and plastic film covering and investment in research and development (R&D) of seeds improvement. However, as the marginal output of industrial inputs also follows the law of diminishing returns, maintaining the necessary growth rate of gross agricultural output requires that the growth rate of material input must be faster than the rate of gross output. This is confirmed by the rising material input ratio shown in Table 4.3. This economic law forces the economy to increase material in-
122
Chapter 4
puts, particularly industrial goods and R&D investment, into agriculture, making the scramble for scarce resources between agriculture and industry increasingly fierce. It can be seen in Table 4.4 that the rural reforms of 1978-84 resulted in a fall in the growth rate of the material input ratio while the annual growth rate of real value added per labourer reached a record level (5.84 per cent). This implies that the increase in effective labour inputs stimulated by the household responsibility system acted as a substitute for material inputs. However, once the reform process was completed this substitution effect marginally disappeared. The rapid rise of material input re-appeared and the higher growth rate of the material input ratio corresponded to the relatively lower annual growth rate of real value added per labourer (2.45 per cent in 1984-93). These trends confirm that the growth of agriculture remains dependent on an increasing material input ratio, as cultivated land per labourer declines or remains small. The scramble for scarce resources between agriculture and industry will also continue.
4.4 The Specific Institutional Setting to Help Minimize Agricultural Fluctuation The institutional setting in China's agricultural sector has significantly helped to stabilize agricultural fluctuation by minimizing sown-area abandonment and yield losses caused by adverse weather conditions and/ or changes in market conditions, with one or more exceptions.3 In order to understand how China's agricultural institutions work against agricultural fluctuation, it helps to show some distinctions between China and the former Soviet Union, between China and India, and between modern and pre-war China. In both the former Soviet Union and pre-reform China, agriculture was characterized by a system of compulsory farm deliveries with the agricultural collective serving as its institutional vehicle. In the former Soviet Union the system remained intact from 1928 onward. China, however, has alternated between the familiar policy approaches of centralizing (from the early to late 1950s, and again from the mid-1960s to the mid-1970s) and decentralizing (in the first half of the 1960s and again during the reform period). Meanwhile, the state's compulsory procurement (by plan or contract) and the collective allocation of land have persisted since 1953, and these physical and administrative controls placed a
Agricultural Constraint to the Insatiable Investment Demand
123
limit on the market profit-seeking motivation of the peasants, and so helped to reduce market-caused agricultural fluctuation. However, the agricultural constraints confronting Soviet planners were never as serious as those in China. In the former Soviet Union, cereal output per capita stood at nearly 500 kg in 1928 when agricultural collectivization was initiated, and at 700 kg in 1989. This is much higher than China's 1952 per capita production of 285 kg of grain (including tuber crops and dry beans)4 and also higher than its 1996 per capita production of 412 kg of grain. Per capita cereal availability on farms, net of forced procurement, was 415 kg in 1932 in the Soviet Union, but only with 269 kg in China in 1957 and still only 366 kg in 1992.5 Such a direct comparison may lack accuracy as it ignores the difference in the food structure between the two countries. Food in the former Soviet Union consisted mainly of meat and dairy products, which require more grain per capita. However, on the basis of the necessary demand for calories per capita per annum, taking roughly 250 kg of grain as the subsistence requirement per person, the grain surplus was more limited in China than in the former Soviet Union. One of the major concerns of the Soviet authorities was how to extract a sufficient workforce from the agricultural sector to meet increasing urban industrial demands for labour, whereas China's leadership has been faced with tremendous population pressures in both urban and rural areas (cf. section 2.3.2). Therefore, the heavyindustry-oriented development strategy in China is specific in its strict control over rural labour mobility by collectivization and a resident registration system, which serves to stave off large-scale rural out-migration and restrict occupational mobility in order to stabilize and increase grain output. Such institutional restrictions on labour movements of both location and occupation not only coerce peasants to fight against floods and drought in order to minimize short-term output fluctuations. They also facilitate mass labour mobilization to construct large-scale reservoirs, dams, and irrigation canals, all of which constitute the infrastructure for installing powered water-siphoning stations, hydropower plants and sprinkle irrigation. This institutional arrangement is in sharp contrast to that of, for example, India and pre-war China. In India and pre-war China the rural setting was dominated by small-scale household holdings of land and market relations, with hardly any direct government control. These poor peasant farmers would be less sensitive than income-maximizing European farmers to changes in market conditions not caused by
124
Chapter 4
weather disturbance, because of the subsistence urge. Nevertheless, in the absence of institutional restrictions, they would also be prompted by this same 'subsistence urge' to join the large-scale rural emigrations normally associated with natural disasters and famines, which would aggravate the degree of fluctuation in agricultural input and output (Kueh 1995, Saith 1995). For instance, in 1931 the disastrous Yangzi and Huai River floods alone forced about 40 per cent of the rural population from the two basins to abandon their farmsteads, either temporarily or permanently.6 This greatly affected farmwork in ensuing crop seasons. Wide tracts of arable land could not be sown, nor could mass mobilization for reconstruction take place. Weather is, of course, an important cause of agricultural fluctuation, as the year-to-year variation in regional grain yields shows (Kueh 1995). These weather effects make the agricultural fluctuation variable weakly exogenous rather than endogenous in the investment determination model established in this book. However, the above analysis indicates that institutional restrictions on population movements, labour accumulation imposed by collectivization and the resident registration system have relatively minimized output fluctuations arising from labour migration prompted by bad weather or natural disasters. Unlike the former Soviet Union, India, and pre-war China, the present institutional arrangements for minimizing the effect of bad weather create a closer link between agricultural fluctuations and policy cycles, particularly distributive shifts of national income between industrial investment and agricultural development. Such close linkages indicate that there is a two-way dependence between agricultural fluctuation and investment adjustment, and imply that the agricultural fluctuation variable is unable to become more than weakly exogenous in the variable system established for modelling investment functions.7
4.5 Selection of Indicator System The selection of an indicator system is based on the following criteria: First, the system should be able to document empirically the input-output relations in agriculture and the distributive relations in the uses of national income, so as to show the two-way dependence between agricultural fluctuations and relevant macroeconomic adjustments. Secondly, the system should include as few indicators as possible. These must not only represent the most sensitive supply constraint of agriculture to the
Agricultural Constraint to the Insatiable Investment Demand
125
macroeconomy, but also indicate (a) the incentives received by the peasants and (b) the marketable surplus of agriculture. The marketable surplus may best reflect the supply capacity of the agricultural sector to other sectors, and may represent the precise supply constraint that agriculture sets to industrial expansion. Thirdly, the system should be dominated by those indicators to which the government authorities have been most sensitive.
Real value added per labourer The costs of inputs per unit of cultivated land is rising rapidly. This is because the output growth of agriculture is increasingly dependent on the use of material inputs. If the prices of agricultural products cannot be revised but remain fixed, and if the incentive awards in kind cannot be increased, the rise in the material input ratio is bound to force the share of value added in per unit product to decrease. It can even cause the marginal net income of farmers to drop to zero or negative value. This situation has recurred and is documented: 'the income doesn't increase when the output rises'. If the growth in output cannot enable the farmers to obtain an additional income, then they will lose the incentive and capacity to afford additional inputs, and stagnation of agriculture becomes inevitable. These analyses apply to both collective and household responsibility systems because the budget constraint has been effective and hard for both. It is therefore more helpful to employ value added as the representative indicator rather than gross output value in estimating farmers' incentive to increase production inputs and in appraising the agricultural situation in general. Given the background of a constantly growing agricultural labour force, only by relying on increasing real value added per labourer can agriculture bear the increase of material input ratio, and contribute a part of its surplus to the non-agricultural sectors. In order to measure the agricultural cycle, a key indicator would be the real value added per labourer (RVAPL) which is 'the actual net output value of agriculture per labourer' in Chinese statistical terms. It is worth noting here that the RVAPL not only indicates the net income or incentive rewards received by a representative labourer, but is also closely linked to the marketable surplus of agriculture than is gross output. As indicated in the literature, marketable surplus of agriculture better represents the supply constraint agriculture places on industrial expansion.
126
Chapter 4
Figure 4.1
The fluctuations of per capita grain output and real value added per peasant
450 -
400 Real value added per peasant 350
300 & O 250-
200 1949 1953 1957 1961 1965 1969 1973 1977 1981 1985 1989
1993
Source: See Table A4.2 in the data appendix of this chapter.
Grain output per capita Perhaps an even more powerful indicator is 'grain output per capita' (GRNPC). Its importance stems from the crucial position of grain both in livelihood and in the national economy (cf. section 4.2). Secondly, grain is a relatively homogeneous, non-substitutable, necessary commodity, and its production takes up a large share of total agricultural production. It has also occupied an unfavourable position within the socialist price and resource allocation systems. This means that grain supply represents the most binding constraint on the economy, and the indicator 'grain output per capita' has been regarded as a barometer of agricultural production. In fact, 'grain output per capita' has been a key phrase in most economic and political documents written by the central government and senior leaders. It also appears in leading articles on the economy and politics in the newspapers controlled by the Communist Party, for instance Renmin Ribao {People's Daily). If one tries to identify a variable to which the decision-makers are most responsive based on the frequency statistics of phrases appearing in the above-mentioned documents, the first candidate would be 'grain output per capita'.
Agricultural Constraint to the Insatiable Investment Demand
127
Besides its political importance, this indicator is statistically sound and relatively error-free. The data is comparable for time series and cross-section figures and its significance is clear and readily understood by both officials and the public. By comparison, other professional indicators designed especially by economists and statisticians, such as 'total factor productivity', or 'shortage index of consumer goods market' are complicated, cumbersome, and controversial. The following analysis reveals that the 'grain output' indicator represents the supply constraint effect of agriculture in terms of gross output, and that the 'grain output per capita' indicator is the most sensitive constraint barometer on which both Chinese leaders and scholars have focused. The fluctuation patterns of both RVAPL and GRNPC, as shown in Figure 4.1, moved in an almost synchronous cyclical pattern before 1984, though the variation in GRNPC is more pronounced. After 1984, grain production compared unfavourably with cash crops and other non-cropping agriculture, in terms of land productivity, material input ratio, and the capacity to absorb labour force per unit area (Lin 1992, Sun 1991, World Bank 1991). As a result the output of grain was relatively stagnant, and the per capita grain output fell visibly until 1990. This is in striking contrast to the continuous high growth rate of the actual gross output value of agriculture, which maintained an annual growth rate of 4 per cent from 1984 to 1989 (see Table A4.1 in the data appendix of this chapter). It was, as emphasized by many official documents and government research reports, the stagnation of grain production from 1985 to 1989 that induced the large-scale adjustment in the national economy from 1989 to 1991. In determining labour productivity in agriculture, five factors are usually documented: (a) levels of material inputs, (b) changes in relationships between land and labour inputs, (c) the technology employed, (d)the economic milieu whose critical dimensions, organization and material incentives oscillate with the policy cycle, particularly the policy shift between agricultural development and industrial investment in the uses of national income, and (e) the weather, usually treated as a random phenomenon.
128
Chapter 4
As discussed in section 4.3, changes in factor (b) lead factor (a) to become the fundamental means of supporting the necessary growth of agriculture. Technology is usually treated as a long-term factor. However, since the marginal costs of both technological progress and maintaining existing technical facilities have increased significantly, the fluctuation in short-run input into technical maintenance and improvement as well as the change in technological selections are bound to induce the fluctuation of output. The former fluctuation shows a close link between the factor (c) and factors (d) and (a), and the latter change brings in a technological shock to agriculture and to the macroeconomy. Although agricultural production is sensitive to climatic changes, at the national level, as pointed out by Kueh (1995: 163), 'it is difficult, if not impossible, exactly to isolate, in quantitative terms, the weathercaused yield losses from the possible implications of policy influence. This is readily understandable, given the vast size of China and the complexity of its agro-climatic and regional grain-yield variations.'8 In addition, we should also take into account the specific institutional setting discussed in section 4.4, which has considerably minimized agricultural fluctuations due to weather, and the significant technological progress in China's agriculture. Therefore it is fair to say that the agricultural fluctuations in China are more closely linked to factors (a) and (d) than to factor (e) in comparison with other developing economies (e.g. India and pre-war China). The following section shows that (a) and (d) cause agricultural fluctuation at the national level, the main factor being (d). In fact, fluctuations of material and labour inputs, (a), are mainly driven by policy shifts in the uses of national income between agricultural development and industrial investment and by changes in the incentive mechanism, namely factor (d), from which fluctuations of RVAPL and GRNPC follow. Of course, the weather (e) may be also an essential factor and can introduce an exogenous shock into the agriculture and in turn into the macroeconomy.
Agricultural material input and the price level of farm products procurement Following section 4.3 it is easy to understand that the agricultural material input level represents not only the material cost of growth, but also the peasants' capability and enthusiasm to increase material input into agriculture. The latter in turn responds to the income policies and incentive measures of the State. Under conditions of shortage, the effective
Agricultural Constraint to the Insatiable Investment Demand
129
supply of material inputs and related award measures also illustrate the agricultural terms of trade in kind when long-term stability of price level was emphasized in the 1960s and 1970s. Among the income policies of the state, price policy on farm products procurement has been the most effective (Lardy 1985, Sicular 1993). Figure 4.2
Accumulation ratio and investment ratio
50 .
Capital accumulation ratio
1949 1953 1957 1961
1965 1969 1973 1977 1981 1985 1989 1993
Source: See Table A4.2 in the data appendix of this chapter.
Capital accumulation ratio and investment ratio Both ratios provide sensitive indicators of the Chinese leadership's shortand medium-term economic policy direction, ambitions, commitments, targets and expectations. According to statistical definitions, capital accumulation in China is that part of the national income that is used for expanded reproduction, non-productive construction and increase of productive and nonproductive stock. Thus, in value terms, capital accumulation is equal to the national income used minus consumption. Its material formation is the newly added fixed assets of material and non-material sectors (less depreciation of the total fixed assets) and the newly acquired circulating fund in kind by the material sectors during the year. The proportion of capital accumulation in the national income used is defined as 'accumu-
130
Chapter 4
lation ratio' (Yearbook 1993: 76). The capital accumulation ratio has been treated as one of most important control variables, that is, one of the 'major relationships' in Mao's vocabulary. This is not only because it reflects the distributive pattern of national income between consumption and accumulation, but also because its main component, accumulation in fixed assets, is determined by fixed investment which is, in turn, controlled by the government and its agents. It should also be pointed out that capital accumulation refers to 'net investment' in fixed capital assets, working capital, and material reserves, as it is a part of national income. The gross investment in fixed capital assets is certainly of independent interest for the purpose of this research. As mentioned in section 1.1, the collection of annual statistics on total fixed investment began only in the early 1980s. However, the main component of the total is the fixed investment by state-owned units, which has been closely monitored by decision-makers and recorded in a consistent and reliable accounting category. State fixed investment dominates because the state has used its monopoly power over fiscal institutions and state banks to re-allocate investable funds from the non-state sector to the state sector. Therefore, the proportion of fixed investment by the state sector in the national income used can be defined as the 'investment ratio' and this ratio can be employed to measure the policy trade-off between consumption and investment, and particularly for the trade-off between agricultural development and industrial expansion. As Figure 4.2 shows, the cyclical plots of both the national capital accumulation ratio and the state sector investment ratio exhibit corresponding swings.9 The growth rate of investment is generally used to measure investment cycles because it is normally a stationary process.10 However, what is more important here is that this variable is not necessarily a good indicator of changes in decision-makers' attitudes towards investment versus consumption. In certain years investment fell (rose) simply because income fell (rose), even if the government maintained an unchanged attitude toward investment and kept it at the same percentage of national income.
4.6 Agricultural Fluctuations and Macroeconomic Adjustment: Empirical Evidence This and the following section outline the essential and internal structural forces which drive the cyclical fluctuations of Chinese agriculture and, to
Agricultural Constraint to the Insatiable Investment Demand
131
a lesser extent, of the national economy. Although mistakes in macro decision-making, choices of economic institutions and selections of development strategy all influence the growth of agriculture and the whole national economy, they may all be secondary features. In fact, mistakes and faults in macro decision-making and strategic choices come from a failure to recognize and adapt to changes in the agricultural factor proportions and the intense competition for scarce resources between agriculture and industry. It is the increasingly intense scramble that leads to the cyclical inflows and outflows of development resources between agriculture and industry. This in turn is what provokes cyclical fluctuations in agriculture and industry. Macro decision-making may also make a stormy sea stormier when decision-makers over-react or act on the basis of a weak and delayed understanding so that the cycles become more pronounced. In order to present empirical evidence of the two-way dependency mentioned above, specific tables for each agricultural cycle and its corresponding investment ratio cycle are presented (Tables 4.5—4.11). The characteristic linkage in each individual cycle will be examined briefly in this section, and a summary of 'stylized facts' will be presented in the next section.
First Cycle, 1954-62 The expansion phase, i.e. the run-up and rush phases described in section 1.3, of this cycle had been accompanied and accelerated by radical institutional change and successive political/economic campaigns. The most famous of these were collectivization and communization, the 'Socialist High Tide', the Anti-Rightist Campaign, the Great Leap Forward and the Anti-Right Deviation. Expansion was characterized by 'politics in command' (zhengzhi guashuai) and ideological fever. However, the material foundation sustaining such successive 'tides' and 'leaps' included impressive growth in agriculture from 1954-58. As presented in Table 4.5, value added per labourer grew by 8.1 per cent annually, and grain output per capita grew at an average rate of 1.9 per cent, while the population increased by 57 million (Table A4.1). 1 1 ' n Unfortunately, record agricultural growth reinforced the politically motivated exaggerations of harvest size (fukuafeng) which became widespread during the Great Leap Forward. This made the initial estimate of the 1958 grain harvest 380 million metric tons (MMT), which was double the 1957 figure and out of all proportion (cf. section 1.5). Even in
Chapter 4
132
1959 and I960, the leaders still believed that they had 100 MMT more grain than they actually had (Bernstein 1984: 13). Based on the mistaken belief that harvests had miraculously broken all records, an excessive procurement of grain was imposed by the government through the commune system in 1958, 1959 and 1960, the total scales of which were, respectively, 22, 40 and 6 per cent more than that in 1957. In 1957 gross procurement was 24.6 per cent of the total grain output; by 1959 it had risen to 39.7 per cent, and in the year of highest mortality, it was 35.6 per cent {Yearbook 1991: 588). Such untenable grain procurement was certainly a prime contributor to severe grain shortages in the countryside (Bernstein 1984). Table 4.5
1954-62: The Great Leap Forward and the agricultural crisis it induced
Starting (index)
Expanding period (avg. annual growth rate, %)
Peak (index)
1958
1. Agricultural cycle
1954
1954-58
RVAPL* GRNPC** Farm products purchasing price level Material input (in 1952 prices)
73.2 92.8
8.1 1.9
91.5 77.8
2.2 6.5
//. Investment cycle
1955
1955-60
Investment ratio Accumulation ratio
39.6 52.3
20.4 17.6C
Contracting period (avg. annual growth rate, %)
Trough (index)
1958-62
1962
100 100
-14.4 -15.4 a
53.7 71.5
100 100
-3.1
129.5 88.3
1960-62
1962
-47.1 -38.0 e
27.9 23.8
1960
100 100
6.7
* Real value added per labourer ** Per capita output of grain Notes: The upper turning point is 100 for all indexes in Tables 4.5-4.11. (a) For 1958-60. (b) For 1960. (c) For 1955-59. (d) For 1959. (e) For 1959-62. Sources: See Tables A4.1 and A4.2.
Parallel to the excessive extraction from agriculture, the national capital accumulation ratio and investment ratio rose from about 25 and 16 per cent in 1957 to about 34 and 25 per cent in 1958. Furthermore, capital accumulation and industrial investment remained high while agri-
Agricultural Constraint to the Insatiable Investment Demand
133
cultural income dropped and both ratios reached the peaks of almost 44 and 33 per cent respectively, in 1959 and 1960. These were the highest ratios of capital accumulation and investment in the history of the People's Republic (cf. Table A4.2). The industrial boom also suddenly extracted a considerable amount of labour force from agriculture. The farm labour force dropped from 193 million in 1957 to 154.8 million in 1958, a decline of 20 per cent, and regained its former level only in 1961 (Table A4.1). The most serious consequence of the Great Leap was an agricultural crisis. Grain output dropped 15 per cent in 1959, and was only about 70 per cent of the 1958 level in 1960 and 1961. In a poor agrarian country, one year of bad harvest may exhaust available stocks of food, while two or three in a row can create famine conditions. In fact, a careful analysis of the newly released demographic data indicates that this crisis caused about 30 million excess deaths and about 33 million lost or postponed births in 1959-61 (Ashton et al. 1984). It was the worst catastrophe in the history of modern China and in the human history of the twentieth century {Cambridge History of China 1987: Chapter 8). The Great Leap also led to a severe shortage of equipment and raw materials, which kept many factories operating with excess capacity and impeded the progress of many projects. This was evident in the sharp rise in the incremental capital-output ratio. During the First Five-Year-Plan period (1953-57) this ratio had averaged around 2.9, but for the Second Plan period (1958-62) it rose to 100.13 The latter figure should be regarded as rhetorical to certain extent in consideration of the income collapse in 1960-61. However, there is no doubt that much investment during the Great Leap Forward period was wasted in that it produced zero or negative income flows (Riskin 1987: Chapter 6). Finally, investment collapse followed the macroeconomic collapse, capital accumulation ratio fell to a low of 10.4 per cent and the investment ratio dropped to 9.2 per cent in 1962 (Table A4.2). Both the severity of the crisis and the logic of its therapy required severe cuts in fixed investment.
Second Cycle, 1962-69 The first three years of this cycle have been named as the 'recovery period'. The guiding slogan of the recovery period was 'readjusting, consolidating, filling out, and raising standards'. Readjustment meant correcting the intersectoral relations to restore balance between agriculture and industry, between raw material industries and processing industries,
134
Chapter 4
etc. As a bitter lesson of the Great Leap, the slogan 'take agriculture as the foundation, industry as the leading factor' of the national economy emerged officially from the Ninth Plenum of the Eighth Central Committee in January 1961. This slogan was more strongly defined at the Tenth Plenum in September 1962 by placing agriculture at the top of the sectoral priority list: agriculture, light industry, and thirdly, heavy industry. These two slogans amounted to the same: recognition of the constraining role of agriculture in China's industrialization. However, the real policy meaning of 'agriculture first' was far from clear, permitting some in the West to assume that agriculture was to receive a significant share of state investment. In fact, as Riskin (1987) noted, no official statement even hinted at a practically meaningful definition of 'agriculture first', nor has such a policy ever been carried out. According to the explanation of then Premier Zhou Enlai (1965: 10), the policy implication of this slogan was that 'the scale of industrial development should correspond to the volume of marketable grain and the industrial raw materials made available by agriculture'. Partly based on this recognition and mainly in response to the severity of the disaster, a shift in resources favouring agriculture and agriculturerelated sectors of industry did take place in the early 1960s. The compulsory deliveries of grain to the state were reduced to 40.5 MMT by 1961, compared to 67.4 MMT in 1959. The quota purchase price for deliveries of grain and other agricultural products was raised by 27 per cent in 1961. Some 20 million workers who had been recruited from the villages during the Great Leap were sent to the countryside in 1961-62. The share of state capital investment allocated to agriculture, water conservancy projects, and the like rose to 23 per cent in 1963, compared with 7.8 per cent in the period of 1953-57 and 9.5 per cent in 1958 and 1959. However, the state commitment to the agricultural sector never became strong enough to prevent it from being the bottleneck of the national economy. In a broader sense the opposite was true. The economic growth based on heavy industry was soon resumed, partly in the name of constructing inland industrial bases in accordance with recovery. The capital accumulation ratio rose sharply to 27.1 per cent in 1965 and 30.6 per cent in 1966. The share of investment allocated to agriculture had dropped to 13.9 per cent by 1965, and the procurement price of agricultural products was reduced, on average, by 6 per cent between 1962 and 1965 {Yearbook 1991: 254, 589; Statistics on Fixed Investment, 19501985 1987: 73-74; Table A4.2).
Agricultural Constraint to the Insatiable Investment Demand Table 4.6
135
1962-69: The readjustment after the crisis and construction of inland industrial bases
Starting point (index)
/. Agricultural cycle RVAPL* GRNPC** Farm products purchasing price level Material input (in 1952 prices) //. Investment cycle Investment ratio Accumulation ratio
1962 76.8 76.9 a 99.9 66.8 1962 55.5 34.1
Expanding period (avg. annual growth rate, %)
Peak (index)
Contracting period (avg. annual growth rate, %)
1962-66
1966
1966-69
6.8 4.8b
100 100
-3.5 -2.4
1969 89.8 92.8
0.03 10.6
100 100
0.0 0.0
100.0 100.0
1966
1966-68
1968
100 100
-19.5 -16.9
64.8 69.1
1962-66 15.9 30.9
Trough (index)
* Real value added per labourer ** Per capita output of grain Notes: (a) For 1960. (b) For 1960-66. Sources: Tables A4.1 and A4.2
A direct comparison of the different outcomes of recovery programmes for industry and agriculture reveals the continued industrial bias rooted in the system. Industrial recovery was very fast and impressive. Light and heavy industries grew at the average rates of 27 and 17 per cent per year, respectively. By 1965, the level of output of such major products as steel, electric power, cement and heavy trucks was more than double that of 1957. In contrast, the agricultural recovery programme was only moderately successful. Grain production had peaked in 1957 and 1958, but then dropped sharply in 1960, reaching the lowest point of 143.5 MMT since 1952. Output then grew, and by the peak year of 1966 nearly reached 1957 levels. With the population having increased by more than 90 million since 1957, per capita grain consumption remained almost 10 per cent below 1957 levels and was also actually below the 1952 level. Rural consumption was even more depressed, remaining at 13.3 per cent less in 1965 than in 1957. Similarly, rural consumption of cotton cloth was 17.1 per cent less in 1965 than in 1957 and the corresponding national figure remained 8.7 per cent below the 1957 level. Rural consumption of edible vegetable oils was only 1.1 kg per person
Chapter 4
136
annually, 40 per cent less than in 1957, while the parallel national figure remained 20 per cent below that of 1957 (Cambridge History of China 1987: Chapter 8; Yearbook 1991: 55, 425-27; Table A4.1). In fact, this pattern of imbalance basically remained unchanged until 1980. The per capita grain consumption of the rural population for 1978-80 was still several percentage points below that of 1955-57, and approximately 100 million peasants were reported to have yearly per capita grain rations of less than 150 kg (Jiang et al. 1980, Walker 1982). Table 4.7
1969-72: Post-Cultural Revolution Advance
Starting (index)
1. Agricultural cycle
Expanding period (avg. annual growth rate, %) 1969-70
Peak (index)
Contracting period (avg. annual growth rate, %)
Trough (index)
1970
1970-72
1972
RVAPL* GRNPC** Farm products purchasing price level Material input (in 1952 prices)
97.0 89.1
3.1 5.9
100 100a
-0.6 -6.0b
98.9 94.0
100.0 94.4
0.0 5.9
100 100
1.2 2.5
102.4 105.2
//. Investment cycle
1968
Investment ratio Accumulation ratio
1969
51.8 62.1
1968-71 24.5 17.2
1971 100 100
1971-73 -3.3 -1.7
1973 93.6 96.6
* Real value added per labourer ** Per capita output of grain Notes: (a) For 1971. (b) For 1971-72. Sources: Tables A4.1 and A4.2
The contraction of fixed investment and industrial production in this cycle has been mainly attributed to the disruption caused by the Cultural Revolution in the years 1967 and 1968 (Cambridge History of China 1991: Chapter 6, Riskin 1987). The impact of interaction between agriculture and industry cannot, however, be ignored. Agriculture nationwide was much less affected by the Cultural Revolution than industry was, but it did experience quite significant contraction in comparison with all following cycles (cf. Tables 4.6-4.11). Some of the decline in agricultural production might be attributed to bad weather, but also due to the lack of
Agricultural Constraint to the Insatiable Investment Demand
137
necessary inputs. In 1968 the supply of chemical fertilizer fell by more than 30 per cent, and the shortfall would have been worse if the sharp drop in domestic chemical fertilizer production had not been matched by a 65 per cent rise in import between 1966 and 1968 {Comprehensive Statistics on China's Rural Economy, 1949-1986 1989: 334-5, 340). In brief, in addition to the general problems posed by bad weather and the negative impact of rapidly rising capital accumulation rate on agriculture, the specific disruption of industry and transport by the Cultural Revolution certainly aggravated this agricultural recession, which in turn worsened the depression of industry and investment.
Third Cycle, 1969-72 The years of 1969 and 1970 have been acknowledged as another recovery period after the chaotic stage of the Cultural Revolution. By 1970, total output of both industry and agriculture had not only recovered to the previous peak levels achieved in 1966, but also surpassed those levels and regained the long-term trend line. The rapid recovery again stimulated investment fever, now characterized by an over-expansion of military projects. The officially published figure on defence spending granted by state fiscal budget rose to 16.9 billion yuan (8.4 per cent of national income) in 1971, compared with its last peak of 10.1 billion yuan in 1966. Western estimates suggest that China's overall defense spending rose from 24 billion yuan in 1965 to more than 40 million yuan in 1971 and accounted for 20 per cent of the national income in that year. As a result of this over-expansion, grain sales, total wage payments and total employment in the state sector went far beyond planned levels. The imbalance of the economic structure again became a major problem, and readjustment and contraction followed {Cambridge History of China 1991: 494, Chronicle of Economic Events 1985: 333-4, Yearbook 1991: 40, 214). The contraction was further aggravated by the great North China drought in 1972. According to Kueh's (1995) measurement against the estimated trend for 1970-84, this led to a 3.9 per cent average grain loss per sown hectare, and may have forced a shift of resources from industry to agriculture.
Chapter 4
138 Table 4.8
1972-77: Initial Attempt at the 'Four Modernizations'
Starting point (index)
Expanding period (avg. annual growth rate, %)
1972
RVAPL* GRNPC** Farm products purchasing price level Material input (in 1952 prices)
90.1 89.6
3.5 3.7
100 100a
-2.0 -1.7
95.9 96.7
98.5 86.3
0.5 5.0
100 100
0.1 2.9
100.2 105.9
1975-77
1977
-2.1 -2.3
95.9 95.5
Investment ratio Accumulation ratio
1973 87.5 97.2
1973-75 6.9 1.4
1975
1975 100 100
1975-77
Trough (index)
/. Agricultural cycle
//. Investment cycle
1972-75
Peak (index)
Contracting period (avg. annual growth rate, %)
1977
* Real value added per labourer ** Per capita output of grain Sources: Tables A4.1 and A4.2
Fourth Cycle, 1972-77 This is a moderate cycle. Lin Biao's fall in late 1971 brought a favourable environment to cut back military investment. Even the officially published military budget fell by several billion yuan in 1972-74 and then stabilized until 1978-79 (Yearbook 1991: 214). The economy, particularly agriculture, made an impressive recovery in 1973 and 1974 following the readjustment. In January 1975, Zhou Enlai put forward his famous call for 'four modernizations' in two stages: the first was to build 'an independent and relatively comprehensive industrial and economic system' by 1980; the second was to 'accomplish the comprehensive modernization of agriculture, industry, national defence, and science and technology before the end of the century, so that our national economy will be advancing in the front ranks of the world' {Peking Review, 24 January 1975). In the summer of 1975, under the charge of Deng Xiaoping, a very ambitious Ten-Year Plan for Economic Development (1976-85) was discussed by the State Council. This was followed by a draft plan, approved by the Political Bureau. This initial attempt at the four modernizations was disrupted in 1976 by the political chaos following the third dismissal of Deng Xiaoping and the deaths of Zhou En-
Agricultural Constraint to the Insatiable Investment Demand
139
lai, Zhu De and Mao Zedong, one after another. The massive Tangshan earthquake in the summer of 1976 and the severe floods of 1977, which affected the most productive zone in South-East China, also forced a shift in resources from industrial investment to disaster relief. However, no readjustment programme was planned and implemented at that time, and the contraction was moderate. Table 4.9
19 77-80: Post-Mao import of technology and the subsequent economic readjustment
Starting (index)
1. Agricultural cycle
RVAPL* GRNPC** Farm products purchasing price level Material input (in 1952 prices) //. Investment cycle Investment ratio Accumulation ratio
1977 88.4 87.4 83.1 79.9 1977 94.8 88.5
Expanding period (avg. annual growth rate, %) 1977-79 6.4 6.9 9.7 11.9 1977-78 5.5 13.0
Peak (index)
Contracting period (avg. annual growth rate, %)
Trough (index)
1979
1979-80
1980
-3.5 -4.6 3.5 6.1
96.5 95.4 103.5 106.1
1978-81
1981
-8.7 -8.2
76.0 77.5
100 100a 100 100 1978 100 100
* Real value added per labourer ** Per capita output of grain Sources: Tables A4.1 and A4.2.
Fifth Cycle, 1977-80 The death of Mao in late 1976 brought to an end a ten-year period of political conflict and administrative deterioration. However, the industrial strategy in the first two years after Mao represented only a limited departure from the industrial policies of the past. The initial response of the new leadership to the degraded state of the economy was to emphasize the rehabilitation of the command economic system. They began to rebuild the institutions that had deteriorated during the Cultural Revolution, and to implement the Ten-Year Plan (1976-85) originally drawn up in 1975, when Deng Xiaoping ran the economy. The Ten-Year Plan
140
Chapter 4
gave heavy industry the priority that had guided China's development since the early 1950s. It centred on 120 mega-projects in industry and transport, including 10 iron and steel complexes, 9 nonferrous metal complexes, 30 power stations, 10 petrochemical plants, 10 synthetic fibre plants, 6 new trunk railways, 5 key harbours and 10 new oil fields on the same scale as Daqing. Investment over the 8 years remaining in the 10 was to 'far exceed' the total of the previous 28 years combined (Hua 1978). Accompanying this massive investment plan was an equally ambitious scale of planned capital and technology imports. With little thought about the overall costs of these import-dominated projects or their funding, and due to the lack of any mechanism to prevent government organs from taking on 'one more' grandiose project, individual ministries and other government organs made increasing claims on the economy's resources and stepped up their negotiation with foreign suppliers in 1978. Because there was no clear guideline on which agencies could take on projects, Chinese buyers were moved to conclude agreements as quickly as possible in order to lock their own governments into commitments. As a result, in 1978, 22 large-scale projects were 'introduced from abroad all at once', which not only exceeded China's ability to pay, but also saddled the economy with an obligation to commit an even greater sum to the provision of ancillary equipment and parts. Moreover, projects were chosen with little regard to relative urgency of need {Commentator 1981). This boom was later pejoratively dubbed the 'foreign leap forward'. As the various government organs jockeyed for position with competing mega-projects, the situation spiralled out of control, in 1978 the year's investment plan was twice revised upward. Planning organs were unable to cope in the 1979 economic plan because of shortfalls in fuel, finished steel, cement supplies and budgetary revenue (Li, Fuchen 1981; Li, Zhining 1987: 444). The imbalance worsened as for almost two decades, more than 25 per cent of the rural population of 800 million had lived below the poverty line by Chinese standards and discontent was mounting in the countryside, where peasants outbreaks were spreading quickly (Chen, Xiwen 1993: 51; Zhang & Yi 1995). In urban areas, the general freeze of wages after 1957 and the entry of new workers into the lowest rungs of the wage ladder caused the real average wage to fall 17 per cent between 1957 and 1977 {Yearbook 1982: 411-12, 435^6). Neglect of 'non-productive' investment over a long period meant that the
Agricultural Constraint to the Insatiable Investment Demand
141
already cramped housing space of 4.3 square metres per urban resident in 1952 had decreased to 3.6 square metres in 1977 (Walder 1984: 24). Millions of 'young intellectuals' also returned from the countryside needing jobs. The resulting widespread dissatisfaction periodically erupted into major strikes and street protests in many industrial cities (Dittmer 1987, Zhang & Yi 1995). Shortages of consumer goods prevailed and most necessities such as foodstuffs, matches and toilet paper were rationed. This crisis was clearly caused by the long-lasting imbalances and disproportions accumulated during the Cultural Revolution decade. This has been nicely summarized in a People }s Daily commentary (9 April 1981), which in the style of the times ascribed them all to 'left' mistakes: 'Left' mistakes manifest themselves mainly in high targets, high accumulation, low efficiency and low consumption; emphasis on capital construction to the neglect of agriculture and light industry; emphasis on production to the neglect of people's livelihood; emphasis on production to the neglect of circulation. This crisis triggered the beginning of the reform era, and brought a new period of'readjustment' from 1979 to 1981. This period of readjustment was not limited to the distribution pattern of the national income and to movement away from rapid growth. The central elements of the programme were to be called 'reform' rather than 'readjustment', and included (a) tackling the over-concentration of authority in economic management; (b) reforming the commune system in agriculture and improving farm incentives; and (c) raising living standards {People's Daily, 24 & 25 December 1978, 22 January 1979). The success of this readjustment has been popularly confirmed and it laid a solid foundation for the impressive economic growth in the 1980s.
Sixth Cycle, 1980-89 The most important four factors which contributed to the remarkable and lasting industrial growth and investment expansion from 1981-87, were the success of rural reform, the remarkable increase of household savings, the development of township and village enterprises (TVEs) and industrial reform and the openness to the world market.
Chapter 4
142 Table 4.10
1980-89: Economic reform and the readjustment induced by stagnation of grain production in 1985-89
Starting (index)
1. Agricultural cycle RVAPL* GRNPC** Farm products purchasing price level Material input (in 1952 prices) //. Investment cycle Investment ratio Accumulation ratio
1980
Expanding period (avg. annual growth rate, %)
Peak (index)
1980-84
1984
Contracting period (avg. annual growth rate, %) 1984-89
Trough (index)
1989
100 72.3 82.8 95.9 73.1 1981 72.0 83.1
8.5 4.8 1.1 8.1
100 a
1981-87
1987
5.6 3.1
1.7 -2.3 a 18.5
100 100
5.0 1987-89
100 100
-11.3 -1.3 C
108.6 90.9 b 233.6 127.6 1989 78.6 96.2 d
* Real value added per labourer ** Per capita output of grain Notes: (a) For 1984-88. (b) For 1988. (c) For 1987-90. (d) For 1990. Sources: Tables A4.1 and A4.2.
The success of rural reform characterized by the household responsibility system, combined with the radical change in distribution of national income in favour of agriculture, brought accelerated agricultural growth from 1980 to 1984, which broke all records since 1954. Grain output per capita grew at the average rate of 4.8 per cent annually, which is the same as the rate of recovery from the famine of 1959-61. The real value added per labourer increased at the average rate of 8.5 per cent annually, the highest increase in any expansion phase of the seven cycles (Tables 4.5-4.11). As presented in section 4.2, agriculture is not only the supplier of wage goods to industry, but also a major source of inputs to the consumer goods industry. Agricultural exports are also a source of foreign exchange. This extraordinary growth of agriculture greatly lessened its constraint on the economy. Chinese grain imports grew from 8.8 MMT in 1978 to 16.1 MMT in 1982 following the radical readjustment introduced in 1979, before falling to only 6.2 MMT by 1985. The decrease in grain imports alone saved more than 2 billion dollars in foreign exchange (Comprehensive Statistics on China's Rural Economy 1949-
795(51989:500-8,535).
Agricultural Constraint to the Insatiable Investment Demand
143
However, experience has shown that this success was a one-off acceleration brought about by the removal of barriers to efficient production. Once a higher level of efficiency had been achieved, the agricultural growth rate could fall back to a more sustainable level. The distributive pattern of national income between agriculture and industry could also play a leading role again in determining agricultural comparative interest and growth. Excited by the reform effects, the policy shift in 1985-88 once more moved the terms of trade (in terms of both price and material reward) against agriculture, particularly against grain production. The relative price margin paid to farmers was lowered and the mandatory procurement contract was reinforced. This policy shift resulted in a swift outflow of the labour force (-8.6 per cent per year) from the cropping sector to more lucrative sectors and to non-agricultural pursuits, causing a sharp fall in the growth rate of chemical-fertilizer input into grain production. As a result, grain production stagnated and there was an impressive growth of other agricultural sectors and of rural industry (cf. Justin Lin 1992, Kung 1992, Weimer 1992, Sicular 1995, among others). Following the rapid increase in demand for higher quality grain, a surplus of low quality grain accompanied a severe shortage of high quality grain (X. Chen 1993). Apart from the effects of rural reform, the other three factors contribute more to expansion than to contraction. As noted in section 3.4.4, the remarkable increase in household savings made a crucial contribution to the impressive economic growth. Increased savings allow continuing high levels of investment, which in turn not only moderate incipient macroeconomic imbalance, but also provide essential funds for the use of new entrepreneurs. In particular, increased household savings allow the state sector to keep a disproportionately high share of fixed investment through the state banking system, along with the continuing decline of state financing capability for investment and the state share of industrial output (cf. Chapter 3, section 4.5 and note 2 of Chapter 1; Naughton 1995b). Secondly, the takeoff of state sector reform in 1984, together with the extraordinary growth of TVEs, produced an unexpected expansion of output in key bottleneck sectors. From 1984-88, the newly increased steel-smelting capacity from state fixed investment was 11.63 MMT, while the real incremental output reached 19.4 MMT. The extra output can be mainly attributed to the improved efficiency and maximizing the
144
Chapter 4
use of existing equipment through the reform, as well as the entry of TVEs. In the energy sector, while the incremental production capacity based on state fixed-investment was only 136 MMT, the real incremental output reached 265 MMT {Yearbook 1991: 171, 190, 424-25; Statistics on Fixed Investment, 1950-85 1987: 266-69). Here the extra output was mainly produced by new TVEs. Thirdly, the 'Open Door' policy brought remarkable expansion in exports as well as access to international financial markets. The foreign exchange constraint considerably loosened. Foreign investment through loans, joint ventures and other means also increased steadily, from 5.9 per cent in 1984 to 9.0 per cent in 1988 in the state sector investment, and from 3.8 per cent in 1984 to 5.8 per cent in 1988 in total national investment {Statistics on Fixed Investment in China, 1950-85 1987, Statistics on Fixed Investment in China 1988-89 1991). Although there were many new ways to finance industrial investment hunger, the state sector, local governments in particular, still tried to extract investable funds from agriculture by every possible means. In 1987 and 1988, the peak years of over-investment, local government issued IOUs {da bai tiaozi) to farmers instead of paying them cash. One report estimates that in 1988, debts to farmers for each 100 yuan-worth of deliveries of farm products averaged 20 to 40 yuan nationwide {China Daily 10 January 1989: 3). These debts remained outstanding for several months and in some cases for up to a year. This trick of IOUs continued in 1989, although its use became less widespread {People's Daily 5 & 8 August 1989: 6; 30 July 1989: 5). In terms of material rewards, the situation remained the same. A nationwide survey of over 10,000 farm households revealed that in 1987 tied sales of fertilizer and diesel fuel for grain contracts were 20 per cent below levels promised in policy documents {Economic Problems of Agriculture 1988: 49). These new methods of extraction from agriculture are not presented in Table 4.10 because of the lack of annual statistical data. However, they played a powerful role in discouraging farmers and in eroding farmers' confidence in government policies. The coexistence of continuing investment expansion and stagnating food production in 1985-88 resulted in a substantial increase in the rate of inflation. Consumer price inflation had been moderate throughout 1984. In 1985, it started to increase rapidly and in 1988 it reached 20.7 per cent in urban areas. The most important aspect of consumer price inflation was the rising cost of food, which increased by 14.9 per cent in
Agricultural Constraint to the Insatiable Investment Demand
145
1987 and by 31.1 per cent in 1988. The long-lasting investment expansion also caused the rapid increase in raw material prices, although the primary cause seems to have been that an increasing share of goods were transacted at the higher market prices. Accelerating inflation throughout 1987 and the early months of 1988 led to widespread panic buying in the consumer goods market and to some extent also in the producer goods market during the summer of 1988. For a few months in the summer of 1988, inflation spiralled up to annual rates of about 50 per cent. Such inflation had not appeared since the early 1960s and led to the adoption of a deflationary austerity programme in September 1988, which aimed to both curb inflation and to strengthen central controls over investment planning and financing. The drive for retrenchment intensified during 1989, especially in the political aftermath of the events in June of that year. The three-year readjustment programme was adopted in November 1989, calling for curbs on aggregate demand, efforts to increase supply and the imposition of controls over the growing free markets. Sectoral imbalance in general and agricultural stagnation in particular were, once again, seen as key obstacles to stable growth {China Economic System Reform Yearbook 1990: 53-65). Seventh Cycle, 1989-96 The effect of the readjustment programme and of the renewal of political discipline was striking. Market demand collapsed in 1989 and 1990. Real investment declined by 16.6 per cent in 1989 and grew only 3.3 per cent in 1990 (cf. Figure 1.1). The total grain output emerged from the stagnation of the last four years and increased by 3.4 per cent in 1989 and by 9.5 per cent in 1990. Inflation decreased in 1990 and 1991. The cost, however, was a dramatic drop in GDP growth rates, which fell to 3.9 per cent in 1989 and to 5 per cent in 1990, well below the long-term trend of around 9 per cent. Furthermore the real purchases of consumer goods declined by 9.5 per cent in 1989 and the growth rate was almost zero in 1990. However, household saving deposit rates rocketed upward to 2 2 23 per cent of money income in 1990—91, up from 16-17 per cent in 1987-88 (Naughton 1995a; Yearbook 1993: 31, 239, 364, 616, 633). All these conditions led to a new round of investment and economic expansion.
Chapter 4
146 Table 4.11
1989-96: Three-year retrenchment, 'Deng Whirlwind' and the 'soft landing'
Starting point (index)
1. Agricultural cycle
1989
RVAPL* GRNPC** Farm products purchasing price level Material input (in 1952 prices)
85.3 92.2a 101.3 72.3
//. Investment cycle
1989
Investment/GDP Accumulation/GDP
69.4 80.1 d
Peak (index)
Contracting period (avg. annual growth rate, %)
Trough (index)
1989-93
1993
1993-96
1996
4.1 1.6 -0.43 b 8.4
100 100 100c 100
5.8 2.3 18.7 14.7
1989-93
1993
1993-96
Expanding period (avg. annual growth rate, %)
9.5 7.7e
100 100
-7.4 -3.3
118.4 107.5 198.2 151.0 1996 79.0 90.5
* Real value added per labourer ** Per capita output of grain Notes: (a) For 1988. (b) For 1989-92. (c) For 1992. (d) For 1990. (e) For 1990-93. Sources: Tables A4.1 and A4.2.
In March 1990, the State Council proposed a number of expansionary measures, including increasing investment and relaxing credit control. At the end of the year, a Ten-Year Development Programme was proposed, and this was eventually approved by the National People's Congress in April 1991. It called for renewed market-oriented reforms and a revival of the long-term trend of economic growth. The renewed momentum stimulated by Deng Xiaoping in an 'imperial tour' of South China at the beginning of 1992, during which he reaffirmed his commitment to both rapid growth and an open economy and endorsed reform-created institutions. In October 1992, the target of building a 'socialist market economy' was adopted at the 14th Party Congress (cf. e.g. Naughton 1995a, Watson 1994). With the wave of reform, economic growth was remarkable in 1992-93. The real GDP growth rate was 13.6 per cent in 1992 and 13.5 per cent in 1993. Real industrial growth was 27 per cent in 1992 and 28 per cent in 1993. The capital accumulation ratio and the investment ratio reached their highest levels since 1961 (Yearbook 1995: 32, 377; Table A4.2).
Agricultural Constraint to the Insatiable Investment Demand
\A1
The highest accumulation and investment ratios since 1961 were also partly due to excessive extraction from agriculture. IOUs had again been issued since late 1992. Some reports estimate that in the autumn of 1993 about half of the planned fiscal funds and bank loans for purchasing agricultural products were diverted to industrial investment and/or real estate investment in coastal cities. This situation forced the central bank to make additional earmarked loans by money creation to allow farmers to cash the IOUs {China Information Daily 18 August 1993: 1; 2 August 1993: 2; Economic Situation 1992-1993: 31-32). Moreover, the fever of 'special economic development zones' spread extremely quickly between March 1992 and the end of 1993. According to the statistics given by Ministry of Agriculture, local governments built about 9,000 'development zones' within one and a half years at and above township levels, which occupied over 15,000 square kilometres farmland {China Investment and Construction, no. 4, 1994: 24; and Economic Situation 19911992 1992: 31). A more general sign of increasing imbalance between industrial expansion and agricultural development was the continued growth in the income gap between urban areas and the countryside. Per capita rural net income was 784 yuan in 1992 and 921 yuan in 1993, and the corresponding real growth rates were only 5.9 and 3.2 per cent, respectively. In contrast, per capita urban income was 1,826 yuan in 1992 and 2,337 yuan in 1993, up 8.8 and 10.2 per cent in real terms. By 1993, the annual per capita net income of peasants fell to about 39.4 per cent that of urban residents, worse than the ratio of 41.7 per cent in 1978 {People's Daily 19 February 1993: 2; 1 March 1994: 2; Economic Situation, 1993-1994 1994: 26). Besides the resentment of this trend in rural areas, it also stimulated farmers to switch to more profitable cash crops and forced the rural economy to rely more on nonagricultural activity to shore up its average income. As a result, in 1993, nonagricultural income accounted for more than two thirds of rural total income. Area sown to grain fell to 110.5 million hectares, at 1.3 million hectares below the minimum stipulated by the Ministry of Agriculture. The purchasing of agricultural producer goods by farmers decreased by 7.8 per cent in real terms {Economic Situation, 1993-1994 1994: 26; World Bank 1994: 6; Yearbook 1995: 344). A decrease in total grain output followed in 1994 despite good weather conditions (Table A4.1). The danger signal of macroeconomic imbalance was reflected once again in accelerating inflation and panic buying. In December 1992, re-
148
Chapter 4
tail price levels were 6.8 per cent above the level of the year before. By June 1993, inflation had doubled to 13.9 per cent, and by June 1994 it had reached 23.5 per cent. In November and December 1993, the second wave of panic buying occurred and created the highest growth rate of retail sale (over the same month last year) in the reform years {Economic Situation 1992-1993 1993: 121; Economic Situation 1993-1994 1994: 122; World Bank 1994: 1). Faced with recurring economic overheating similar to that in 1987 and 1988, the government introduced a 16-point programme in July 1993, which comprised a mixture of administrative and indirect measures designed to slow the economy to a sustainable rate of growth. For the next two years, government engaged in a difficult readjustment quest for an economic 'soft-landing' (cf, among others, World Bank 1994: 19-23; Naughton 1995a, 1995b; Oppers 1997). In order to avoid further decrease of grain production, central government initiated a new programme named 'the provincial governor's "grain bag" responsibility system' (Midaizi shengzhang fuzezhi) in late 1994. Hereafter we refer to this programme as the 'grain bag system'. The basic points of the 'grain bag system' are as follows. Firstly, central and provincial governments should manage grain production and distribution through a responsibility system. Secondly, the system aims at balancing overall grain supply and demand, stabilizing areas sown to grain, increasing yields and production, stabilizing grain reserves, using grain for disaster releasing and stabilizing prices (Ministry of Agriculture, 1996). Concretely speaking, the programme requires that each provincial governor has to sign a responsibility contract with the state council, and that the governor has to take the responsibility to (a) stabilize areas sown to grain crops, (b) guarantee investment in supply of agricultural inputs like chemical fertilizer to support grain production, (c) raise the level of grain self-sufficiency, (d) insure that certain quantities of grain are added into stock, (e) stabilize the prices of grain and edible oils, (f) control 70-80 per cent of marketable grain sales by government, and (g) develop means to control grain market (Ministry of Agriculture, 1996). In this way, the centre hopes that responsibility is established at the provincial level alone and to keep regional comparative advantages at lower levels. In practice, provincial governments pass the programme down their administrative structure to prefecture, county, even township levels.
Agricultural Constraint to the Insatiable Investment Demand The 'grain bag system' seemed to work effectively during 1995-97. The areas sown to grain crops increased by 0.47 per cent in 1995 and 2.26 per cent in 1996. An increase in 1997 was also reported. Arable land was also reported to have increased in 1995 and 1996 (Table 4.3). Chemical fertilizer supplies from both domestic production and imports increased by 8.1 per cent in 1995 and 6.7 per cent in 1996. Grain output increased from 445 million tons in 1994 to 467 in 1995, reaching a record of 504 million tons in 1996. The overall grain self-sufficiency rate also increased, reaching 96.7 per cent in 1995 and over 99 per cent in 1996 (Crook 1998). Although the 'grain bag system' looks like a new version of the old administrative package, this time the government has the ability to raise procurement prices to a level higher than free market prices in 1996 and 1997. We hope that this signals a turning point in the history of China's agricultural policy-making: a genuine effort to give priority to agricultural stabilization and development.
4.7 Agricultural Fluctuations and Macroeconomic Adjustment: Stylized Facts For the sake of clarity, the key messages in Tables 4.5-4.11 have been plotted on Figures 4.3a to 4.3g. Their common features can be presented as follows. (l)It appears that during the growth period of RVAPL and GRNPC there is an expansion of the investment and capital accumulation ratios. When the growth of RVAPL and GRNPC continues for two or three years, the peak of the investment ratio must emerge. (2) The annual growth rates of the investment and capital accumulation ratios are usually nearly in double digits before, during, and even one or two years after (compare the years of the starting points, peaks, and troughs for each cycle) the peaks of RVAPL and GRNPC. After this, they become negative. However, changes in the purchasing price levels of farm products are often the opposite (with the exception of 1978-79 when the reform was initiated). This means that before the peaks, the annual growth rate of the purchasing price level is very low. This reflects the favouritism decision-makers show to the industrial sector in the distribution of economic resources. After the peaks, the growth rates increased, indicating readjustment in favouritism to
149
150
Chapter 4
the industrial sector in the use of national income. Due to the emphasis on the long-term stability of price levels in the 1960s and 1970s, other policies are more powerful and effective in implementing favouritism to the industrial sector and in the consequent readjustment to the economy. These policies include the assignment of materials and labour, credit policy, fiscal policy, recruitment and repatriation of casual labourers and contract workers from and to rural areas. Unfortunately, due to data limitations it is impossible to establish indicators measuring the changes in these policies. (3) The variations in material input in agriculture are similar to the cycles of RVAPL, GRNPC, and the investment ratio. This reflects the strong dependence of agricultural growth on incremental material inputs. The tables and figures show that during the expansion phase, the pressure to increase material input into agriculture rises rapidly. However, once the pressure becomes unbearable for farmers, stagnation and shrinkage of actual labour productivity and a falling growth rate of material inputs occur. Forced readjustments in the distributive structure of uses of national income follow. Increases in purchasing prices of farm products, reduction in investment outlays, and ensuring the return of casual labourers and contract workers to agriculture soon strengthen the agricultural sector in order to meet the demand for incremental material inputs. (4) The bigger the increases in investment ratio, capital accumulation ratio and material input of agriculture before the peaks, the more serious the recession in agriculture will be. In other words, when the material input ratio of agriculture rises rapidly, the margin in which resources are transferred from the agricultural to the industrial sector widens, and the depression in agriculture worsens. One of the reasons for agricultural stagnation is the favouritism decision-makers show towards the industrial sector in the distribution of national income. The successive cause-effect processes in the empirical patterns above can be summarized in three stages.
Agricultural Constraint to the Insatiable Investment Demand Figure 4.3
151
The correlations between agricultural fluctuation and investment adjustment (a) First cycle (peak = 100)
190 170-
Capital accumulation ratio \ (+100)
150
130 110 90 70
Real value added per peasant 50 1954
1955
1956
1957
1958
1959
1960
1961
1962
(b) Second cycle (peak = 100) 200
Investment ratio (+100) \
180
160
Capital accumulation ratio (+100) 140
120 -
Material input ratio 100 -
Grain per capita \ Real value added per peasant
80
60 1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
152
Chapter 4
First Stage: This starts from the trough of agricultural fluctuation and corresponds to the 'slowdown' phase of an investment cycle as described in section 2.2.2. This trough always marks the beginning of a new adjustment to the national economy. This readjustment consists of decision-makers assigning an incremental share of development resources to agriculture and other bottleneck industries by financial, credit, price and allocation policies, which signify that the investment outlays in the industrial sector have to be reduced. Therefore, following readjustment it is necessary that both capital accumulation ratio and investment ratio are cut down. Casual labourers and contract workers (sometimes even 'permanent workers' registered in the city system) are also ordered to return to agriculture.14 Troughs in the investment ratio and accumulation ratio always follow recessions in agriculture.
Second Stage: This stage is regularly accompanied by the 'run-up' and 'rush' phases of investment cycles. Allocating an incremental share of development resources to agriculture relieves the growth barrier to agriculture formed by the rise in the material input ratio. Once the pattern of macro-allocation of development resources is adapted for the growth mode determined by the current agricultural factor proportions, agriculture will recover and real value added per labourer will rise again. When agricultural production maintains a steady growth for two or three years in succession, the decision-makers consider that the bottleneck of agriculture has been relieved, or that agricultural products are in surplus. This brings a reverse in the policies of resource allocation. The increase in purchasing prices of farm products is slowed down and the price level may even be lowered. Investment outlays are increased at a high rate and the investment ratio and capital accumulation ratio rise once more. There is a large-scale movement of agricultural labourers into the industrial sector as casual labourers and contract workers. The prosperity of agriculture always stimulates the decision-makers to speed up industrialization and the growth peak of agriculture is often followed by the peak of investment in industry.
Agricultural Constraint to the Insatiable Investment Demand
153
Figure 4.3 (continued)
(c) Third cycle (peak = 100) 190
Capital accumulation ratio (+100)
170-
Investment ratio (+100)
150130 110
Material input ratio Real value added per peasant
90 -
Grain per capita
70 1968
1969
1970
1972
1971
1973
(d) Fourth cycle (peak = 100) 125 n
Capital accumulation ratio (+20)
120 -
~~—^—=_
115
Investment ratio (+20)
110 105
^-^~^^
100 -
Grain per capita
95 90
Material input ratio
Real value added per peasant
85
1972
1973
1976
1977
Chapter 4
154
Third Stage: The growth peak of agriculture and the 'halt' of the investment cycle are present almost simultaneously. The high-speed growth in investment and industry requires using vast amounts of cultivated land to expand the infrastructure and to increase building. It also requires releasing agricultural labour - as casual labourers and contract workers who remain registered in the rural system - to the industrial sector, especially to the building industry. This forces the agricultural sector to further depend on incremental material inputs, leading to the rapid rising of the material input ratio. However, it is just in this period that an increased share of development resources is allocated to the industrial sector. The macroallocation of economic resources runs counter to the objective requirement of ensuring agricultural growth. The cost of agriculture rises but the share of net income in total output value falls. These changes then result in agriculture losing its comparative profitability. Agriculture then will stagnate once more, and the national economy will be confronted with a new round of readjustment.
Figure 4.3 (continued) (e) Fifth cycle (peak = 100) 155 145
135125
Capital accumulation Investment ratio (+50) / ratio (+50)
115
Material input ratio
105 - Real value added per peasant
\
95
Grain per capita
85 75
1977
1978
1979
1980
1981
Agricultural Constraint to the Insatiable Investment Demand
155
Figure 4.3 (continued) (f) Sixth cycle (peak = 100) 150 140 -
Capital accumulation ratio (+50) \
130120110
\ ' Real value added per peasant
100
i——_
90
Grain per capita
80 70
1980
1981 1982 1983 1984 1985 1986 1987 1988 1989
(g) Seventh cycle (peak = 100) 200 -
Capital accumulation
\
180 - ratio (+100)
160140 120 100 80- Real 60
value added per peasant
40 1988
1989
1990
1991
1992
1993
1994
1995
1996
1990
156
Chapter 4
There is one exception that needs to be mentioned. This is the 'soft landing' period of 1994-97. As indicated in the subsection of 'the seventh cycle', the 'grain bag responsibility system' was established in this period. For the first time, the government started to use 'protecting prices' to procure agricultural products. As a result, the possible stagnation in grain production was successfully avoided. It is hoped by many that this policy may mark the beginning of a genuine effort to make agriculture the priority sector in China.
4.8 Summary Dealing with interactions between agricultural development and industrial accumulation in the process of industrialization is a recurring subject in both development economics and socialist political economy. This chapter shows that China's practice supplies a typical case for such a study. Based on a carefully selected indicator system and using elaborately tabulated data analysis this chapter reveals a clear two-way interaction between agricultural fluctuation and macroeconomic adjustment. The main reasons for this are both the conflicts between an active industrialization drive and the reactive, unstable agricultural policy of decisionmakers at different levels, and between the response of the peasants to economic incentives and the agricultural policy in view of the increasing dependence of agricultural growth on modern factor inputs. As a result of these acting forces, swings in agricultural production and marketing have been followed by dramatic swings in industrial accumulation and investment. Agricultural growth diminishes the difficulties caused by shortage and thus provides stimulus to investment expansion. In other words, agricultural harvests have been followed by major shifts in macro-distributive policy and significant reductions in incentives and price growth for agriculture. When agricultural shortfalls occur, a passive reversion follows. The distributive policy shifts to increase incentives and prices, and to make cutbacks in industrial capital accumulation and investment. Although recent agricultural policy has appeared to depart from the stylized facts analysed above, the sustainability of the new policy package is not yet clear (Crook 1998). It should be emphasized again that this chapter focuses on showing the two-way interaction between agricultural fluctuation and investment adjustment and does not discuss the full determination of agricul-
Agricultural Constraint to the Insatiable Investment Demand
157
tural fluctuation itself through a qualitative analysis, although the research depends heavily on data analysis of time series. The finding does not imply that agricultural fluctuation is fully determined by such interactions. Weather has also been an important factor in accounting for short-run agricultural fluctuation, although there has been a strong institutional hedge against natural disaster (cf. section 4.4). Free-market forces may also play a role in causing agricultural fluctuation as indicated by the Cobweb Theorem although their effect was very limited before the reform. During decentralization, it may also be difficult to mobilize peasants to cope with bad weather, as short-run income benefits cannot be easily defined and provided for such mobility. The twoway dependence between agricultural fluctuation and investment adjustment supplies a sound starting point for modelling investment cycles in China. Other factors such as weather and market effects make the causality from agriculture to investment more significant than that from investment to agriculture. Using statistical terms, the agricultural fluctuation is weakly rather than strongly exogenous in this selected information system for modelling investment cycles. This is also what needs to be revealed and tested in the modelling exercise.
Chapter 4
158
Data Appendix
Table A4.1
Basic data on output, labour and population
Year
GOVA
1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
32.6 38.4 42.0 46.1 51.0 53.5 57.5 61.0 53.7 56 6 49.7 45.7 55.9 58.4 64.2 72.0 83.3 91.0 92.4 92.8 94.8 102.1 106.8 107.5 117.3 121.5 126.0 125.8 125.3 139.7 169.8 192.3 218.1 248.3 275.0 321.4 361.9 401.3 467.6 586.5 653.5 766.2 815.7 908.5 1099.6 1575.0 2034.1 2342.9
AGOVA 31.07 36.56 40.01 46.10 47.53 49.14 52.88 55.55 57.53 58.92 50.89 44.44 43.38 46.05 51.40 58.41 63.20 68.69 69.75 68.04 68.78 72.75 75.10 74.31 80.44 83.30 85.88 85.52 85.19 92.11 99.02 100.45 106.26 118.25 127.47 143.09 147.98 152.96 161.81 168.22 173.43 186.66 193.57 205.97 222.04 241.16 267.45 292.59
NOVA 24.5 28.7 31.6 34.0 37.4 38.8 41.7 43.9 42.5 44.0 37.6 33.2 43.2 44.4 48.8 54.9 64.1 69 2 70.3 71.4 72.2 77.8 80.8 80.8 88.6 92.2 94.6 94.0 91.3 98.6 122.6 132.6 150.9 172.3 192.1 225.1 249.2 272.0 315.4 381.8 420.9 500.0 526.9 579.5 688.21 945.72 1199.30 1388.42
ANOVA 22.88 26.79 29.51 34.00 34.54 35.12 37.91 39.61 40.83 40.90 34.20 28.42 28.80 30.16 33.63 38.05 41.79 44.85 45.63 44.74 44.95 47.53 48.28 47.77 52.05 54.13 55.18 54.09 52.73 54.81 58.31 57.26 61.34 68.54 74.36 83.98 86.26 88.88 92.89 95.00 98.02 105.37 107.78 113.19 117.72 122.42 128.54 135.09
GRAIN
ALAB
POPUL
113.18 132.13 143.69 163.92 166.83 169.52 183.74 192.75 195.05 200.00 170.00 143.50 147.50 160.00 170.00 187.50 194.53 214.00 217.82 209.06 210.97 239.96 250.14 240.48 264.94 275.27 284.52 286.31 282.73 304.77 332.12 320.56 325.02 354.50 387.28 407.31 379.11 391.51 402.98 394.08 407.55 446.24 435.29 442.66 456.49 445.10 466.62 504.54
163.33 166.44 169.76 173.16 177.44 181.47 185.85 185.35 193.00 154.80 162.57 169.96 197.29 212.59 219.48 227.78 233.72 242.73 251.41 260.38 270.92 277.86 283.65 282.48 288.20 291.80 294.15 293.98 292.94 283.13 286.29 291.17 297.71 308.53 311.45 308.62 311.05 312.12 316.14 321.97 331.70 340.49 348.76 347.69 339.66 333.86 330.18 329.10
541.67 551.96 563.00 574.82 587.96 602.66 614.65 628.28 646.53 659.94 672.07 662.07 658.59 672.95 691.72 704.99 725.38 745.42 763.68 785.34 806.71 829.92 852.29 871.77 892.11 908.59 924.20 937.17 949.74 962.59 975.42 987.05 1000.72 1016.54 1030.08 1043.57 1058.51 1075.07 1093.00 1110.26 1127.04 1143.33 1158.23 1171.71 1185.17 1198.50 1211.21 1223.84
Agricultural Constraint to the Insatiable Investment Demand Notations: GOVA AGOVA NOVA ANOVA GRAIN ALAB POPUL
159
Gross Output Value of Agriculture at current prices, and in billions yuan; Actual GOVA at 1952 constant prices, and in billions 1952 yuan; Net Output Value of Agriculture at current prices, and in billions yuan; Actual NOVA at 1952 constant prices, and in billions 1952 yuan; Output of grain, in millions tons; agricultural labour force, in millions; Total Population, in millions.
Sources and Notes: GOVA and NOVA: The data for 1952-96 are taken from Yearbook (1993: 50, 33; 1994: 330, 33; 1997: 42, 369), for 1949-51 are from Rural Statistics 1989: 52, 64). AGOVA and ANOVA are generated based on their indices at comparative prices (Yearbook 1993: 5 1 , 34; 1994: 330, 35; 1997: 42, 369; Rural Statistics 1989: 56, 66) and relevant numbers of 1952 at current prices. According to the production approach in China's statistics, the national income of agriculture (before and in 1993) and agricultural GDP (after 1993) are taken as the net output value of agriculture (Yearbook 1993: 75; 1997: 42). ALAB: The data for 1952-96 are taken from Yearbook (1993: 101; 1994: 83; 1997: 98), and for 1949-51 are estimated by using the natural growth rates of national population from 1949 to 1952, which are 2 % for 1951-52 and 1950-51, and 1.9% for 1949-50 (Yearbook 1993: 81), as substitutions for the growth rate of the agricultural labour force in these years. It is possible that the labour force in 1949, 1950 and 1951 was slightly over-estimated since after a long-drawn-out war period the natural growth rate of population may be a little higher than the growth rate of labour force. GRAIN and POPUL: All data are directly taken from Yearbook (1993: 364, 8 1 ; 1994: 345, 59; 1997: 69, 383).
Chapter 4
160 Table A4.2
The source of Tables 4.5-4.11
Year
RVAPL
1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
140.10 160.97 173.85 196.35 194.68 193.54 203.98 213.70 211.58 264.22 210.40 167.24 145.97 141.86 153.21 167.03 178.79 184.76 181.49 171.84 165.91 171.06 170.21 169.11 180.62 185.50 187.60 184.01 180.02 193.58 203.67 196.64 206.03 222.16 238.75 272.11 277.31 284.75 293.82 295.05 295.51 309.45 309 04 325.54 346.58 366.67 389.29 410.49
GRNPC 208.95 239.38 255.22 285.17 283.74 281.29 298.93 306.79 301.69 303.06 252.95 216.74 223.96 237.76 245.76 265.96 268.18 287.09 285.22 266.20 261.52 289.14 293.49 275.85 296.98 302.96 307.86 305.50 297.69 316.61 340.49 324.77 324.79 348.73 375.97 390.30 358.15 364.17 368.69 354.94 361.61 390.30 375.82 377.79 385.17 371.38 385.25 412.24
PPIAP
MIA
AMIA
A C C R (%)
INVR (%)
na. 100.0 119.6 121.6 132.5 136.7 135.1 139.2 146.2 149 4 152.1 157.4 201.4 200.1 194.4 189.5 187.9 195.8 195.5 195.2 194.9 195.1 198.3 201.1 202.8 204.5 208.7 209.7 209.2 217.4 265.5 284.4 301.2 307.8 321.3 334.2 362.9 386.1 432.4 531.9 611.7 595 8 583.9 603.8 684.7 957 9 1148.5 1196.8
8.1 9.7 10.4 12.1 13.6 14.7 15.8 17.1 11.2 12.6 12.1 12.5 12.7 14.0 15.4 17.1 19.2 21.8 22.1 21.4 22.6 24.3 26.0 26.7 28.7 29.3 31.4 31.8 34.0 41.1 47.2 59.7 67.2 76.0 82.9 96.3 112.7 129.3 152.2 204.7 232.6 266 2 288.8 329.0 411.2 629.3 834.8 954.4
8.19 9.77 10.50 12.10 12.99 14.02 14.97 15.94 16.70 18.01 16.69 16.02 14.58 15.90 17.78 20.36 21.42 23.84 24.12 23.30 23.83 25.21 26.82 26.54 28.39 29.17 30.70 31.42 32.46 37.30 40.71 43.20 44.92 49.70 53.11 59.11 61.72 64.08 68.92 73.22 75.41 81.29 85.79 92.79 104.32 118.75 138.91 157.50
n.a. n.a. n.a. 21.4 23.1 25.5 22.9 24.4 24.9 33.9 43.8 39.6 19.2 10.4 17.5 22.2 27.1 30.6 21.3 21.1 23.2 32.9 34.1 31.6 32.9 32.3 33.9 30.9 32.3 36.5 34.6 35.1 28.3 28.8 29.7 31.5 35.0 34.7 34.1 34.5 33.8 (36.0) 32.8 (34.7) 32.8 (34.8) 34.4 (36.2) 38.7 (43.3) (41.2) (40.8) (39.2)
n.a. n.a. n.a. n.a. 12.60 13.42 13.04 18.11 16.17 24.98 28.89 32.96 15.41 9.21 11.14 14.01 16.10 16.60 13.15 10.76 16.07 19.62 20.78 20.12 19.45 20.22 22.23 21.61 21.31 22.48 20.84 20.18 17.09 19.70 19.92 20.79 22.39 23.29 23.73 22.52 18.65(15.40) 20.29(15.93) 22.24(17.05) 26.46 (20.39) 29.43 (22.20) (19.79) (18.35) (17.60)
Agricultural Constraint to the Insatiable Investment Demand
161
Notations: RVAPL GRNPC PPIAP MIA AMIA ACCR INVR
Real Value Added Per Labourer, i.e. per labourer actual NOVA, in 1952 yuan; Per Capita Grain Output, in kilogrammes; Procurement Price Index of Agricultural Products, using 1950 as the base year; Material Input of Agriculture at current prices, and in billions of yuan; Actual MIA at 1952 constant prices, and in 1952 billions of yuan; Capital Accumulation Ratio; Fixed Investment Ratio.
Sources and Notes: RVAPL and GRNPC: Are directly derived from Table A. PPIAP: Data are taken from Yearbook (1993: 238; 1994: 231; 1997: 279). MCA is generated by GOVA minus NOVA in Table A. AMCA: Is obtained by subtracting ANOVA from AGOVA in Table A. ACCR: Figures are from Yeanboo/c(1993: 43; 1994: 40; 1997:46). INVR is generated through the Fixed Investment of State Sector (Yearbook 1993: 149; 1994: 144; 1997: 154) being divided by the National Income Used (Yearbook 1993: 43; 1994: 40; 1997: 46). The 'n.a.' indicates that the relevant data are not available. Numbers in the parentheses in the columns ACCR and INVR are Total Capital Formation over GDP' and 'State Fixed Investment over GDP.
Energy as the Representative of Producer Goods Constraints
5.1 Introduction Energy and transport have long been considered the two major bottleneck constraints to China's economic development. This chapter examines the extent to which investment expansion and economic development have been constrained by shortfalls in energy supplies. It highlights the close linkage between effective energy supply and long-distance transport, particularly railway freight transport. Energy in this research refers to primary modern (commercial) energy, which includes coal, crude oil, natural gas and hydro-power, as opposed to traditional (noncommercial) fuel such as crop by-products, firewood, grass, and dung. The basic reasons why traditional energy is not included in the research are (a) traditional energy is not directly linked with industry; (b) the required database is absent; and (c) more importantly, the long-lasting shortage of traditional energy in rural areas makes it impossible to substitute further traditional fuel for modern energy so as to relieve the energy shortage in the modern sector. In fact, the energy shortage in rural areas, where two thirds of the population reside, has perhaps been more severe than in urban areas, where traditional fuel accounts for about 80 per cent of the energy used in households. Several Chinese reports estimate that in the early 1980s about 300 million metric tons (MMT) of crop by-products were burned for cooking and heating by rural households every year. The real demand for such biomass fuels was about 500 MMT and was met by collecting roots, twigs, brush, leaves, grass, dung and firewood (Xu, J. et al. 1980; Xu & Huang 1981; Liu, F. et al. 1992). This indicates that China consumes more of these traditional fuels than any other nation, at a rate not sustainable in the long run. Nearly half of
162
Energy as the Representative of Producer Goods Constraints
163
all crop by-products are burned, gradually draining humus from the soil, and firewood consumption is more than twice the sustainable harvest (Liu, F. et al. 1992: 38, 39, 67; World Resources 1994: 67). As a consequence, the sustainable development of agriculture in China has required the increasing inputs of modern rather than traditional energy. Commercial energy is omnipresent in modern production, service and consumption processes. It supplies fuel, power and raw materials and is the life blood of modern industries. In a developing economy like China the significance of commercial energy is further intensified by the following factors. Firstly, due to rapid industrialization, a speedy increase in the more energy-intensive manufacturing industries and a relative decline in the less (modern) energy-intensive traditional industries such as agriculture becomes inevitable. This puts constant pressure on energy supply. Secondly, infrastructure development is often a precondition for industrialization and social development. This too requires large quantities of energy-intensive materials, such as steel and cement. Thirdly, developing economies typically experience rapid urbanization. This shifts production activities formerly undertaken in the home with little or no commercial energy, to outside producers who do use commercial energy. This means increased demands for transportation and fuel. Fourthly, besides the increasing demand for modern energy, agricultural stability and development become increasingly dependent on modern material inputs such as chemical fertilizer, irrigation, machinery and tractors, all of which are energy-intensive (cf. Chapter 4). Fifthly, although the use of traditional energy is widespread in many developing economies, higher incomes resulting from economic development tend to facilitate the choice for commercial rather than traditional energy. All these factors have played important roles in China's economic development, and produced strong demand pressure on the economy. China's resource endowments are fairly limited. Although China's proven recoverable coal reserves are the third largest in the world (about 62.2 billion tons in 1990), they equal only 55 per cent of those of the United States and 60 per cent of the former Soviet Union. Reserves of crude oil and natural gas are also limited, and the proven recoverable volumes account for only 2.4 and 0.9 per cent of the world total {World Resources 1994: Table 21.3). The proportions of high quality reserves are also small, the costs of exploiting them are high and the required transport distances are enormous (World Bank 1983, 1985).
164
Chapter 5
The intensive demand pressure, combined with limited resources endowments and supply possibilities, have constantly caused energy shortages, leading to recurring energy crises in the economy. In fact, in China's modern sector, commercial energy can be identified as the key production factor which constrains the processes of planning, production, distribution and consumption. Compared with the softer financial constraint, energy has played a significant role in the refusal of many lowerpriority projects. Many of these are refused because there is no power quota and/or other energy quota for them in their regions. Some projects are stopped halfway through their construction (banjiezi gongcheng) as the power quota or some other quota in their region has been exhausted. It is extensively reported in the national media and confirmed by both SOEs' managers and government officials that these quotas are crucially important. For example, one of the most critical and difficult negotiations in the process of project approval (cf. Chapter 3) is the scramble for power quota because power cannot be arbitrarily 'printed out' by government and bank officials in the short run. Therefore energy shortage adds to government power at different levels, giving it control of the SOEs, TVEs, and even of private-owned enterprises through the mechanism of distributing power and other energy rationing quotas (Young 1992). In fact, it should rather be investment goods such as steel, cement and plate glass that act as bottleneck constraints to fixed investment. The reasons why energy can comprehensively represent a bottleneck is not only because it includes power, gasoline, diesel fuel, coal and other such key investment goods, but also because the long-lasting heavy-industryoriented development strategy (cf. Chapters 3 and 4) has overdeveloped the producer goods sector. Over 20 per cent of this sector's equipment lies idle mainly due to energy shortage. Power stations frequently have even had to turn off generators because of coal shortages, although power shortage has been the number one problem in the economy. The media has repeatedly reported that factories (even steel plants) have to stop part of production due to power shortage. Power stations also have to switch off generators because of lack of coal and transport does not run due to oil shortage. These facts clearly indicate that on the production side the most significant cause of long-lasting shortage of investment goods is energy shortage. A second reason to select energy as the representative of the large set of investment goods is in order to avoid measurement difficulties.
a
1953
1.2 0.9 3.1 3.5
0.8 0.9 2.6 2.7
10.3
14.1 14.7
81.6 80.9
309.90 292.91 1.01 15.98
1970
8.6
88.0 86.5
188.24 189.01 1.26 -2.03
1965
4.4 4.6
2.4 2.5
22.6 21.1
70.6 71.9
487.54 454.25 14.41 18.88
1975
3.8 4.0
3.0 3.1
23.8 20.7
69.4 72.2
637.35 602.75 27.97 6.19
1980
4.3 4.9
2.0 2.2
20.9 17.1
72.8 75.8
855.46 766.82 54.34 -25.09
1985
4.8 5.1
2.0 2.1
19.0 16.6
74.2 76.2
1039.22 987.03 45.65 -32.19
1990
6.2 6.1
1.9 1.8
16.6 17.5
75.3 74.6
1290.34 1311.76 13.20 4.91
1995
(a) Excludes bio-energy, solar, geothermal and nuclear energy, all fuels are converted into standard coal equivalent (SCE) with thermal equivalent of 7,000 kilocalories per kg. The conversion is as follows: 1 kg of coal = 0.714 kg of SCE; 1 kg of crude oil = 1.43 kg of SCE; 1 cubic metre of natural gas = 1.33 kg of SCE. The conversion of hydropower into SCE is calculated on the basis of the consumption quota of standard coal for thermal power generation of the given year, which obviously over-estimates the SCE value of hydropower (cf. Statistical Yearbook of China (English version) 1990:459). (b) Before 1980, stock changes are defined as production minus consumption and net export, hence the statistical balances are included. For the data in and after 1980 the statistical balances are excluded.
2.9 3.2
0.1 0.1
2.9 3.0
0.9 0.9
2.1 4.6
1.7 3.8
2.0 1.8
4.8 6.6
94.9 92.3
96.3 94.3
91.4 89.2
171.85 165.40 0.94 5.60
1962
Sources: The figures of net exports before 1980 are generated based on Chinese Trade and Price Statistics: 1952-1983 (1984: 503, 515). All others are taken from Yearbook (1995:199-200; 1997: 215 & 216).
Notes:
Percentage of Total Energy Coal Production Consumption Crude Oil Production Consumption Natural Gas Production Consumption Hydropower Production Consumption
98.61 96.44 1.29 0.88
1957
Primary commercial energy production and consumption, 1953-95
Total Energy (million tons of SCE) 51.92 Production 54.11 Consumption 0.68 Net exports -2.87 Stock changesb
Table 5.1
ON
2
I
o
1
1
1
1
§
166
Chapter 5
As in Chapter 4, this chapter will focus on the short-run two-way interaction between investment expansion and energy constraint, which is substitutive in one way and complementary in the other. The long-run dominance of complementarity will be analysed in Chapter 6 based on the equilibrium comovement equation. Section 5.2 of this chapter presents an overview of China's energy production and consumption. Section 5.3 will show how the widespread and chronic shortage of primary energy has impeded the industrialization drive and investment boom. Section 5.4 reveals that due to the constraint of the transport bottleneck, there has been a significant gap between effective energy supply and energy output. As a consequence, the difference between energy consumption and effective energy supply becomes negligible, and the effective energy supply becomes capable of representing the key bottleneck constraints in the producer goods sector. Some primary econometric evidence of the chronic energy constraint to investment hunger is given in section 5.5. Section 5.6 summarizes the chapter.
5.2 Energy Situation in China: An Overview This section provides a background on the energy situation in China and compares it with other countries. In order to reveal the specific features of the energy situation in China, both the mix of energy sources and the sectoral structure of energy consumption will be examined based on an international comparison. The four most prominent features of energy production and consumption in China are its heavy dependence on coal, the industrial sector's dominance in final energy consumption, the country's very high energy intensity, and its long-lasting quantitative restrictions and price controls. In addition, although China's energy exports and imports have maintained an insignificant share in comparison with total production and consumption, the absolute magnitude of China's net energy export has not been inconsiderable and has played an important role in lessening the foreign exchange constraint since the mid 1970s (cf. Table 5.1). A unique characteristic of the energy situation in China is the high share of coal in both energy production and consumption. Although there has been a declining trend since the early 1950s, in the early 1990s coal still accounted for three-quarters of primary commercial energy production and consumption (cf. Table 5.1), the highest share in any major country.1 In the United States and the former Soviet Union, where coal
Energy as the Representative of Producer Goods Constraints
167
reserves are almost double those of China, coal accounts for about 25 per cent of commercial energy (World Resources 1994: 66). Even in India, where the most important indigenous energy source is coal, it accounted for only 57 per cent of the total primary energy supply in 1990 (Ishiguro & Akiyama 1995: 15). Furthermore, about 76 per cent of the power supply is generated by burning coal in thermal plants, which may also be the highest proportion in any major country, so that the shortage of coal has adversely affected electricity supplies in a decisive way {World Development Report 1994: Box 1.2). In the 1950s and early 1960s, coal accounted for more than 95 per cent of energy production and more than 91 per cent of energy consumption. During this period, the absence of any substantial hydrocarbon reserves made the massive expansion of coal mining the only reasonable choice for satisfying the huge demand for energy driven by industrialization ambitions, in particular by the strategic orientation toward iron, steel and heavy metallurgical industries. Coal production was doubled during the economic recovery period of 1949-52, and doubled again during the First Five-Year Plan period of 1953-57. During the 'Great Leap Forward' of 1957-59, raw coal output expanded to 2.3 times its 1956 level. But this expansion was mainly based on shallow open pits, using the heavy application of labour input with traditional tools, and produced mostly inferior fuel (most of which was wasted in the irrational 'Mass Steel- and Iron-Smelting' campaign). It was clearly an unsustainable record. Orderly development recommenced after the famine and economic collapse induced by the 'Great Leap Forward', but the basic mechanism of quantity drives has persisted. In the 1970s raw coal production in China almost doubled. From 1980 to 1994 it almost doubled again {Yearbook 1993: 477; People's Daily, 2 March 1995: 2). This impressive growth was achieved by the development of large mines under the central control and by the widespread establishment of small, local-government run and/or township and village-owned pits, whose production was considered to be crucial for increasing output and diminishing the chronic dependence of the southern provinces on large-scale coal shipments from the north. Despite rapid expansion, it has been consistently pointed out that too many large mines have ignored the proper balance between tunnelling and excavation. Too many accidents have happened in mines through neglecting even basic safety precautions. Expansion has been achieved at the expense of longterm development potential and conservation of resources. The quality of
168
Chapter 5
output has been substandard because of the headlong quantitative rush. And last but not least, the living conditions of miners and their families have been overwhelmingly poor (Li & Zhu 1992: 3; Lu 1993: 8-9; Zheng 1992: 6, 8; Zhou & Chu 1992: 5). The use of coal as the foundation has several disadvantages (Zhou & Chu 1992), three of which are most relevant here. First, coal transportation is more difficult, not only because of much lower efficiency in collecting and distributing processes, but also because of its burden on transport. About 40 per cent of all railway tonnage (and 50-70 per cent on major north-south trunk railways such as Beijing-Guangdong and Beijing-Shanghai) has been used by coal transportation (Smil 1988: 89; World Development Report 1994: Box 1.2). According to an official forecast, an increasing share of coal transportation in the total railway tonnage and a longer transportation distance can also be expected in the near future (Lu 1993: 19-20, Wu et al. 1988). This will continue to put great pressure on China's over-intensively used railway networks. Secondly, the dominance of coal in energy supply and consumption implies a lower energy efficiency. Coal has been the major energy source for electricity, steam and heat for industry, and is extensively used in cities as the major fuel for cooking and residential heating. As summarized in Ishiguro & Akiyama (1995: 28), the average thermal efficiency of coal boilers used in the industrial sector is estimated at 50-60 per cent, while the efficiency of the oil and gas boilers used in industrialized countries is 80-90 per cent. The thermal efficiency of coal stoves used in the household sector is estimated at only 20-25 per cent, whereas that of modern gas stoves is between 55-60 per cent. Thirdly, the high dependence on coal has great environmental costs. In 1989, over 700 million tons of raw coal was burned as a heat source, of which about 90 per cent was consumed in urban areas. There are tens of millions of small and medium-size boilers, furnaces, locomotives, and home cooking stoves that emit all kinds of pollutants into the atmosphere. Treatment of the flue gases from such a huge number of small, scattered coal-burning devices is too expensive to be practical. Therefore, air pollution caused by coal combustion has been one of the main headaches of China's urban environmental protection, and urban air pollution is becoming steadily worse as the economy grows (Drysdale & Huang 1995, Lu 1993, World Resources 1994: Chapter 4). The disposal of the huge amounts of ash and cinder produced by coal combustion also creates a serious problem.
Energy as the Representative of Producer Goods Constraints
Table 5.2
169
Percentage shares of final energy consumption, 1980-95
BY SECTOR Material production sectors (= national income) Agriculture Industry . Light industry . Heavy industry Construction Transport and communication Commerce Non-material production sectors Household consumption BY USE Final uses Industry Losses in processing & transform. Coking Oil processing Other Losses In oil fields In electricity transformation
1980
1985
1990
1992
1995
81.1
79.4
80.5
81.7
84.5
5.8
5.3
4.9
4.6
4.2
68.0
66.6 13.2 53.4
68.5 13.9 54.6
69.9
73.3
1.6 4.8 0.9 3.0
1.7 4.8 1.0 3.2
1.2 4.6 1.3 3.5
1.3 4.6 1.3 4.0
1.0 4.5 1.5 3.4
15.9
17.4
16.0
14.3
12.0
95.4 63.5
96.0 62.6
95.5 64.1
95.3 65.3
94.7 68.2
2.3 1.1 0.2 2.3 0.6 1.7
1.9 0.7 0.1 2.1 0.4 1.6
2.3 0.9 0.3 2.2 0.3 1.7
2.3 0.7 0.5 2.3 0.3 1.9
2.8 2.5
Sources: Statistical Yearbook of China (1991 • 45R-1993:478, 482; 1995: 200; 1997::?1B)
Table 5.3
Energy intensity by sector, 1980-94 (gram of standard coal equivalent per yuan of GDP in 1980 prices)
Agriculture Industry . Light industry8 . Heavy industry8 Construction Transportation & communication Commerce Note:
1980
1985
1992
1994
255.3 2054.1 795.2 3181.0 489.0 1415.6 242.5
200.3 1595.7 617.8 2471.1 393.4 1108.7 139.1
189.0 1106.6 436.9 1747.5 232.2 768.4 173.9
176.5 892.6 345.6 1382.3 170.2 688.5 196.5
(a) Based on the assumption that the shares of light and heavy industry in the industrial GDP are the same as those in the industrial gross output value.
Sources: Yearbook 1995:200 (energy consumption by sector), 1995: 32 (GDP and its indices by sector), and 1995:27 (shares of light and heavy industry in the industrial gross output value).
170
Chapter 5
Table 5.4
China India Indonesia Thailand Source:
Indexes of energy intensity in relation to physical unit of output in energy intensive industries (average in industrialized countries = 100), 1990 Iron & Steel
Cement
Fertilizer
135 140 na na
170 165 120 110-120
160 160 135 na
Pulp & Paper 160 200-500 130 120
Ishiguro & Akiyama (1995: Table 3.3). The original source is from the Commodity Policy and Analysis Unit of the World Bank.
China's industrial sector dominates primary commercial energy consumption, accounting for over 66 per cent of the total in the period of 1980-95 (Table 5.2). The rapid growth of the industrial sector and its high energy intensity (see Table 5.3) have been the major contributory factors underlying the rapid growth of energy consumption and low energy efficiency. One reason for energy inefficiency in the industrial sector is that most industries in China are still using old equipment in smaller plants that do not permit economies of scale. Although impressive progress in lowering energy intensity in industry has been made during the reform period (cf. Table 5.3), in 1990 the amount of energy required to produce a unit of steel, cement, ammonia or paper is still considerably more than is required in industrialized countries (see Table 5.4) where substantial progress in fuel-saving techniques has been made over the past three or four decades. A more important reason may be the lack of incentive to economize on energy. Very low energy prices (to be discussed below) and the soft budget constraint (cf. Chapter 3) have, until recently, left enterprise managers and workers with too little incentive to limit energy or to demand new and more fuel-efficient models from their suppliers. These suppliers also have been under little pressure to undertake appropriate research and innovations. In the rural industrial sector, low energy efficiency is mainly caused by the use of outdated, even secondhand equipment and backward technology, and also due to a lack of economies of scale. In the state sector, which in 1993 still produced the dominant share of gross output in such high energy intensity industries as petroleum and refining (over 90 per cent), thermal power (over 85 per cent), iron and steel (70 per cent), coal and coke (over 65 per cent), metallurgy and chemicals (over 50 per cent)
Energy as the Representative of Producer Goods Constraints
171
(Naughton 1995a: Figure 2), the reason for low energy efficiency can be traced back to the very low energy prices and soft budget constraint. The SOEs have been getting energy at low, state-planned prices and have enjoyed soft budget constraints (cf. Chapter 3). Because of the different methods used in output value statistics, different relative prices, and problems resulting from use of official exchange rates to convert monetary statistics to comparable units, the direct international comparisons based on official exchange rates and unadjusted data sometimes show an improbable figure of energy consumption per US dollar of GDP for China. A more realistic comparison, based on comparable energy intensities in physical rather than value terms is presented in Table 5.4. It can be seen that in 1990, after a remarkable decline in the energy intensity, which was measured in kg of standard coal equivalent per Chinese RMB of the real GDP, between 1980 and 1990 (X. Lin & Polenske 1995), China's energy intensity in the four, energyintensive industries, iron and steel, cement, fertilizer, and pulp and paper, was still more than 35-70 per cent higher than the averages in industrialized countries. The intensity was also higher than in Indonesia and Thailand, and similar to the situation in India. According to the World Bank (1983), which presented a more comparable picture of energy intensity in value terms based on price adjustment, in 1979 China's energy intensity in terms of kg coal equivalent per US dollar of GDP was about 2.5 times the average for other developing countries or for industrialized market economies, and was about 1.5 times the average for other centrally planned economies at that time (World Bank 1983 (2): 198). Such an exceptionally high energy intensity has certainly aggravated the tension of energy shortage and crises. On the other hand it also implies a considerable scope for energy conservation.3 Here, as in most centrally planned economies, energy pricing has aggravated shortage and waste. For many years, the government fixed energy prices well below the cost of production, while establishing quotas for industrial and residential use. The prices of all primary modern energy and energy products were fixed until the early 1980s, when a double-track pricing system was introduced. Under the double-track system, the government sets output quotas for energy industries and requires the sale of a quota of output at in-plan prices. This enables energy-producing firms to sell above-quota output at much higher market prices (close to international market prices). In 1989, about 70 per cent of steam coal was sold at in-plan prices and the other 30 per cent
Chapter 5
172
at market prices, when the market prices averaged three times the inplan prices (Table 5.5). This pricing system, with some variation over time, was practised until the end of 1993. In 1993, about 15.4 per cent of crude oil was sold at 'low level' in-plan price (one-sixth of market prices), 76.9 per cent at 'high level' in-plan prices (one-third of the market prices), and only 7.7 per cent at market prices (Table 5.5). About two-thirds of the oil products were sold at the in-plan prices, and 30 per cent of steam coal was sold at planned prices (Ishiguro & Akiyama 1995: 31). It is widely believed that the double-track system had some marginal effects on resource allocation and on providing incentives for energy producers. But at the same time, it was also blamed for having its negative effects such as widespread rent-seeking, smuggling and corruption. Table 5.5
Plan versus market prices for steam coal (1989) and crude oil (1993) (price: yuan Aon) Market
Plan Steam Coal
-
3
Price Beijing Shenyang Jiangsu Shanxi
55-70 45-70 65-80 34-50
b
Coverage (%)
i-
8
Price
140-150 80-180 220-280 110-140
Plan
Coverage (%)b ^ I [ J
JU
Market
Crude Oil Price Nationwide (Two in-plan prices) Notes:
550 270
Coverage (%) 76.9 15.4
Price 1,500 (in Shanghai)
Coverage (%) 7.7
(a) These are indicative prices as mentioned by Albouy (1991), which were gathered by Albouy during his visits to the cities cited and in discussions with coal-purchasing companies and users. Shenyang is located in a coal-producing area but needs more coal than it produces. Beijing has limited coal production but is relatively close to a major coal centre in Shanxi, i.e. 500 km from Datong. Jiangsu is one of the major coal-deficit provinces although some coal is produced in the north at Xuzhou and in the neighbouring province, (b) These coverage data were collected by Ishiguro & Akiyama through communication with several officials of the government of China (Ishiguro & Akiyama 1995: 31).
Sources: Albouy (1991: Table 2.1), Ishiguro & Akiyama (1995: 31).
Energy as the Representative of Producer Goods Constraints
173
In 1994 the government started to decontrol energy prices. The double-track system for crude oil and electricity remained, but the complex pricing formulas for electricity have been somewhat simplified. The double-track prices for coal and oil products were nominally abolished but price ceilings for major products were imposed. The real effect of this new round of price reforms cannot be easily spelt out. The immediate effect was a significant increase of losses made by state coal producing firm because of a highly inflationary environment, transport limitations and a further sharpening of transport constraints. Finally, retail prices (including taxes) are still lower in China than in most countries.4 Quantitative restriction is bound to follow price control. In China, there have been numerous quantitative restrictions on energy use, which are summarized by Hussain (1995) as follows: (a) Coal and oil products are not always available even at higher market prices, and quantitative rationing in the margin has been more or less a permanent feature even in the reform years. (b) Planned and unplanned electricity outages are commonplace and households are usually supplied low amperage electricity to discourage the use of household electrical appliances. The countryside, where 80 per cent of the Chinese population lives, has a very limited supply of electricity {World Resources 1994: 67). (c) Heating of public buildings and the government- and enterprisesupplied houses in winter is banned south of the Yangtze river, even though the average winter temperature can be lower than in southeast England, and is only allowed for a limited period of the year north of the river. (d) International trade in energy is controlled by central government and national self-sufficiency in energy has until now been the strategic policy, at the cost of limiting the use of imported oil and natural gas. The initial rationale of energy price control and quantitative restriction is to serve the heavy-industry-oriented development strategy (section 2.3.3). Inefficiency caused by energy price control has been widely recognized. On the demand side, with a high value placed on gross output and with energy a negligible part of total operating costs, there has been little incentive to consume less than the full energy quota or to demand new and more energy-efficient technology. On the supply side, energy
174
Chapter 5
producing firms have constantly faced financial difficulties and have thus lost the capability for self-development. Investment in the energy sector has become unattractive to local investors. There are inefficient patterns of energy production and transport (Albouy 1991: 6-9). All of this has exacerbated energy shortages in China. In addition, along with economic development and industrial structure changes, the economy has responded to the price structure and formed a low-input price-induced interest structure. This vested interest rigidity has manifested itself through the mechanism referred to as 'the restoration of previous relative prices' in the reform period, and thus has effectively limited the progress of price reform.
5.3 Widespread and Chronic Shortage of Energy in China China's chronic and widespread shortages of primary energy are likely to continue for a long period into the future. Effective energy supply has been so insufficient that no less than one-fifth of industrial capacity is idle and the output of a large number of essential enterprises is determined by their fuel and electricity rations. Power stations have to turn off generators due to coal shortages (dengmei fadian), factories have to stop production due to power shortages (tingchan daidian) and transport sometimes comes to a standstill due to lack of oil (queyou tingyun). Such problems are widespread and reach a crisis at the peaks of the investment cycles. Reports indicate clearly that decision-makers, researchers and the public have viewed effective energy supply as a key bottleneck constraint to the economy.5 Under these conditions of long-term energy shortage, China's macroeconomy has been accustomed to the normal shortage level as analysed in Kornai (1980, 1982).6 That is to say, if the energy supply increases along a norm path or equilibrium steady growth path, and the normal shortage intensity is neither strengthened nor alleviated, the adapted normal demand of the economy for energy can be satisfied, and the economy thus progresses smoothly along with the normal shortage path (cf. section 2.3.4). As a result, even the state plan used to leave a significant 'plan gap' for energy distribution. For example, according to the annual plan of the national economy for 1993, the planned growth rate for GDP was 9 per cent and overall primary energy production 2 per cent. The planned gap for energy was 40-50 million tons SCE, or 4-5 per cent of
Energy as the Representative of Producer Goods Constraints
175
realized supplies. For steel, shortages of 9-10 million tons, or 5.1-5.6 per cent of supplies were planned. For nonferrous metals, the planned gap was 0.25 million tons (State Planning Commission 1993; Yearbook 1995: 199,412). The government considers that the gaps can be balanced by active measures of microeconomic units such as economizing on energy, screening out some lower priority projects, promoting energy-saving technology, increasing energy output based on local and enterprise motivations. The realized output pattern in 1993 was that the growth rate of GDP reached 13.4 per cent, the primary energy production grew 2.2 per cent, the price level of fuels and power increased by 36.7 per cent, the energy stock in consuming areas decreased significantly and the transportation crisis became more serious.7 These danger signals in turn forced central government to initiate the retrenchment campaign in July 1993, and to reinforce the contraction measures in March 1994 by sending central delegations to each province to review and supervise the implementation of these retrenchment policies. The main purpose of this campaign was to control the scale of fixed investment {People's Daily, overseas version, 12 July 1993: 1; China Investment and Construction 1994 (3): 60; 1994 (4): 4). The negative correlation between the shortage intensity of power supply and the capital accumulation ratio of the national economy may provide further insightful evidence for understanding how the energy shortage has checked the ambitious investment drive. Because the power shortage intensity can be relatively easily defined and estimated, Chinese power engineers have done a series of fruitful investigations. Although the relevant database is unpublished, the research logic and results can still be readily understood. First there is a relatively stable relationship between real fixed investment and resultant incremental capacity of power-consuming equipment. Table 5.6 presents the incremental capacity of power-consuming equipment induced by one million yuan of real fixed investment in the period of 1982-88. From this table it can be seen that the incremental capacity of power-consuming equipment induced by one million yuan real fixed investment ranges between 74-116 kWe, and the average is 95 kWe. In comparison to the remarkable growth and fluctuation of the real investment, this indicator has maintained its stability. It demonstrates how
Chapter 5
176
and to what extent the fixed investment can generate demand pressure for power production. Table 5.6
Incremental capacity of power-consuming equipment induced by fixed investment (a) Total fixed investment (yuan billions)
1982 1983 1984 1985 1986 1987 1988 Mean
Current prices
1980 prices
123.04 143.01 183.29 254.32 301.96 364.09 449.65 259.90
116.56 132.21 147.40 193.30 213.80 130.80 285.10 188.45
(b) Increase of power-
Ratio of (b) to (a) (kWe per million yuan)
consuming capacity (GWe)
Current prices 98.0 98.0 66.0 50.0 71.0 74.0 56.0 69.0
12.00 13.59 12.11 14.13 21.53 26.84 25.19 17.96
1980 prices 103.0 105.0 82.0 74.0 101.0 116.0 88.0 95.0
Note:
The capacity of power-consuming equipment is based on the statistics of Energy Ministry. GWe indicates gigawatt (10* watt), and kWe represents kilowatt electricity. Source: Hu, Zhaoyi (1990: Table 1).
Table 5.7
Stock and increment of power-generating and consuming equipment (GWe) Total capacity
Capacity Power-generating equipment Power-consuming equipment Ratio of consuming to generating equipment
Increment
1985
1988
87.06 212.58
112.90 289.95
25.84 77.37
2.44
2.58
2.99
Annual growth rate
9.05 10.90
Source: Lu (1993: Table 11.4).
Table 5.7 presents the supply response of power-generation based on both stock and increment. Chinese energy experts have considered that the balanced ratio of consuming to generating equipment should be 2:1 according to an empirical survey of the average comprehensive utilization hours per equipment per year and technical calculations.8 Based on
Energy as the Representative of Producer Goods Constraints
177
this reference indicator we can see from Table 5.7 that in 1985 and 1988 the hard power shortage gaps are 22 (44 per cent -*• 2) and 29 (58 per cent -s- 2) per cent, respectively. Since the load demand has been significantly greater than the power-generating capacity, it is inevitable that many enterprises are bound to be ordered to close for as much as one to three days each week {tingsan kesi, tinger kewu, and so on). In peak load hours the power supply will have to be cut back, without advance warning {lazha xiandian). Table 5.7 shows us, although only partly, the reason why Chinese media, official documents, and research papers frequently mention that over 20 per cent of industrial equipment is idle due to power shortage. Another surprising finding from Table 5.7 is that although a remarkable power shortage gap existed in previous periods, the ratio of incremental capacity of power-consuming to power-generating equipment does not move to narrow the gap, and from 1985 to 1988 at least did just the opposite. Because of the investment boom in this period the incremental power shortage gap was as large as 49.5 (99 per cent •*• 2) per cent in this period. The general pattern of changes in power shortage intensity is as follows. The investment peak of 1953 was accompanied by power shortages in the region of Beijing, Tianjin and Tangshan. The height of investment in 1956 caused power shortages to spread into 23 of the total 29 provinces. The 'Great Leap Forward' of 1958-60 further induced the country-wide power shortage. In the 1960s, 1970s and 1980s, along with the general trend of increasing power shortage, in the years when the capital accumulation ratio was higher than 30 per cent, country-wide power shortages appeared and persisted. When capital accumulation ratio or investment ratio was exceptionally high, power shortage crises followed immediately. This was what really happened in 1970, 1975, 1978, 1982 and 1985-88, when the hard power shortage gaps were over 20 per cent and even reached around 30 per cent in certain months of the year (cf. Hu, Zhaoyi 1990; Yu, Weizhou 1992; Zhou & Chu 1992, among others). At first glance, one may attribute this deterioration to the insufficient power-generating equipment, which is unable to keep up with rapidly increasing power-consuming machinery. However, even the existing, insufficient power generators have constantly suffered coal hunger. Whenever a power crisis occurs, the provincial governors and mayors are dutybound to help their negotiation team and/or purchasing teams to get coal by every possible means. This situation is caricatured in the Chinese media as 'the governors and mayors lead the charge, with hundreds of thou-
178
Chapter 5
sands of soldiers, to secure coal'. Unsustainable development measures, such as accelerating extraction at the cost of the proper balance between tunnelling and excavation, conservation of resources, and necessary safety protection, are re-employed. A large amount of overstocked coal gangue will also be purchased as normal raw coal if the transport problems can be temporarily solved. Such over-heated purchasing campaigns and unsustainable extraction are often accompanied by a series of social disturbances such as inflation, corruption and production being held up due to lack of energy. All are seen by central government as danger signals and are used to persuade local officials to initiate retrenchment campaigns. The administrative control on oil and oil products has been much stricter than on coal and coal products because of their scarcity in China. Although since the early 1980s the government has reinforced the policy to comprehensively restrict 'burning oil', that is, using oil directly as a heat source or as fuel for power stations, the demand pressure for burning oil (induced mainly by power shortage) has in fact increased rather than decreased. As a result, along with the intensified power shortage in the latter half of 1980s, the quantity of oil burned has been rising steadily. According to state energy policy, oil and oil products should be mainly used as a power source for motor vehicles and other equipment such as ships, airplanes and tractors. In this area alone, however, excessive demand has also risen. For example, in 1980-88, the supply of gasoline rose by 79.2 per cent, and the supply of diesel oil by 55.3 per cent. However, motor vehicles, mostly trucks and buses, increased by 133.2 per cent, vessels and railway diesel locomotives increased by 142.7 per cent, and other oil-consuming equipment also increased significantly (Zhou & Chu 1992). This means that the gap between demand for and supply of oil and oil products had been greatly widened. In 1990-94, this general trend continued. The supply of gasoline and diesel increased by 41 per cent whereas motor vehicles and railway diesel locomotives rose by 70.8 and 34.4 per cent respectively {Yearbook 1993: 487, 489; 1995: 205, 473, 480). There was clearly tension between demand for and supply of oil and oil products and this has increased since 1980. Natural gas has played an insignificant role, providing only about 2 per cent of primary modern energy (Table 5.1), because of quite limited reserves. In view of its remarkable cost, convenience and environmental advantages over coal and oil, the supply possibility of natural gas seems to be next to nothing compared with the significant potential demand for
Energy as the Representative of Producer Goods Constraints
179
it. China has also begun to set up nuclear power plants with French and Russian aid, although this will remain a small part of the nation's total energy supply in the near future {WorldResources 1994: 69).
5.4 Transport Bottleneck and Effective Energy Supply China's economic geography, combined with its vast size, puts particularly intensive pressures on the transport sector. Most energy resources are located in the north (mainly coal) and northeast (notably oil and natural gas) of the country. Most other natural resources are located in the west and centre (for instance, metal ores and ingredients for cement). However, the vast majority of industry and of the population are located in the east and southeast. There is thus a continuing need for long-distance transport of bulk resources. In 1992 the average rail transport distances of coal, oil, metal ores, cement and iron were 558 km, 689 km, 481 km, 479 km, and 960 km, respectively. The transportation of coal and oil accounted for as much as 48 per cent of total railway freight traffic {Yearbook 1993: 525). Furthermore, the regional distribution of China's industry is also significant, with heavy industry - mainly powerplant boiler-making and heavy electrical industries - concentrated in the northeast, and textiles in the east of the country. In addition, striking agricultural variations exist across the climatic zones from north to south, and from the coastal irrigated areas to the arid interior. According to Yenny & Uy's (1985) estimation based on the World Bank's database, in 1980 freight intensity (expressed in ton-km of freight per US dollar of GNP) in China is almost twice as high as in the USA, India and Brazil, and eight times the level in Japan and South Korea. Howard (1989) estimates that freight intensity in China is around 60 per cent higher than in the USA, and considerably higher than in Japan or the South Korea. When the service sector is excluded from the estimation, freight intensity in China still remains higher than in other developing countries, though less than in the USA. Many factors contribute to China's high freight intensity. These include China's size, location of resources and population, the level of processing of raw materials such as ores, coal, lumber and agricultural products. The structural characteristics of the economy, the administrative structure and vertical integration of industry often result in circuitous routings and excessive transport.
180
Chapter 5
Table 5.8
Average railway traffic density in selected countries, 1979 and 1990 Net ton-km/route-km (millions) 1979
Former Soviet Union China United States Romania India Poland Brazil Notes:
24.7 10.9
1990 22.2 a
Passenger-km/route-km. (millions) 1979
1990
2.4 2.4 0.1 2.1
6.5 a 0.05
7.7 6.9
7.4
4.5 b
3.1
4.4 b
3.4
5.4
1.7 1.1
0.8
4.2 1.0
(a) 1993 Figures, (b) Broad gauge only: On the meter gauge system, 0.9 m. for freight and 1.6 m. for passenger traffic.
Sources: The data for 1979 are all taken from World Bank (1983 (2): 287). The 1993 figures for China are based on Yearbook (1995:467,783). The data on railway freight traffic mileage and passenger for Brazil in 1990 are from Statistical Yearbook of United Nations (1992: 630). All others are generated based on Yearbook (1995: 783) and the World Bank's World Development Report (1994:143-5).
China's transport system is based predominantly on railways. In 1980, if ocean shipping with average hauls of 7,500 km was excluded, 67.3 per cent of China's freight traffic of 849.4 billion ton-km was moved by rail, 17.9 per cent by coastal shipping and inland waterways, 9.0 per cent by road and 5.8 per cent by pipelines. In 1992, the corresponding shares were 57.4 per cent by rail, 20.9 per cent by coastal shipping and inland waterways, 12.3 per cent by road and 3.1 per cent by pipelines (Yearbook 1993: 519). China is now the world's third largest rail freight transport country, after the former Soviet Union and the USA (Yearbook 1995: 783). Yet in comparison with the size of the country, the rail network at 41,000 km in 1970, 50,000 km in 1980, and 56,700 in 1996, is relatively small (Yearbook 1993: 514; 1997: 518). This has resulted in very high and rapidly increasing traffic density (net ton-km million/route-km) by rail, which is much higher than in other countries, except in the former Soviet Union (cf. Table 5.8). This high traffic density, combined with backward technology and equipment, means that China's railways are operating at or near full capacity in large parts of the system. Compared to many rail networks, use of equipment is very intense. On the Shenyang-Beijing line, for example, there are 206 trains a day, or a scheduled train every 7 minutes, leaving
Energy as the Representative of Producer Goods Constraints
181
little time for maintenance. Nanjing Yangtse River Bridge is one of the most important bridges over the world's third longest river since it was opened in 1968. In 1995, the traffic density on it was five times the originally designed capacity, with a scheduled train every six minutes and a motor vehicle every six seconds, but with no overhauls in the last 27 years. Freight traffic densities of up to 30 million ton-km per route km per annum are common on the coastal and northeast lines. Such intensive use of the rail capacity is still far from balancing demand and supply. In 1993, the latest investment peak year, the requested number of trains was about 120,000 per day, but the maximum supply was about 70,000 per day. The gap between demand and supply is very large. This has caused serious railway bottlenecks in the investment peak years on major trunk lines, particularly those used for the transport of coal.10 Energy development in China has been heavily dependent on transport. In particular, the effective energy supply is not determined so much by production as by the transport bottleneck. As shown in Table 5.9, most of the coal resources are concentrated in northern China, and most centres of consumption are located in the east and south. In addition, the recent reform and 'open door' policy have further accelerated the economic development of the coastal regions, and thus further increased this discrepancy. According to the forecasts of both Chinese energy experts (Table 5.9) and the World Bank (1985: 208), the share of coal production in northern China between 1980 and 2000 is expected to increase from 33 per cent in 1980 and 37.7 per cent in 1985 to 46 per cent of the national total by 2,000. The quantity of coal shipped from the north to eastern and southern China will increase from about 100 million tons in 1985 to 400 million tons in 2000. Thus it can be predicted that transportation will continue to be the determinant bottleneck of China's effective energy supply. The construction of more railways has been, and will continue to be, obstructed by the massive mountain ranges between the coal bases and the coastal region. The limited capacity of available harbours has been restricted and will further restrict the huge amount of required coal transportation. It has been reported that although coal transport accounted for near 40 per cent of 'daily average loaded freight wagons' in the 1980s, both the shortage of coal in major consuming areas and the coal stock going beyond the limit of storage capacity in the mining centres have usually coexisted.11 This is not only predicted to continue, but is likely to become even worse. The coal-rich areas of Shanxi, Shaanxi and Inner
(40.0) (16.0) (13.3) (6.0) (10.7) (9.3) (4.7) 1500 (100.0)
600 240 200 90 160 140 70
1453
150 315 390 155 235 148 60
(100.0)
(10.3) (21.7) (26.8) (10.7) (16.2) (10.2) (4.1)
Consumption
2000 (forecast)
450 -75 -190 -65 -75 -8 8
Transp.
Lu (1993:19) cites Wu et al. (1988).
(100.0)
127 -24 -54 -30 -22 7 2
Production
Source:
865
(15.0) (19.8) (20.9) (11.7) (17.7) (10.2) (4.7)
Transp.
(1) This table is based on the output of a very influential state priority research project conducted in 1980-88. (2) The figures in parentheses are the relevant percentage share in the total. (3) Northeast includes three provinces of Heilongjiang, Jilin, and Liaoning; East represents Anhui, Fujian, Jiangsu, Jiangxi, Shandong, Shanghai, and Zhejiang, seven provinces; South-central refers to Guangdong, Guangxi, Henan, Hubei, and Hunan; Southwest consists of Guizhou, Sichuan, Yunnan, and Xizang; and Northwest covers Gansu, Ningxia, Qinghai, and Xinjiang. Usually, Shanxi, Shaanxi, Inner Mongolia, Beijing, Tianjing, and Hebei together are the North.
(100.0)
871
769,200 (100.0)
Total
130 171 181 101 153 88 41
Consumption
1985 (actual)
Notes:
(29.5) (16.9) (14.6) (8.2) (15.0) (10.9) (4.9)
257 147 127 71 131 95 43
(62.4) (8.3) (6.1) (2.3) (3.0) (9.7) (8.2)
480,000 64,100 46,900 17,800 23,400 74,300 63,000
Production
Raw coal reserves and production by region, 1985-2000 (million tons)
Shanxi, Shaanxi & Inner Mongolia Northeast East Beijing, Tianjing, Hebei South-central Southwest Northwest
Region
Table 5.9
I
oo to
Energy as the Representative of Producer Goods Constraints
183
Mongolia house over 60 per cent of China's coal reserves and almost all of its higher-quality coal (Table 5.9; World Resources: 1994: 69). Due to a poor water supply there is insufficient water to provide either slurry pipes or facilities for coal washing in these areas, either of which would certainly help rationalize coal transportation. Perhaps it is due to water shortage that only 18 per cent of China's coal is washed at all (Smil 1988: 87). In order to better understand how effective energy supply is determined by transport bottlenecks, it helps to look at the shifting patterns of coal production and transportation in the 1980s as summarized in Zhou, Fengqi et al. (1990). Between 1980 and 1988, the ratio of incremental freight volume to output increment decreased from 66.9 per cent in 1980 to 41.2 per cent in 1988. This indicates that the growth of freight volume lagged significantly behind that of production. In the meantime the number of provinces with net coal-import increased from 14 in 1980 to 20 in 1988, and over 90 per cent of the interprovincially exportable coal was concentrated in five provinces, mainly in Shanxi province. Since 1980, the state has given key priority to coal transport railway projects. As a result, the coal volume hauled from Shanxi by rail almost doubled from 1980 to 1988, increasing from 72 million tons in 1980 to 140 million tons in 1988 (with additional 21 million tons transported by road). Meanwhile, however, the coal output in Shanxi rose from 121 million tons in 1980 to 232 million tons in 1988. The incremental ratio of output to freight volume is 1.64. Because of such a binding and persistent limit of transport capacity, 'planning output according to the transport possibility' (yiyun dingchan) has been a rule for coal mines in Shanxi. The general inference of the previous analysis is that under the conditions of chronic energy shortage and transport constraints, the normal stock of coal and other energy in consuming sites has remained insignificant in comparison with annual total consumption. As a result, the indicator 'energy consumption' becomes indistinguishable from the effective energy supply in terms of annual measurement, although 'energy output' may be significantly different from energy consumption in some cases. For example, coal used for power generation accounts for about 25 per cent of the total (Peng 1989). The normal coal stock in those power plants which have access to state directly planned coal supply is equal to their consumption for one week, accounting for only 1.7 per cent of their annual consumption. During an energy crisis, coal stocks are often only sufficient for 2 or 3 days, less than 1 per cent of annual consumption
184
Chapter 5
(Liang 1989; Zhang J. 1991). Against this background it is reasonable to take 'annual energy consumption' as a proxy for the effective energy supply. The annual energy consumption can also indirectly represent the shortage intensity of energy, raw materials and other investment goods because of energy's role as discussed in section 5.1. There are three more advantages of using energy consumption as the proxy for effective energy supply. First, energy consumption measurement can be readily understood by officials, scholars and the public, in comparison with other complicated and controversial shortage indicators designed by economists and econometricians.12 Second, the statistics on energy consumption already exist. Third, as the most important output of the transport bottleneck, the statistics available on energy consumption measurement have encompassed the impact of the transport constraint. Therefore, the econometric analysis in section 5.5 and Chapter 6 will use 'energy consumption' as the proxy for 'effective energy supply' and as the representative indicator for the supply possibility in the producer goods sector, in terms of annual statistics.
5.5 Energy Constraint to Investment Demand: Some Primary Econometric Evidence The purpose of this section is to apply the technique of Granger causal analysis (Granger 1969) in the primary investigation of the causal relationship between energy consumption and real investment growth in China. I do not employ correlation analysis here because, as we all know, correlation does not necessarily imply causation in any meaningful sense of the word, and the econometric graveyard is full of magnificent correlation, which are simply spurious or nonsensical. The Granger approach can only supply a primary econometric investigation since it has to be short-run in order to avoid spurious regression on non-stationary variables. In addition, it measures only precedence and information content but on its own does not indicate causality in the more common use of the term. The more detailed analysis done in the next chapter is based on cointegration and error correction approaches. In the Granger sense, the procedure dealing with the question of whether X causes Y is to see how much of the current Y can be explained by past values of Y and then to see whether adding lagged values of X can improve the explanation. Y is said to be Granger-caused by X if Y is better predicted by using the past history of X than by not doing so. It is
Energy as the Representative ofProducer Goods Constraints 185 important to note that the statement 'X Granger causes Y' does not imply that Tis the effect or the result of X. The concrete test can be performed by setting the following regressions: Y =f(n past Y, m past X)
(5.1)
Y=f{n past?)
(5.2)
X=f(n past X, m past Y)
(5.3)
X=f(npzstX)
(5.4)
Equations (5.1) and (5.3) are in unrestricted forms, and equations (5.2) and (5.4) are in restricted forms. Equations (5.1) and (5.2) are paired to show whether all the coefficients of the lagged X's may be considered to be zero. Similarly, equations (5.3) and (5.4) are paired to show whether the coefficients of the lagged Y 's can be zero as a whole. The relevant Ftest for these two null hypotheses can be established as:
URSS/df2
(5.5)
in which RRSS is the residual sum of squares of the restricted regression, URSS is the residual sum of squares of the unrestricted regression, and df represents the degree of freedom. The basic data set for exercising the Granger Causality test consists of the annual real investment level and energy consumption as the proxy for the effective energy supply (cf. section 5.4). The sample period is 1952— 95. Both data and their sources are listed in Table A6.1 of the Data Appendix of Chapter 6. In order to avoid the problem of spurious regression on non-stationary variables, the first difference of the natural logarithms of real investment level and energy consumption (i.e. growth rate) is employed. Another problem, namely that the Granger test is sensitive to the lag structure of the independent variables (Hsiao 1979,1981), may not be relevant here, because the results given in Table 5.10 show a consistent conclusion within the interval consisting of the most popular lag structures for annual series with limited degrees of freedom.
186
Chapters
Table 5.10
Order of lags
3 4 5 6 7 Notes:
Granger causality between real investment and effective energy supply (1952-95) A/ Is not Granger caused by Ae 8.61 5.00 5.02 4.88 4.75
(0.000) (0.003) (0.002) (0.002) (0.003)
Ae is not Granger caused by A/ 3.01 2.43 6.13 4.20 2.92
(0.044) (0.069) (0.001) (0.005) (0.027)
A/ and Ae stand for the first differences of logarithms of real investment and energy consumption, respectively. The figures in parentheses are the significant level of the relevant F-test (eq. 5.5).
Data Source: Table A.6.1.
The main empirical results are presented in Table 5.10, which shows, as just mentioned, a consistent conclusion that there is bi-directional Granger causality between both growth rates of real investment and the effective energy supply. This has two important implications. First, within the information system for explaining investment cycle in this research, the effective energy supply can be weakly exogenous but cannot be strongly exogenous. Second, the growth of effective energy supply will release the energy bottleneck and stimulate investment growth. In the short run, the expansion of real investment will place greater pressure on energy supply and thus cause an increase in the energy supply, based on drawing on stock and overloading the supply capacity of both the transport and energy sectors. However, once the stock is exhausted and the overloading of supply capacity cannot be continued, the effective energy supply will be bound to decrease. Real investment will be forced to cut back through ad hoc economic and administrative measures and/or by a retrenchment campaign.
5.6 Summary This chapter uses empirical evidence to show how the energy shortage has checked China's economic growth and played a representative role among producer goods constraints. Widespread shortage of primary energy has been chronic and this may continue far into the future. Due to power shortages, over 20 per cent of power-driven equipment remains idle. Due to coal shortages power stations have to turn off gen-
Energy as the Representative of Producer Goods Constraints
187
erators. Finally, due to transport constraints, coal shortages in industrial centres and the coal stocks above storage capacity in the mining bases have coexisted in most months. Although since 1980 the government has given high priority to coal transport and railway development, the incremental gaps of coal demand over coal freight volume, and of coal output over the freight volume, have steadily increased. This leads to continued tension between energy demand and effective supply. In an economy such as China, where there is shortage of energy but surplus of labour, these empirical findings justify that the best substitution for the labour factor in the Kaleckian labour constraint equation (2.4) is energy. This chapter also uses the Granger causality test, which shows an evident two-way dependence between the growth rates of real investment and effective energy supply. This finding supports further exploration, and shows that within an information system for modelling investment cycle in this research, the effective energy supply could be weakly exogenous but cannot be strongly exogenous. The weak exogeneity of the effective energy supply can be intuitively understood as follows. The geographical condition of energy bases, the distribution patterns of energy production and consumption, and the geographical difficulty of constructing railways and roads are beyond the policy trade-off of how to allocate investment funds. In other words, the effective energy supply and transport capacity cannot be easily planned for in the short run. However, officials can write out an expanding credit schedule relatively easily. It should also be emphasized that because of the binding transport constraint, there has been a significant gap between energy output and energy consumption (mainly coal). This makes energy consumption the best proxy for effective energy supply and a way to encompass the impact of the transport constraint. In addition, because primary commercial energy is needed for all modern industries, and the energy shortage has caused overall constraint, it is necessary to employ the effective energy supply as the representative for the supply possibility of producer goods sector. In comparison to the agricultural constraint on investment demand, energy constraints are more direct.
.
Estimating Investment Functions Based on Cointegration
6.1 Introduction The purpose of this chapter is to model both investment growth and cycle without artificially imposing a separation between them. This will be done by establishing both a long-run equilibrium investment level (adjustment) function and a short-run investment growth rate cycle (adjustment) equation based on the recently developed cointegration and error correction approaches. This chapter is an extended version of Sun (1998a). The next section introduces such concepts as unit roots, long-run equilibrium relationship (cointegration), and the error correction (disequilibrium adjustment) mechanism. Unit root tests can be applied to determine whether the variables in a regression are stationary or nonstationary. A long-run equilibrium relationship entails a systematic comovement among non-stationary economic variables that an economic system exemplifies precisely in the long run. The corresponding error correction representation shows that the resultant dynamic adjustments of agents' behaviour are attracted by this equilibrium relationship. Based on these new notions we can summarize the elementary findings presented in the previous chapters in a rigorous way: There is a long-run stable equilibrium relationship among real fixed investment, effective energy supply and agricultural output (crucially, grain output) in China. Because of the chronic pressure and incentives to maximize the possibility of investment growth, much of its cyclical pattern can be explained by the adjusting force toward the long-run equilibrium path and by the relevant changes in energy supply and agricultural output.
188
Energy as the Representative of Producer Goods Constraints
189
The concrete modelling framework and steps are sketched in section 6.3. Within the LSE methodological framework (cf. section 1.5), the most comprehensive econometric modelling strategy is the one recently proposed in Hendry & Mizon (1993). Following this strategy, the expected linear dynamic structural econometric models (SEMs) are derived by sequential reduction of an underlying well-specified and unrestricted vector autoregressive (UVAR) representation of the data, without any pre-imposed market clearing conditions and other behaviour assumptions (cf. section 1.5). In sections 6.4 and 6.5 the integration and cointegration properties of the data are analysed. A long-run investment-level function and a conditional error correction model of investment growth rate are presented. The relevant weak exogeneity of the conditional variables is examined. The possibility of the presence of structural break is also tested by the one-step-ahead chow test in recursive estimation. Section 6.6 summarizes the economic interpretation and theoretical implications of the equations. It discusses the advantages of the comovement attractor notion over the traditional 'norm' concept of planners' and/or agents' behaviour adjustment. The relevant technical details are discussed in the technical appendix to this chapter.
6.2 Unit Roots, Equilibrium Relationship and Error Correction Mechanism A significant re-evaluation of the statistical basis of econometric modelling has taken place since the early 1980s. Its analytical basis has been extended from the assumption of stationarity to the inclusion of non-stationary (integrated) processes. The impact of this evolution is already radical, though it is far from complete. It has significantly influenced the choice of model forms, modelling strategy, statistical inference, distribution theory and the interpretation of many traditional concepts such as simultaneity, collinearity and exogeneity. This section attempts to introduce the most relevant notions for the research in an intuitive rather than a precise, mathematical way and considers only the bivariate case. Generalizations on more variables is straightforward but more complex mathematically. The standard classical methods of statistical estimation are based on the assumption that the means and variances of the time series variables are well-defined constants and independent of time. Thanks to the in-
190
Chapter 6
vention of unit root tests, it is found that this classical assumption is not satisfied by a large number of macroeconomic time series. A variable with varying mean and variance over time is classified as non-stationary or unit root variable. Using classical estimation methods such as the ordinary least squares (OLS) to estimate relations involving non-stationary variables may lead to spurious regression. As the means and variances of the non-stationary variables vary over time, all the computed statistics in a classical regression model, which are based on these first and second moments (means and variances), may also be time-dependent and fail to converge to their true values. The conventional tests of hypothesis may be seriously biased here and give misleading inferences. The unit roots notion also has important economic implications. One of the implications particularly relevant to this research is that it calls for a fundamental change in modelling fluctuations in economic activities. The long tradition in macroeconomic paradigms is to treat economic fluctuations as temporary deviations from a more or less stable trend path (and/or moving average norm) of growth in terms of output, income, investment and consumption (cf. section 2.2.3). Unit root tests reveal that the assumption of the stable long-run growth trend is untenable because most aggregate macroeconomic variables are found to be non-stationary. Artificially separating a non-stationary time series into long-run trend component and short-run cyclical component can be very misleading, because random shocks may have a permanent effect on the system and thus fluctuations may not be transitory (Fischer 1988, Granger 1981, Nelson & Plosser 1982, Perron 1994). Time-series econometrics focuses on the estimation of relationships among a group of variables observed at a number of consecutive time points, most of which are non-stationary. The relationships among these series may be complicated. The value of each series may be influenced by others in several previous time periods, and thus the impact of a change in one variable on another depends upon the time horizon we consider. In order to deal with such dynamic relationships, it helps to understand what are usually called 'long-run' equilibrium relationships (cointegration) and 'short-run' adjustments (error correction mechanism).1 The long-run equilibrium relationship and the accompanying error correction mechanism can be understood by considering the following example. Let PNt and PSt be the prices of an agricultural product, for
Energy as the Representative of Producer Goods Constraints
191
example tomatoes, in the north and south of a country. At a given time t, values of these prices will be a point in the plane with axes PN and PS. In this plane the line PN = PS may be considered to be a long-run equilibrium relationship. If the prices are quite different and thus off the line, then some profit-seeking body would be able to make a profit by purchasing tomatoes in one region, transporting them to the other region and selling them there. The operation of the market would tend to raise demand (thus prices) in the region where the extra purchases occur, and to raise supply (thus lower prices) where they are sold. Once the market pressure brings the prices sufficiently close, further profits cannot be made because of transportation costs, risks etc. In this example, the line PN = PS is similar to an attractor, with the existence of some adjustment mechanism such that if the prices deviate from this line, there will be a tendency (error correcting preference) to return to it. Because of uncertainties, incomplete information, longterm contracts, etc., the mechanism may not immediately bring the points exactly to the attractor. Sometimes shocks to the economy may take the prices away from the line, but there will be an overall tendency towards it. If the economy lies on the line, a shock may take it away; and if there is an extended period without exogenous shocks, the economy will definitely go to the line and remain there. The line acts as an 'equilibrium' to which the specific system is attracted, other things being equal. Expressed differently, the line shows the trace of a systematic co-movement among the specific economic variables in the long run. This example indicates that the equilibrium attractor acts as a natural intuitive 'norm' of economic agents' behaviour adjustment. Cointegration can be viewed as the statistical expression of the nature of such long-run equilibrium relationships. It allows us to describe the existence of an equilibrium, or stationary, relationship among two or more time series, each of which is individually non-stationary. That is, while the component time series may have moments such as means, variances, and covariances changing with time, some linear combination of these series, which defines the equilibrium relationship, has time-invariant linear properties. The concept of cointegration can be more strictly explained using characteristics of time series as being either 7(0) or 7(1), that is, being integrated of order either 0 or 1. An 7(0) series can be regarded as a stationary, trend-free series whereas an 7(1) series is such that its difference is 7(0). For an 7(0) series, an old shock to it has virtually no effect on its current value if the shock occurred
192
Chapter 6
long enough ago. For an 7(1) series this is not true; an old shock will still have a noticeable impact on the current value of the series. In other words, under reasonable assumptions, the variance of an 7(0) series is bounded whereas the unconditional variance of an 7(1) series increases without bound as / increases. A pair of time series xt , yt are said to cointegrate if they are each 7(1) but there is a linear combination zt = xt - $yt which is 7(0). A generating mechanism that produces cointegrated series is
where e\t and e2t are 7(0) series and ut is 7(1). Any pair of cointegrated series must have a representation such as (6.1), in which the cointegration property is produced by the single 7(1) component ut. Another generating mechanism that must also occur is the error correction model, of the form A Xt = oti zt-\ + lags of A Xt > A yt + su (6.2) A .y, = ct2 zt-i+ lags of A xt > A y, + 82/ in which at least one of 014 and 0,2 is non-zero, zt—xt — (3y, as before, and both 81/ and s2/ are white noises and hence 7(0) processes. While xt - $yt is considered to be an equilibrium relationship, equation (6.2) represents the corresponding disequilibrium adjustment mechanism that produces this equilibrium relationship by the 'error correction' terms a\zt.\ and a 2 z M . If xt andy, are generated by (6.2) they will be cointegrated and if they are cointegrated then they will certainly have an error correction representation (Engle & Granger 1987).2
6.3 Modelling Strategy and Steps: A General Framework There are three major steps in applying the recently developed modelling strategy characterized by cointegration and error correction. Firstly, it is necessary to use unit root tests to pretest the stationarity and integrating order of each variable before considering the long-run comovement among the relevant variables. Although there is more than one method of unit root tests, the augmented Dickey-Fuller test (ADF)
Energy as the Representative of Producer Goods Constraints 193 is widely used to provide an initial test for the presence of unit roots in the annual time series with an evident trend (Dickey & Fuller 1979). In the case of the presence of important structural changes in the trend, Perron's innovational outlier and additive outlier models are to be recommended (Perron 1993, 1994). A further joint test for unit roots can be conducted by cointegration estimate (Dickey et al. 1994). Secondly, a cointegration relationship is estimated based on a selected modelling strategy and if certain conditions of system specification are satisfied. Thirdly, the short-run or the dynamic disequilibrium adjustment equations are estimated within the error correction framework and by using the estimates of the long-run parameters. The robustness of the estimated dynamic disequilibrium adjustment equations is evaluated by the standard diagnostic tests. Since the ADF unit root test is technically straightforward, the corresponding technical details are presented in the appendix of this chapter. The rest of this section will focus on the discussion of the modelling strategy and the notions of weak, strong and super exogeneities based on a new and more precise setting. Following the LSE methodological framework summarized in Spanos (1990, 1995), it can be said that the most comprehensive modelling strategy over a non-stationary time-series system is the one recently proposed in Hendry & Mizon (1993). The starting point of this modelling strategy is an unrestricted VAR (UVAR) representation for all endogenous variables in the system under analysis: k
Z, = £r/Z/./ + |i, + e/
(6.3)
where Z, is an (n x 1) vector of 1(1) variables, Z_*+i, ..., Zo are fixed, F, (i = 1, ..., k) are (n x n) coefficient matrixes, \it is an (n x 1) constant vector, and e, is independently, normally distributed with mean 0 and covariance matrix Q, namely, s, ~IN(0, Q). Since we plan to distinguish between stationarity by linear combinations of raw variables and by differencing we can write the model in error correction form: k-\
AZ, = nz,-i + Eri/AZ/-/ + li, + e/ i=\
where
(6.4)
194
Chapter 6
k — \^
n ~ 2^i T/ "" * y T-
T
n
=
k \~*
r-«
11/ — ~~ 2-i r J y
/=1
V —1
~ '''''
1r
\
"
7=/+l
One advantage of this reformulation is that the hypothesis of cointegration can be formulated entirely as a restriction on the matrix n i.e. Hr: II = a p ' where P consists of the coefficient vectors of r cointegration relationships and a is the relevant adjustment coefficient matrix, leaving the other parameters unrestricted (thus overparameterized heavily). So that we can use the Gaussian full-information maximum likelihood (FIML) cointegration procedure proposed by Johansen (1988) and Johansen & Juselius (1990) to find these r cointegrating vectors and the corresponding adjustment coefficients matrix. Another advantage, in terms of modelling strategy, is that one can formulate a partial system as a conditional model and discuss its properties. In other words, for Zt = (yt, X/)' the VAR model (6.4) under the cointegration hypothesis Hr can be factorized into a conditional equation for yt given Xt and a marginal model forX, as follows (Johansen 1992b).
t!
(6.5)
k-\
+ Jijr + ej»
(6.6)
where a, I I i , ... , n*_i, e, and Q, are decomposed correspondingly, and co = ClyxQxx1- The stochastic properties of the conditioning variables are well defined, and the various relevant exogeneities also become testable in the factorized system. In such a setup, we can interpret the weak, strong, and super exogeneities more precisely. If the parameter vector of interest denoted 0, for example, the coefficient vector of a unique cointegration relationship or, in our case, of the conditional equation, is a function only of the parameter of the conditional equation, and if the parameters of the conditional and marginal distributions (X\ , Xi) belong to the product of their individual parameter spaces (Ai x A2) and are thus variation free, the variables in Xt are weakly exogenous for 0. If Xt is weakly exogenous for 0 and there is no feedback from yt (and Ay,) to Xt (and AXt),
Energy as the Representative of Producer Goods Constraints
195
thenX, is strongly exogenous for 0. If Xt is weakly exogenous for the parameter 0 and 0 is invariant to changes in the parameters of marginal distribution for Xt, then Xt is super exogenous for the parameter 0, to the class of changes in the marginal process. A formal test of weak exogeneity for the parameter P in n = aP' is found in Johansen (1992a, 1992b) and Johansen & Juselius (1990), and an extended test of weakly exogeneity for all of the long-run and short-run coefficients in the conditional equation as parameters of interest is given by Urbain (1992). The weak exogeneity for all the conditional parameters can also be tested via the tests of super exogeneity of Engle & Hendry (1993). Weak exogeneity implies that the conditional equation contains as much information as the full UVAR system about the parameters of interest in the equation. In particular, if there is only one cointegrating relation, the analysis of model (6.4) can be reduced to the well-known single-equation analysis. In our case it means that the derived conditional investment determination equation does not lose any information on the parameters of interest i.e. both long-run and short-run coefficients in the conditional equation, when compared to the corresponding completely specified vector system (UVAR), provided r = 1.
6.4 Estimate of Cointegration and Investment Level Equation Based on the analyses of Chapters 4 and 5, the data set used in the modelling exercise consists of annual data on fixed investment by state sector (/), the deflator of/(P), a lagged 3-year moving average of grain output per capita (GM) as the representative series of the basic consumer goods, and energy consumption per capita (E) as the representative indicator of producer goods. We use a lagged 3-year (t, t-t, t- 2) moving average of grain output per capita because, as discussed in Chapter 4, at least two or three years of consecutive good (bad) harvests can create the general expectation that the bottleneck of agriculture has been significantly relieved (narrowed) and thus investment booming can be guaranteed (contraction is urged). However, a single good (bad) harvest alone is not sufficient to stimulate such general expectations.
196
Chapter 6
Table 6.1
Tests for unit roots Variables
gm Levels ADF Equation form MacKinnon critical value (10%)
-2.513 with trend t -3.190
-2.983 with t -3.187
-3.416 with t -3.514(5%)
-2.490 with t79 -3.61 (Perron)
-5.122 with drift c -3.597
-5.301 with c -3.585
-3.859 without c -2.616
-5.751 with t79 -3.91(5%) (Perron)
First differences
ADF Equation form MacKinnon critical value (1%) Notes:
'With trend t' means to allow for the presence of a linear trend under the stochastically stationary alternative, whereas 'with constant c' means to allow for a non-zero mean only. The 't79' is a special time trend starting from 1979 (=1, and then, 2, 3,4 ...), which indicates the structural changes in the trend of price level since 1978. The generating procedure of MacKinnon critical values is given in MacKinnon (1991). For /, gm and e, the inferences about the significance of trend t, drift c, and augmented lags AyM (/ = 1, 2, .... p) are based on standard t-test (Harvey 1990: 812; Engle & Yoo 1989; Holden & Perman 1994: 64-5). The Perron critical values are directly taken from Perron 1993: Table 1, and correspond to X = 0.4 there.
Table 6.2 Equation
Residual misspecification tests of the UVAR model (6.4)
Standard deviation
oKewness
A/
0.189
-0.608
0.721
Ae
0.104
0.150
3.258
Agm
0.018
-0.049
-0.812
Ap
0.029
1.872
7.296
Notes:
Excess kurtosis
Normality testx2(2) 3.549 (0.170) 22.962 (0.000) 0.766 (0.803) 18.608 (0.000)
ARCH(1) testX2(1) 0.515 (0.473) 1.390 (0.238) 0.023 (0.880) 0.891 (0.345)
Autocorr. test x2(2) 1.669 (0.434) 3.635 (0.163) 6.349 (0.042) 11.190 (0.004)
Figures in parentheses are the probability values of the test statistics under the relevant null hypothesis.
The sample period is 1953-95. The data appendix of this chapter details the data and their sources. Throughout the paper lower-case letters represent natural logarithms of capitals, and A denotes the first-difference operator. It should be mentioned here that the sample size of 43 observations is admittedly small for the integration and cointegration analyses. While attempts have been made to correct for small-sample biases, the limitations inherent in a small sample needs to be borne in
Energy as the Representative of Producer Goods Constraints
197
mind when interpreting the findings. The findings should be treated as instructive rather than definitive. On the other hand, compared with existing research, the sample used here is one of the largest. More importantly, the results lend themselves easily to economic interpretation and exhibit well-specified and well-fitted statistical properties. Following the general steps mentioned at the beginning of the last section, it is necessary first to test the stationarity and integrating order of each time series before considering the long-run co-movement among them. To test for the existence of autoregressive unit roots in the non-deterministic components of the individual series, the ADF procedure for /, gm and e and Perron's additive outlier model forp are used (Dickey & Fuller 1979; Perron 1993, 1994)).4 Test results presented in Table 6.1 show that all /, gm, e and/? appear to be 7(1). A UVAR cointegration analysis of the joint integration properties of the data will confirm this once more. Model (6.4) together with the reduced rank hypothesis II = ccP' is the base of the empirical analysis. The misspecification tests for this model with k = 2 and unrestricted constant are reported in Table 6.2.5 The results are satisfactory, although there are some indications of excess kurtosis in equations Ae and Ap, causing the Jarque-Bera test statistic for normality to become significant, and of error autocorrelation in equation Agm and Ap. This is not surprising since the variables e9 gm andp are chosen to explain the variation of/ but not vice versa. In other words, the selected information set is probably not sufficient to account for the variation in e, gm andp. Additionally, e, gm andp might well be weakly exogenous for the log-run parameters of interest, which would make the deviation from normality less important. In fact, the error autocorrelation in equation Agm comes directly from the lagged 3-year moving average in the construction of GM. The error autocorrelation in equation Ap may link with the inflation expectation in the post-reform period. By graphic analysis, the large residuals in equation Ae were found to coincide with the high waste of energy in the 'Mass Steel- and Iron-Smelting' campaign during the 'Great Leap Forward' period (1958-60) and the consequent collapse in 1961-62. Those in equation Ap were found to correspond with high growth rates of investment prices in 1985, 1992, 1993 and 1994, as well as with the stagnation during the 'Rectification' period in 1989-91. When introducing event-specific dummy variables into Model (6.4) to capture these exogenous disturbances, it is found that the specification of equa-
198
Chapter 6
tions Ae and Ap can be improved, but the /-statistics on the dummy variables are not significant in equation A/. Here, to save degrees of freedom is more important, we thus decide not to introduce these dummy variables into the cointegration analysis. In order to examine the cointegration property of the basic system (/, e, gm and p), the Gaussian full-information maximum likelihood (FIML) cointegration procedure proposed by Johansen (1988) and Johansen & Juselius (1990) is applied. The dimension of the cointegration space could be determined by means of maximal eigenvalue (A^ax) and trace (A^-ace) likelihood ratio (LR) test statistics. However, in the case of a small sample, the Johansen procedure may over-reject when the null is true, as pointed out by Reimers (1992). Thus we employ Reimers' small-sample correction by using {T-nk)log(\-Xi) instead of Tlog(l-Xi) for the test statistics, which typically improves the properties of LR tests for cointegration in moderately sized samples (Reimers 1992, Reinsel & Ahn 1992). The outcomes of the test using both T-nk and Tare presented in Table 6.3. The LR tests by using T-nk indicate that we can reject r=0 at the 5 per cent level and accept the other hypotheses, whereas the original LR test indicates that we should reject r=\ at the 5 per cent level. The graphic test for cointegration relations is used to reveal the improved properties of the LR test, using T-nk. Figure 6.1 shows that the relationship P'(Z', 1)\ which corresponds to the first eigenvector p in Table 6.3, looks fairly stationary but all of the others appear to be distinctly non-stationary. Based on these findings we can proceed on the assumption of one cointegration vector and three unit roots. Table 6.3 also reports the standardized eigenvectors and corresponding adjustment coefficient vectors derived from the maximization of the likelihood function of Model (6.4). The cointegration vector p will characterize a long-run solution to a real investment level determination equation if the relevant cointegration relation is homogeneous of degree one in p. Table 6.4 presents the results of testing unitary elasticity of energy per capita, price homogeneity of degree one, double elasticity of grain output per capita, as well as the clarified long-run investment determination relation, which are based on Johansen's (1992c) LR tests of linear restrictions on p. The results suggest that we should accept a well-performed long-run investment adjustment function for China as follows:
Energy as the Representative of Producer Goods Constraints Table 6.3
199
The cointegration analysis of model 6.4
Hypothesis
r=0
r=1
r=2
r=3
Eigenvalue
0.607 30.85 [38.393] 28.10 65.37 [81.22] 53.10
0.425 18.28 [22.71] 22.00 34.52 [42.89] 34.90
0.272 10.50 [13.04] 15.70 16.24 [20.18] 20.00
0.160 5.75 [7.14] 9.20 5.75 [7.14] 9.20
bi
b2
b3
1.000 -1.000 -2.015 -1.175 13.800
-0.728 1.000 3.366 2.035 -21.88
-0.086 -0.156 1.000 0.041 -4.423
-0.139 -0.218 -9.487 1.000 56.03
a
ai
a2
a3
-0.939 -0.027 0.026 -0.050
0.171 -0.058 0.021 0.034
1.048 0.733 -0.027 -0.012
-0.009 0.002 0.005 -0.009
?tmax by
T-nk
95% Quantile ;Wace by T-nk 95% Quantile Eigenvector
P
/ e gm p Constant Adjustment coefficient M Ae Agm Ap Note:
'By T-nW indicates the Reimers' (1992) small-sample correction by replacing Tby T-nk in both A™ax and A^ace statistics, where n is the dimension of the vector system and k is the lag length in Model (6.4). The figures in square brackets are the original ^max and W e statistics. The critical values are taken from Table 6.1 (case 1) in Osterwald-Lenum (1992).
~ Pt
=
2.0 gm t
(6.7)
In order to confirm this result for those readers who prefer explicit standard errors presented in parentheses together with the estimated equation, an unrestricted autoregressive distributed lag of / on p, e and gm with two successive lags has also been estimated. The solved 'static' long-run solution is: / =
1.19p+ 1.01e+ l.95gm+ 13.48 (0.11) (0.06) (0.27) (1.36)
(6.7*)
which is statistically not different from equation (6.7). What is new about equation (6.7) is that it is based on Johansen procedure and dem-
200
Chapter 6
onstrates empirically the cointegration property of the basic system (i, e,gm and/?).
Figure 6.1
The graphic tests for cointegration relation P'Z, 1.5 X.2 .9
1965
1975
1985
1995
vectop3=
veotor4=_
1965
Note:
1975
1985
1995
Vector 1 represents the relationship p'(Zf' 1)\ Vectors 2-4 correspond to relationships b\(Z{ 1)\ b'2(Zt 1)',and b' 3 (Z/1) \ Please refer to equation (6.4) and Table 6.3.
Table 6.4
Tests for linear restriction on p and weak exogeneity
Hypothesis (Ha)
LR(Ha, Ho)
Hypothesis (Ha)
LR(H a , Ho)
p2 = -1.0(pi = 1.0) p3 = -2.0(Pi = 1.0) p4 = -1.0 (ft = 1.0) P = (1, - 1 , - 2 , - 1 , * ) '
0.000(1.00) 0.003 (0.95) 2.586(0.11) 2.832 (0.42)
on = 0 a2 = 0 a3 = 0 a4 = 0 a 2 = oi3 = a 4 = 0
4.89 (0.03) 0.06(0.81) 0.47 (0.49) 0.19(0.66) 0.87 (0.83)
Notes:
LR(Ha> Ho) is Johansen's (1992c) LR statistics for testing H a in Ho: r = 1, asymptotically x2(3) distributed on the null of p = (1, - 1 , -2, -1)' and a 2 = a 3 = a 4 = 0; and x 2 (1) distributed on the null of the others. The figures in parentheses are the probability of the LR test.' * ' in p means no restriction on the constant element.
Energy as the Representative of Producer Goods Constraints
201
Equation (6.7) indicates that the long-run real investment level in China has moved along the supply possibilities frontier of bottleneck sectors and along the distributive barrier between industrial expansion and necessary agricultural growth, represented by the effective energy supply and per capita grain output, respectively. In other words, the key binding constraints on the long-run level of real investment demand are the effective supply of representative consumer goods and producer goods as well as the distributive barrier which industrial expansion poses to necessary agricultural development. Another important economic implication of equation (6.7) is that it serves as a statistical expression of an equilibrium or stationary relationship between real investment (corresponding to the demand side in a typical equilibrium equation in an ex ante sense) and energy supply and grain output (corresponding to the supply side). This forms an empirical proof of the existence of the equilibrium relationship without the aid of first modelling separately the demand and supply functions. As a consequence, the error term, {z -p - e - 2.0g7w},_i, represents the previous disequilibrium and should be a useful explanatory variable for the next direction of movement of // or (z -p)t. More concretely, when the error is positive (negative), z,_i and (z -p)t-\ are too high (low) relative to the equilibrium comovement path and the economic agents will in general reduce (increase) / in future periods relative to the comovement path given in equation (6.7). It is the error-correcting behaviour on the majority of economic agents that induces the cointegrating relationship among the corresponding economic time series. The estimated adjustment coefficient vector a suggests that the disequilibrium error P'(Z'/-i 1)' has an important impact on A/, while its influence on Ae, Agm and Ap is less significant. This is confirmed by running Johansen's (1992c) LR tests of linear restrictions on a. Table 6.4 also records these tests for the hypothesis a, = 0 (z = 1, 2, 3, 4), where a, is the z-th element of a. As proved by Johansen (1992a, 1992b) and Urbain (1992), a, = 0 is equivalent to the z-th component of Zt being weakly exogenous for the long-run parameters in P as well as for the short-run parameters in the conditional equation (6.5). The test statistics in Table 6.4 show consistently that the weights of the cointegration relationship in the Ae, Agm and Ap equations are not significantly different from zero, implying that energy supply, grain output and inflation are weakly exogenous in the UVAR system. Thus it becomes possible to construct a well-specified conditional investment determi-
202
Chapter 6
nation equation based on (6.5) to fully capture the dynamic of investment determination represented in Model (6.4).
6.5 Estimate of Conditional Investment Growth Rate Equation Based on the findings in the last section, the conditional representation (6.5) of the data can now be used directly as the starting point for sequential reductions and reparameterization aiming at the development of the well-specified conditional investment determination equation. It should be pointed out that the UVAR model (6.4) has been mapped into 7(0) space through differencing and cointegration transformation (restricting n = ccp'), thereby allowing OLS estimation and statistical inference to proceed along the lines of standard Gaussian asymptotic theory. This being the case, and taking into consideration the price homogeneity of degree one, we can rearrange model (6.5) into real investment terms so that we can conveniently interpret the economic implication of the induced error correction equation. The rearranged Model (6.5), as an 7(0) equation, is estimated by OLS, and is subsequently simplified in order to obtain a more concise, yet congruent data characterization. The results of estimation with the usual standard errors are reported in Table 6.5. The statistics listed in Table 6.5 indicate that the goodness of fit measured by R2 and a is highly significant, and all of diagnostic tests suggest the equation appears to be well-specified. The overall parameter constancy of this conditional equation is confirmed by recursive estimation. Figure 6.2 shows the sequence of one-step-ahead Chow tests, that is, the standard structural break test, none of which are significant. Figure 6.2 also records the one-step residuals and the corresponding ±2 standard errors to show that the standard error a is almost constant over the sample. The constancy of each coefficient, except the Ap's, is also confirmed by the same recursive estimation; the relevant figures have been omitted here for simplicity. The recursive ^-statistics for Ap's coefficient show that the Ap's impact is dominated by the striking outlier after 1985 when the industrial price reform was initiated; before 1985 the impact was not statistically significant. Together these suggest that the equation given in Table 6.5 appears to be a wellspecified and well-fitted investment growth rate determination equation
Energy as the Representative of Producer Goods Constraints
203
with parameter constancy (i.e. no structural break) over the sample period (1953-95).
Figure 6.2
Note:
One-step residuals, ±2 standard errors, and one-stepahead Chow tests for the investment growth equation
Res1 Step is the one-step residual; S.E. is the standard error. 11 CHOW is the one-step-ahead Chow test; 5% crit. is the critical value at the 5% significant level.
The investment growth rate equation has a clear economic interpretation in all of its explanatory variables. First, the estimated coefficient of the cointegration relationship (i.e. error correction terms) reveals a large and significant adjustment to disequilibrium deviations of real investment level from their norm level determined by the long-run investment determination function. In other words, on average the investment agents decrease (increase) their real investment outlays by about 83 per cent of the last year's over- (under-) investment. Second, notwithstanding this disequilibrium impact, the short-run bottleneck multipliers of effective energy supply (representative producer goods) and grain output (necessary consumer goods) are significantly greater than zero; their values are fairly close to the corresponding long-run multipliers in equation (6.7). Third, previous real investment growth has a significant short-run inertia as discussed by Bauer (1978) and others, which is indeed consistent with certain characteristics of a socialist economy. Fourth, the significant positive correlation between A(i-p) and Apt-\ reflects the basic fact that, in a bureaucratically coordinated
204
Chapter 6
economy, an increase in the price level is not only a result of overinvestment but also, perhaps more significantly, a reason to request extra investment at least in the short run because shifting the blame of extra expenditure to rising input prices is straightforward and appears to be more reasonable (Kornai 1992: 548-52). This effect was relatively limited before reform, but has played an increasingly important role during reform, as shown by the recursive /-statistics mentioned in last paragraph. The variable A/?,_i also represents a special short-run shortage signal translated by the market component of the economy. However, the estimated investment growth rate equation implies that this effect is overbalanced by the impact of offering bargaining reasons for requesting additional investment. Finally, the parameter constancy indicates that there is a lack of structural break induced by the reform. It should be emphasized that the deeper constancy prevailing in the determination of the investment cycles does not imply the concrete and immediate adjustment mechanism remaining unchanged over time. In fact, as sections 2.4.2, 3.4.5 and 4.6 reveal, the concrete adjustment processes can be quite different due to the changing patterns in distribution of power among different political groups and between the central and local governments. Several significant changes have taken place during the reforming years. First, the inflation barrier has played an increasingly important role in the investment adjustment process. This may make Kaleck's growth theory more relevant to China. Second, since credit has begun to play a major role for financing investment, the investment cycles have shown a new scenario during the reform period. Over-investment first induces over-expansion of bank credit, followed by excessive money creation, high inflation in the market component and shortage in the planned component of the economy. This in turn justifies retrenchment based on ad hoc administrative measures (cf. section 3.4.5). Third, the rigour/chaos characterizes the reform cycle and the widespread rent-seeking activities (cf. section 2.4.2, Lin et al. 1996). However, a key feature has been persistent. This is the investment hunger and the resulting tensions between investment expansion and the supply and distributive barriers to the expansion. This fundamental feature is exactly what has been captured by both the long-run investment level equation and the short-run investment growth equation.
Energy as the Representative of Producer Goods Constraints 205
Table 6.5
The estimation of investment growth rate equation in China Dependent Variable: A(/-p) t
Fxnlanatorv \7ariablp ^A.L/ICII I d l l / I V VCIIICIUIw
A(/-P)M
Aet &gmt Apf-i
(/-p-e-2.0fifm;M Constant Type of Test 2
Coefficient
Standard Error
0.233 1.547 2.311 1.154 0.833 11.437
0.072 0.110 0.424 0.277 0.124 1.696
Test Value
Probability
R
0.90
G
0.09 1.86 65.70 0.42 0.61 0.06 1.11 1.24 1.63
DW F(5, 35) Normality %2(2) AR 1-2 F(2, 33) A R C H 1 F(1,33) HeteroF(10, 24) Functional Form F(20,14) RESET F(1, 34) Notes:
— 0.00 0.81 0.55 0.82 0.40 0.34 0.21
d is the estimated standard deviation of residuals. The normality x2(2) is the Jaris the Lagrange Multiplier test for /th- to que-Bera statistic. AR i-j F(q, T-K-q) yth-order residual autocorrelation (q=j-i+ 1; T = the number of observations, and K = the number of regressors in the equation). ARCH 1 - q F(q, T-K-2q) is the qfth AutoRegressive Conditional Heteroscedasticity test. Hetero F(q, T-K-q - 1) is White's test for heteroscedasticity and tests the joint significance in a regression of the squared residuals on the regressors and their squares. Functional Form F(q, r - K - c / - 1 ) i s a general test for functional form mis-specification/ is Ramsey's test for specification heteroscedastic errors. RESET F(q, T-K-q) error. Probability in test-panel is the probability values of the test statistics under the relevant null hypothesis.
6.6 Theoretical and Empirical Implications: A Summary Industrial capital accumulation and investment cycles is a recurring subject in the literature which seeks to explain the working of a socialist and/or developing economy. This chapter advances the standard investment cycle theory of a socialist economy by simultaneously modelling both investment level and growth rate based on the framework of
206
Chapter 6
cointegration and error correction, and by incorporating the distributive barrier theory related to a typical dual developing economy. It is shown empirically that a persistent tension exists between system-generated investment ambitions and the supply and distributive barriers to the ambitions for both pre- and post-reform periods and in the presence of significant policy changes. As a result, first, the longrun investment function is characterized by equilibrium co-movement among the real fixed investment level, grain output per capita as a proxy for necessary consumer goods and effective energy supply per capita as a proxy for basic producer goods. Because the annual real investment level has persistently moved along the supply possibilities frontier of bottleneck sectors, it has been possible to maintain a constant high level of real investment. Second, much of the cyclical pattern of investment growth rate can be explained by the adjustment to the comovement path and by the relevant change rates of energy supply and agricultural output. Third, the fact that both functions maintain parameter constancy over the sample period indicates that there is a lack of structural break induced by the reform. This constancy, together with the structural rigidity of the state industrial sector (cf. section 1.3), may suggest that China's state investment system has still followed its own logic. Until very recently, it has not actively responded to the significant changes in demand occurring in a rapidly growing and transforming economy. Furthermore, the investment decision-making process is still dominated by bureaucratic negotiation based on vested sectoral interests and structural inertia, although the concrete mechanisms have changed. This finding is consistent with the recognition of China's top officials. For example, in his annual report to the 1997 National People's Congress, Chen Jinhua, head of the State Planning Commission, once again blamed the poor performance of state-owned enterprises on overheated fixed investment. He claimed that the persistent overheated investment led to excessive production capacity and insufficient use of facilities in most of the state industrial sectors, as well as to low efficiency, high costs and low competitiveness of the whole state sector {People's Daily, 4 March 1997). The concepts of 'norm' and 'control by norm' play a central role in the traditional modelling practice of a socialist investment cycle. A fundamental assumption in the practice is that control is based on 'negative feedback'. Because of excessive intended investment and output,
Energy as the Representative of Producer Goods Constraints
207
tensions arise, leading to reduced investment and output. Within the framework of negative feedback, a 'normal' value needs to be defined for each variable. The deviations of the control variables from their norms depend on the deviations of the controlled variables from their norms, simultaneously and/or with lags. The difficulties in defining and measuring 'norm' has been widely acknowledged by modellers.6 As a compromise, univariate mean or moving average has been employed as the norm for each variable (e.g. Kornai 1982, Kornai & Martos 1981, Roland 1987, Simonovits 1992). The question of what determines the interaction among such norms and how can we deal with it has been left untackled. In this regard, this chapter may make an important advancement. The co-movement equation (6.7) depicted by the cointegration relation acts as an equilibrium attractor of the investment adjustment behaviour of the economic agents toward the dynamic equilibrium, as shown by the error correction equation. This cointegration attractor is determined by tensions between investment ambition and the corresponding barriers rather than by habit or convention. The attractor has an intuitive interpretation of disequilibrium adjustment behaviour and can be modelled. The disequilibrium error correction is, in terms of economic behaviour, also more informative and instructive than univariate negative feedback. This new framework avoids the spurious deviation from the 'norm' by a filter such as a moving average (Fishman 1969), and determines naturally the long-run interaction among relevant univariate levels.
208
Chapter 6
Appendices
Al Data See Table A1 on next page.
Notations: Investment
=
Grain Energy Population Price index
= = = =
Fixed investment by state-owned units at current prices, and in billion yuan. Total output of grain, in million tons. Overall energy consumption, in million tons coal equivalent. Total population, in millions. Price index of fixed investment by state-owned units.
Sources and Notes: Investment, grain, energy, and population: Data for 1951-96 are directly taken from Yearbook 1993: 149, 364, 477, 81; 1994: 140; 1995: 137; 1997: 150, 383, 215, 69). Price index: Since the official statistics of this index started in 1989, we have to look for a proper proxy for it before 1989. Figures for 1952-85 are the deflator of accumulation on fixed assets of state-owned units (AFAS). The deflator can be generated based on AFAS's value series at current prices and index series at comparative prices taken from Statistics on National Income: 1949-1985 (1987: 37, 40). Because capital accumulation is the newly added fixed assets (less depreciation of the total fixed assets) plus the newly acquired circulating fund, and the latter only accounts for about 25 per cent of the total (Yearbook 1993: 76, 48), we claim that this deflator is the best proxy for the price index of fixed investment by state sector in China's statistics. The most suitable proxy data for 1986-88 are price indices of state-owned construction output values which were published after 1982 (Survey 1987: 75; 1988: 73; 1990: 82; 1994: 89), because construction accounts for about 60 per cent of the investment (Yearbook 1993: 150). The figure for 1989 is from Chinese Economic Yearbook (1990: 11-32). Data for 1990-93 are from Economic Situation and Prospect of China, 1991-92 (1992: 202, 209); 1992-93 (1993: 207); 1993-94 (1994: 3) and for 1994-96, data are from People's Daily (2 March 1995: 2; 5 March 1996: 2; 5 April 1997). a
According to Yearbook (1997: 215), 1996 figure of energy consumption was estimated one.
Estimating Investment Functions Based on Cointegration
209
Table A1 Year 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996
Investment
__ 9.159 10.268 10.524 16.084 15.123 27.906 36.802 41.658 15.606 8.728 11.666 16.589 21.690 25.480 18.772 15.157 24.692 36.808 41.731 41.281 43.812 46.319 54.494 52.394 54.830 66.872 69.936 74.590 66.751 84.531 95.196 118.518 168.051 197.850 229.799 276.276 253.548 291.864 362.811 527.364 765.797 932.249 1089.820 1205.620
Grain
Energy
143.69 163.92 166.83 169.52 183.94 192.75 195.05 200.00 170.00 143.50 147.50 160.00 170.00 187.50 194.53 214.00 217.82 209.06 210.97 239.96 250.14 240.48 264.94 275.27 284.52 286.31 282.73 304.77 332.12 320.56 325.02 354.50 387.28 407.31 379.11 391.51 402.98 394.08 407.55 446.24 435.29 442.66 456.44 445.10 466.62 504.54
__ __ 54.11 62.34 69.68 88.00 96.44 175.99 239.26 301.88 203.90 165.40 155.67 166.37 189.01 202.69 183.28 184.05 227.30 292.91 344.96 372.73 391.09 401.44 454.25 478.31 523.54 571.44 585.88 602.75 594.47 620.67 660.40 709.04 766.82 808.50 866.32 929.97 969.34 987.03 1037.83 1091.70 1159.93 1227.37 1311.76 1388.11*
Population
Price index
563.00 574.82 587.96 602.66 614.65 628.28 646.53 659.94 672.07 662.07 658.59 672.95 691.72 704.99 725.38 745.42 763.68 785.34 806.71 829.92 852.29 871.77 892.11 908.59 924.20 937.17 949.74 962.59 975.42 987.05 1000.72 1016.54 1030.08 1043.57 1058.51 1075.07 1093.00 1110.26 1127.04 1143.33 1158.23 1171.71 1185.17 1198.50 1211.21 1223.89
«... 100.00 98.64 97.79 94.82 91.95 88.74 90.48 95.85 99.36 100.66 101.17 100.59 99.37 97.39 96.57 96.51 94.49 94.10 93.21 94.22 95.01 95.25 95.16 96.09 96.97 98.13 98.78 100.78 102.66 106.30 108.85 111.29 116.17 129.22 142.14 154.08 175.34 192.87 214.86 235.27 271.27 352.38 394.67 425.45 448.85
210
A2
Chapter 6
Cointegration Analysis of the Vector System
Unit Root Test Most econometric tests are built upon the assumption of stationary, ergodic stochastic processes. In general, the statistical properties of regression analysis using non-stationary time series are dubious. On the other hand, although most economic time series are not stationary, many of them, in fact, can at least be approximated by stationary processes if they are differenced. If a non-stationary series must be differenced d times to make it stationary, it is said to be integrated of order d. This is expressed by writing yt ~ I(d) (Engle & Granger 1987). Clearly, it may be important to pretest the stationarity and integration order of each relevant time series before considering their long-run stable relationships. Moreover, the unit root test has its own independent value in terms of detecting some basic features of an individual series. Various tests have been suggested for testing stationarity. A straightforward procedure for annual time series is to test p = 1 in the following equations: y^ai
+ axt + x,
x,= PxM + e,
(f = 0 , 1 , 2 , - ) ,
(t=l,293,.~).1
(A2J) (A2.2)
If the error term in equation (A2.2) is a white noise process then (A2.1) and (A2.2) can be interpreted as a random walk about a linear trend when P = 1, and as an asymptotically stationary first-order autoregressive (AR(1)) process about a linear trend when | P | < 1. However, in this case we are not using a conventional Student's f-test. An appropriate and simple approach is suggested by Dickey & Fuller (1979, 1981). The so-called Dickey-Fuller test (DF) relies on the OLS estimates of any of the following regressions, which are equivalent to equations (A2.1) and (A2.2), namely: Ay, = Ty M + £,
(A2.3)
Ay, = T0 + Ty M + e,
(A2.4)
Ay, = T0 + ixt + xy M + e,
(A2.5)
AV/ = T0 + 1\t + T2t2 + Ty M + £,
(A2.6)
Estimating Investment Functions Based on Cointegration
111
where A is the difference operator (Ay, = yt-yhi). x = (M. Equation (A2.3) makes sense for x < 0 only if yt has (population) mean zero, (A2.4) allows yt to have a nonzero mean, (A2.5) allows it to have a trend, and (A2.6) allows it to have a trend that changes over time.2 The DF test consists of testing that x is negative and significantly different from zero; x < 0 implies that p < 1 and that the series is stationary (with or without drift, trends, etc.). Under the unit root hypothesis, the DF statistics have nonstandard distributions. Especially, the conventional ^-statistic is not asymptotically distributed as the standard normal distribution. We have to use the special tables of critical value given by Dickey (1976) and Fuller (1976), and particularly, by MacKinnon (1991). MacKinnon implemented a much larger set of replications than did Dickey and Fuller, and he estimated response surface regressions more accurately over these replications. Asymptotic critical values can be read off directly from these regressions, and critical values for any finite sample size can be easily computed with a hand calculator. The main drawback of the original DF test is that it does not take account of possible autocorrelation in the error process. If et is autocorrelated then the OLS estimates of equations (A2.3)-(A2.6) are not efficient. In order to overcome the above problem we can employ AR(p) model: p
A
T
yt=
+
*A yt-i+*
yt-\ Z
(A2-7)
i=\
which corresponds to equation (A2.3). We also can add the second term Ex, Ay,./ into equations (A2.4)-(A2.6) to construct relevant AR(p) models. Where p should be relatively small in order to save degrees of freedom, but large enough to allow for removing the autocorrelation in et. The distribution of the '/-statistic' associated with>v-i is the same as that given in the DickeyFuller tables for the AR(1) model, and the test for a unit root is known as the 'augmented Dickey-Fuller test' (ADF). On the other hand, it can be shown that the /-statistics associated with AyM (/ = 1, 2, ..., p) are asymptotically standard normal distributions (Harvey 1990: 81-2, Engle & Yoo 1987). In the case of important structural changes in the trend function of a time series, Perron's innovational outlier model and additive outlier model have shown an evident comparative advantage (Perron 1994). For a given series {yt, / = 0, 1, ..., T}, the approach is generalized to allow a one-time change in the structure occurring at time Tb (1 < Tb < T). Three different models are considered under the null hypothesis: one that permits an exogenous change in the level of the series (a 'crash'), one that allows an exogenous change in the rate of growth and one that permits both changes. For the structural changes to the trend function one can view them as 'big shocks' or infrequent events that have permanent effects on the level of the series. In order to
212
Chapter 6
distinguish the way these 'big shocks' affect the level of the series, two other models are introduced (Perron 1989). The first, the 'additive outlier model', specifies that the change to the new trend function occurs instantaneously. The second, the 'innovational outlier model', specifies that the change to the new trend function is gradual. The corresponding test procedures are based on simple autoregressions (estimated by OLS), which are appropriately augmented with trend and dummy components. For the innovational outlier models, the regressions are as follows:
i=i
and
p + 8D(Tb ) t + x yt_x + X T, A yM + 8 ,
(A2.9)
/=i
where DUt = 1 and Dft = {t- Tb) if t > Tb (0 otherwise), and D(Tb)t = 1 if t = Tb + 1 (0 otherwise). Equation (A2.8), the crash model, allows for a one-time change in the intercept of the trend function. Equation (A2.9) permits both a change in the intercept and the slope of the trend function to take place simultaneously. The statistic of interest is the /-statistic for test that T = (J-l = 0 as before. In the case where the break point Tb can be treated as known, it allows the /-statistic for testing x = (}-l = 0 to be invariant (in finite samples) to the value of the change in the intercept under the null hypothesis. For the additive outlier models, the procedures consist of two-step approaches. In the first step, the trend function of the series is estimated and removed from the original series through the following regressions: yt = \i + r|/ + yDUt + xt yt = \i + x\t + QDUt + yDT*t + x, ^ = ^1 + 11/ + yDT*t + xt
(A2.10) (A2.ll) (A2.12)
where xt is defined as the detrended series. The second step also differs according to whether or not the first step involves DUt, i.e. the dummy associated with a change in intercept. For equations (A2.10) and (A2.11), the test is exercised in the following autoregression applied to the estimated detrended component xt:
Estimating Investment Functions Based on Cointegration
A
^ = iVt-1+ £ dj D(Tb)t.j 7=0
+ E T/ A yt_t + s,
213
(A2.13)
/=1
For equation (A2.12) where no change in intercept is involved and the two segments of the trend are joined at the time of break, there is therefore no need to introduce the dummies in the second step regression. The second step form of the regression is thus the same as equation (A2.7), but applied to the estimated detrended series xt. Under the null hypothesis of a unit root, the ^-statistic for x = (M = 0 also has a non-standard distribution. Furthermore, the critical values to be used depend on the particular model selected. For models (A2.8), (A2.9) and (A2.13) the critical values can be found in Perron (1989: Tables IV.B, VLB) and are the same in the innovational or additive outlier versions. For the additive outlier model based on (A2.12), the critical values can be found in Perron (1993).
Cointegration and Error Correction Model The recent development associated with testing the long-run equilibrium behaviours of time series that are difference stationary processes is inextricably related to the concept of cointegration. The notion that in the long run two or more variables might have convergent values, i.e. the deviations from this long-run co-movement path are stationary, has received considerable empirical testing following the work of Granger (1981), and others (see Hendry 1986, Banerjee et al. 1993), on cointegration. Cointegration is defined as follows (Engle & Granger 1987): The components of the ^-dimensional vector Z, are said to be cointegrated of order d, b, denoted Zt ~ CI(d, b), if (a) all components of Z, are I(d); (b) there exists a vector p (* 0) so that ut = p>, ~ l{d-b\ b>0. The vector p is called the cointegration vector. In the d = b > 0 case, cointegration would mean that although the components of Z, were all I(d), the equilibrium error would be 7(0), and ut will rarely drift far from zero if it has zero mean and ut will often cross the zero line. If Zt was not cointegrated, then ut could wander widely and zerocrossings would be very rare, suggesting that in this case the equilibrium and/or co-movement concepts have no practical implications. There may be r cointegration vectors, with r < n. When 1 < r < n, the cointegrating vectors may be denoted p1? ..., pr and may be gathered together into the nxr matrix B = (p b ..., pr). By construction, the rank of B is r, which is termed the cointegrating rank of Z,.
214
Chapter 6
It is apparent that in the case of three or more variables it is possible to have a subset of the variables which are integrated at a higher order than remaining variables and still have a valid cointegrating relationship if one linear combination of all variables in the subset is integrated at the same order as the remaining variables. There are many test procedures for cointegration. The computationally simple one is the Engle-Granger two-step estimator. In the first step, the parameters of the cointegrating vector are estimated by running the static regression in the levels of the variables. In the second step, these are used in the error correction form. Both steps require only OLS. The central feature is testing whether the error series ut of the static regression is 7(0) based on the Dickey-Fuller unit root test procedure but using different critical values (Engle & Granger 1987). The results may be shown to be consistent for all the parameters. However, the Engle-Granger approach suffers from several problems. First, inference about the p vector depends upon nuisance parameters, and large finite-sample biases can result when p is estimated by a static regression (Banerjee et al. 1986). Second, unit-root tests applied directly to ut usually lack power because the approach ignores the dynamics of the system (Kremers et al. 1992). Third, the number of cointegrating vectors is often of interest, and many hypotheses of interest relate to the complete conditional model specification and concern speeds of adjustment and the constancy of p over time, but the Engle-Granger two-step approach lacks means to deal with these issues. Finally, the choice of normalization in regression affects the finite-sample properties of the Engle-Granger procedure. Johansen (1988) and Johansen & Juselius (1990) develop a maximumlikelihood-based testing procedure for estimating and testing multiple cointegrating vectors based on the dynamics of the system. The Johansen procedure shows significant advantages over the Engle-Granger technique in terms of all issues mentioned in last paragraph, and therefore, has attracted the closest attention of empirical researchers. The Johansen procedure starts from the ^-dimensional vector autoregressive process Zt as defined by the model (6.3). The error correction form of model (6.3) is model (6.4). The model (6.4) is expressed as a traditional first difference VAR model except for the term n Z M . The main purpose of the Johansen procedure is to investigate whether the coefficient matrix II contains information about long-run relationships between the variables in the data vector. For r = 0, 1, . . . , « , the hypothesis of at most r cointegrating vectors is defined as the reduced rank condition: # r : I I = <xP'
(A2.14)
Estimating Investment Functions Based on Cointegration
215
where a and p are n x r matrices. Thus, (a) Hn specifies rank (II) = n, i.e. the matrix II has full rank, indicating that the vector process Z, is stationary; (b) Ho specifies rank (II) = 0, i.e. the matrix II is the null matrix, model (6.4) corresponds to a traditional differenced vector time series model, and there are n unit roots in | II(z ~l)\ = 0; and (c) 0 < rank (II) = r < n indicates that there are r cointegrating vectors and n-r unit roots in | II(z~l)\ = 0. P' is the matrix of cointegrating vectors, and a is the matrix of 'weighting elements'. Each 1 x n row pi of p1 is an individual cointegrating vector. Each 1 x r row ay of a is the set of weights for the r cointegrating terms appearing in theyth equation of model (6.4). The rank r is exactly the number of cointegrating vectors in the system. While a and p themselves are not unique, p uniquely defines the cointegration space, and suitable normalization for a and p are available. The essence of aP'ZM is that it contains all the long-run (levels) information on the process for Z,. The only other terms in model (6.4) are current and lagged AZ,. The vector P'ZM measures the extent to which observed data deviate from the long-run stationary relation(s) among the variables in the system. The statistical analysis of model (6.4) under the restriction of reduced rank of the matrix II = aP' can be performed by reduced rank regression (Johansen 1988). The variables AZ, and Z,_i are regressed on the lagged values AZ,_i, AZ,_2, ..., AZt-k+i and 1 to form residuals /?0/ and Rlh and residual product moment matrices:
The cointegrating relationships are then estimated as the eigenvectors corresponding to the r largest eigenvalues of the equation: lA.Sn-S'io SoUoil = O
(A2.1S)
The likelihood ratio test statistic of the hypothesis Hr in Hn is given by the so-called trace statistic: n
Qr = -T Y,
ln
H ~ h ) S i v e n t h e eigenvalues ij>.-.>in.
(A2.16)
i=r+l
Similarly the likelihood ratio test statistic for testing Hr in Hr+\ is given by so-called 1^^ statistic: ir+l)
(A2.17)
216
Chapter 6
Under the assumption that the number of cointegrating vectors is r and that the coefficient a'ljx * 0, such that there is a linear trend in the data, the limit distribution of both Qr and Qr/r+u which only depends on the degree of freedom n — r, is nonstandard and tabulated by simulation in Johansen & Juselius (1990: Table Al) and in Osterwald-Lenum (1992: Table 1). The hypothesis H*r that the trend is absent (a'±(i = 0) can be analysed by another reduced rank regression, the relevant trace and / l ^ statistics have the same forms as in equations (A2.16) and (A2.17), and their limit distribution is given by Table A3 in Johansen & Juselius (1990) and by Table 1* in Osterwald-Lenum (1992). The idea of the error correction model is relatively simple, i.e. that a proportion of the disequilibrium from one period is corrected in the next period. Error correction mechanisms can be derived as optimal behaviour with some kinds of adjustment costs or incomplete information. Engle & Granger (1987) defined a general error correction representation for a vector system as follows: A vector time series Z, has an error correction representation if it can be expressed as A(B)(l-B)Zt = -awM + 8,
(A2.18)
where 8/ is a stationary multivariate disturbance, with A(0) = /, ,4(1) has all elements finite, B is the backshift operator (BZt = ZM), utA = $Zt_x presents the cointegrating error, and a * 0. It is clear that by rearranging terms, any set of lags of the u can be written in this form, therefore equation (A2.18) permits any type of gradual adjustment toward a new equilibrium. As has been proven in Engle & Granger (1987) as the Granger Representation Theorem and again in Engle & Yoo (1991) using polynomial matrix functions, cointegration implies that the system follows an error correction representation, and conversely an error correction system has cointegrated variables. The error correction model provides a way of modelling, simultaneously, the dynamics both of short-run (changes) and long-run (levels) adjustment processes. Particularly, the idea of incorporating the dynamic adjustment to steady-state targets in the form of error correction seems to have introduced a quite useful approach to modelling dynamic adjustment behaviour of economic agents.
Estimating Investment Functions Based on Cointegration A3
217
Exogeneity
Concepts and Structure A clear understanding of exogeneity is critical for analysing the implications of cointegration for statistical inference, policy analysis and forecasting. Whether a variable is exogenous depends upon whether the variable can be treated as 'given' without losing information for the purpose concerned. The distinct purposes of statistical inference, policy analysis and forecasting define the concepts of weak, super and strong exogeneity. Ericsson (1992) presents a detailed and clear overview on these concepts. The following explanations are mainly based on Engle, Hendry & Richard (1983) and Ericsson (1992). Weak exogeneity is one of the essential concepts in this research, and provides the basis for efficient inference in our conditional model. For simplicity I would start from a bivariate normal case. Consider two variables, yt and xh which are jointly normally distributed and serially independent:
zMyt\~
IN (ii, Q)
(A3.1)
in which ' - IN(\i, £2)' denotes 'is distributed independently and normally, with mean vector \i and covariance matrix Q\ Without loss of generality, equation (A3.1) can be factored into the conditional density of yt given xt and the marginal density of xt ,3 as follows:
Fz(zt; v) = Fy\x(yt I xt; W • Fx(xt; x2)
(A3.2)
where F?(-) represents the density function for variable '?'. Thus, Fz(Zt; \j/) is the joint density of Z{, Fy\x(yt\ xt; Xx) is the conditional density ofyt given xh and Fx(xt; X2) is the marginal density of xt. The parameter vector \|/ is the full set of parameters in the joint process whereas X{ and X2 are the parameters of the conditional and marginal processes, and their respective parameter spaces are *F, A h and A2. Defining X as (V, V)' and denoting its parameter space as A, then there is a one-to-one function g(-) such that X = g(\|/). It should be noted that though this factorization (A3.2) is without loss of generality, analysing the conditional density Fy\x(yt\xt; Xi) while ignoring the corresponding marginal density Fx(xt; X2) is with the loss of generality, and in general implies a loss of information about the conditional process being modelled, unless xt is weakly exogenous over the sample period for the parameters of interest.
218
Chapter 6
Definition: The variable xt is weakly exogenous over the sample period for the parameters of interest 0 if and only if there exists a reparameterization of \|/ as X, with X = (V> V)\ such that (a) 8 is a function of A! alone, and (b) the factorization in equation (A3.2) operates a sequential cut, that is, (Xu A^) belong to Ai x A2, the product of their individual parameter spaces. (Engle, Hendry & Richard 1983: 282) That (Xh X2) belong to A! x A2 is called 'variation free'. It means that the parameter space Ai is not a function of the parameter X2, and the parameter space A2 is not a function of the parameter Xx. In other words, knowledge about the value of one parameter provides no information on the other parameter's range of potential values. If weak exogeneity holds, then efficient estimation and testing can be conducted by analysing only the conditional model (e.g. equation (6.5) in our case), ignoring the information of the marginal process (e.g. equation (6.6) in our case). Strong exogeneity is the conjunction of weak exogeneity and Granger noncausality, and it insures valid conditional forecasting. Consider a first-order vector autoregressive as a simple situation extended from equation (A3.1) in following form: Z/ = n Z / - i + e , ,
s,~/W(0,Q)
(A3.3)
Model (A3.3) carl be factored into the conditional model of yt given xt and the marginal model of xt as follows:
yt = box(
+ V M
+btyt_x +vlt , vl(
xt = T I 2 2 X M +TI 2 1 ^ M + e 2/ ,
~IN(0,G2)
e 2/ ~/AT(0,CD 2 2 )
(A3.4a) (A3.4b)
where Ti/, presents the (/,y)th element of n, and (b0, bh b2) are derived from II and £1 with b0 = co12/co22, bx = nn - (COI2/CO22)TC22, and b2 = nn - (COI2/CO22)TC21 (see, Engle, Hendry & Richard 1983: 297). Valid prediction of yt from its conditional model (A3.4a) requires more than weak exogeneity. With weak exogeneity alone, >V-i influences xt if n2X * 0 in the marginal model (A3.4b). In this case xt in the conditional model (A3.4a) cannot be treated as 'given' for prediction of yt. The additional restriction required is that 7i21 = 0, or in general that y does not Granger-cause x. Weak exogeneity plus Granger noncausality generates strong exogeneity. Concretely speaking, strong exogeneity permits valid multi-step ahead prediction of y from (A3.4a), conditional on predictions of x generated from (A3.4b) with n2X = 0, where the predictions of x depends upon their own lags only.
Estimating Investment Functions Based on Cointegration
219
Super exogeneity is the conjunction of weak exogeneity and 'invariance', and insures valid policy simulations. The concept of invariance can be introduced as follows. The reduced form (A3.3) may be empirically nonconstant because of changes in policy rules, unexpected shocks, innovations, etc. The factorization (A3.2) may aim to isolate those nonconstancies into the sub-set of parameters, X2, leaving the parameters of the conditional model Xx invariant to the changes that have occurred. Thus we can define the parameter Xx as invariant to a class of interventions to the marginal process of xt if Xi is not a function of X2 for the class of interventions. Policy analysis (or counterfactual analysis) often involves changing the marginal process of xt. Valid analysis of a conditional model under such changes requires that the parameters X\ be invariant to those changes (or interventions). The relevant concept is super exogeneity. Super exogeneity means that xt is weakly exogenous for the parameters of interest 0, and X\ is invariant to the class of interventions to X2 under consideration. It should be understood that when a variable is super exogenous with respect to a specific class of interventions, the variable need not to be super exogenous with respect to those interventions outside this class, although it may be.
Testing and Inference Following the definition of weak exogeneity, if we are interested in the longrun cointegration parameters given by p in model (6.4), it can be easily seen, based on models (6.5) and (6.6), that the weak exogeneity of Xt with respect to p is equivalent to the condition that ax=0.Jn this case p and the remaining adjustment coefficients Vy enter only in the conditional model (6.5), and the properties of the Gaussian distribution show that the parameters in the models (6.5) and (6.6) are variation free (see Johansen & Juselius 1990; Johansen 1992a, 1992b; Boswijk 1990). Thus the hypothesis of weak exogeneity of Xt for p in n = aP' can be formulated as: H: ax = 0 This hypothesis is a linear restriction on a and is discussed in Johansen & Juselius (1990), in which it is shown that under the hypothesis H the maximum likelihood estimation of the parameters could be performed by reduced rank regression and that the test ofHinHr consists of comparing the eigenvalues calculated without the restriction to those with the restriction (cf. equation A2.16). The test statistic is: QWH=T
220
Chapter 6
It is asymptotically distributed as y?(rnx), where «;r denotes the dimensions of Xt. Urbain (1992) proves that the H above is also the hypothesis of weak exogeneity of Xt for all of the long-run and short-run coefficients in the conditional equation (6.5) as parameters of interest. Thus equation (A3.5) can also serve this extended purpose. In addition, the weak exogeneity for all coefficients in the conditional equation can be tested via the tests of super exogeneity of Engle & Hendry (1993). Two common tests for super exogeneity are as follows. (1) If the conditional equation has constant parameters, but the marginal models have nonconstant parameters, then the conditional equation parameters could not depend upon the marginal model parameters. This is the test of Hendry (1988). (2) If the marginal processes are constant, we can use Wu-Hausman tests for independence between the conditioning variables and the residuals; that is, to test the significance of the residuals from the marginal models, or reduced forms, in the conditional equation. If the marginal processes have non-constancy we can further develop the marginal model for Xh by adding dummy and/or other variables, until it is empirically constant. Then we can test for the significance of those dummy and/or other variables when they are added to the conditional model. Their insignificance in the conditional model demonstrates invariance of the conditional model's parameters to the changes in the marginal processes; thus we test both weak and super exogeneity (Engle & Hendry 1993). It is clear that tests of parameter constancy are central to the tests of super exogeneity.
7
Conclusions
7.1 Introduction This closing chapter presents the general conclusions and explores further the theoretical and policy implications of this research. The theoretical implication is related to the recognition that we have reached a stage in which we are able to advance the standard investment cycle theory of a socialist economy by integrating both the Hungarian School's standard investment cycle theory and Kalecki's distributive barrier theory into a new framework of growth cycle. The policy implication involves such issues as the investment inefficiency of the state sector, capital accumulation strategy, structural adjustment and reform of the state investment system. Section 7.2 presents the theoretical contribution of the research. Section 7.3 summarizes the stylized facts about aggregate investment behaviour in China both from a perspective that features the sources of investment hunger at the central, local and enterprise level, and from a perspective that emphasizes the link between key bottleneck constraints and retrenchment campaigns. Section 7.4 discusses the inefficiency caused by investment hunger and bureaucratic coordination, and examines how China's state sector has suffered from investment inefficiency induced by the rigid state investment system. Section 7.5 deals with the difficulties and possible options for reforming the state investment system with reference to the renewed reform package of late 1993 and the evolution of the reform in 1994-98. Section 7.6 reviews what can be perceived as the major limitations of the research. Finally, section 7.7 makes some further comments on the critical issues highlighted by the research.
221
222
Chapter 7
7.2 Major Theoretical Contributions of the Research The first major theoretical contribution of the research is that it supplies a new framework for conceptualizing the aggregate investment behaviour and modelling investment cycles in a socialist and/or developing economy. The new framework is established by integrating the existing theories of investment cycle and bottleneck constrained growth in a socialist and/or developing economy with the cointegration and error correction mechanism. Three advancements brought about by this integration are worth mentioning. First, although the concrete investment adjustment processes have been changed following the changing patterns in power distribution among different political apparatuses and in particular, the significant changes which have taken place in the reforming years (cf. section 6.6), the investment hunger and its resultant tensions between investment expansion and the supply and distributive barriers have persisted. This fundamental feature has been highlighted by the both long-run investment level function and the short-run investment growth function. As long as the annual real investment level has unceasingly moved along the supply possibilities frontier of bottleneck sectors, a relatively high level of real investment has been constantly maintained. Secondly, this integration highlights the importance of coordination mechanisms that prevail among economic institutions and their agents. There have been different investment approval and brake apparatuses operating at different levels in a complex, interactive fashion, and the operation and interaction have been altered by dramatic political and economic changes (sections 1.3 and 1.4 and Chapter 3). However, one fundamental feature of these operations and interaction, interest conflicts and accompanying bureaucratic coordination, has persisted. All bureaucrats are rational economic agents, and they have tried to maximize the interests of their own institutions subject to certain internal and external supply and distributive constraints. They cannot be considered as homogeneous actors. If business cycles in the market economy can be partly attributed to market co-ordination failure according to Keynesians and new Keynesians (Fischer 1988), investment cycles in China can be also considered as a type of bureaucratic coordination failure.
Conclusions
223
Thirdly, the traditional univariate 'norm' and 'negative feedback' concepts can be updated, respectively, by introducing the notions of the equilibrium co-movement between real investment level and the supply and distributive barriers to investment expansion, and of the disequilibrium adjustment behaviour of economic agents toward the equilibrium attractor. The comparative advantage of the attractor notion is that a comovement attractor is directly linked with the real tension between system-generated investment ambition and the supply and distributional barriers to the ambition, and thus it can be modelled. The advantage of the error correction notion is that it represents the disequilibrium adjustment behaviour of the economic agents toward an equilibrium comovement path, which is economically more active and intuitive in the context of there being many non-homogeneous bureaucrats. As mentioned before, the study of industrial accumulation mechanisms and investment cycles has been a recurring and popular subject in the literature on the dynamics of the socialist and/or developing economy. In practice, the determination of a 'rational scale of investment' has also been a critical question, which has perplexed economic authorities and scholars in economies like China's. Following the typical textbook framework, scholars might need to focus on building up demand and supply functions to address the investment issue. They may ignore the question of how to bridge the gap between theoretical (usually unobservable) variables and the observed (non-experimental) data (cf. sections 1.5). With reference to the literature dealing with investment cycles in the former Soviet Union, Eastern Europe and China, it seems that no one has tried to set up a demand/supply function. Instead, most scholars have concentrated on building up and testing planners' response functions to one or several representative shortage signals. This departure from the standard supply and demand functions framework comes out of learning from data and direct economic experiences. In this connection, the framework suggested by this book can be also justified by learning from data and direct experiences. The variable system is selected based on a political economy analysis of China's state investment system and on comprehensive data analyses. The probabilistic structure exploration of the selected time series is established through modelling an unrestricted vector autoregression system, without any pre-imposed equilibrium conditions or behavioural assumption. The empirical findings can also be interpreted in the terminology of probabilistic structure exploration. First, there statistically exists a long-
224
Chapter 7
run stable comovement among investment outlays, grain output, effective energy supply and the price level of investment. Within these four variable information systems investment is endogenous whereas all the others are weakly exogenous. The cointegration relation is homogenous of degree one in price level. The cointegration relation thus represents a typical equilibrium path shaping real investment level. Secondly, the conditional error correction model developed from the equilibrium level equation is a well-defined and well-fitted investment growth rate adjustment equation. Much of the cyclical pattern of investment growth rate can be explained by the error correction towards the co-movement path and by the relevant change rates of energy supply and agricultural output. Thirdly, the standard one-step-ahead Chow test does not detect any statistically significant structural break induced by policy changes or reform within the sample period. This indicates that both investment level and growth rate equations have parameter constancy over the sample period. Besides the contributions presented above, the research provides an alternative way to investigate the interaction between agricultural fluctuation and investment adjustment without econometric modelling (cf. section 2.4.1 and Chapter 4). It shows that in China's case, common linear regression based on the notion of the 'negative feedback' between variations of agricultural harvests, procurement and variations in investment outlays may be unable to find the expected 'negative feedback'. However, once a comprehensive indicator system is selected based on an examination of the main proportional relations in the economy to which the decision-makers are most sensitive, and once the major input-output relations in agriculture and the distributive relations in the uses of the national income are figured out, a picture of two-way interaction (not only 'negative feedback') emerges. The two-way interaction indicates that the interaction between investment expansion and agricultural development cannot be simplified as substitution alone. Investment expansion may be at the expense of agricultural development, but the good harvests of agriculture have stimulated and supported investment expansion. This appears to be true for both the short and long run, though it is more likely that complementarity may be dominant in the long run (cf. sections 1.2 and 4.1).
Conclusions
7.3
225
Aggregate Investment Behaviour in China: Stylized Facts
This section intends to summarize the stylized facts on aggregate investment behaviour in China, which have been analysed in detail in Chapters 2 to 5, and represent the main contribution of this research to the literature on the issue of investment behaviour analysis in a socialist developing economy like China. The facts are grouped under two headings. The first is that 'system-generated insatiable investment demand exists at all levels', and the second reviews supply and distributive barriers to investment expansion and consequent retrenchment campaigns.
7.3.1 System-generated insatiable investment demand exists at all levels The process of investment decision-making in the state sector is in fact a distribution process of rights to possess and use certain scarce state assets such as budget funds, bank loans, land, quotas of power, oil and other key materials in shortage. The primary intention of ministries, local governments and firms is to try to obtain as much investment and corresponding other properties from the bureaucratic distribution process as possible so that they can benefit in the future and justify their existence and power base. A simple but insightful example is that if a state enterprise is assigned a building or a piece of land in the commercial centre of a city by negotiation or just by chance, it can obtain more benefit and its employees can get more bonuses just from renting the building or land. Such intention is the essential source of investment hunger at each level. It is based on these motivations that the relevant decision-makers accustomed to pay only secondary attention to the future profitability of any new investment project. It is fine if the project turns out to be profitable, but if not, the loss will be borne by the state anyway. Such investment expansion drive, combined with the soft budget constraint, will always lead to investment hunger and inefficiency. At central government level, the most immediate reason for expansion drive and investment hunger is the internal and external pressure to provide evidence of socialist superiority, to catch up with the industrialized powers as fast as possible, and to create a modernized army force for national security. In addition to these general motivations, one more China-specific and increasingly important stimulation comes
226
Chapter 7
from the divergent priorities of the central and local governments. Such divergence, generally linked to the great regional differences and diversity, has increased considerably during the reform period, following the general trend of decentralization. Before the reform, the central government used to depend on political/economic campaigns and its ability to control the state budget to implement and finance its investment ambitions. During the reform period, besides using various direct and indirect administrative measures to bring local investment decisions in line with national priorities, and with its greater access to information and a wider range of tools at its disposal, the central government created and has since expanded a special programme of national 'priority' projects mainly related to energy, transport and communication. These projects have been planned with a 'rational time-table and guaranteed supply of materials', and thus have first claim on materials still under state control. In order to ensure the implementation of these priority projects, the central government employs such strategies as concentrating its own resources on these projects, harnessing the financial resources of the state bank to them, and drawing on local financial resources through a version of 'matching funds' for them (cf. section 3.4.2). This shows how the central government has played a dual role as both the highest administrator of the economy and as an active investor with its own interests. When employing ad hoc administrative measures to suppress the overall over-investment, the central government has forced local authorities to cut their investment outlays but allowed its own plans to go ahead. At local level, besides expansion motivations such as providing evidence of socialist superiority, catching up with more developed regions and expanding the power, prestige and material benefits of local bureaucrats, local governments have to bear responsibility for their own regional development, employment, stability and welfare. In fact, local governments have been faced with much stronger pressure to promote development than the central government. For instance, they have to accept liability for social disturbances arising from unemployment, housing shortage, infrastructure deficiency and growing dissatisfaction in consumer sector. They must struggle to meet expenditure obligations imposed by central policies and commands, and they have to promote a more rapid rate of local economic growth so as to raise their own negotiating position within the bureaucracy. In order to promote local economic development and extend employment, local governments
Conclusions
227
have been eager to set up new factories under their jurisdiction and to expand their own enterprises, without regard for duplication of construction or economy of scale at the national level. While they rely on their own state enterprises (i.e. local SOEs) for providing a critical social safety net which will provide grants for housing, schooling, health care and retirement benefits for the urban population, local governments are willing to prevent their enterprises from bankruptcy even in the event that chronic losses are occurring. Faced with intense expenditure pressure and a tax system that depends on industry for the generation of over two-thirds of total revenues (both in- and extra-budget), local governments seem to have little choice but to engage in industrial expansion at the cost of agricultural and infrastructure development. At enterprise level, state-owned enterprises (SOEs) have received various forms of external assistance from different government bodies, both before and after reform. They can request, through bargaining and lobbying, soft appropriations and subsidies from national or local budgets, ad hoc exemptions and postponements of tax payments from taxation authorities and soft credit from state banks and their local branches. They can also enjoy soft administrative pricing. It is widely acknowledged that in the pre-reform period the budget constraint of SOEs had been quite soft. Surprisingly, the empirical evidence presented in Table 3.8 shows that the core of SOEs, 'the industrial SOEs with independent accounting systems', had also made increasing losses in the period of industrial reform. The ratios of their losses to their pre-tax profits increased dramatically, from 2.4 per cent in 1985 to around 20 per cent in the period of 1991-96. Such an overall level of loss is not possible without 'soft' subsidies from the government budget and 'soft' credits from the state banks, in view of the fact that the share of the planned component has rapidly declined. It is the lasting soft budget constraint, combined with the increasing autonomy, that has stimulated SOEs' expansion drive and investment hunger, and has encouraged SOEs to take too much risk in their investment processes during the transition period (cf. Chapter 3, Zou & Sun 1996). In addition to the soft budget constraint, SOE expansion drive and investment hunger can also be explained partly by the managers' bureaucratic motives such as policy and moral conviction, identification with the job, expanding power, prestige and material benefits. However, the soft budget constraint is the most essential and system-specific source of SOEs' investment hunger.
228
Chapter 7
Official documents indicate the existence of a strict project approval system and a sound distribution of project approval limits among local governments at different levels. In practice, however, the situation is very different. SOEs used to work together with their immediate supervisory agency on many issues and could therefore easily find ways to obtain approval or even to bypass the limitations. Meanwhile, in order to promote their own regional growth and create employment, the local governments at lower levels are willing to collude with their SOEs and local state bank branches to circumvent project approval requirements at higher levels. The most popular collusion in the investment planning process is separating projects into smaller components with underestimated costs so as to evade or simplify project review and approval procedures. The superficially strict credit plan system has functioned in a similar way. Under the institutional arrangement of 'dual subordination' (shuangchong lingdao) along the lines of both vertical and regional accountability, the local branches of state banks have been defenceless against pressure from local governments. Additionally, the local branches have had to depend on their local governments for basic welfare such as housing, children's schooling and employment for bank employees. Therefore the local branches have strong incentives to follow the local credit plan and to collude with local authorities to extend credit, formally or informally. Succinctly, the expansion drive and investment hunger are endogenously generated within the political/economic system, and exist ubiquitously at all levels. In the meantime, there is no genuine fear of a financial failure nor any other internal restraint to resist the expansion drive. The credit system is far from independent, and financial obligation is not binding unless some strict ad hoc administrative measures are imposed by political/economic campaign or through the hierarchy. The first significant outcome of such an institutional set-up and incentive mechanism is that real investment is bound to move along the supply possibility frontier of the bottleneck sectors and the distributive barrier between investment expansion and necessary agricultural growth, as discussed in detail in previous chapters. The second important consequence is widespread investment inefficiency, which will be discussed in section 7.4.
Conclusions
229
7.3.2 Supply and distributive barriers to investment expansion and retrenchment campaigns As summarized above, at all of the central, local and enterprise levels, the political/economic system constantly produces pressure and incentives to maximize the investment possibility, but it cannot initiate internal and/or self-imposed restraint to resist the expansion drive. As a result, real investment is bound to expand until it comes up against the shortage and inflationary barriers of the bottleneck sectors. Meanwhile, the intuitive data explorations in Chapters 4 and 5 and the cointegration analysis in Chapter 6 indicate that the grain supply as the necessary consumer good and effective energy supply as the representative producer good have repeatedly checked investment booming and industrial expansion. The decision-makers at different levels understand and are constrained by the shortage and inflationary barriers of basic consumer goods and producer goods. A bad project may be able to ensure the funds it needs but may be screened out at the very beginning because of the shortage. The interactive relationship between agricultural fluctuation and investment adjustment has attracted major attention from Chinese leadership, scholars and the public. As shown in Chapters 4 and 6, swings in agricultural production and marketing have been followed by dramatic swings in industrial accumulation and investment. The increase of agricultural surpluses has stimulated and supported rapid investment expansion, but will be followed by major shifts in the distribution of national income and significant reductions in incentives (e.g. material rewards) and prices (or the margin of raising prices) for agriculture. On the other hand, agricultural shortfalls have led to a passive reversion to increases in incentives and prices, and to cutbacks in industrial accumulation and investment. The basic force which determines these interactions is the conflict between the active industrialization drive and the reactive agricultural policy at different levels, which has been characterized by China's basic capital accumulation mechanism. This accumulation mechanism had successfully served the heavy-industry oriented development strategy in the pre-reform period, by depending on the distorted macroeconomic environment, planned allocation system, and induced institutional arrangements to guarantee the high profits of industry and low consumption of both workers and peasants. During the reform period,
230
Chapter 7
however, the essential characteristic of the accumulation mechanism extracting surplus from agriculture by coercion - seems to have persisted. Although agriculture's terms of trade significantly improved between 1979 and 1984, the coercive procurement system in rural areas and the food rationing system in urban areas have continued. The government tried to abolish low-price rationing in 1993, but revived it in most cities in 1994 and 1995. Moreover, the urban and industrial biases of local governments have been reinforced by the intense expenditure pressure they face and by the tax system, which depends on industry for the generation of over two thirds of total revenues (cf. section 3.7). As a consequence, the agricultural policy has been, in practice rather than according to official documents, guided by budgetary considerations and local industrialization drives. The combination of budgetary and inflationary pressure encourages the governments to lower agriculture's terms of trade and to reduce other material rewards whenever possible, and to increase procurement prices and other incentive rewards only when it is absolutely necessary. Although signs appeared during the 'soft landing' period of 1994-97, that the Chinese government started to give agriculture priority through the enforcement of 'provincial governor's "grain bag" responsibility system' and of protecting procurement prices, the duration of the policy package is not clear. Therefore, we may conclude that the agricultural problem in China is not only a problem of theoretical understanding of agriculture's extreme importance in economic development, but also, and perhaps more significantly, a practical issue which is related to state investment and accumulation systems, and to the political negotiation position of the peasantry. The energy shortage in China is, to some extent, similar to labour shortages in former Eastern European socialist economies. In China the widespread shortage of primary energy has been chronic and cannot be anticipated to ease in the near future. It is widely acknowledged that because of power shortages over 20 per cent of power-driven equipment remains idle. Meanwhile, the coal shortage forces power stations to stop generators, and the transport constraint means that coal shortage in industrial centres and coal stocks above storage capacity in the mining bases have usually co-existed. During the reform period, the government gave first priority to coal transport railway construction and improvement. However, along with the continuous and high-speed
Conclusions
231
economic growth, the incremental gaps between coal demand over coal freight volume and between coal output and freight volume continued to grow. This indicates that the growth rate of demand for coal has been greater than the growth of effective coal supply during rapid economic growth. As a result, the tension between demand for and effective supply of basic energy (mainly coal) has remained and even been aggravated during the reform period. The econometric evidence given in Chapters 5 and 6 (based on the Granger causality test and cointegration analysis) shows that there is an evident two-way dependence between real investment and effective energy supply in terms of both level and growth rate, and that within the information system consisting of nominal investment, deflator of investment, grain output and effective energy supply, energy is weakly but not strongly exogenous. The weak exogeneity of effective energy supply here can be attributed to such factors as the location of energy bases, distribution patterns of energy production and consumption, the geographical difficulty of constructing railways and roads, as well as other factors which are beyond the policy trade-off of how to allocate investment funds. Due to the persistent energy shortage, China's macroeconomy has progressed steadily within a normal state of shortage. Usually, even the state annual plan would leave a significant 'plan gap' for energy distribution, which was normally equal to 5 per cent of the realized supply (section 5.3). The government used to assume that the gap would not damage the steady progress of the macroeconomy if the planned economic growth targets were not exceeded. Indeed, in the interval around trough point of each cycle, these 'plan gaps' did not undermine the progressing and shortage was even relieved. However, in the interval around the peak point of each cycle, the realized growth rates of output and investment have certainly exceeded the planned targets by a significant margin. Meanwhile, neither the effective energy supply nor energy production are able to grow so dramatically. As a consequence, energy stocks in consuming areas are bound to decrease considerably. A transportation crisis emerges and becomes increasingly severe. The supply of energy and raw materials for the implementation of ongoing investment projects and even for maintaining simple reproduction becomes irregular and uncertain. Inflation pressure re-emerges and increases. These danger signals in turn cause social disturbances and force the central government to initiate a retrenchment campaign and to reinforce this by ad hoc administrative measures, such as sending cen-
232
Chapter 7
tral delegations to each province and using the Party hierarchy to review and supervise the implementation of these retrenchment policies. The first theme of such retrenchment campaigns has been to control the scale of fixed investment (cf. Chapter 3, section 5.3).
7.4 Inefficiency as a Consequence of Investment Hunger and Bureaucratic Coordination Investment inefficiency becomes fundamentally important when the discussion moves from theoretical modelling to policy concern. China's state sector has suffered from investment inefficiency which, as has been mentioned, is a consequence of its state investment system. This system is characterized by complex bureaucratic co-ordination at and across different levels. When dealing with investment inefficiency in the state sector, the most famous and impressive evidence is the lasting and large-scale duplication of construction (chongfu jianshe) at national level. The situation has deteriorated since the fiscal reform and relevant decentralization in the early 1980s (Wong 1992, World Bank 1994). There are numerous examples of this. By the end of 1990 China had built up 167 production lines for colour television sets with an annual production capacity of 20 million sets. The actual output was only 10 million, which means that half of the production capacity was idle. In 1993, China had 126 automobile factories and 5,000 re-equipping automobile factories theoretically capable of producing one million automobiles annually. However, most of these factories had no economy of scale by any standard, and the average utilization ratio of production capacity was again about 50 per cent. Similar situations exist in almost all sectors. For instance, in the raw material sector about 40 per cent of soda ash production capacity was idle in 1992. In the textile sector, the wool spinning industry suffered from a remarkable cycle of redundant construction from 1980 to 1990. The most impressive cycles of such construction seem to appear in durable consumer goods production such as wrist watches, radio sets, sewing machines (in the early 1980s), TV sets, washing machines and tape recorders (in the late 1980s). While most such construction projects were completed, over 50-90 per cent of production capacity came to a standstill. The fundamental mechanism behind the recurring repetitionary construction cycles is that each local government is competing with others
Conclusions
233
to obtain more investment from the centre and/or state bank to establish more seemingly profitable construction projects under its jurisdiction when they perceive the possibility of making quick profits. However, by the time the projects are completed, it is often discovered that there are already too many such projects at the national level, and that there is not enough market for the products. The opportunity cost of such duplicated construction is also high. Initiation of too many projects is certainly at the expense of technical updating of existing assets, inducing greater prominence of gestation lags between the initiation and completion of investment projects. An international comparison of industrial concentration by major industrial sectors in China's state sector, USA and Japan in 1985 (Table 7.1) indicates that many of China's SOEs remain small in scale compared with their counterparts in the major industrialized countries. It is well known that the concentration of production is much higher in former socialist economies of Eastern Europe than in capitalist ones (Kornai 1992: Chapter 17), Table 7.1 thus indicates that China's economy, in terms of economy of scale, is remarkably different from its socialist counterparts in Central and Eastern Europe, where under communism large-scale state monopolies and central direct control were generally favoured. The existence of a large number of small, duplicated production facilities with few plants able to reap full economies of scale may represent a unique feature of China's state investment system. Table 7.1
Industrial concentration by major industrial sectors in China, USA and Japan, 1985 (percentages) USA
China Sector
Largest firms
Iron & steel Motor vehicles Chemicals Engineering Textiles Note: Source:
38 7 7 6 2
Market shares 46.7 35.6 10.2
2.9 1.5
Largest firms
7 13 19 9 5
Japan
Market shares
Largest firms
83.7 94.6 84.9 58.4 40.7
Largest firms mean that they have the largest market shares. Lin, S. (1990:123).
20 10 15 10 10
Market shares 84.0 74.1 48.0 53.9 68.4
234
Chapter 7
At the aggregate level the investment inefficiency in the state sector is partly evident in the incremental capital-profit ratio. During 1985— 92, the net value of fixed assets of industrial SOEs with independent accounting systems increased from 398 billion yuan to 1,098 billion yuan, an increase of 700 billion yuan, while the realized pre-tax profits only increased 61 billion yuan from 133 billion yuan to 194 billion yuan. The incremental ratio of fixed assets to pre-tax profit is 11.5, indicating only a one yuan increase in pre-tax profits for every 11.5 yuan increase in fixed assets (net of depreciation). The ratio of pre-tax profits to total capital decreased from 23.8 per cent to 9.7 per cent, and the realized pre-tax profits per 100 yuan fixed assets went down to 12.4 yuan in 1992 from 22.4 yuan in 1985 {Yearbook 1993: 430, 437). As in the case of shortage and inflation pressures, the over-large scale of duplicated construction is also interpreted as a danger signal, because it not only indicates the deteriorating investment inefficiency but also the impending scale of over-commitment and wasting outputs. In official reports on the retrenchment campaign, the number of smallscale and duplicated investment projects (usually half-constructed) that have faced forced closure has been treated as an important measurement to evaluate retrenchment {A Collection of Documents on Fixed Investment and Construction). In this sense, the retrenchment campaigns do help improve investment efficiency through curbing further waste. However, given the basic investment system, the retrenchment campaigns are essentially reactive rather than active. A large number of these closed projects will sooner or later, be revived. During the reform period, alongside increasing decentralization, the problem of smallscale and duplicated construction seems to have been exacerbated by local governments and SOEs using their greater autonomy to build even more small-scale plants. This has led to yet more duplicated production facilities and thus to over-capacity in many industries {People's Daily, 26 April 1994; Yearbook, 1997: 454-5). The abovementioned facts indicate that a state investment system characterized by bureaucratic negotiation may create 'the wrong kind of wealth' as analysed in Naughton (1995a). That is, the bureaucratic investment system builds many factories and other facilities each of which, if taken in isolation, appears viable and productive in the context of the bureaucratic component of the economy, but turns out to have little value once its products have to face market demand and competition. In spite of the obvious improvement in the overall incen-
Conclusions
235
tive environment during the reform periods, China's rigid state investment system seems to continue to build and expand factories that ultimately prove to be unviable and have to be closed. This 'wrong kind of wealth' implies that a significant part of the impressive economic growth of the last decade is offset by ill-conceived and wasteful investment projects, and will not only turn out to be useless but will also impose conversion costs on the next stage of reform and development. In addition, the remarkable resistance of the state investment system to reform may create serious problems for the future macroeconomic stability and the reform of SOEs and the state banking system.
7.5 The Difficulties and Possible Selections of Reforming the State Investment System The economic reform of the last two decades has undoubtedly helped China achieve her remarkably high rates of economic growth, and shows ex post coherence to some extent although such coherence is not the result of a carefully planned reform strategy (Lin et al. 1994, 1995; Naughton 1995a & b). On the other hand, some profound shortcomings remain and a series of uncompleted essential tasks still confront the reformers. Among the most obvious shortcoming has been the continuing failure to develop institutions and impartial rules that apply to all economic organisations. One of the most essential tasks is to remove the resistance of the state investment system to reform. The difficulties of reforming the state investment system involve not only determining how to tackle the twin problems of central-local relations and relations among government, enterprises, and banks, but also how to deal with the soft budget constraint of the state sector, particularly with regard to state enterprises (SOEs). On the one hand, SOEs continue to have access to the bulk of capital under conditions in which their liability and ability to repay are unclear, and the efficiency of their capital utilization is obviously low. On the other hand, SOEs have born the responsibility of supporting an extensive work force (over 100 million) by keeping redundant workers on their payroll, and of providing a number of goods and services (including housing, health care, child care and schooling, etc.) with positive externalities to society. Therefore, the SOE sector should not be allowed to collapse, despite the fact that a significant number of them need to rely on soft subsidies and credits to remain in business. The social consequences of
236
Chapter 7
the their collapse would be too dire and the political and economic costs too great. How to harden SOEs' budget constraint while avoiding SOEs' collapse and/or large-scale unemployment will continue to be the most difficult theoretical and practical challenge that Chinese reformers must face in the transitional period (Gu 1999, Jefferson 1998, Perotti et al. 1998, Sun 1999, World Bank 1997). In terms of the central-local relations, China's reform has not yet found a way of maintaining central effective control over macroeconomic policy while also preserving the benefits of decentralized decision-making and greater sub-national autonomy. In particular, on the issues of profit-making economic management, China's central and local governments have no clear division of functions. During the reform period, through gradual decentralization, local governments have obtained increasing economic management power from the central authorities. At the same time, the central government has retained some control over industry, commerce, trade, prices, wages and credit allocation. As the central and local governments undertake the same functions, their differences have devolved mainly to ownership, production types, and quota allocation. For instance, the central government is responsible for managing large-scale coal enterprises at the central level, and local governments are responsible for those smaller-scale coal enterprises at local levels. In foreign trade, providing the foreign exchange quotas set by the centre are fulfilled, local governments can take measures to encourage export on their own. Import and export of some kinds of products are monopolized by the central agents, while other categories are open to both central and local foreign trade agents, and so on. Such division in turn determines the composition and scope of expenditure of the central and the local authorities, which is usually known as responsibility for fiscal expenditures. While local governments have certain administrative intervention powers in many economic issues and in their own enterprise businesses, they must be provided with certain sources of revenue as an incentive to use their power to good effect. In other words, if local governments exercise their power inappropriately, they are punished by a reduction of revenues. Similarly, the central government also holds concrete administrative intervention powers and needs its own independent sources of fiscal revenue.
Conclusions
237
Against this background we can understand why the arrangement and structure of the fiscal system has played a leading role in the distribution of administrative power among the governments at different levels, and why other measures of administrative decentralization are gradually built around it. This indicates that tackling central-local relations requires a broad effort to define systematically the respective economic roles of central and local governments at each level. Basically, this requires the development of national consensus on a delineation of the economic responsibilities of central and local governments. As long as the local governments are fearful that their discretionary authority may be revoked, and as long as there remains widespread opportunities for bargaining, it would be difficult to reverse the tendency of local governments to focus on quick revenues without being concerned about the negative externalities of their actions on the national economy, and/or to depend on ad hoc measures for retaining control over local resources (Lou 1992, World Bank 1994). The challenging issue with regards to the relations among government, enterprises and banks is how to cope with their collusion in investment and financing affairs (cf. Chapter 3), which also forms one of most important avenues for realizing the soft budget constraint enjoyed by industrial ministries, local governments, and their enterprises. The co-mingling of government and enterprise activities, and the pervading interference of governments at all levels in the activities of the state banks, have created incentives and opportunities for collusion. As a result, the leveraged financial position of SOEs is increasing at a good pace.2 The governments at different levels and their SOEs monopolize the benefits from soft borrowing while distributing the associated costs over the rest of the economy. For example, in 1995, the SOE sector consumed about 80 per cent of state band credit funds but created less than 45 per cent of China's GDP (Perotti et al. 1998). Soft lending to the SOE sector has resulted in a rapid accumulation of bad debts in the state bank system. It is estimated that China's four main state banks had bad debts equal to a crippling 22 per cent of their lending by mid1997 (The Economist, 13 September 1997: 24). Tackling this problem is as difficult as dealing with how to harden SOEs' soft budget constraint and transforming most of the state banks into pure commercial banks. To design a detailed reform schedule is beyond the scope of this research. Instead of dealing with specific reform programmes, in the sue-
238
Chapter 7
ceeding paragraphs we will follow a 'positive-predictive approach' (Kornai 1995) to examine the most relevant or feasible strategies for the future reforming of China's state investment system. The core of the strategies may be significantly different from the type of reform that has prevailed in Europe. The impact of the SOEs' soft budget constraint can be reduced, and a major part of the problem can be solved by gradual 'denationalization' in the future. Denationalization in a narrow sense represents a transitional process that mainly includes successful non-state enterprises taking over or merging with poorly performing SOEs. SOEs are converted into joint ventures with either domestic non-state or foreign enterprises. SOEs are reorganised into joint stock companies or transformed into joint stock cooperatives, and small-sized SOEs and/or some medium-sized SOEs are sold or leased. In a broad sense, denationalization means that the relative reduction of the state sector can also be achieved through the growth of the non-state sector, composed mainly of township and village enterprises (Sun 1989, Lin et al. 1995, Qian&Xu 1993a). In fact, denationalization in a broad sense is already practised and has shown a significant achievement in China since the early 1980s. In recent years, reorganization of large and medium SOEs into joint stock companies and transformation of small and medium SOEs into joint stock cooperatives has become a focus of SOEs' reform. Takeovers and mergers of SOEs by non-state enterprises have emerged as well (Parker & Pan 1996; Sun 1998b, 1999). Certain limitations exist and some difficult outstanding issues remain, which are mainly related to the SOEs' unrelieved social burden. This in turn encourages more far-reaching social security and housing reforms. As a result, the total output proportion of pure, state-owned industry decreased from 77.6 per cent in 1978 to 34.1 per cent in 1994 and further to 28.5 per cent in 1996 {Statistical Yearbook of China 1997: 413).3 In comparison with the high social and economic cost of massive and fast privatization in the former Soviet Union and Eastern European countries, denationalization may be an easier, less costly, and more suitable reform strategy in China. In addition, denationalization has led to a natural industrial structural adjustment in China, whereas the privatization process per se does not automatically bring about economic structural modification (Lavigne 1995, Sun 1998b).
Conclusions
239
A better and more feasible framework for re-configuring the relations between central and local authorities may be termed 'market reinforcing decentralization' (McKinnon 1993, Montinola et al. 1995, Qian & Roland 1999, Qian & Weigast 1997, World Bank 1994). Such a framework includes the following four essential characteristics: (1) Shared authority, which means that the locus of economic regulatory authority is shared by the different government levels so that no one level can monopolize the regulatory authority over the entire economy; in order to constrain central government from taking actions that may undermine market development this principle is necessary; (2) Local openness, that is, regional governments, are prevented from using their regulatory authority to build barriers against the development of interregional common markets; this principle is extremely important in terms of breaking down regional monopolies, reducing regional and rural/urban inequality, and promoting healthy interjurisdiction competition; (3) Hard budget constraint to all governments, which ensure that revenue sharing among governments at different levels is limited and borrowing by governments is constrained; this requirement is essential for rationalizing the state investment system because in most cases the recurring over-investment is local- and/or central-government-led (cf. Chapter 3); (4) Institutionalized scope and durability of authority and responsibility, which ensures that each level of government has a delineated and institutionalized scope of authority and responsibility so that it can act autonomously within its own well-defined domain of authority; and secondly, that the durability of authority allocation is also institutionalized so that the allocation cannot be altered by the central government either unilaterally or under pressure from local governments. Based on such an institutionalized framework, the economy is most likely to enjoy the benefits of decentralization while avoiding its most serious negative consequences such as dukedom economies, local-led investment hunger and the resultant macroeconomic imbalance. 'Market reinforcing decentralization' as a fundamental mechanism to achieve efficient institutional structures and to regulate fair competi-
240
Chapter 7
tiveness is bound to be a long historical process. However, its institutional formulation can be constitutionalized much earlier, based on the emerging development of national consensus on the delineation of the economic responsibilities of central and local governments. China started to commercialize the operations of the state bank in 1994, in part through the creation of 'policy banks'4 that are expected to take over government-directed lending for fixed investment and to help the central government finance its 'priority' expenditures without having funds diverted to unintended uses. Within the design and functioning of the policy bank system, however, several potential problems remain (World Bank 1994, 1997). Among these the most important is that the creation of the policy banks alone cannot make a clear break between policy and commercial lending of the state banks, and thus will not stop the deterioration of the loan portfolio of the state banks. It is impossible for the State Development Bank and Agriculture Development Bank to take over responsibility for all fixed investment loans in their respective sectors. A significant part of fixed investment for 'basic' and 'pillar industries' has still been and will continue to be financed by other state banks at preferential interest rates earmarked by the People's Bank of China (the central bank) specifically for these sectors. Moreover, all working capital loans, including those to loss-making SOEs, have been and will continue to be part of the portfolios of all the state banks. Therefore, more profound reforms need to be undertaken in this area. In particular, denationalization in both the narrow and broad senses is required. In the broad sense, denationalization of the banking system means encouraging the urban credit cooperatives to develop into non-state banks by alliance or merger, prompting the development of other non-state banks, and allowing domestic and foreign joint-ventured banks and/or foreign banks to undertake domestic financial business step by step. This would break the persistent monopoly of state banks and introduce essential competition into the financial area. Denationalization in the narrow sense calls for converting some state banks into joint ventures with either domestic or foreign financial institutions, and to reorganize parts of state banks into joint stock banks (Sun 1989). Such a denationalization of the banking system can in the short term effectively limit the impact of the state banks' soft budget constraint. In the medium term, it can introduce and reinforce fair competition among those different financial institutions characterized by multiple owner-
Conclusions
241
ship, and finally, it can generate an efficient and pluralized financial market.
7.6 Limitations of the Research Several limitations of the research can be identified, a majority of them attributable to the inherent data shortage. First of all, discussion of political campaigns is far from complete or sufficient. Discussion has involved the economic conditions for initiating, and the economic consequences of those political campaigns, but does not pay so much attention to the corresponding political reasons and processes. Although a good economic condition usually serves as a necessary premise for initiating a political campaign, each political campaign has its own political reasons and logic. Such political determinants in turn bring in exogenous consequences to the economic processes, which have, to a great extent, shaped the amplitude of investment cycles. A further exploitation in this direction would add new insights into the politics and political economy of investment cycles in China; however, this would be beyond the focus and scope of the research. Secondly, as mentioned earlier, during the pre-reform period overinvestment usually resulted in supply shortages of energy, raw materials and agricultural products through direct material conduction. In the reform period, over-investment led first to over-expanding of credit, followed by shortages in the planned component and inflation in the market component of the economy through both material and value conduction. This research mainly discusses the persistent forces in both periods which shape the investment cycles. It has paid relatively less attention to the particularity of the conducting mechanism in the transition. Thirdly, annual time series data limits the ability of this research to go deeply into detecting the dynamic relationships between investment, credit scale and structure, inflation, and consumer goods supply for the reform period because of too few degrees of freedom. And finally, again, because of the limitation of degrees of freedom based on the annual time series data, the research only deals with the fixed investment made by the state sector. In fact, the investment made by collectives (mainly, township and village enterprises) and individuals, and foreign direct investment are playing an increasingly important role in shaping the macro-scale of investment levels. In this non-state sector the central
242
Chapter 7
government has to rely heavily on indirect policy instruments to carry out its industrial and other macro-policy objectives. Some of these limitations can be overcome by focusing on the transition period and using quarterly time series data, although the data collection is not at all easy. During the transitional period, the annual series for the fixed investment in the non-state sector has become available, and quarterly data for credit scales and structure, cash supply, retail sales of consumer goods, as well as for fixed investment in the state sector can be collected. Based on this extended database it is possible and inviting to tackle the recurring cycles of over-investment (inflation) and retrenchment in China. To stylize the characteristics of the state investment system in the transition and of the conducting mechanism (from excessive investment demand to over-expanding of credit to inflation and then to the bottleneck constraints) could make a worthwhile contribution to the literature of transitional economics. The establishment of a seasonal norm function of the investment level based on cointegration and the corresponding error correction model of investment growth rate cycles would offer a formal framework in which deeper insight into the cyclical problem might be obtained, and more practical policy implications might be spelt out.
7.7 Summary This book has sought to establish a new framework for conceptualizing the aggregate investment behaviour and modelling investment cycles in a socialist and/or developing economy like China. The new framework is characterized by the integration of the existing theories of the investment cycle and of the bottleneck-constrained growth in a socialist and/ or developing economy with the cointegration and the error correction mechanism. Based on this framework, it is found that although the concrete investment adjustment processes have been subject to changes brought about by power redistribution among different political apparatuses and by the succeeding reform, the investment hunger and its generated tensions between investment expansion and the supply and distributive barriers have persisted. As long as the annual real investment level has unremittingly moved along the supply possibilities frontier of bottleneck sectors, a high level of real investment can be constantly maintained.
Conclusions
243
The importance of the coordination mechanisms that prevail among economic institutions and their agents is highlighted by the research. Different investment approval and brake apparatuses operate at different levels in a complex, interactive fashion, and these are also subject to political and economic changes. However, their fundamental feature remains, which is the persistent interest conflicts and accompanying bureaucratic coordination. We have assumed that all bureaucrats are rational economic agents, that they maximize the interests of their own institutions subject to certain internal and external supply and distributive constraints, and that they are not homogeneous. We have found that the investment cycle in China is a type of bureaucratic coordination failure, rather than a result of the planner's myopic intertemporal irrationality. The long-run investment level function and short-run investment growth rate function are established through modelling an unrestricted vector autoregression system, without any pre-imposed equilibrium conditions or behaviour assumption. The investment level function is represented by a long-run equilibrium comovement among investment outlays, grain output, effective energy supply and the price level of investment. The investment growth rate function is characterized by a conditional error correction model developed from the equilibrium level equation. A large proportion of the cyclical patterns of investment growth rate can be explained by the error correction behaviour towards the co-movement path and by the relevant change rates of energy supply and agricultural output. In the modelling process, we employ the standard one-step-ahead Chow test to detect any statistically significant structural break induced by policy changes or reform within the sample period, and the conclusion is negative. This indicates that both investment level and growth rate functions have parameter constancy over the sample period. In other words, investment hunger and its resulting tension between investment expansion and supply/distributive barriers have persisted within the sample period. The complicated bureaucratic co-ordination and bottleneck constraints have caused widespread investment inefficiency. This has been evident in large-scale duplication of construction at national level, initiation of too many new projects at the expense of technical updating of existing assets, greater prominence of gestation lags between the starting and completing of investment projects, and in other kinds of projects turning out to have little value once their products have to face
244
Chapter 7
market demand and competition. In order to eliminate the scourge of investment inefficiency, greater reform of the state investment system and related areas is required. Reforming the state investment system involves several difficult issues such as how to tackle the problems of the central-local relationships and relationships among government, enterprises and banks, and how to solve the soft budget constraint of the state sector. These difficulties can be overcome gradually by such reform strategies as the denationalization of state enterprises and state banks, and market reinforced decentralization. Besides reform, four other policy implications can be spelt out: (a) The policies and measures aimed at releasing agricultural, energy and transport bottlenecks and at promoting transformation of traditional agriculture are essentially anti-cyclical. (b) In order to overcome the bottleneck constraints it is necessary to slow down the extensive expansion of general industry, so that more economic resources can be allocated to agriculture, energy and transport.5 (c) The remarkable rigidity of the state investment system, which stands in striking contrast to successful reforms in other areas, indicates that reforming the state investment system is indeed one of most difficult aspects of the transition process. Consequently, this system is and will remain a dominant force in the state sector in the medium term at least. How it is changing and what can be done to adapt it to changes in the dynamic reformed economy remain major policy questions. Because markets and market-related institutions are still in their infancy, complete reliance on them is not possible. However, policies which give greater priority to the development of agriculture, energy, transport, communication, infrastructure and the raw materials sectors, which allow investment to anticipate changes in domestic consumer preferences and demand, promote joining international competition, which attempt to separate investment plans from credit plans and policy lending from commercial lending of the state banks, and which continue to support the growth of collective, township and village enterprises and other non-state sectors, are significantly helpful to the transition of the state investment system in the medium term. (d) With enormous population pressure in both the rural and urban sectors, China must create a high value-added but labour-intensive
Conclusions
245
modern agriculture so as to generate full employment at the same time as building modern agriculture. Following this strategy, Chinese agriculture should stimulate the good tradition of intensive and meticulous farming, promote land-saving techniques, and produce processed agricultural products. Clearly, the most problems addressed here are quite common to other transitional economies from central planning to market domination, and a part of the problems is also very typical in other developing economies.
Notes
Chapter 1 1 The average proportions of capital accumulation and state sector fixed investment in the national income used, i.e. the accumulation ratio and investment ratio respectively, are 29.9 with a standard deviation (SD) of 6.65 and 19.5 with a SD of 5.27 during the 1953-93 period; and the same figures as a percentage of GDP are 27.6 (SD 2.45) during 1978-93 and 18.3 (SD 2.19) during 1978-95, when the relevant statistical figures were available (source: Statistical Yearbook of China, 1993: 31, 43, 149; 1995: 137; 1996: 42,139). National income is the sum of net output of agriculture, industry, transportation, construction, and commerce, the five material production sectors of the economy. The coverage of national income in China's statistics excludes value added in 'non-material production sectors.' Thus, national income in China is approximately equivalent to the United Nations' net material products (NMP). The above-defined national income is also known as national income produced. National income used = national income produced - export + import {Statistical Yearbook of China, English version, 1990: 38). While GNP is not available, national income is the best proxy. 2 The percentage shares of state-sector fixed investment in the total are 69.5 in 1981, 65.6 in 1990, 60.6 in 1993 and 54.4 in 1995 {China Statistical Yearbook on Investment in Fixed Assets, 1950-1995: 27). The fact that statistics on total fixed investment covering all ownership types started only in 1981 prevents us from dealing with the patterns of total fixed investment for the pre-1980 period. However, in consideration of the strict restriction on development of the non-state-owned sector before the reform, it is safe to say that the shares of state fixed investment were higher than 70 per cent of the total from 1953 to 1979. 3 There are seven cycles of investment growth rates in China with lengths of three to eight years in the 1953-96 period. The maximum and minimum values of real investment growth rates in China from 1953 to 1996 are
246
Notes
247
84.53 and -62.54 per cent, respectively; in the former Soviet Union 18.2 and 0.7 percent; in Hungary 36.0 and -23.3 percent; in Poland 25.4 and -22.3 percent; and in Bulgaria 51.2 and -2.2 per cent from 1951 to 1989, respectively (source: data for China are based on the first column in Data Appendix of Chapter 6; for the others see Mihalyi 1992: 120). 4 For an extensive survey on this subject, see Mihalyi (1992). For other surveys, see Ickes (1986) and Simonovits (1992). Attempts to apply such response functions to China include Naughton (1986, 1987) and Imai (1994a &b). 5 The most direct reason behind this argument is that various econometric exercises, including several relevant international trade indicators and capital flows, show that none of them are statistically significant. The economic logic behind this non-significance will be analysed in sections 1.6, 4.2, 4.6 and 5.2. 6 This World Bank country study, China: Socialist Economic Development, is reviewed as the earliest and most comprehensive study of China's economy, based on the Bank's privileged and unprecedented access to Chinese data and officials, and from an internationally comparable perspective (Cyril Lin 1988). The statistical data appearing in the following paragraph are taken from this report. 7 The details about this most serious consequence of the Great Leap will be presented in section 4.6. 8 This section involves the ongoing methodological debate in econometric modelling. For those readers who have no interest in or are not familiar with this debate, you can skip over this section. 9 For the detailed discussions on this issue, see Maddala (1983) and Spanos (1990,1995), among others. 10 Roughly speaking, we say that an equilibrium relationship/fo, x2) = 0 holds between two variables X\ and x2 if the amount s, =f(xu,x2t) by which actual observations deviate from this equilibrium is a median-zero stationary process (Banerjee et al. 1993: 4). This implies that the equilibrium relation acts as an 'attractor', of the centre of gravity type. Along the attractor, no further incentive for, say, making extra benefit, can be produced (Engle & Granger 1991: 1-2). 11 The main data limitation for sectoral analysis is lack of a sectoral breakdown of technical updating and transformation investment for the pre-1980 period, although for most years the sectoral data of capital investment is available.
248
Notes
12 Lardy (1995: 1073) raises a strong argument: 'contrary to what one might believe from observing the flood of foreign capital flowing into China, on a net basis such inflows have not contributed to domestic capital formation in China', although they have made a major contribution to China's openness to the world.
Chapter 2 1 The terms of 'self-restraint' in investment decision-making and 'relatively symmetrical relationship' between demand and supply are used here to explain why investment hunger becomes so impressive in socialist economies in comparison with market economies. In a market economy there may also be imbalances between desired investment and supply capacity of the economy, which cannot be fully explained by such 'self-restraint' and 'symmetrical relationship'. The relevant adjustment mechanisms have been discussed intensively in macroeconomic literature. 2 In this section Bauer's four-phase description is employed to sketch the basic and most general features of investment cycle in China. The specific features of a boom-bust reform cycle will be discussed in section 2.4.2. Investment hunger and bottleneck constraints have persisted in both pre- and post-reform periods. 3 In fact, Mihalyi (1992: 81-7) compares four similar indicators in measuring investment dispersion: 'uncompleted investment stock/current investment', 'current investment/current completed investment', 'uncompleted investment stock in / year/investment in t-\ year in uncompleted projects', and 'current investment in unfinished projects/current investment', by using data for Hungary (1950-89) and the former Soviet Union (1958-63). He finds that movements in the four indicators do not exhibit parallel variations; he questions which indicator planners should use (ibid.: 86). 4 In fact, as pointed out by Wuyts (1988), FitzGerald's reconstruction (related to price stability and the food balance) is a little vague. Equation (2.6) has the advantage of being able to be intuitively understood, although it still remains vague. 5 For an excellent survey of Kalecki's research on the socialist economy and of the debates around Kalecki's growth theory, see Osiatynski (1988). 6 Identifying the investment category with Kalecki's industry sector in this section is reasonable when we deal with the planners' trade-offs. In China over 90 per cent of fixed investment by state-owned units has gone into industry (over 60 per cent of the total), construction, transportation, communication, commerce, banking and other social service sectors, which toge-
Notes
249
ther form the so-called modern sector or industrial sector, leaving little over for the traditional agricultural sector. About half of the remaining 10 per cent of fixed investment is shared by water conservation and irrigation projects, with the rest shared by farming, forestry, animal husbandry and fishery. In the period 1981-85 only 3.9 per cent of total fixed investment went to the agricultural sector including water conservancy and from 1986-95, the figure was only 2.5 per cent (Statistics on Fixed Investment in China, vols: 1950-85, 1986-87, 1988-89, 1990-91; China Statistical Yearbook on Investment in Fixed Assets, vol. 1950-95: 42-43). 7 The economic logic behind equation (2.6) for a semi-industrial market economy is well summarized in Jansen (1990: 10-11), where the price mechanism in general and agriculture's terms of trade (price scissors) in particular play a central role in, and export and import also contribute to, avoiding inflation and stagnation of the domestic economy. In China's case, however, coercive procurement of key agricultural products, binding rationing of foodstuffs and other basic consumer goods, and the adjustment of material reward to peasants (rather than price) have played the leading role in determining agriculture's terms of trade in a much broader sense. In addition, for a huge economy like China, the export and import of grain and energy do not play a significant role in influencing the balance between industry and agriculture. This will be demonstrated in Chapters 4 and 5. 8 In a Kaldorian sense, forced saving might be realised by shifts in the distribution of national income from wages to profits at the expense of a tolerable inflation rate. Also industrial concentration may help raise profits through depressing the real wage. However, in the meantime, the reaction of workers to the reduction of the real wage is a demand for higher money wages based upon trade-union bargaining power, causing a price-wage spiral to follow (Kaldor 1955, Kalecki 1976: 44, Thirlwall 1974: 86-96). In China such forced saving is directly collected by the state sector through industrial nationalization and agricultural collectivization. The cost is the stagnation of the urban real wage rate and of consumption by both the urban and the rural population for a long time, as well as inflation. 9 In addition, rural/urban inequality is high in comparison to other nations in Asia (Khan et al. 1993). In 1988 the official estimated ratio of urban to rural income was 2.19, and Khan et al. (1993) put it at 2.42 by taking into account in-kind income and subsidies. Considering that the rural population accounts for almost 80 per cent of total population, it is fair to say that the agricultural sector has provided a significant share of accumulation through its extremely low consumption, although part of this income difference may be explained by the productivity gap. 10 The agricultural constraint to investment demand and the linkage between agricultural fluctuation and macroeconomic adjustment in China will be
250
Notes
discussed from the perspectives of both history and data analysis in Chapter^ 11 The term 'job-off in China indicates those employees who still keep the employment title and corresponding welfare benefits, but have lost their jobs. 12 The dependence ratio in the less developed northwest region is certainly higher than the national average. 13 According to Lin et al., the economic causality in general and development strategy choice in particular is decisive. Once a capitalist economy adopts the same 'leap forward' development strategy, it also has a similar macro policy environment, resource allocation system, and micro management institution, as well as a development performance similar to that of the corresponding socialist economy (Lin et al. 1996: Chapters 2 & 3). 14 There are a large number of official documents on setting up and implementing these five-year plans, most of them unpublished. For published materials, see, for instance, 'Report about the First Five-Year Plans on the Second Conference of the First National Peoples's Congress' by Li Fuchun {Documents of the Second Conference of the First National Peoples's Congress of the People's Republic of China, Beijing: People's Press 1955), 'Recommendation about the Second Five-Year Plan of Developing National Economy' approved by the Eighth National Congress of the Communist Party of China {Documents of the Eighth National Congress of the Communist Party of China, Beijing: People's Press 1956), the news reports and leading article about the design and implementation of the Third FiveYear Plans in People's Daily (28 & 30 November, 3 December 1965; 1 January, 14 August 1966), Compendium of a Ten-Year Plan from 1976 to 1985 about the Developing National Economy approved by the First Conference of the Fifth National People's Congress (Beijing: People's Press, 1978), The Sixth Five-Year Plans of National Economy and Social Development of the People's Republic of China approved by the Fifth Conference of the Fifth National People's Congress (Beijing: People's Press, 1982), and others. 15 For two intensive surveys on this subject, see Sheng (1993a, 1993b). For other surveys, see Knight (1995) and Karshenas (1993). 16 For a rough definition of the equilibrium relation, see note 8 of Chapter 1. For a strict definition and explanation, see section 6.2. 17 A full exploration will depend on cointegration and error correction mechanism analysis. 18 Other statistical weaknesses include that
Notes
251
(a) all behaviour equation estimates show non-constancy of parameter (by one-step-ahead Chow test), (b) the money demand equation presents heteroscedasticity [ARCH test: F(l, 35) = 16.82; heteroscedastic error test: F(4, 31) = 4.50] and misspecification of functional form [White's general functional form misspecification test: F(5,29) = 3.43], (c) the consumption goods demand equation exhibits serial correlation [LM test for 1 to 2 order serial correlation: F(2, 33) = 2.98], heteroscedasticity [ARCH F(l, 33) = 11.56; heteroscedastic F(6, 27) = 5.39], and misspecification of functional form [White test: F(9, 23) = 3.55], and (d) the investment function estimate shows significant autoregressive conditional heteroscedasticity [ARCH F(l, 32) = 5.93]. 19 For relevant English surveys and research on this issue, see Garnaut & Ma (1992), Cyril Lin (1988, 1989), Oppers (1997), Sung & Chan (1987), Watson (1994) and World Bank (1990,1994). 20 In Fischer (1988: 304) four sets of stylized facts and in Chapter 2 of Dore (1993), ten stylized facts (empirical regularities) associated with business cycles in the modern market economies are stated, together with empirical evidence. 21 For the dominant control of investment resource by the state, see note 2 of Chapter 1. The dominance of bureaucratic coordination over investment decision-making in the post-reform period will be dealt with in Chapter 3.
Chapter 3 1 According to the official statistics {China Statistical Yearbook on Investment in Fixed Assets, 1950-1995: 458), China's government set the classification standard of large, medium and small construction projects in 1953 and revised this in 1962, 1977 and 1979, respectively. In principal, the construction projects are classified according to the total designed capacity or the total amount of investment in the design specification or preliminary design approved by authorities at higher levels of the government hierarchy. Usually, the size of a project is defined according to the designed capacity of the main product of the project. If there are many types of products and it is difficult to determine which one is the main product, the size of a project is defined according to its total planned investment. 2 For example, the family farming system, one of most important reforms, now known as the household responsibility system, was first secretly adopted by some peasants in Anhui Province in 1978 and was fully officially recognized in late 1981, when 45 per cent of the former collectives had already been dismantled and the household responsibility system had
252
Notes
been instituted (X. Chen, 1993, Lin et al. 1995). The first steps in enterprise reform, namely introducing profit retention for enterprises, performancerelated bonuses for workers and permitting state enterprises to produce outside the mandatory state plan, were pioneered by Sichuan Province in 1978 and 1979 under the then governor Zhao Ziyang. More recent examples include stock exchanges in Shanghai and Shenzhen, which were also initiated by the local authorities and only later became accepted by the central authority (Chen et al. 1992, Zhang & Yi 1995). 3 Fixed investment by a state-owned unit is divided into 'capital construction', 'technical updating and transformation', other investments (mainly oil field development and repair of highways), and 'commercialized housing construction'. Theoretically, the capital construction investments emphasize the creation of new enterprises or major expansion of existing enterprises while the technical updating and transformation investments are intended for the modification of existing facilities. Historically, capital construction investments related to projects were financed through the government's budget and thus more tightly controlled by the central government, whereas technical updating and transformation investments were financed mainly from depreciation funds and usually controlled at local level. Now the distinction is less clear-cut as budget funds are also channelled through state banks, and technical updating and transformation investments are funded to a large degree by bank loans. But the basic feature has been maintained, i.e. that the former is more under central control and the latter more under local control. 4 The relevant details on the 'national priority projects', will be presented in section 3.4.2. 5 Before 1979, bank loans and working capital were closely interrelated. Working capital was divided into two categories: 'quota working capital' was the amount required by an enterprise for the maintenance of a normal rate of business turnover and was provided by government budgets (to which all profits had to be remitted). 'Non-quota working capital' was the capital requirement allowed above the quota, and consisted of temporary and seasonal needs, plus the financing of goods in transit (Chen & Niu 1990: 1213). 6 For instance, the inflation rates of construction materials are as follows: Year 1988 1989 1990 1991 1992
Inflation rate (%) 16.3 13.0
6.6 8.6
Year 1993 1994 1995 1996
18.8
Source: A Statistical Survey of China 1994: 89; 1997:69.
Inflation rate (%) 61.1 12.5
2.6 2.8
Notes
253
7 For a detailed introduction to the history and reform of China's material supply system, see Eckstein (1977: 134-39) and Wong (1985). 8 Data from a World Bank survey of over 900 SOEs, summarized in Table 2 of Jefferson & Rawski (1994), show the increasing and significant positive correlations between profits and retained earnings and bonuses in the periods 1980-83, 1984-87 and 1988-90. 9 This empirical research is based on the surveys conducted by the Institute of Economics, Chinese Academy of Social Science with assistance of four provinces' System Reform Commissions (Sichuan, Jiangsu, Jilin and Shanxi). Annual data for 1980-89 for 769 SOEs in the surveys give details of the firms' internal incentives, the firms' cost and revenue accounts, and the nature of the relationship between the firms and the state. Because large firms are over-represented in comparison to SOEs in general, the sample may cover the core of the traditional state-run enterprises.
Chapter 4 1 The general meaning of agriculture in China, especially in official statistics, is that it consists of cropping (i.e., agriculture in a narrow sense), forestry, animal husbandry, sideline activities and fishery. The output value of cropping accounted for over 74 per cent of all agriculture before 1979 {Comprehensive Statistics of China's Rural Economy, hereafter Rural Statistics, 1989: 114-5), and in 1992, this share decreased to 55.5 per cent {Yearbook 1993: 335). 2 In Chinese official statistics, 'grain output' covers all types of grain such as rice, wheat, corn, sorghum, millet and other miscellaneous grains, as well as tuber crops and soybeans. The output of beans refers to dry beans without pods. Tuber crops (sweet potato and potato, not including taro and cassava), are included as a grain item converted on a 4:1 ratio, that is four kilograms of fresh tubers was considered equivalent to one kilogram of grain until the end of 1963. Since 1964 a 5:1 ratio has been used {Yearbook 1986: 743). 3 One obvious exception is the agricultural collapse during 1960-62 caused by the ultra communization and the Great Leap Forward, together with the high compulsory grain delivery quotas and bad weather. For the detailed and insightful analyses and the debates on the reasons for this collapse, see, among others, Kueh (1995), Lin (1990, 1993), Liu, M. (1993) and Putterman & Skillman (1993). 4 See note 2 of this chapter.
254
Notes
5 The Soviet figures for 1928 and 1932 are cited by Kueh (1995) from Hoeffding (1959) and Kueh (1984). The other data are from Yearbook (1991: 819, 836; 1993: 81, 364, 609; 1994: 59, 345). 6 The example, cited by Kueh (1995), is from 'The 1931 flood in China: An economic survey', in Nanjing Academic Journal (Jinling Xuebao) 1932 (2/1): 216. 7 Roughly speaking, weak exogeneity means that if the variables in the righthand side of an equation are weakly exogenous, the single equation analysis for explaining the left-hand variable maintains the full information of interest as does the unrestricted vector autoregressive system, when incorporating both long-run and short-run dynamic. The strong exogeneity is much closer to traditional exogeneity than is weak exogeneity. Weak exogeneity plus Granger noncausality from the left-hand variable to the righthand variables generates strong exogeneity. 8 Justin Lin (1989,1990) put the point more bluntly when he tested the commonly accepted explanations of the collapse of agricultural production in 1959-61: The official explanation for the disaster of 1959-61 was bad weather, but it is unlikely that in three successive years bad weather hit every part in a country with a vast territory. In fact, bad weather is the easiest excuse for local officials in China to explain poor agricultural performance caused by other reasons such as the deprivation of the right to withdraw from a collective, mistakes in policy or management, incentive problems, etc. National statistics show a clearly increasing trend in the "areas reported to be hit by natural disasters', which is contradictory to the fact that more acreage was irrigated each year and most of the increased irrigation came from modern engine-powered irrigation. Lin (1989: 15) employed the indicator 'areas hit by drought/areas by flood' to reveal that there is a clear trend for more areas to be reported to be hit by drought than flood. The reason is that floods are harder to fabricate and easier to identify than droughts. Therefore, when agricultural production falls, local officials often used 'drought' as an excuse. 9 The difference between the accumulation ratio and the investment ratio is caused by the fact that the accumulation also includes changes in inventories and non-state investment. Here the non-state investment dominates the difference. For example, during 1981-94, non-state fixed investment accounts for over 30 per cent of the national total, and over 9 per cent of GDP {Statistics on Fixed Investment in China, 1950-85, 1987: 5; Yearbook 1993: 145, 31; 1995: 137, 32). In the retrenchment phase non-state investment is usually cut back first by the state bank's much stricter control over non-state credit than over the state sector; the bank may stop lending to non-state investment projects and so on. This activity goes beyond the scope of this research.
Notes
255
10 Pragmatically speaking, consider the following specification of a time series >>, (without trend and nonzero mean): yt = qyt.j +s(, where the term et is a sequence of uncorrelated disturbances with mean zero and constant variance. If cp is less than one in absolute value, yt fluctuates around a mean of zero, and thus is said to be stationary. The detailed discussion and test procedure of stationarity will be given in the technical appendix of Chapter 6. 11 The remarkable success of China's collectivization movement from 1952 to 1958 and the sudden collapse of agricultural production in 1959-61 have been the topics most discussed among students of the Chinese economy. The reasons for the sudden collapse are hotly debated in the recent literature (cf. Justin Lin 1990, 1993; a Symposium issue of the Journal of Comparative Economics June 1993; Kueh 1995; Riskin 1987, among others). The focus of this chapter is to deal with the correlation between agricultural fluctuation and investment adjustment, which may indirectly show some reasons for the boom and collapse from the perspective of macroeconomic policy. 12 There is a subcycle of agriculture in 1954—57 as shown in Figure 4.1, which is accompanied by an investment cycle in 1955-57 as indicated in section 1.3. The basic macro-mechanism behind this correlation is fundamentally same as that of the major cycle from 1954-62; therefore I do not discuss it separately (for a detailed discussion about the subcycle, see Eckstein 1968 and Riskin 1987: Chapter 5). 13 Dong (1980), cited by Riskin (1987). 14 On average, at least ten million casual and contract workers worked in state-owned mines, construction sites and other non-farm industries (Lardy 1985: ii), while registered in the rural system. This means they are ineligible for rationed food, the distribution of low-rent houses, retirement insurance and other welfare systems which are only open to 'permanent workers'. They must also purchase food in non-state markets where prices are normally two to four times the prices of rationed commodities and must finance housing and medical costs themselves.
Chapter 5 1 If the conversion of hydropower into SCE is calculated on the basis of an internationally used equivalent coefficient as done by the China State Statistical Bureau in 1993 and 1994, the result will show that the share of hydropower will decrease by more than 3 percentage points and the proportion of coal will increase correspondingly by more than 3 percentage points, reaching 76.8 in 1993 and 77.7 per cent in 1994 (cf. Yearbook 1995: 199).
256
Notes
2 For example, based on the official exchange rate, China's energy intensity (unit of energy use/US$GDP) in 1990 was 2.65 kg oil equivalent per dollar of GDP, whereas the OECD average was 0.25 kg oil equivalent per dollar (Ishiguro & Akiyama 1995: 28). 3 For the detailed discussions of the reasons for the high energy intensity in China, see, among others, Lu (1993: section 3.2) and World Bank (1985: 10-16). 4 A detailed discussion of the 1994 energy price reform and its effect goes beyond the concern of this research. For relevant information, please see Economic Situation and Prospect of China 1994: 40-5; 1997: 61-6, 8 0 ^ ; China Information Daily, 16 Feb. 1994: 1, 3. 5 For relevant reports on this issue written by Chinese officials and researchers, see Energy of China (Zhongguo Nengyuan) (1989 (2): 6-9; (3): 1-4; (6): 31-4; 1990 (3): 4-9; (5): 18-21; 1991 (7): 37-40; (10): 5-15; 1992 (2): 4-5; (3): 5-10; (7): 23-8), China Investment and Construction (Zhongguo Touziyu Jianshe) (1988 (8): 25-7; (12): 18-19; 1989 (3): 10-12). For the relevant literature in English, see Ledic (1989), Jing-Tong Liu (1987), Lu (1993), Smil (1988), and World Bank (1985). 6 It should be noted that in this research the normal shortage level is determined by the co-movement among relevant variables rather than by a univariate moving average. 7 See, for instance, China Information Daily (2 July 1993: 1; 16 February 1994: 1, 3), People's Daily (overseas version) 14 June 1993: 2; 1 March 1994: 2), Statistical Survey of China (1994: 8, 47). 8 A symbolized technical consideration is that, on average, the comprehensive utilization time of representative power-consuming equipment is 2,500 hours a year, whereas that of representative power-generating equipment with the same capacity is 5,000 hours a year. This implies that the reasonable ratio should be 2:1 in terms of technical efficiency (Hu, Zhaoyi 1990). 9 China had the smallest service sector and the largest industrial sector in comparison with the USA, India, Brazil, Japan and South Korea in the early 1980s. This contributes significantly to the high freight intensity. Assuming that the service sector does not generate any freight transport, the freight intensity related to total GNP of an economy like China, where the service sector is 20 per cent, would be double that of an economy like the USA where the service sector accounts for 60 per cent of GNP, with the same intensity of freight for agricultural and industrial sectors (for details, see Yenny & Uy 1985: 9-10). 10 For relevant reports on transport bottleneck and crisis, see China Information Daily (2 July 1993: 1), People's Daily (9 March 1989: 5; People's
Notes
257
Daily (overseas version), 14 June 1993: 2), European Time (18 November 1995: 5) cites Economic Daily (November 1995), and World Bank (1983). 11 For relevant reports on this issue, in addition to those given in note 4 of this chapter, see People's Daily (overseas version) (9 March 1989: 5), China Information Daily (2 July 1993: 1; 16 February 1994: 1, 3), China Investment and Construction (1992: (12): 25-6). 12 In the literature, 'unfinished construction' is used as a proxy for shortage in the investment goods sector. However, as discussed in section 2.2.3, it may be a misleading indicator. Furthermore, this indicator is not available for China.
Chapter 6 1 The following discussion about long-run equilibrium relationship and error correction mechanism is based on Banerjee et al. (1993: Chapters 1 & 5), Engle & Granger (1991: Introduction), Granger & Hallman (1991). 2 The relevant formal definition and discussion will be provided in the technical appendix of this chapter. 3 It is a basic fact that the sample size of macroeconomic variables is relatively small in most developing economies. This, however, should not become an insurmountable barrier to further exploration based on the interaction between theoretical and statistical analysis. One must be cautious to obtain economically meaningful insights. 4 For the technical procedures of the unit root test and other tests discussed in following paragraphs and sections, see the technical appendix of this chapter. 5 Both Schwarz and Hannan-Quinn information criteria are used to choose the proper number of lags. For the technical details, see Doornik (1994: 151-2). I also estimate the relevant system in which the |i is restricted to lie entirely in the space spanned by a, thus only allowing for an intercept in pzt. The test statistics indicate that r = 2 (after Reimers' small-sample correction). Following Johansen's (1992d) sequential testing procedure, which estimates both r and the presence or absence of trends, and guarantees correct size asymptotically, one should accept the null hypothesis associated with the largest eigenvalue for which rank (n) < r is not rejected against the unrestricted alternative rank (n) = 4 in either of the two systems. Therefore, I chose the unrestricted constant in Model (2) because it suggests that r=l, and allows for there being a linear trend in the data.
258
Notes
6 Among others, see Hare (1982, 1989), Kornai (1982, 1992), Kornai & Martos (1981), Kornai & Simonovits (1986), Martos (1990), and Simonovits (1992).
Chapter 6 Appendices 1 Substituting equation (A2.2) into (A2.1) and rearranging it gives yt = \i + bt + By,_i + £t (/ = 1, 2, 3, ...), where ^i = Oo(l-B) and 5= a 1(1—B). Dickey (1976), Fuller (1976) and Dickey & Fuller (1981) introduced tests of 6 = 1 based on statistics obtained from applying ordinary least squares (OLS) to this rearranged equation of the quasi-first-difference transformation of equation (A2.1). The distribution of the /-statistic for 6 changes if a time trend is excluded in the regression, and if a constant and a time trend are excluded, which was first tabulated by Dickey using the Monte Carlo method and is given in Table 8.5.2 of the book by Fuller (1976). Recently, a formalized generating procedure of these critical values, MacKinnon (1991), has been popularly accepted. 2 Equation (A2.6) is less commonly encountered than (A2.3)-(A2.5), but it may well be a plausible specification in some cases, as argued by Ouliaris etal.(1988). 3 The word 'marginal' in the modelling process is used in its statistical sense as usually employed in textbooks on the probability theory (see, e.g. Kendall & Stuart 1977: 22, Spanos 1986: sections 5.3 and 5.4). 'Marginal' as used in statistics should not be confused with its economic sense, as in 'marginal versus average cost'.
Chapter 7 1 For the relevant reports and surveys on the issue of repetitionary construction, see, among others, Almanac of China's Finance and Banking (1993: 265-6), Statistical Yearbook of China (1997: 454-5), China's Youth Daily (3 February 1993: 2), and Survey Report by the research group (1992). 2 At the beginning of the reform, state industrial enterprises owed banks a sum of credits equal to only about 11 per cent of their book value (depreciated fixed assets plus the value of all inventories). At that time SOEs relied basically on fiscal appropriation for investment funds (cf. sections 3.2 and 3.4, Table 3.4). During the reform period, SOEs have become increasingly dependent upon bank funds for both fixed investment and working capital. By 1988, external liabilities were about 45 per cent of the book value of state industrial enterprises (Naughton 1995a: 264). In June 1995, according to a recent survey by the People's Bank of China (the central bank) of 248 SOEs in various parts of China, the SOEs' average total debt to asset ratio
Notes
259
stood at 69 per cent, while their average debt-to-working capital asset ratio was 98 per cent {Shenzhen's Securities Times, 22 January 1996). 3 When including the output value of state-joined shareholding companies, the output proportion of industrial SOEs in total was 40 per cent in 1994 {Statistical Yearbook of China 1995: 377). 4 The three newly created 'policy banks' are the State Development Bank of China (SDBC), the Agricultural Development Bank of China (ADBC) and the Export Import Bank (EIB). 5 It should be noted that in terms of identifying sectoral priorities, policy concern is clearly different from theoretical modelling. According to the former, transport has been the first bottleneck in the chain of resource and goods circulation. In the modelling experiments, it seems that effective energy supply (energy consumption) is the best representative of these bottlenecks (cf. Chapter 5).
Bibliography
I.
Sources of Data
Almanac of China industry (Zhongguo Gongye Nianjian) (1991) Beijing: China Labour Press. Almanac of China's Finance and Banking (Zhongguo Jinrong Nianjian) (various years) Beijing: Editorial Department of China's Finance and Banking. Almanac of China's Finance and Banking (1994) 'State Council's Decisions on Financial System Reform, 1993', Almanac of China's Finance and Banking {Zhongguo Jinrong Nianjian). Beijing: Editorial Department of China's Finance and Banking. China Fiscal Statistics {Zhongguo Caizheng Tongji): 1950-1988 (1989) edited by the Ministry of Finance. Beijing: China Fiscal and Economic Publishing House. China Population Statistics {Zhongguo Renkou Tongji Nianjian) (various years) Beijing: Social Science Press. China Statistical Yearbook on Investment in Fixed Assets: 1950-1995 (1997) (in Chinese and English) (various years) Beijing: China Statistical Publishing House. Chinese Agricultural Yearbook {Zhongguo Nongye Nianjian) (various years) Beijing: Agricultural Press. Chinese Economic Yearbook {Zhongguo Jingji Nianjian) (various years) Beijing: Jingji Guanli Zazhishe. Chinese Encyclopedic Yearbook {Zhongguo Baike Nianjian) (various years) Beijing: Zhongguo Da Baike Quanshu Chubanshe. Chinese Rural Statistical Yearbook {Zhongguo Nongcun Tongji Nianjian) (various years) Beijing: China Statistical Publishing House. Chinese Trade and Price Statistics: 1952-1983 {Zhongguo Maoyi Wujia Tongji Ziliao): 1952-1983 (1984) Beijing: China Statistical Publishing House.
260
Bibliography
261
Comprehensive Statistics on China's Rural Economy (Zhongguo Nongcun Jingji Tongji Daquari) (1989) Beijing: Agriculture Press. State Statistical Bureau (1987) 1985 Industrial Census of China. Beijing: China Statistical Publishing House. Statistical Yearbook of China {Zhongguo Tongji Nianjian) (various years) Beijing: China Statistical Publishing House. Statistical Yearbook of China (English Version) (1990) Beijing: China Statistical Publishing House. Statistical Yearbook of Rural China {Zhongguo Nongcun Tongji Nianjian) (various years) Beijing: China Statistical Publishing House. United Nations (1993) Statistical Yearbook of the United Nations, 1992. New York: United Nations. Statistics on Fixed Investment in China 1990-91 {Zhongguo Guding Zichan Tongji Ziliao) (1993) Beijing: China Statistical Publishing House. Statistics on Fixed Investment in China 1988-89 {Zhongguo Guding Zichan Tongji Ziliao) (1991) Beijing: China Statistical Publishing House. Statistics on Fixed Investment in China 1986-87 {Zhongguo Guding Zichan Tongji Ziliao) (1989) Beijing: China Statistical Publishing House. Statistics on Fixed Investment in China 1950-85 {Zhongguo Guding Zichan Tongji Ziliao) (1987) Beijing: China Statistical Publishing House. Statistical Survey of China {Zhongguo Tongji Zhaiyao) (various years) Beijing: China Statistical Publishing House. Statistics on Labour and Wages of China 1949-85 (Zhongguo Laodong Gongzi Tongjiziliao) (1987) Beijing: China Statistical Publishing House. Statistics on National Income: 1949-85 {Guominshouru Tongji Ziliao Huibian) (1987) Beijing: China Statistical Publishing House. UNIDO (1990) Handbook of Industrial Statistics. Vienna: United Nations Industrial Development Organization. World Bank (1995) World Development Report. Washington DC: World Bank. World Bank (1994) World Development Report. Washington DC: World Bank. World Resources Institute (1994) World Resources, 1994-95 Oxford: Oxford University Press, for the World Resources Institute in collaboration with the United Nations.
II.
Chinese Magazines, Newspapers, and Official Documents
16-Point Programme (1993) Beijing: State Planning Commission (unpublished document).
262
Bibliography
Beijing Review (various issues) (former Peking Review). China Daily (various issues) Beijing. China Economic System Reform Yearbook (Zhongguo Jingji Tizhi Gaige Nianjian) (various issues). China Information Daily {Zhongguo Xinxibao) (various issues). China Investment and Construction {Zhongguo Touzi yu Jianshe) (various issues) Beijing: State Planning Commission, Bureau of Investment. China Youth (various issues). China Youth Daily {Zhongguo Qingnianbao) (various issues) A Collection of Documents on Fixed Investment and Construction {Guding Zichan Touzi he Jianzhuye Wenjian Huibian) (1992) compiled by the State Statistical Bureau, Department of Investment Statistics. Beijing: China Statistics Press. Compendium of a Ten-Year Plan from 1976 to 1985 about Developing National Economy (1978) Beijing: People's Press. Documents of the Eighth National Congress of the Communist Party of China (1956) Beijing: People's Press. Documents of the Second Conference of the First National Peoples's Congress of the People's Republic of China (1955) Beijing: People's Press. Economic Daily (various issues). Economic Problems ofAgriculture {Nongye Jingji Wenti) (various issues). Energy of China {Zhongguo Nengyuan) (various issues). European Time (various issues) Chinese newspaper published in Europe. Peking Review (see Beijing Review, above). People's Daily (Renmin Ribao) (various issues). Shenzhen's Securities Times {Shenzhen Zhenjuan Shibao) (various issues). The Sixth Five-Year Plans of National Economy and Social Development of the People's Republic of China (1982) Beijing: People's Press. 'The Standards for Classifying Large, Medium and Small Investment Projects' (1979) Beijing: State Planning Commission (unpublished document). State Planning Commission (1993) 'A Plan on the Comprehensive Utilizing and Economizing of Energy in 1993' (Document N° 25)', Energy of China {Zhongguo Nengyuan), 2: 17-19.
III. Books, Articles and Reports Albouy, Yves (1991) 'Coal pricing in China: issues and reform strategy', World Bank Discussion Papers, N° 138. Washington DC: World Bank.
Bibliography
263
Ash, Robert F. (1992) 'The agricultural sector in China: Performance and policy dilemmas during the 1990s', China Quarterly, 131: 545-76 (September). Ashton, Basil, Kenneth Hill, Alan Piazza & Robin Zeitz (1984) 'Famine in China, 1958-61', Population and Development Review, 10(4): 613^5. Banerjee, A., JJ. Dolado, J.W. Galbraith & D.F. Hendry (1993) Cointegration, Error Correction, and the Econometric Analysis of Non-stationary Data. Oxford: Oxford University Press. Banerjee, A., JJ. Dolado, J.W. Galbraith, D.F. Hendry & G.W. Smith (1986) 'Exploring equilibrium relationships in econometrics through static models: Some Monte Carlo evidence', Oxford Bulletin of Economics and Statistics, 48(3): 253-77. Bauer, T. (1978) 'Investment cycle in planned economies', Ada Oeconomica, 21: 243-60. Bauer, T. (1988) 'From cycles to crisis?' Eastern European Economics, 27 (1): 5-44 (autumn). Bell, Michael W., Hoe Ee Khor, and Kalpana Kochhar (1993) 'China at the Threshold of a Market Economy,' IMF Occasional Paper, No. 107, Washington DC: IMF. Benassy, J.-P. (1986) Macroeconomics: An Introduction to the Non-Walrasian Approach. Orlando USA: Academic Press. Bernstein, Thomas P. (1984) 'Stalinism, famine, and Chinese peasants', Theory and Society, 13(3): 339-78 (May). Bhalla, S. (1978) 'The role of sources of income and investment opportunities in rural savings', Journal of Development Economics. 58(3): 578-82. Boswijk, H.P. (1990) 'On the scope of conditional dynamic modelling of cointegrated variables', Tinbergen Institute Research Bulletin, 2: 97-111. Brada, J.C. & A.E. King (1992) 'Central planners as market stabilizers: Evidence from Poland and the Soviet Union', The Review of Economics and Statistics. 74(1): 1-13. Cambridge History of China (1991) vol. 15. Cambridge: Cambridge University Press. Cambridge History of China (1987) vol. 14 Cambridge: Cambridge University Press. Central Rural Policy Research Office, Rural Survey Division (1988) Economic Problems in Agriculture (Nongye Jingji Wenti), 8: 49. Chang, H. & P. Nolan (1996) 'Structural inheritance or policy choice? Explaining the results of post-Stalinist reform in Europe and Asia', pp. 1334 in: D.H. Brown & R. Porter (eds) Management Issues in China, vol. 1.
264
Bibliography
Chen, Dong & Zhili Niu (eds) (1990) Practical Manual on Economic Statistics of China (Zhongguo Jingji Tongji Shiyun Daquari). Beijing: Chinese Peoples' University Press. Chen, Dunjin (1980) 'A few opinions on the choice between bank loans and budgetary allocations in capital construction investment', in Jingji Yanjiu (compiled by an editorial board) An Exploration of Reform of the Economic Management System {Jingji Guanli Tizhi Gaige de Tantao). Jinan: Shandong People's Press. Chen, Shanlin (1985) Shangye Tongji {Commercial Statistics). Shanghai: Shanghai People's Press. Chen, Chien-Hsun (1989) 'Monetary aggregates and macroeconomic performance in Mainland China', Journal of Comparative Economics, 13(2): 314-24. Chen, Kang, G. Jefferson & I. Singh (1992) 'Lessons from China's Economic Reform,' Journal of Comparative Economics, 16(2): 201-25. Chen, Xiwen (1993) China's Rural Reforms: Review and Perspectives {Zhongguo Nongcungaige: Huigu yu Zhanwang). Tianjin: Tianjin People's Press. Chen, Yun (1984) Selections from Chen Yun: 1949-1956 {Chen Yun Wenxuan: 1949-1956). Beijing: People's Press. China Material Economic Society (1983) A Draft History of China's Socialist Material Management System {Zhongguo Shehuizhuyi Wuzi Guanli Tizhi Shilue). Beijing: Wuzi Press. Chow, Gregory C. (1997) 'Challenges of China's economic system for economic theory', American Economic Review, 87 (2): 321-27. Christin, D.E. & N.C. Short (1991) 'Testing a planners' reaction function using disequilibrium model excess demand indicators: The case of Poland', Journal of Comparative Economics. 15(1): 122-31. Chronicle of Economic Events of the People's Republic China (October 1949-September 1984), (Zhonghua Renmin Gongheguo Jingji Dashiji) (1985) (edited by the official Compiling Group of the Chronicle) Beijing: Beijing Press. Commentator (1981) 'Straighten out guidelines for economic work: On leftist mistakes in economic construction', People's Daily {Renmin Ribao), 23 May, p. 1. Crook, Frederick W. (1998) 'China's "governor's grain bag policy": Concerns about food security', China Information, 12 (3): 87-103. Cui, Zhiyuan (1993) 'The Dilemmas of Soft Budget Constraint: Three Institutions That Challenge the "Invisible Hand" Paradigm'. PhD dissertation, Department of Political Science, University of Chicago, Chicago, USA.
Bibliography
265
Dahlstedt, Roy (1980) Cyclical Fluctuation under Central Planning: An Inquiry into the Nature and Causes of Cyclical Fluctuation in the Soviet economy. Helsinki: The Helsinki School of Economics. Dickey, D.A. (1976) 'Estimation and hypothesis testing for nonstationary time series', PhD dissertation, Iowa State University, Ames, USA. Dickey, D.A. & W.A. Fuller (1979) 'Distribution of the estimators for autoregressive time series with a unit root', Journal of the American Statistical Association, 47: 427-31. Dickey, D.A. & W.A. Fuller (1981) 'The likelihood ratio statistics for autoregressive time series with a unit root', Econometrica, 49: 1057-72. Dickey, D.A., D.W. Jansen & D.L. Thornton (1994) 'A primer on cointegration with an application to money and income', pp. 9-46 in: B. Bhaskara Rao (ed.) Cointegration: For the Applied Economist. Basingstoke & London: Macmillan. Di Matteo, M. (1984) 'Alternative monetary policies in a classical growth cycle', pp. 14-24 in: R.M. Goodwin et al. Nonlinear Models of Fluctuating Growth. New York: Springer-Verlag. Dittmer, Lowell (1987) China's Continuous Revolution: The Post-Liberation Epoch 1949-1981. Berkeley, USA: University of California Press. Doornik, J.A. (1994) PcFiml 8.0: Interactive Econometric Modelling of Dynamic Systems. London: International Thomson Publishing. Dore, Mohammed H.I. (1993) The Macrodynamics of Business Cycles: A Comparative Evaluation. Oxford: Basil Blackwell. Drysdale, Peter & Yiping Huang (1995) 'Growth, energy and the environment: New challenges for the Asian-Pacific economy', Asian-Pacific Economic Literature. 9 (2): 1-12. Eckstein, Alexander (1977) China's Economic Revolution. Cambridge: Cambridge University Press. Eckstein, Alexander (1968) 'Economic fluctuations in communist China's domestic development', pp. 691-729 in: P.T. Ho & T. Tsou (eds) China in Crisis. Chicago: University of Chicago Press. Economic Situation and Prospect of China (Zhongguo jingji xingshi yu zhanwang) 1996-1997. (1997) Beijing: China Development Press. Economic Situation and Prospect of China (Zhongguo jingji xingshi yu zhanwang) 1993-1994. (1994) Beijing: China Development Press. Economic Situation and Prospect of China (Zhongguo jingji xingshi yu zhanwang) 1992-1993. (1993) Beijing: China Development Press. Economic Situation and Prospect of China (Zhongguo jingji xingshi yu zhanwang) 1991-1992. (1992) Beijing: China Development Press. Engle, R.F. & C.W.J. Granger (1987) 'Cointegration and error correction: Representation, estimation, and testing', Econometrica, 55(2): 251-76.
266
Bibliography
Engle, R.F. & C.W.J. Granger (1991) Long-Run Economic Relationships: Readings in Cointegration. Oxford: Oxford University Press. Engle, R.F. & D.F. Hendry (1993) 'Testing super exogeneity and invariance in regression models', Journal of Econometrics. 56: 119-39. Engle, R.F. & B.S. Yoo (1987) 'Forecasting and testing in co-integrated systems ', Journal of Econometrics, 35:143—5 9 Engle, R.F. & B.S. Yoo (1991) 'Cointegrated economic time series: An overview', pp. 237-66 in: R.F. Engle & C.W.J. Granger, Long-Run Economic Relationships: Readings in Cointegration. Oxford: Oxford University Press. Engle, R.F., D.F. Hendry & J.-F. Richard (1983) 'Exogeneity', Econometrica, 51(2): 277-304. Ericsson, Neil R. (1992) 'Cointegration, exogeneity, and policy analysis: An overview', Journal ofPolicy Modelling, 14(3): 251-80. Fischer, Stanley (1988) 'Recent developments in macroeconomics', Economic Journal, 98 (June): 294-339. Fishman, G.S. (1969) Spectral Methods in Econometrics. Cambridge, USA: Harvard University Press. FitzGerald, E.V.K. (1993) The Macroeconomics of Development Finance: A Kaleckian Analysis of the Semi-Industrial Economy. Basingstoke and London: Macmillan. Friedman, M. (1957) A Theory of the Consumption Function. Princeton, USA: Princeton University Press. Frisch, Ragnar (1928) 'Changing harmonics and other general types of components in empirical series', Skandinavisk Aktuarietidskrift, 11: 220-36. Fisch, Ragnar (1933) 'Propagation problems and impulse problems in dynamic economies', in Economic Essays in Honor ofGustav Cassel, pp.
171-205. Fuller, W.A. (1976) Introduction to Statistical Time Series. New York: Wiley. Garnaut, Ross & Guonan Ma (1992) 'Economic growth and stability in China', Journal ofAsian Economics, 4(1): 5-24. Goodwin, R.M. (1953) 'The problem of trend and cycle', Yorkshire Bulletin of Economic and Social Research, reprinted pp. 112-21 in: R.M. Goodwin (1982) Essays in Economic Dynamics. Basingstoke & London: Macmillan. Goodwin, R.M. (1967) 'A growth cycle', in: C.H. Feinstein (ed.), Socialism, Capitalism and Economic Growth: Essays Presented to Maurice Dobb. Cambridge: Cambridge University Press, reprinted pp. 165-70 in: R.M. Goodwin (1982) Essays in Economic Dynamics. Basingstoke and London: Macmillan.
Bibliography
267
Goodwin, R.M. (1982) Essays in Economic Dynamics. Basingstoke & London: Macmillan. Goodwin, R.M., M. Kruger & A. Vercelli, (eds) (1984) Nonlinear Models of Fluctuating Growth. New York: Springer-Verlag. Granger, C.W.J. (1969) 'Investigating causal relations by econometric models and cross-spectral methods', Econometrica, 37: 424-38. Granger, C.W.J. (1981) 'Some properties of time series data and their use in econometric model specification', Journal ofEconometrics, 16: 121-30. Granger, C.W.J. & J. Hallman (1991) 'Long memory series with attractors', Oxford Bulletin of Economics and Statistics. 53(1): 11-26. Granger, C.W.J. & P. Newbold (1974) 'Spurious regressions in Econometrics', Journal of Econometrics, 2:111-20. Grosfeld, I. (1987) 'Modelling planners' investment behaviour', Journal of Comparative Economics. 11(1): 180-91. Groves, T., Y. Hong, J. McMillan & B. Naughton (1994). 'Autonomy and Incentives in Chinese State Enterprises,' Quarterly Journal of Economics, 109(1): 183-209. Gu, Edward X. (1999) 'From permanent employment to massive lay-offs: The political economy of "transitional unemployment" in urban China (1993-1998)', Economy and Society (forthcoming). Guan, S. & H. Jiang (1990) Economics of the 'Blocks': An Analysis on Economic Behaviour of China's Local Governments (Kuaikuai Jingjixue). Beijing: Haiyang Press. Gurley, J.G. & E.S. Shaw (1967) 'Financial structure and economic development', Economic Development and Cultural Change, 15(3): 257-68. Hare, P.G. (1982) 'Economics of shortage and non-price control: A review article', Journal of Comparative Economics, 6: 406-25. Hare, P.G. (1989) 'Theoretical foundations: The economics of shortage', pp. 49-81 in: C. Davis & W. Charemza (eds), Models of Disequilibrium and Shortage in Centrally Planned Economies. London: Chapman and Hall. Harvey, A.C. (1990) The Econometric Analysis of Time Series (2nd edn) London: Philip Allan. Howard, Maurice (1989) 'Transport and communications', pp. 77-93 in: David S.G. Goodman (ed.) China's Regional Development. London: Routledge. He, Jianzhang (1984) Problems of Planned Administration in China (Zhongguo Jihua Guanli Wenti). Beijing: China Social Science Press. Hendry, D.F. (1986) 'Econometric modelling with cointegrated variables: An overview', Oxford Bulletin of Economics and Statistics, 48(3): 20112.
268
Bibliography
Hendry, D.F. (1988) 'The encompassing implications of feedback versus feedforward mechanisms in Econometrics', Oxford Economic Papers, 40: 132-49. Hendry, D.F. & G.E. Mizon (1993) 'Evaluating dynamic econometric models by encompassing the VAR', pp. 272-300 in: P.C.B. Phillips (ed.) Models, methods, and applications of econometrics: Essays in honour of Rex Bergstrom. Oxford: Basil Blackwell. Hoeffding, Oleg (1959) 'State planning and forced industrialization', Problems of Communism, 8: 38-46 (November/December). Holden, D. & R. Perman (1994) 'Unit roots and cointegration for the economist', pp. 97-112 in: B. Bhaskara Rao (ed.) Cointegration: For the Applied Economist. Basingstoke & London: Macmillan. Houthakker, H. & L. Taylor (1970) Consumer Demand in the United States (2nd ed.). Cambridge, USA: Harvard University Press. Hsiao, C. (1979) 'Autoregressive modelling of Canadian money and income data', Journal of the American Statistical Association, 47(3): 553-60 (September). Hsiao, C. (1981) 'Autoregressive modelling and money-income causality detection', Journal of Monetary Economics, 20(1): 85-106 (January). Hu, Xinli (1994) '1994 Reforms have no romantic melody', Reform (Gaige), 1 (January). Hu, Zhaoyi (1990) 'Causes and the relaxation countermeasures for the serious shortage of electricity', China Energy, 5: 18-21 (October). Hua, Guofeng (1978) 'Report on the work of the government', Beijing Review, 10 March. Hussain, Athar (1995) 'Discussion on "Energy pricing for sustainable development in China"', pp.230-35 in: Ian Goldin and L. Alan Winters (eds), The Economics of Sustainable Development, Cambridge: Cambridge University Press. Hussain, Athar & Nicholas Stern (1991) 'Effective Demand, Enterprise Reforms and Public Finance in China,' Economic Policy, 12: 141-78. Hussain, Athar & Nicholas Stern (1992) 'Economic Reforms and Public Finance in China,' Public Finance, 47 (Supplement): 289-317. Ickes, B.W. (1986) 'Cyclical fluctuations in centrally planned economies: A critique of the literature', Soviet Studies, 38(1): 36-52. Imai, H. (1990) 'Business Cycles in China: 1956-1986', PhD dissertation, Baltimore, USA: Johns Hopkins University. Imai, H. (1994a) 'China's endogenous investment cycle', Journal of Comparative Economics, 19(2): 188-216. Imai, H. (1994b) 'Inflationary pressure in China's consumption goods market: Estimation and analysis', Developing Economies, 32(2): 127-53.
Bibliography
269
Ishiguro, M. & Akiyama, T. (1995) 'Energy demand in five major Asian developing countries: Structure and prospects', World Bank Working Papers, N° 277. Washington DC: World Bank. Ishikawa, S. (1983) 'China's economic growth since 1949: An assessment', China Quarterly, 94: 242-81 (June). Jansen, Karel (1990) 'Finance, Growth and Stability: Financing Economic Development in Thailand', 1960-1984, PhD dissertation, Free University of Amsterdam, The Netherlands. Jefferson, Gary H. (1998) 'China's state enterprises: Public goods, externalities, and Coase,' American Economic Review, 88 (2): 428^132. Jefferson, G.H. & T.G. Rawski (1994) 'Enterprise reform in Chinese industry', The Journal of Economic Perspectives. 8(2): 47-70. Johansen, S. (1988) 'Statistical analysis of cointegration vectors', Journal of Economic Dynamics and Control, 12: 231-54. Johansen, S. (1992a) 'Cointegration in partial systems and the efficiency of single-equation analysis', Journal of Econometrics, 52: 389^02. Johansen, S. (1992b) 'Testing weak exogeneity and the order of cointegration in UK money demand data', Journal of Policy Modelling, 14(3): 313-34. Johansen, S. (1992c) 'Testing structural hypotheses in a multivariate cointegration analysis of the PPP and the UIP for UK', Journal of Econometrics, 53: 211-44. Johansen, S. (1992d) 'Determination of cointegration rank in the presence of a linear trend', Oxford Bulletin Of Economics and Statistics, 54: 383-97. Johansen, S. & K. Juselius (1990) 'Maximum likelihood estimation and inference on cointegration - With applications to the demand for money', Oxford Bulletin of Economics and Statistics, 52: 169-210. Kaldor, N. (1955) 'Model of distribution', Review of Economic Studies, 23(1): 94-100. Kalecki, M. (1970) 'Theories of growth in different social systems', Scientia, 40: 1-6. Kalecki, M. (1972) Selected Essays on the Economic Growth of the Socialist and the Mixed Economy. Cambridge: Cambridge University Press. Kalecki, M. (1976) Essays on Developing Economies. Hemel Hampstead: Harvester Press. Kambara, Tatsu (1992) 'The energy situation in China', China Quarterly, 131: 608-36 (September). Karshenas, M. (1993) 'Intersectoral resource flows and development: Lessons of past experience', pp. 179-238 in: A. Singh & H. Tabatabai (eds) Economic Crisis and Third World Agriculture. Cambridge: Cambridge University Press.
270
Bibliography
Kendall, M. & A. Stuart (1977) The Advanced Theory of Statistics: Distribution Theory, vol. 1,4th edn, London: Charles Griffin. Khan, A.R., K. Griffin, C. Riskin & R. Zhao (1993) 'Household income and its distribution in China', pp. 25-73 in: K. Griffin & R. Zhao (eds) The Distribution ofIncome in China. Basingstoke & London: Macmillan. Knight, John (1995) 'Price scissors and intersectoral resource transfers: Who paid for industrialization in China?' Oxford Economic Papers. 47: 11735. Kornai, J. (1979) 'Resource-constrained versus demand-constrained systems', Econometrica, 47(4): 801-19. Kornai, J. (1980) Economics of Shortage. Amsterdam: North-Holland. Kornai, J. (1982) Growth, Shortage and Efficiency. Oxford: Basil Blackwell. Kornai, J. (1992) The Socialist System. Oxford: Oxford University Press. Kornai, J. (1995) Highway and Byway: Studies on Reform and Post-Communist Transition. Cambridge, USA: The MIT Press. Kornai, J. & B. Martos (eds) (1981) Non-Price Control. Amsterdam: NorthHolland. Kornai, J. & A. Simonoviets (1986) 'Investment, efficiency and shortage: A macro growth model', MATEKON, 24: 3-21. Kremers J.J.K., N.R. Ericsson & J.J. Dolado (1992) 'The power of cointegration tests', Oxford Bulletin of Economics and Statistics, 54(3): 32548. Krueger, A.O. (1974) 'The Political Economy of Rent-Seeking Society', American Economic Review. 64: 291-303 (March). Kueh, Y.Y. (1984) 'China's food balance and the world grain trade', Asian Survey, 24(12): 1251 (December). Kueh, Y.Y. (1995) Agricultural Instability in China, 1931-1991: Weather, Technology, and Institutions. Oxford: Oxford University Press. Kung, James K. (1992) 'Food and agriculture in post-reform China: the market surplus problem revisited', Modern China, 18: 138-170 (April). Lacko, Maria (1980) 'Cumulating and easing of tensions', Ada Oeconomica, 24(3-4): 357-77. Lacko, Maria (1984) 'Behaviourial rules in the distribution of sectoral investment in Hungary, 1951-1980', Journal of Comparative Economics, 8(3): 290-300. Lardy, Nicholas R. (1983) Agriculture in China's Modern Economic Development. Cambridge: Cambridge University Press. Lardy, Nicholas R. (1985) 'Agricultural prices in China', World Bank Staff Working Papers, N° 606, Washington DC: World Bank. Lardy, Nicholas R. (1994) China in the World Economy. Washington DC: Institute for International Economics.
Bibliography
271
Lardy, Nicholas R. (1995) 'The role of foreign trade and investment in China's economic transformation', China Quarterly, no. 144 (Dec): 1065— 82. Lavigne, Marie (1995) The Economics of Transition: From Socialist Economy to Market Economy. Basingstoke & London: Macmillan. Ledic, Michele (1989) 'The energy sector', pp. 94-111 in: David S.G. Goodman (ed.) China's Regional Development. London: Routledge. Lees, Francis A. (1997) China Superpower: Requisites for High Growth. New York: St. Martin's Press. Li, Chengrui (1984) 'Are the 1967-1976 statistics on China's economy reliable?' Beijing Review, 12: 21-26 (19 March). Li, Fuchen (1981) 'Control the scale of capital construction', China Finance (Zhongguo Jinrong), 6: 25. Li, Jingwen (1981) 'Material Balance in the National Economy', in Guoguang Liu (ed.) Several Theoretical Problems of the Comprehensive Balance of the National Economy (Guomin Jingji Zonghe Pingheng de Ruogan Lilun Wenti), Beijing: People's Press. Li, Kaixin (1981) 'Management of China's materials and resources', Almanac of China's Economy (Zhongguo Jingji Nianjian), 1981: iv & 124. Li, Kaixin (1983) 'Concentrate materials to ensure keypoint construction', Red Flag (Hongqi), 17: 16. Li, Ziyun & Yuenian Zhu (1992) 'Retrospect the past and prospect the future: An overview of coal industry in 1991', Energy of China (Zhongguo Nengyuan), 8: 1-3. Li, Zhining (1987) A Dictionary of Major Economic Events in the People's Republic of China: October 1949-January 1987. (in Chinese). Changchun: Jilin People's Press. Liang, Weilie (1989) 'Rethinking and analysis of the energy defect', Energy of China (Zhongguo Nengyuan), 2: 6-9 (February). Lin, Cyril Z. (1988) 'China's economic reforms II: Western perspectives', Asian Pacific Economic Literature, 2(1): 1-25. Lin, Cyril Z. (1989) 'Open-ended economic reform in China', pp. 95-136 in: V. Nee & D. Stark (eds) Remaking the Economic Institutions of Socialism: China and Eastern Europe. Stanford: Stanford University Press. Lin, Justin Y. (1987) 'The household responsibility system reform in China: A peasant's institutional choice', American Journal of Agricultural Economics, 69(2): 410-15. Lin, J.Y. (1989) 'Collectivization and the 1960s agricultural crisis in China', paper presented in the International Conference on 'Methods of Planning and Policy Analysis for Mixed Economies', Indira Gandhi Institute of Development Research, Bombay, 5-7 January.
272
Bibliography
Lin, Justin Y. (1990) 'Collectivization and China's agricultural crisis in 1959-1961', Journal ofPolitical Economy, 98(6): 1228-52. Lin, Justin Y. (1992) 'Rural reforms and agricultural growth in China', American Economic Review, 82(1): 34-51. Lin, Justin Y. (1993) 'Exit rights, exit costs, and shirking in agricultural cooperatives: A reply', Journal of Comparative Economics, 17(2): 504-20. Lin, Justin Y., F. Cai & Z. Li (1996) The China Miracle: Development Strategy and Economic Reform. Hong Kong: The Chinese University of Hong Kong Press. Lin, Justin Y., F. Cai & Z. Li (1995) 'Why China's economic reforms have been successful: Its implications for other reforming economies', China Centre for Economic Research Working Paper, N° E1995002, Beijing: Peking University. Lin, S. (1990) China's Industrial Structure and the Present Industrial Adjustment (in Chinese). Beijing: Institute of World Economics and Politics, Chinese Academy of Social Science. Lin, Xiannuan & Karen R. Polenske (1995) 'Input-output anatomy of China's energy use changes in the 1980s', Economic Systems Research, 7(1): 67-84. Liu, F., W.B. Davis & M.D. Levine (1992) An Overview of Energy Supply and Demand in China. Berkeley: Lawrence Berkeley Laboratory. Liu, Guoguang & Ruisun Wang (1982) Reform of China's Economic System (Zhongguo deJingji Tizhi Gaige). Beijing: Social Science Press. Liu, Jing-Tong (1987) 'China's energy and economic development', Asia Pacific Economies: Promises and Challenges — Research in International Business and Management, 6(B): 53-63. Liu, Minquan (1993) 'Exit right, retaliatory shirking, and the agricultural crisis in China', Journal of Comparative Economics. 17(2): 540-59. Liu, R. & B. Zou (1992). 'We must smash the problem of enterprises taking responsibility for profits and not for losses,' Economic Research (jingji Yanjiu), 12: 22-24. Lou, Jiwei (1992) 'On division of economic power between central and local levels', Chinese Economic Studies, 25(4): 25-34. Lu, Y. (1993) Fuelling One Billion: An Insider's Story of Chinese Energy Policy Development. Washington DC: Washington Institute Press. Lucas, Robert E. Jr. (1973) 'Some international evidence on output-inflation tradeoffs'. American Economic Review, 63 (June): 326-34. Lucas, Robert E. Jr. (1981) Studies in Business Cycle Theory. Oxford: Basil Blackwell. Lucas, Robert E. Jr. (1987) Models of Business Cycle: Yrjo Jahnsson Lectures, May 1985. Oxford: Basil Blackwell.
Bibliography
273
Luo, Rucheng (1979) 'Make full use of small-scale loans', Economic Administration (Jingji Guanli), 12: 28-30. MacKinnon, J. (1991) 'Critical values for cointegration tests', pp. 267-76 in R.F. Engle & C.WJ. Granger (eds) Long-Run Economic Relationships: Readings in Cointegration. Oxford: Oxford University Press. Maddala, G.S. (1983). Limited-dependent and Qualitative Variables in Econometrics. Cambridge: Cambridge University Press. Mao, Zedong (1956) 'On the ten major relationships', pp. 284-307 in Mao, Zedong (1977) Selected Works of Mao Zedong, vol. 5. Beijing: Foreign Languages Press. Mao, Zedong (1977) Selected Works ofMao Zedong, vol. 5. Beijing: Foreign Languages Press. Mankiw, N.G. (1985) 'Small menu costs and large business cycles: A macroeconomic model of monopoly', Quarterly Journal of Economics, 100 (2): 329-33. Mankiw, N.G. & D. Romer (1991) New Keynesian Economics. Cambridge, USA: The MIT Press. Marrese, M. & J.L. Mitchell (1984) 'Kornai's resource-constrained economy: A survey and an appraisal', Journal of Comparative Economics, 8(1): 74-84. Martos, B. (1990) Economic Control Structures. Amsterdam: North-Holland. McCallum, B.T. (1989) 'Real business cycle models', pp. 16-50 in: R.J. Barro (ed.) Modern Business Cycle Theory. Cambridge, USA: Harvard University Press. McKinnon, R.I. (1993) 'Market-preserving fiscal federalism', Working Paper, Department of Economics, Stanford University. McKinnon, R.I. (1994) 'Financial growth and macroeconomic stability in China, 1978-1992: Implications for Russia and other transitional economies', Journal of Comparative Economics, 18(3): 438-69. Mihalyi, P. (1992) Socialist Investment Cycles: Analysis in Retrospect. London: Kluwer Academic. Milgrom, Paul (1988) 'Employment contracts, influence activities, and efficient organization design', Journal ofPolitical Economy, 96(1): 42-60. Milgrom, Paul & John Roberts (1988) 'An economic approach to influence activities in organizations', American Journal of Sociology, 94 (supplement): si 54-79. Ministry of Agriculture - Study Group of Rural Cooperative Economy (1991), 'A Survey of Contractual Arrangements and Cooperative Organisation in Rural China.' Issues in Agricultural Economy (Nongcun Jingji Wenti), no. 8 (Aug.): 33-40.
274
Bibliography
Ministry of Agriculture (1996) A Report on China's Agricultural Development in 1996. Beijing: China Agriculture Press (in Chinese). Morgan, Mary S. (1990) The History of Econometric Ideas. Cambridge: Cambridge University Press. Montinola, Gabriella, Yingyi Qian & Barry R. Weingast (1995) 'Federalism, Chinese style: The political basis for economic success in China,' World Politics, 48: 50-81 (October). Naughton, B.J. (1986) Saving and Investment in China: A Macroeconomic Analysis. PhD dissertation, New Haven, USA: Yale University. Naughton, B.J. (1987) 'Macroeconomic policy and response in the Chinese economy: The impact of the reform process', Journal of Comparative Economics, 11(3): 334-53. Naughton, B.J. (1992) 'Hierarchy and the bargaining economy: Government and enterprise in the reform process', pp. 245-82 in: K.G. Lieberthal & D.M. Lampton (eds) Bureaucracy, Politics, and Decision-Making in Post-Mao China. Berkeley, USA: University of California Press. Naughton, B.J. (1995a) Growing out of the Plan: Chinese Economic Reform, 1978-1993. Cambridge: Cambridge University Press. Naughton, B.J. (1995b) 'China's macroeconomy in transition', China Quarterly, 144: 1083-104 (December). Nelson, C. & H. Kang (1981) 'Spurious periodicity in inappropriately detrended time series', Econometrica, 49: 741-51. Nelson, C. & C. Plosser (1982) 'Trends and random walks in macroeconomic time series', Jo urnal of Econometrics, 10: 139-62. Oppers, Erik S. (1997) 'Macroeconomic cycles in China', IMF Working Paper, WP/97/135. Washington DC: International Monetary Fund. Osiatynski, J. (1988) Michal, Kalecki on a Socialist Economy. Basingstoke & London: Macmillan Press. Osterwald-Lenum, M. (1992) 'A note with quantiles of the asymptotic distribution of the maximum likelihood cointegration rank test statistics: Four cases', Oxford Bulletin of Economics and Statistics, 54: 461-71. Ouliaris, Park & Phillips (1988) 'Testing for a unit root in the presence of maintained trend', College Park MD, USA: University of Maryland, Department of Economics working paper series 1988-37. Parker, David & Weihwa Pan (1996) 'Reform of the state-owned enterprises in China', Communist Economies and Economic Transformation, 8(1): 109-27. Patnaik, Prabhat (ed.) (1995) Macroeconomics. Delhi: Oxford University Press. Peebles, G. (1991) Money in the People's Republic of China: A Comparative Perspective. Sydney: Allen & Unwin.
Bibliography
275
Peebles, G. (1992) 'Why the quantity theory of money is not applicable to China, together with a tested theory that is', Cambridge Journal of Economics, 16(1): 2 3 ^ 2 . Peng, Fangchun (1989) 'Important coimtermeasures for the relaxation of the current shortage of electricity', Energy of China (Zhongguo Nengyuan), 6: 31-4 (June). Perkins, D. (1994) 'Completing China's move to the market', The Journal of Economic Perspectives, 8(2): 23-46. Perotti, Enrico, Laixiang Sun & Liang Zou (1998) 'State-owned versus Township and Village Enterprises in China,' Working Paper No. 150, UNU/ WIDER, Helsinki. Perron, Pierre (1989) 'The great crash, the oil price shock and the unit root hypothesis', Econometrica, 57: 1361-1401. Perron, Pierre (1993) 'The great crash, the oil price shock and the unit root hypothesis: Erratum', Econometrica, 61(1): 248-9. Perron, Pierre (1994) 'Trend, unit root and structural change in macroeconomic time series', pp. 113-46 in: B. Bhaskara Rao (ed.) Cointegration: For the Applied Economist. Basingstoke and London: Macmillan. Phillips, P.C.B. (1986) 'Understanding spurious regressions in Econometrics', Journal of Econometrics, 33: 311^0. Portes, R. & A. Santorum (1987) 'Money and the consumption goods market in China', Journal of Comparative Economics, 11(3): 354-71. Prescott, Edward C. (1986) 'Theory ahead of business cycle measurement', in: K. Brunner & A.H. Meltzer (eds) Real Business Cycles, Real Exchange Rates and Actual Policies. Carnegie-Rochester Conference Series on Public Policy, vol. 25, Amsterdam: North Holland (autumn). Putterman, L. & G. Skillman (1993) 'Collectivization and China's agricultural crisis', Journal of Comparative Economics, 17(2): 530-39. Qian, Yingyi (1994) 'A theory of shortage in socialist economies based on the "soft budget constraint" ', The American Economic Review, 84(1): 145-56. Qian, Yingyi (1995) 'Reforming Corporate Governance and Finance in China', in: Masahiko Aoki & Hyung-Ki Kim (eds), Corporate Governance in Transitional Economies: Insider Control and the Role of Banks, Washington, DC: The World Bank. Qian, Yingyi & Gerard Roland (1999) 'Federalism and the soft budget constraint', American Economic Review, 88(5): 1143-62. Qian, Yingyi & Barry R. Weingast (1997) 'Federalism as a commitment to preserving market incentives', Journal of Economic Perspectives, 11 (4): 83-92.
276
Bibliography
Qian, Y. & C. Xu (1993a) 'Why China's economic reform differ: The Mform hierarchy and entry/expansion of the non-state sector', Discussion Paper N° 154, Centre for Economic Performance, London: London School of Economics and Political Science. Qian, Y. & C. Xu (1993b) 'The M-form hierarchy and China's economic reform', European Economic Review, 37: 541-48. Rakshit, M. (ed.) (1989) Studies in the Macroeconomics of Developing Countries. Delhi: Oxford University Press. Reimers, H.-E. (1992) 'Comparisons of tests for multivariate cointegration', Statistical Papers, 33: 335-59. Reinsel, G.C. & S.K. Ahn (1992) 'Vector autoregressive models with unit roots and reduced rank structure: Estimation, likelihood ratio test, and forecasting', Journal of Time Series Analysis, 13: 353-57. Riskin, Carl (1987) China's Political Economy: The Quest for Development since 1949. Oxford: Oxford University Press. Roland, Gerard (1987) 'Investment growth fluctuations in the Soviet Union: An econometric analysis', Journal of Comparative Economics. 11(2): 192-206. Sachs, I. (1995) 'Russia's struggle with stabilization: Conceptual issues and evidence', Proceedings of the World Bank Annual Conference on Development Economics 1994. Washington DC: World Bank. Saith, Ashwani (1985) 'Primitive Accumulation, agrarian reform and socialist transitions: An argument', pp. 1-48 in: A. Saith (ed.) The Agrarian Question in Socialist Transitions. London: Frank Cass. Saith, Ashwani (1995) 'From collectives to markets: Restructured agricultural-industrial linkages in rural China: Some micro-level evidence', Journal of Peasant Studies, 22(2): 201-60. Semmler, W. (ed.) (1989) Financial Dynamics and Business Cycles: New Perspectives. Armonk NY, USA: M.E. Sharpe. Sheng, Yuming (1993a) Intersectoral Resource Flows and China's Economic Development. New York: St. Martin's Press. Sheng, Yuming (1993b) 'The capital sources of China's industrialization', The Developing Economies, 31(2): 173-207. Shih, Chih-Yu (1992) 'Chinese managers' motivation for investment: Reflections on some interviews', China Information, 7(2): 47-55. Sicular, T. (1989) 'China: Food pricing under socialism', pp. 243-88 in: T. Sicular (ed.) Food Price Policy in Asia: A Comparative Study. Ithaca NY, USA: Cornell University Press. Sicular, T. (1993) 'Ten years of reform: Progress and setbacks in agricultural planning and pricing', pp. 47-96 in: Y.Y. Kueh & R.F. Ash (eds) Eco-
Bibliography
277
nomic Trends in Chinese Agriculture: The Impact of Post-Mao Reforms. Oxford: Clarendon Press. Sicular, T. (1995) 'Redefining state, plan and market: China's reforms in agricultural commerce', China Quarterly, 144: 1020^6 (December). Simonovits, A. (1991) 'Investment, starts and cycles in socialist economies: A mathematical model', Journal of Comparative Economics, 15(3): 460— 75. Simonovits, A. (1992) Cycles and Stagnation in Socialist Economies: A Mathematical Analysis. Oxford: Basil Blackwell. Singh, I. (1992) China: Industrial Policies for an Economy in Transition. World Bank Discussion Paper, N° 143. Washington DC: World Bank. Skinner, G. & E. Winkler (1969) 'Compliance succession in rural communist China: A cyclical theory', pp. 4110-38 in: A. Etzioni (ed.) A Sociological Reader on Complex Organizations. New York: Holt, Rinehardt, & Winston. Slutsky, Eugen E. (1927) 'The summation of random causes as the source of cyclic processes', The Problems of Economic Conditions, (ed. by the Conjuncture Institute, Moscow), 3(1): 34-64 (English summary, pp. 15661) Slutsky, Eugen E. (1937) 'The summation of random causes as the source of cyclic processes', Econometrica, pp. 105^6. Smil, Vaclay (1988) Energy in China's Modernization Advances and Limitations. London: M.E. Sharpe. Soos, Karoly A. (1975) 'Causes of investment fluctuations in the Hungarian economy', Eastern European Economics, 14(2): 25-36. Spanos, Aris (1986) Statistical Foundations of Econometric Modelling. Cambridge: Cambridge University Press. Spanos, A. (1990) 'Towards a Unifying Methodological Framework for Econometric Modelling', pp. 335-64 in: C.W.J. Granger (ed.) Modelling Economic Series: Readings in Econometric Methodology. Oxford: Oxford University Press. Spanos, Aris (1995) 'On theory testing in Econometrics: Modelling with nonexperimental data', Journal of Econometrics, 76(1): 189-226. Stock, J.H. & M.W. Watson (1988) 'Variable trends in economic time series', Journal of Economic Perspectives, 2(3)147-74. Sun, Haishun & Frank B. Tipton (1998) 'A comparative analysis of the characteristics of direct foreign investment in China, 1979-1995', Journal of Developing Areas, 32 (2): 159-86. Sun, Laixiang (1999) The Evolutionary Dynamics of China's Small and Medium Enterprises in the 1990s. World Development Studies Series, UNU/WIDER, Helsinki (forthcoming).
278
Bibliography
Sun, Laixiang (1998a) 'Estimating investment functions based on cointegration: The case of China', Journal of Comparative Economics, 26 (1): 175-91. Sun, Laixiang (1998b) 'Unorthodox ownership and governance structures in East Asia', MOCT-MOST: Economic Policy in Transitional Economies, 8 (1): 33-62. Sun, Laixiang (1996) 'Rationing, coercion, and price scissors: An extended framework for understanding primitive socialist accumulation', mimeo. Institute of Social Studies, The Hague, The Netherlands. Sun, Laixiang (1991) 'Agricultural development and industrial accumulation in China', Lecture Notes (in Chinese), Economics College of Peking University, Beijing, China. Sun, Laixiang (1989) 'The special contradictions in our economy and the key choices for our reform', Qiushi (Theoretical Journal of the Central Committee of CPC), 4: 27-31 (February). Sun, Laixiang, Gunther Fischer & Peter Albersen (1998) 'Establishment of Agricultural Production Relations in LUC Model: The case of China', Interim Report, International Institute for Applied Systems Analysis, Laxenburg, Austria. Sung, Ynwing & Thomas Chan (1987) 'China's economic reforms I: The debates in China', Asian Pacific Economic Literature, 1(1): 1—24. Survey Report by the Research Group of Policy Orientation in Vitalizing Large and Medium SOEs (1992), China Social Sciences (Zhongguo ShehuiKexue),(3). Tang, A.M. & CJ. Huang (1982) 'Changes in input-output relations in the agriculture of the Chinese mainland, 1952-1979', pp. 319^8 in: C. Hou & T. Yu (eds) Agricultural development in China, Japan and Korea. Taipei: Academia Sinica. Tang, A.M. & B. Stone (1980) 'Food production in the People's Republic of China', Research Report 15, Washington DC: International Food Policy Research Institute. Thirlwall, A.P. (1974) Inflation, Saving and Growth in Developing Economies. Basingstoke and London: Macmillan. Urbain, J.-P. (1992) 'On weak exogeneity in error correction models', Oxford Bulletin ofEconomics and Statistics. 54(2): 187-207. Vellupillai, K. (ed.) (1990) Nonlinear and Multisectoral Macrodynamics: Essays in Honour of Richard Goodwin. New York: New York University Press. Walder, Andrew (1984) 'The remarking of the Chinese working class, 19491981', Modern China, 10(1): 3-^8 (January).
Bibliography
279
Walker, Kenneth R. (1982) 'Interpreting Chinese grain consumption statistics', China Quarterly, 92: 575-88 (December). Wang, Jiye & Yuanzhen Zhu (eds) (1987) Handbook on Reform of the Economic System (Zhongguo Jingjitizhi Gaige Shouce). Beijing: Planning Press. Watson, Andrew (1994) 'China's economic reforms, 1987-1993: Growth and cycles', Asian Pacific Economic Literature. 8(1): 48-65. Weimer, Calla (1992) 'Price reform and structural change: distributional impediments to allocational gains', Modern China, 18: 171-196 (April). Wong, Christine (1985) 'Material allocation and decentralization: Impact of the local sector on industrial reform', pp. 253-78 in: Elizabeth J. Perry & Christine Wong (eds) The Political Economy of Reform in Post-Mao China. Cambridge USA: Harvard University, Council on East Asian Studies. Wong, Christine (1987) 'Between plan and market: The role of the local sector in post-Mao China', Journal of Comparative Economics. 11(3): 38598. Wong, Christine (1991) 'Central-local relations in an era of fiscal decline: The paradox of fiscal decentralization in post-Mao China', The China Quarterly. 128: 691-715. Wong, Christine (1992) 'Fiscal reform and local industrialization', Modern China, 18(2): 197-227. World Bank (1983) China: Socialist Economic Development (3 vols), World Bank Country Study. Washington DC: World Bank. World Bank (1985) China: The Energy Sector. Washington DC: World Bank. World Bank (1988a) China: Finance and Investment. Washington DC: World Bank. World Bank (1988b) China: External Trade and Capital. Washington DC: World Bank. World Bank (1990) China: Between Plan and Market. Washington DC: World Bank. World Bank (1991) China: Options for Reform in the Grain Sector. World Bank Country Study, Washington DC: World Bank. World Bank (1994) China: Macroeconomic Stability in a Decentralized Economy. World Bank Country Economic Memorandum, Washington DC: World Bank. World Bank (1997) China's Management of Enterprise Assets: The State as Shareholder. Washington DC: World Bank. Wu, Zongxin, with members of the INET (1988) China's Energy Demand by 2050 (in Chinese). Beijing: Institute of Nuclear Energy Technology.
280
Bibliography
Wuyts, Marc (1988) 'The food balance and economic growth: An appraisal of FitzGerald's reformulation of Kalecki', Development and Change, 19(4): 649-68. Xu, Guohua & LJ. Peel (1991) The Agriculture of China. Oxford: Oxford University Press. Xu, J.5 Z. Zhang, Z. Yang & B. Zhu (1980) 'On energy construction for China's modernization', Journal of Natural Dialects (Ziran Bianzhengfa Tongxun),4: 16-22. Xu, J. & Z. Huang (1981) 'The way to solve rural energy problems', People 's Daily (Remin Ribao), 27 October, p. 5. Xu, Xiaonian & Yan Wang (1997) 'Ownership structure, corporate governance, and firm's performance: The case of Chinese stock companies', Policy Research Working Paper, World Bank, Washington DC. Yang, Jianbai (1991) 'Velocity, structure, and efficiency', Jingji Yanjiu {Economic Research), 9 (September, 1991). Yeh, K.C. (1984) 'Macroeconomic changes in the Chinese economy during the readjustment', China Quarterly, 100: 691-716 (December). Yenny, J. & L.V. Uy (1985) 'Transport in China: A comparison of basic indicators with those of other countries', World Bank Staff Working Papers, N° 732. Washington DC: World Bank. Young, Susan (1992) 'Ownership and obligation in China's rural enterprises: Community interests in a non-legalistic environment', pp. 42-9 in: Y. Wu and X. Zhang (eds), China Economy in Transition. Canberra: National Centre for Development Studies, Australian National University. Yu, Weizhou (1992) 'Historic review on the fluctuation of the development of China's power industry', China Energy, 4: 1-4 (April). Yu, Xiaogu (1980) 'Implement the merger of planned and market adjustment to liven up market circulation', Economic Management (Jingji Guanli), 2:4. Yule, G.U. (1926) 'Why do we sometimes get nonsense correlations between time series? A study in sampling and the nature of time series', Journal of the Royal Statistical Society, 84: 1-64. Yusuf, S. (1994) 'China's macroeconomic performance and management during transition', The Journal of Economic Perspectives. 8(2): 71-92. Zhang, Jiwu (1991) 'Receive lessons from the Seventh Five Years Plan period, and make good coordinative development', Energy of China (Zhongguo Nengyuan), 10: 6-8 (October). Zhang, Shushan (1979) 'Some views about reforms in commodity control', Economic Management {Jingji Guanli), 8: 15—18.
Bibliography
281
Zhang, Weiying & Gang Yi (1995) 'China's gradual reform: A historical perspective', China Centre for Economic Research Working Paper N° E1995001, Beijing: Peking University. Zheng, Mingyue (1992) 'Attach importance to economy of coal resource', Energy of China (Zhongguo Nengyuan), 10: 6-8 (October). Zhou, Dadi & Ming Chu (1992) 'Potential crisis of China's energy', Energy of China (Zhongguo Nengyuan), 2: 4—5 & 3: 5-10. Zhou, Enlai (1965) 'Report on the work of the government to the First Session, Third National People's Congress', Peking Review, 1(1): 6-20 (January). Zhou, Fengqi, Shiyuan Qu, Wenke Han & Zhengming Su (1990) 'Countermeasures on alleviation of the overall tightness of national energy supply', Energy of China (Zhongguo Nengyuan), 3: 4-9 (April). Zou, Liang & Laixiang Sun (1996) 'Interest rate policy and incentives of state-owned enterprises in the Transitional China', Journal of Comparative Economics, 23 (3): 292-318.
Index
'Above-norm' projects 79 Adj ustment and re-adj ustment 1 5,13,18 macro-economic 1-5, 30, 48, 57,70,112-8,124-57,249 structural 13, 18, 108-10 Adverse distributive impact 2, 46 consequences 3, 28,30 Agricultural development and industrial expansion 130 Agricultural terms of trade 46-7, 54-5, 70, 129, 143, 230 in kind 129 Agricultural value-added 6 Agriculture first 134 Anti-Rightist Campaign 131 Anti-Right Deviation 131 Asymptotically stationary 210 AutoRegressive Conditional Heteroscedasticity (ARCH) test 205,251 Attractor 5-6,21-4,189-91,207, 223, 247 Autocorrelation 197, 205, 211 residual 205
model 18-24,35-8 theory 18-24,31,35-8 Behaviour assumption 4, 189, 243 Bottleneck constraint 2-7, 23-5, 28, 30, 38, 45-8, 55-6, 623,70, 113, 162-6,174,221, 242-4, 248 type of growth theory 2, 301,55 Bottleneck sectors 2, 7, 11, 15, 23,28,38,56,88,97,111, 143,201,206,222,228-9, 242 Brazil 179-80 Bureaucratic co-ordination 5, 678,81,221-2,232,243,251 failure 222,243 mechanism 5, 68 Business cycle theory 30, 64, 70 Capital accumulation 1, 6, 10, 27,36,52,54-5,70,117-8, 129-34,156,205,221,246 levels 1 mechanism 10, 45-7, 51, 54, 70,117,229 ratio 6,113, 129-34,137, 146-52, 161, 175, 177 Capital construction 73-83, 141, 252 projects 82
Bargaining power 8, 249 Barriers to growth 3, 46 Bauer 18,29,35-8,42-3,69,203 Bauer's four phase description 18-24,29,69, 248
282
Index Central Bank (People's Bank of China) 89-90, 96, 147, 240, 258 Central-local relations 235-7, 244 Centrally planned economy 1 -2, 10,38,41,72-3,81,99100, 102, 171 ChenJinhua 206 China International Engineering Consulting Company (CIECC) 80 China's development model 8, 11 China's state investment system 27,71,206,223,233,238 China's traditional economic system 60 Chinese economy 1, 52, 255 Chow test 28, 189, 202-3, 224, 243,251 one-step ahead 189,202-3, 224,243,251 'Clambering into the plan' 81 Cobweb Theorem 157 Cointegration, see also Attractor 2, 4-6, 20-2, 28, 56, 68, 113,184, 188-203,206-7, 213-9,222,242,250 analysis 22, 196-8,210,229, 231,250 approach 4-5,20-2,28,184, 188-203 norm 24 relation 5,23, 193-4,200-3, 207, 224 vector 198,213 Cointegrated of order 213 Collectivization 9, 47, 53, 123-4, 131, 249,255 Collinearity 189 Collusion 71, 82, 89-90, 105, 110,228,237
283
Commune and brigade enterprises 101 Co-movement, see also Attractor and Cointegration 3, 5-6, 11, 15,188,191, 197,2067,213,223-4,256 Conditional density 217 distributions 194 equation 194-5,201-2,21820 error correction model 23, 28, 189,194,224,243 process 217 variables 189 'Control by norm' 206 Constraints a key factor- 48-50 bottleneck See Bottleneck constraints labour- 46-48,50 Construction Bank of China 74, 86 Co-ordination 2, 5, 67-8, 81, 221-2,232,243,251 mechanism 2, 5, 68 see also Bureaucratic co-ordination Covariance matrix 193,217 Credit plans 72,76,85,244 Cultural Revolution 19, 26, 38, 101, 136-9, 141 Daqing 140 Data-generating process (DGP) 20-1 Decentralization 4, 12, 27, 60, 71,73-7,81,100-1,107, 157,226,232,234,236-9, 244 and re-centralization 4, 27, 71,73,76,96
284 Degrees of freedom 185,198, 211,241 Deng Xiaoping 138-9,146 Deng Whirlwind 19, 36, 146 Denationalization 238, 240, 244 in a broad sense 238, 240 in a narrow sense 238, 240 Developing economy 1-3,30-1, 45-6, 55, 65-6, 163, 205-6, 222-5, 242 countries 8-10,51,171,179 low-income 9 middle-income 9 socialist 3,30-1,46,65,2225,242 Development drive 72, 106 Dickey-Fuller test (DF) 192, 210-1,214 augmented- (ADF) 192-3, 196-7,211,214 Disequilibrium 3, 5, 38-9, 42, 188,192-3,201,203,207, 216,223 adjustment behaviour 3, 5, 39, 203, 207, 223 adjustment mechanism 188, 192-3,203 Distributive barrier 4-5, 35, 47, 201,204,222-3,225,2289, 242-3 theory 206,221 Domestic loans 77, 84 Dual subordination 110, 228 Dual track economy 61 price system 61 Dummy variables 59,197-8 Duplication of construction 227, 232, 243 Eastern Europe 1, 29, 67, 223, 230,233,238 Eckstein's theory 58,112
Index Economic reform 8, 11, 15, 63, 73, 76, 94, 142, 235 Effective energy supply 28, 47, 162, 166, 174, 179, 181, 183-7 Eigenvalue 198-9,215,219,257 Endogenous cycle approach 64-5 Energy commercial 10,163-4, 167 constraints, and shortage 7, 11,28,46-9,88,96,10910,113, 162,164,166, 174-5,183-7,230-1,241, 244 consumption, see also Effective energy supply 10,27,166,169-71,183-7, 195 intensity 169-71 modern 162-3, 178 rationing quota 164 traditional 162-3 England 173 Engle-Granger two-step estimator, and approach 214 Enterprise contract responsibility system 102 Equilibrium business cycle approach 63-4 Error correction, see also Cointegrtion approach 2,20, 184, 188 mechanism 22, 24, 48, 18890,216,222,242,250,257 model 23,68,192,213,216, 224, 242-3 representation 188, 192, 216 Excess kurtosis 196-7 Exogeneity strong 217-8,254 super 195,219-20, weak 22,30,44,187, 189, 195,200,217-9,231,254
Index Expansion drive 3, 7, 28-9, 31-4, 38, 42, 69, 86, 105, 108, 110-1,225,227-9 Factor proportions 118, 131, 152 Factorization 217-9 Fallacy of composition 40, 69 Feedback 39, 69, 194, 206-7, 223-4 Till the gap' 90,96 'Five small industries' 101 Five-year plans 52 first to eighth 75, 106, 167, 250 Foreign investment 24-5, 144 loans 77,84 'Foreign Leap Forward' 19, 36, 140 Formal procedure versus real practice 79,81 Former Soviet Union 1, 29, 45, 73, 98-100, 122-4, 163, 166,180,223,238,247-8 Four Modernizations 19,138 Fundamental tension 5, 7, 15 Gaussian asymptotic distribution 219 asymptotic theory 202 full-information maximum likelihood (FIML) cointegration procedure 194, 198 General to specific 20, 22 Gestation lags 41,233,243 Goodwin's growth cycle model 30-1,68,70 'Grain bag responsibility system' 156 Grain output per capita (GRNPC) 7, 126-7, 131, 142, 195, 198,206 Granger 28, 184,254
285
approach 184 causal analysis 184-5 Causality, and test 28,185-7, 231 see also Strong exogeneity Graphic test 198,200 Great Leap Forward 10, 19, 35, 38,59,74, 131-3,167, 177, 197,253 great famine 10 Growth cycle 11,13,20 framework 6, 20, 29-31, 221 model 3,70 theory 6-7 see also Goodwin's growth cycle model Hannan-Quinn information criteria 257 Hard budget constraint 4, 32, 239 Heavy industry-oriented development strategy 27, 51-2,54,61-2, 117 Heteroscedasticity test 205, 251 Homogeneity 198,202 Hungarian School 3-4, 6, 29-31, 38,42,48,55-6,69-70,221 India 51, 122-4, 128, 167, 170-1, 179-80,256 Indonesia 170-1 Industrial and Commercial Bank of China 86 Industrial concentration 233, 249 Inference 183, 189-90, 196, 202, 214,217,219 In-plan prices 171-2 Input-output relations 6, 66, 112, 124, 224 Integrated of order 191,210 Interest conflicts 71,222,243 rigidity 63, 174
286
vested 89, 174 Intersectoral distributive policy 55 Intertemporal inconsistency 3 Invariant, and invariance 4, 191, 195,212,219-20 Investment behaviour 2, 4, 7, 29, 66,221-2,225,242 aggregate 2,221-2,225,242 Investment commitments 30 engagement 35-6, 42 starts 5,39-41,43 vintages 39 Investment cycle first to seventh 17,19,13155 theory 30-1,205,221 Investment growth rate cycle 7, 17,188,242 Investment fluctuation 2,15-8, 38,58,112-3 Investment hunger 1, 3, 7, 15, 24, 27,29,31-4,38,42,62-3, 67,69,71-2, 144,166,204, 221-2,225-8,232,239, 242-3, 248 of the central government 868, 109-11 of local governments 106-11 of state-owned enterprises 102-5,110-1 of township and village enterprises 24 Investment plans 85, 97, 105, 244 Investment ratio 1, 6, 17-8, 35-7, 69, 113,129-55,177,246, 254 aggregate 1, 17-8, 129-55, 246 Investment/saving relation 3 Institutional
Index changes 8,26,131 pressure and incentives 5, 20 Interest rates 24, 52, 61-2, 78, 90-5 nominal 90-1,94 market 52 real 24,90-5 preferential 240 regulated, or administered 92, 103 IOUs 55, 144, 147 Jarque-Bera test, and statistics 197, 205 Johansen's procedure 5, 194-5, 198-9, 214,257 LR tests 198,200-1 Kaleckian, or Kalecki's distributive barrier theory, see also Distributive barrier 221 growth theory 2-3, 6, 30-1, 45-6,51,55,70 labour constraint equation 48, 187 model 2, 46 profit formation equation 55 'Labour accumulation' 118,124 Lagrange multiplier test 205 Learning from data 7, 223 'Leaving gaps' 99 Lending outside the credit plan 89-90 Likelihood function 198 ratio (LR) test 198,215 Limited liability 29,34 Lin Biao's fall 138 Linear restriction 198, 200-1, 219 Local openness 239
Index Locality's evasion 82 Long-run equilibrium relationship See Attractor, Cointegration, Co-movement LSE methodology, and framework 20-3, 189, 193 Lucas critique 68 monetary misperception model 64 Philips curve model 64 MacKinnon critical values 196, 211,258 Macroeconomic adjustment 28, 48,57,70, 112, 114, 118, 124,130,149, 156,249 Management accountability 104 Mandatory credit plan 85-6 investment plan 85-6, 97 loans 84-5 plans 76,78, 101,252 procurement 143 Mao, or Mao Zedong 130, 139 Marginal density 217 distributions 194-5 model 23, 194,218,220 process 195,217-9,220 Market component of the economy 8, 12,96,102,204,241 criteria 2, 15 co-ordination failure 222 reinforcing decentralization 239 Matching funds 78, 87, 109, 226 Material input level 128 input ratio 121-2, 125, 127, 150, 152, 154
287
supply system 72, 78, 97101,253 Maximal eigenvalue and trace See Eigenvalue Ministry of Finance 85 Misspecification, and test 196-7, 205,251 Modelling strategy, and approach 7,22-3,29,31,189,192-4 Motives of bureaucrats, and governments 33, 109-10, 227 Moving average 5, 22, 39-40, 42, 56,190,195,197,207,256 Multi-layer, multi-regional system 107 Nanjing Yangtse River Bridge 181 National income, and used 2-3, 6, 17,27,36-7,45,47-8,5051,54,63, 112,114, 117, 124, 127-30, 137, 141-3, 150,159,161, 169,224, 229, 246, 249 National People's Congress 146, 206, 250 National priority projects, and programme of 78, 86-8, 90,109,252 Negative feedback See Feedback New Keynesian approach 64 'Nodding approval' 80 Non-bank financial institutions 89-90, 96 Nonexperimental data 20 Non-stationary variables, time series, and process 5, 22, 68,184-5, 188-93, 198,210 Norm (path or normal state) 5, 22-4,56, 174, 189-91,203, 206-7, 223, 242 see also 'Control by norm'
288
Norms (of project approval) 79, 82-3,110 see also 'Above-norm' projects Normal growth path 22, 55-6 Normality test 196-7,205 One-step-ahead Chow test See Chow test One-step residuals 28, 202-3 'Open door' policy 144, 181 Ordinary least squares (OLS) 190,202,210-2,214,258 Over-investment, and expansion 8, 24, 93, 96, 144, 204, 226,239,241-2 Parameter constancy 202-4, 206, 220, 224, 243 Parameterization 20,202 People's Bank (of China) 84-6, 89-90,96, 104,240,258, see also Central Bank People's Daily 11, 26, 49, 83, 88, 115,126,141, 144,147, 167,175,206,234,250, 256-7 Perron critical values 196 Perron's innovational outlier model 193 additive outlier model 193, 197 Philips curve 64 Planned component of the economy 8,96,204,227, 241, see also Dual track economy Planner's rationality 3, 243 reaction or response function 44,58 Policy banks
Index State Development Bank 240, 259 Agricultural Development Bank 240,259 Export Import Bank 259 Policy trade-off 113,130,187, 231 Political economy 2, 18, 20-2, 63,71,117,156,223,241 Political campaigns 19, 26, 38, 241 'Politics in command' 131 Positive-predictive approach 238 Post-Cultural Revolution Advance 19, 136 Power-consuming equipment 175-6,256 Power-generating equipment 176-7,256 Pre-reform, and post-reform 4, 811,15,62,67,72,75,82, 96, 108-9, 122, 197,206, 227,229,241,248,251 Pre-war China 122-4, 128 Privatization 238 Probabilistic information 20 reduction 20,22 structural model 20 structure of observed data 20, 223 Procurement contract 143 price 55,134, 149, 161,230 Project approval 34, 71-2, 78-82, 105,110,164,228 Railway traffic density 180 Ramsey's test for specification error 205 Rational scale of investment 1, 223 Real business cycle
Index approach 31,63-4 models 64, 68 theory 30 Real investment level 2-3, 5, 15, 113,185,198,201,203, 206, 222-4, 242 Real value added per labourer (RVAPL) 112,122,125, 127-8, 132, 135-6, 138-9, 142, 146, 149-50, 152, 1601 'Recovery period' 133, 137, 167 Recursive estimation 28, 189, 202 'Reference rate' 90 Reform cycle theory 31,59, 70 Reimers' small-sample-correction 198-9,257 Rent-seeking 12, 32, 63, 93, 172, 204 Reparameterization 202, 218 Representative planner 3, 5, 39, 69 Resident registration system 1234 Resource-constrained (RC), versus demand-constrained (DC) 37,66 Resource endowments 163 Retrenchment 7, 17-9, 67, 95-6, 71,73,82,95-6, 145-6, 175,178, 186,204,221, 225,229,231-2,234,242, 254 Schwarz information criteria 257 Self-imposed restraint 2, 20, 28, 33, 229 Self-raised funds, and investment 84 Self-sufficient, or sufficiency 10, 148-9, 173 Shared authority 239
289
Shortage index 58, 127 signals 4, 30, 38, 42, 44, 223 Simultaneity 189 Skewness 196 Social relationship 32 'Socialist High Tide' 19,131 ' Socialist market economy' 146 SOE autonomy 104 Soft administrative pricing 32, 227 budget constraint 3-4, 7, 29, 31-5,66,69,72, 102,105, 110, 170-1,225,227,235, 237-8, 240, 244 credit 32,108,227 subsidies 32, 235 taxation 32,227 'Soft landing' 146,156,230 Solow method 57 Special economic development zones 147 Spurious regression 184-5,190 Standard errors, or deviation 28, 58,91,196,199,202-3, 205 State budget 84,226 State compulsory procurement 47 State Council 74, 79, 82-5, 93, 98, 138, 146, 148 State Investment Corporation (SIC) 79-80 State Planning Commission (SPC) 27, 74, 79-83, 85, 90, 97-8, 175,206 State investment system 2, 5, 15, 21,27,71-2,75,79,86, 110,206,221,223,232-5, 238-9, 242, 244 State-owned enterprises (SOEs) 32, 72, 75, 93-5, 101-2,
290
104-6,108-110,164, 171, 206, 227-8, 233-240 Stationarity 210 Stochastic analysis, or approach 30, 64 processes 210 Stock exchanges, of Shanghai and Shenzhen 89 Structural break test See Chow test Structural inertia 15 rigidity 2, 15 unbalance 9 Stylized facts 10, 64-6, 68, 131, 149,156,221,225 Subsistence urge 123-4 Supply constraints 2, 6-7, 11, 24, 47, 111, 124-5, 127 Technical updating (and transformation) investment 74-5, 77-82, 84, 233, 243 Ten-Year Development Programme 146 Ten-Year Plan 138-9 Terms of trade, agriculture's 467, 54-5, 70, 129, 143, 230 Thailand 170-1 Three-year readjustment programme 19, 145 Total factor productivity 10-1, 57,61,112,127 Township and village enterprises (TVEs) 11, 24, 141,143-4,164,238,241, 244 industry 13 Transformation of economic systems 8
Index Transport bottleneck 113, 166, 179,181,183-4,244 t-statistics 58,211 Two-way interaction, and dependence 5-6,28,30,70, 166, 187,224,231 Uncertainty 33 Unfinished construction 42-3, 69 Unit root test 210-1,213-5,257, see also Dichey-Fuller test Unitary elasticity 198 United States 163,166,180 Univariate mean, moving average, or norm 5, 22, 40, 56, 69, 207, 223, 256 Unrestricted vector autoregressive (UVAR) representation 20,23,189,193,195-7, 201-2,254 Urban Credit Cooperatives 89, 240 Variances 189-93,255 'Variation free' 194, 218-9 Walras' economics 22, 65 Well-specified 20, 23, 189, 197, 201-2, White noise process 192, 210 White's test for heteroscedasticity 205 Working capital 84, 90-1, 105, 130,240,252,258-9 Wu-Hausman tests 220 'Young intellectuals' 114, 141 ZhouEnlai 134,138 ZhuDe 139