SHIPPING ECONOMICS
RESEARCH IN TRANSPORTATION ECONOMICS Series Editor: Martin Dresner Recent Volumes: Volumes 1–6:
Research in Transportation Economics – Edited by B. Starr McMullen
Volume 7:
Railroad Bankruptcies and Mergers from Chicago West 1975–2001: Financial Analysis and Regulatory Critique – Michael Conant
Volume 8:
Economic Impacts of Intelligent Transportation Systems: Innovations and Case Studies – Evangelos Bekiaris & Yuko Nakanishi
Volume 9:
Road Pricing: Theory and Evidence – Georgina Santos
Volume 10:
Transportation Labor Issues and Regulatory Reform – James Peoples & Wayne K. Talley
Volume 11:
Interurban Road Charging for Trucks in Europe - Jose Manuel Viegas
RESEARCH IN TRANSPORTATION ECONOMICS VOLUME 12
SHIPPING ECONOMICS EDITED BY
KEVIN CULLINANE School of Marine Science and Technology University of Newcastle, UK
2005
Amsterdam – Boston – Heidelberg – London – New York – Oxford Paris – San Diego – San Francisco – Singapore – Sydney – Tokyo
ELSEVIER B.V. Radarweg 29 P.O. Box 211 1000 AE Amsterdam The Netherlands
ELSEVIER Inc. 525 B Street, Suite 1900 San Diego CA 92101-4495 USA
ELSEVIER Ltd The Boulevard, Langford Lane, Kidlington Oxford OX5 1GB UK
ELSEVIER Ltd 84 Theobalds Road London WC1X 8RR UK
© 2005 Elsevier Ltd. All rights reserved. This work is protected under copyright by Elsevier Ltd, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier’s Rights Department in Oxford, UK; phone: (+44) 1865 843830, fax: (+44) 1865 853333, e-mail:
[email protected]. Requests may also be completed on-line via the Elsevier homepage (http://www.elsevier.com/locate/permissions). In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) 7508400, fax: (+1) (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 20 7631 5555; fax: (+44) 20 7631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of the Publisher is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier’s Rights Department, at the fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. First edition 2005 British Library Cataloguing in Publication Data A catalogue record is available from the British Library. ISBN: 0-7623-1177-0 ISSN: 0739-8859 (Series) ∞ The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in
The Netherlands.
CONTENTS LIST OF CONTRIBUTORS
vii
ACKNOWLEDGMENTS
ix
1.
2.
3.
4.
5.
6.
EDITORIAL: KEY THEMES IN SHIPPING ECONOMICS RESEARCH Kevin Cullinane
1
A SURVEY OF THE MODELLING OF DRY BULK AND TANKER MARKETS D. R. Glen and B. T. Martin
19
ECONOMETRIC MODELLING OF NEWBUILDING AND SECONDHAND SHIP PRICES H. E. Haralambides, S. D. Tsolakis and C. Cridland
65
CROSS-INDUSTRY COMPARISONS OF THE BEHAVIOUR OF STOCK RETURNS IN SHIPPING, TRANSPORTATION AND OTHER INDUSTRIES Manolis G. Kavussanos and Stelios N. Marcoulis
107
THE FISCAL TREATMENT OF SHIPPING: A CANADIAN PERSPECTIVE ON SHIPPING POLICY Mary R. Brooks and J. Richard Hodgson
143
DETERMINANTS OF VESSEL FLAG Jan Hoffmann, Ricardo J. Sanchez and Wayne K. Talley
173
v
vi
7.
8.
THE CONTAINER SHIPPING INDUSTRY AND THE IMPACT OF CHINA’S ACCESSION TO THE WTO Kevin Cullinane
221
LINER SHIPPING STRATEGY, NETWORK STRUCTURING AND COMPETITIVE ADVANTAGE: A CHAIN SYSTEMS PERSPECTIVE Ross Robinson
247
AUTHOR INDEX
291
SUBJECT INDEX
00
LIST OF CONTRIBUTORS Mary R. Brooks
Faculty of Management, Dalhousie University, Halifax, Canada
C. Cridland
Braemar Seascope, London, U.K.
Kevin Cullinane
School of Marine Science & Technology, University of Newcastle, U.K.
D. R. Glen
Centre for International Transport Management, London Metropolitan University, U.K.
H. E. Haralambides
Center for Maritime Economics and Logistics, Erasmus University Rotterdam, The Netherlands
J. Richard Hodgson
Faculty of Management, Dalhousie University, Halifax, Canada
Jan Hoffmann
UNCTAD, Geneva, Switzerland
Manolis G. Kavussanos
Athens University of Economics and Business, Athens, Greece
Stelios N. Marcoulis
Laiki Investments, Cyprus Popular Bank, Nicosia, Cyprus
B. T. Martin
E A. Gibson Shipbrokers Ltd., London, U.K.
Ross Robinson
Australian Centre for Integrated Freight Systems Management, Melbourne University Private, Melbourne, Australia
Ricardo J. Sanchez
Universidad Austral, Buenos Aires, Argentina
Wayne K. Talley
Old Dominion University, Norfolk, USA
S. D. Tsolakis
Center for Maritime Economics and Logistics, Erasmus University Rotterdam, The Netherlands vii
ACKNOWLEDGMENTS A work such as this cannot be undertaken without incurring numerous debts. In partial repayment of what is owed, I would take this opportunity to express my heartfelt appreciation of the efforts of certain key individuals without whom this book would not have come to fruition. In the first instance, I would like to thank Professor Martin Dresner of the Robert H. Smith School of Business of the University of Maryland for his unstinting efforts in reviewing and improving the contents of the book in his role as series editor. I would also express my appreciation of Elsevier’s publishing editor, Chris Pringle, for not only accepting my proposal for this volume, but also for the support and advice he has given throughout the editing and publication process. The patience of both Martin and Chris in waiting for the final manuscript is much appreciated. The contribution of others that have provided help and advice behind the scenes should also be acknowledged. In this respect, I am grateful to the following for their insight and timely support when needed: Roar Adland of Clarksons in London, Giuseppe Alessi of the University de L’Aquila in Italy, Anthony Beresford and Peter Marlow of Cardiff University, Stephen Gong of the Hong Kong Polytechnic University and Photis Panayides of the Cyprus Institute of Management. Obviously, in any work of this nature, however, the greatest expression of gratitude must be reserved for the authors that have committed time and effort to the project. In all cases, contributing authors were unswerving in meeting deadlines, undertaking recommended revisions and in supporting whatever requests the editor laid before them. This work would not have been possible without their outstanding and untiring commitment to see it completed. Kevin Cullinane Chair in Marine Transport & Management School of Marine Science & Technology University of Newcastle
ix
1.
EDITORIAL: KEY THEMES IN SHIPPING ECONOMICS RESEARCH
Kevin Cullinane 1. INTRODUCTION In order to fully comprehend the scope of this volume in the Research in Transportation Economics series, it is important to recognise a distinction between Shipping Economics and Port Economics. While the area covered by the discipline of Port Economics may appear to be intuitively obvious, the scope of what is entailed within Shipping Economics is rather more difficult to define. As can be seen in the content of the chapters in this volume, the coverage of the latter is extremely eclectic and, in common with the area of Port Economics, draws upon many concepts, theories and methods that are ubiquitously applied in other branches of economics. Together, they may be considered to comprise Maritime Economics. As with many taxonomies that are developed and utilized within the social sciences, on most occasions it is very clear what specific issue may be termed Shipping Economics and what Port Economics. However, on some occasions, the distinction may not be absolutely categorical. Many issues of relevance to the port industry simply cannot be analysed without taking into account the economic behaviour of their main customer, the shipping industry. Similarly, it is easy to visualize areas of concern to the shipping industry that are significantly affected by the economic behaviour of one of their main service suppliers, the port industry. Thus, at the interface between Shipping Economics and Port Economics, there exists a grey area where one “discipline” may impinge on the territory of the other.
Shipping Economics Research in Transportation Economics, Volume 12, 1–17 Copyright © 2005 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0739-8859/doi:10.1016/S0739-8859(04)12001-5
1
2
KEVIN CULLINANE
2. MARKET MODELLING AND ANALYSIS Research in the field of shipping economics has, throughout its history, been very much concerned with various aspects of the functioning of markets. Since this is undeniably a common characteristic of many branches of economics, this certainly cannot be claimed to be a unique feature of shipping economics. Within the context of transportation economics, however, it may be true to say that market modelling is further advanced in shipping economics than in the study of any other of the transportation modes. This is undoubtedly due to the great range of markets which may be encompassed within the broad sphere of what may be considered to be shipping economics. International shipping can be distinguished from intra-regional and domestic (or coastal) shipping, as well as from freight movements on inland waterways. In terms of mere revenue contribution, the shipping industry can be somewhat simplistically divided between the bulk and liner sectors. The former is broadly characterized by large single shipments of loose cargo in whole ships that are operated on tailormade voyages and the latter by mixed shipments of containerised cargo in ships that are operated to a regular schedule on pre-defined routes. The nature of the cargoes carried within each of these two major dichotomous sectors dictates that radically different designs of ship are deployed in each. Other more specialized markets play a relatively minor, but nonetheless important, role. Examples that readily spring to mind are the cruise and ferry passenger markets, the roll-on/rolloff freight markets, the specialized bulk markets (e.g. car carriers) and the smaller non-containerised general cargo trades. In turn, the mainstream bulk cargo market can be divided between the tanker and dry bulk sectors (the “wet” and the “dry” bulk cargo markets) and the liner market into the three major East-West trades, the three major North-South trades, the important intra-Asian trades, feeder trades etc. In the bulk market, the size of ship, the cargo carried, the trade route and the contract of carriage all provide yet another basis for further segmentation of the market. In order to serve the primary shipping markets that relate in a straightforward manner to the carriage of freight, there are important secondary markets that also provide potential avenues for research in shipping economics. The ships which carry the cargoes may be bought new or secondhand. They may also, at some point in time, need to be scrapped. This alludes to the importance of analysing the shipbuilding, ship sale and purchase (S&P) and scrap markets. Since all these shipping markets function in a totally international arena where national political boundaries pose only a minor irritation to the smooth conduct of trade and commerce, there is also a need for a worldwide focus on generic markets that are of critical import to the shipping industry, such as those for money, currencies, labour and fuel.
Key Themes in Shipping Economics Research
3
The potential for conducting market analysis in shipping economics is, therefore, very great; a fact not lost on the comparatively large number of consultancy companies, such as Clarksons, Drewry, Fearnleys, Platou etc., that produce a plethora of very detailed and expensive market analyses on a regular and ongoing basis. The academic community engaged in research in shipping economics also recognises this potential to conduct market analyses. Adopting a very different perspective and rationale than the consultancy companies, the analysis of shipping markets has constituted a key strand in shipping economics research over the last few decades. In this respect, it is certainly the case that some markets have received greater attention than others. Because of the extremely volatile nature of price movements within it, the bulk market has repeatedly been put under the microscope, although not always at a completely aggregate level; as well as general or holistic analyses of how the bulk market works, there have also been several more specific studies, conducted at a greater level of disaggregation. The first chapter in this volume is by David Glen and Brendan Martin. They provide a comprehensive survey of the corpus of work done on modelling the dry bulk and tanker markets. By its very nature, inclusivity is a necessary characteristic of any effort to undertake such a task and it is this which justifies its position as the inaugural chapter of the volume. They highlight the seminal contributions of Tinbergen (1931, 1934) and of Koopmans (1939) and summarise the ensuing evolution of alternative approaches to modelling these markets. In so doing, they highlight the pivotal and seminal contribution of Beenstock and Vergottis (1993); a work that provides a reference point for much of the discussion in which Glen and Martin go on to engage. Having reviewed the relevant literature, Glen and Martin point out that since Beenstock and Vergottis (1993), efforts to model either or both of the dry bulk and tanker markets have eschewed an approach based on hypothesized causality and structural modelling. Instead, recent preferred modelling methodologies have revolved around data-driven approaches that focus on the statistical properties of market data and determining reduced form dynamic relations therein. As the authors indicate, this contemporary approach has been motivated by data-analytic and modelling innovations that have their origins in the discipline of financial economics (e.g. Dickey & Fuller, 1979, 1981; Engle & Granger, 1987; Johansen, 1988). It has also been coincidentally facilitated by the wider availability and enhanced accessibility of better data that is characterised by higher frequency and longer duration. Glen and Martin go on to conduct their own empirical analysis. Firstly, by collecting supplementary data to February 2003, they extend and enlarge the database originally analysed by Veenstra (1999). They then repeat the VAR modelling approach adopted in the original. Interestingly, their findings so closely
4
KEVIN CULLINANE
approximate those of the original that the VAR modelling methodology adopted by Veenstra (1999) is very much vindicated and the results validated. This is so much the case, in fact, that Glen and Martin continue with their empirical analysis by estimating a greatly simplified, “reduced form” model of spot market rates that actually utilizes an output from their own original VAR analysis as an input into this later model. As Glen and Martin freely admit, this market model is “somewhat unusual” in utilizing the spread relation estimated in the VAR context in tandem with a range of “structural” variables. As such it represents a hybrid model form which spans the methodological divide between the structural and VAR modelling methodologies by very simply relating first differences in spot rates to: the lagged values of the difference between time charter and spot rates; a vector of exogenous demand variables; the existing fleet that is relevant to an individual segment of the overall market and; the unit price of the fuel used by ships. They conclude that the forecasting performance of their model is, in most cases, only “marginally better” than those derived from a “na¨ıve” model. It is claimed, however, that forecasting performance will improve as variables are added from the estimating equation for the first differences of the spot rate from inside the original VAR model. One of the most poignant conclusions that Glen and Martin draw from their survey of the literature and ensuing empirical analysis is that the contemporary reliance on data-driven methodologies does not yield the same insights or depth of understanding that may be derived from approaches based on structural modelling. As such, they attest that it may be time to revert again to a more traditional approach, a` la Beenstock and Vergottis (1993), that is based on the modelling and testing of hypothesised causal relationships that are derived directly from economic theory. This “crie de coeur” is one that will resonate with many shipping economists and to which many would certainly lend their support. In their paper on the “Econometric Modelling of Newbuilding and Secondhand Prices,” Hercules Haralambides, Stavros Tsolakis and Colin Cridland attempt to put into practice exactly the sort of “fundamental” approach, based upon the specification and testing of a structural model, that is advocated by Glen and Martin. In common with the authors of the first chapter, they too build upon the significant body of work concerned with the modelling of prices in the various markets that are pivotal to the shipping industry. As previously mentioned, such analyses have their origins in the pioneering work of Tinbergen (1931, 1934) and Koopmans (1939). Both these Nobel laureates made early seminal contributions to shipping economics and can be attributed with the original exposition of the much referred to, and often taught, shipping market cycle; a concept that makes explicit reference to the complex interdependence of freight and asset markets. In their review of the literature on the modelling of newbuilding and secondhand ship
Key Themes in Shipping Economics Research
5
prices, the authors identify and summarise the main contributions to this field over the ensuing years. Although the focus of this paper would appear to be limited to the market for both new and for used ships, it conforms to precedent in acknowledging that each of these markets cannot be analysed in isolation. Not only does previous research suggest that the potential exists for prices in each of these markets to be dependent on the other, but also that they may be causally dependent on many other factors, not least: the charter rates (prices) which prevail in the market for the carriage of goods by sea; exchange rates; interest rates; worldwide shipyard capacity and utilization; national industrial and fiscal policies etc. One of the most interesting aspects of this contribution lies with the authors’ exposition of the difficulties faced in variable definition and data collection, as well as the varied, sometimes imaginative and sometimes downright devious, methods employed to resolve them. Those of us that have undertaken data analytic research will share the sense of frustration that can be detected in the authors’ cataloguing of problems such as: the absence of appropriate data across all the variables at anything other than an annual frequency; inconsistency in the definition of market segments over time; changes to Lloyds Register ship categorization criteria in 1995 (such that certain ships that had been classified as bulk carriers until that time were suddenly found to be classified as general cargo vessels); the inconsistency of data between sources and the consequent need for the “triangulation” of data values between those sources; the incomparability of newbuilding prices under different terms of payment and; the dearth of comprehensive and reliable information to allow precision in measuring shipyard capacity. By integrating what they consider to be the most appropriate elements of previous models, Haralambides, Tsolakis and Cridland espouse interdependent, causal model specifications for both the newbuilding and secondhand ship markets. They then test the significance of the relationships they have hypothesised. A disaggregate approach to the analyses of the two markets is adopted. This facilitates a focus on the price of new and secondhand ships for specific shipping market segments and the comparison of differences and similarities between them. It also simultaneously reduces the potential for the overall picture to be obscured by variations in price behaviour between market segments. Even for the uninitiated in the development of complex causal models of market price behaviour, the results of the analysis are interesting. As is often the case with such analyses, this is as much due to the inconsistency of findings as it is to their conclusiveness. For instance, in drawing inferences from the estimation of their newbuilding model, the authors find that shipbuilding costs have a significant impact on the prices of both tankers and bulk carriers across all market segments and in both the long and short run. This is unsurprising since it is this factor that
6
KEVIN CULLINANE
largely explains the recent and continuing migration of shipbuilding contracts to the Chinese mainland and similar migrations in times past to Japan and Korea. It is also a finding that features prominently in the results of previous studies. Another conclusive result is that exchange rate fluctuations are found to have no significant impact on the price of newbuildings in any market segment. Since this would appear to be counterintuitive, it is a rather more surprising result. While the authors point to the potential for covariance amongst the independent variables as a possible explanation, there certainly remains scope for further research. At the same time, however, the results for the newbuilding model suggest that, in the long-run, the freight market is an even more important influence on the prices of handy-sized bulk carriers than the cost of building these ships (although the latter remains a significant influence on prices). While the authors speculate as to why this should be the case, there is again significant scope for further investigation.
3. SHIPPING FINANCE Shipping Finance is an extremely important aspect of the shipping economics discipline. Apart from air transport (see Morrell, 2002), there is no other area of transportation economics, where it plays so vital a role. The most obvious reasons why this should be the case are: (a) the capital intensity of the shipping industry; (b) the availability and cost of finance for ships is dependent upon a highly volatile and, therefore, risky freight market and; (c) the international mobility of the assets means that the issue of financial security takes a greater priority than usual. The uniqueness of the market for shipping finance is illustrated by the fact that many banks and finance houses employ specific shipping finance expertise and that sometimes these groupings are even recognised formally within an entity’s organisational structure. Similarly, the importance of shipping finance has been recognised not only by a significant heritage of books (Cheng, 1979; Grammenos, 1979; Paine, 1990; Slogett, 1999; Stephenson, 1995; Stokes, 1997) and industry guides on the subject (e.g. Euromoney, 2004/2005), but also by the continuing growth in scholarly publications desseminating the results of shipping finance research. For example, see Haralambides (1993); Grammenos and Marcoulis (1996), Sjogren (1999), Leggate (1999), Leggate (2000), Cullinane and Panayides (2000), Akatsuka and Leggate (2001), Panayides and Gong (2002), Grammenos and Arkoulis (2002), Kavussanos and Visvikis (2004), Chen and Wang (2004). The chapter by Manolis Kavussanos and Stelios Marcoulis is again another form of specific market analysis. In this case, the work is concerned with the performance of listed shipping companies across the world and the comparison
Key Themes in Shipping Economics Research
7
of this performance against that of the transportation and other industries. Given the trend in recent years for shipping companies to access the funds of the general public, particularly through stock market flotation and the widening of the share ownership base, it is not surprising that the authors should focus on this form of finance. Specifically, they attempt to specify and estimate alternative models for the valuation of the shares of “water transportation companies.” The models they test range from single-index models, where a market index alone is assumed to be driving returns, through to multifactor models that incorporate micro- and macroeconomic factors as determinants of the rate of return on shares. By comparing the stock market performance of different shipping markets against that of other industries, a clear vision of the relative risk-return trade-offs can be derived that provides investors with a sound basis for making rational decisions. The authors’ literature review suggests that previous analyses have largely found that differences between industries are highly significant in explaining differences in returns from shares. The reported empirical work of the authors suggests that, in addition to the general market movement, there are microeconomic factors at the level of the company that tend to influence the returns from shipping and other industries. For shipping, in fact, the level of gearing is found to have a particularly strong negative relationship to stock market performance. When it comes to testing the impact of macroeconomic factors, however, their results might be perceived to be rather counterintuitive. While both monthly industrial production and oil prices are found to have a significant influence on the rate of return on shipping company shares, the former is found to be a negative relationship and the latter positive. There would seem to be a rather more complex relationship between these macroeconomic factors and stock market returns than one would expect on the basis of elementary economic theory. As such, this would again seem to justify further in-depth investigation. As potentially the case with all research, the investigation conducted by Kavussanos and Marcoulis seems to have raised more questions than answers. However, one categorical conclusion that does seem to emerge from this work is that factors other than simply the general market movement do seem to have an influence on the prices of shipping industry shares. As such, the authors advocate the use of multifactor models in future research so that these other influential factors can be identified and the level of their influence evaluated. Implicitly, this would appear to provide yet further support for a more fundamental approach to model development. After all, the factors that are hypothesised to have some bearing on shipping share prices, and whose significance is subsequently tested within the model, are actually posited on the basis of the economics that underpins the operation of the physical market for shipping.
8
KEVIN CULLINANE
4. FISCAL POLICY At first sight, a paper on the fiscal treatment of shipping in Canada may appear to be rather a specialised and esoteric topic for inclusion in a work representing the major themes of contemporary research in shipping economics. The contribution by Mary Brooks and Richard Hodgson, however, addresses one of the most controversial and oft-debated policy issues affecting the maritime industry. In the light of increasing competition from nations with a lower cost base and in the absence of appropriate policies that ensure its continuation, the combined merchant fleet of the world’s developed countries has experienced inexorable decline. The greatest impetus to this phenomenon has been the establishment of open registries, most usually in low cost nations, that deliberately set out to attract the registration of ships that are owned and controlled by overseas shipowners. Until the early to mid-1990s, in fact, private-sector shipping companies of the developed world had exhibited a consistently greater propensity to register their ships offshore with open registries. This trend was facilitated by the inherently international outlook of the decision makers involved and by the mobility of the assets – the ships themselves. This decline in the registered fleets of many of the world’s traditional maritime nations was allowed to approach crisis point. In Europe, where most of the countries affected were clustered, what appeared to bring this issue to the forefront of the political agenda was not the prospect of any form of economic loss that may or may not have been sustained as the result of the loss of a flag fleet. Nor was it the deleterious impact on the employment levels of nationals on board ships (see Goss, 1993). What really seemed to bring the matter home was the seemingly sudden recognition that: (a) a continuing decline in a nation’s shipping fleet would greatly reduce the seafaring knowledge, skills and experience that would be available and could be put to good use ashore, and that this would constitute a significiant loss to the national economy; and (b) flag administration remains the most viable mechanism through which safety and environmental regulation can be reliably implemented, especially where this relates to freight transport within national waters and to the domestic shipping industry. Following decades of political and academic debate (Gardner & Marlow, 1983; Gardner & Richardson, 1973/1974; Goss, 1985, 1993; Marlow, 1991a, b, c, 2002), a decision had to be made as to whether some form of intervention was warranted. Either the traditional maritime nations had to face up to the reality that they would no longer be significant players in world shipping (in terms of ship registrations at least) or policies would need to be implemented to stem, and possibly reverse, the continuing decline in their national registered fleets.
Key Themes in Shipping Economics Research
9
A range of possible options were available to political decision makers. In Europe in the early 1990s, a tranche of second registers had been developed that offered shipowners of these high cost countries the opportunity to remain loyal to their home nation, enjoy the cultural affinity to which they were accustomed and yet to simultaneously reap certain economic advantages over and above registering with their main domestic flag. The so-called “Second (or International) Registries” provided a nation’s shipowners with the potential to make savings (particularly in respect of the fixed costs of registration, taxation on profits and the labour cost savings that could accrue from employing overseas nationals on board ships) that were competitive with those offered by open registries (see Cullinane & Robertshaw, 1996; Sletmo & Holste, 1993; for more in-depth analyses of this phenomenon). From the perspective of developed world shipowners, it was unfortunate that Second Registries incurred the wrath of the International Transport Worker’s Federation (ITF). Claiming to be protecting its members’ interests, the ITF took the view that “Second Registries” undermined their efforts to ensure equity in the international market for shipboard labour and that they constituted merely a guise for concealing a flag administration that, to all intents and purposes, was tantamount to a “Flag of Convenience (FoC).” Although the veracity of this view could be regarded as contentious to say the least, in a few notable cases the ITF has actually carried out its threat to classify “Second Registries” as FoCs. The difficulties and additional costs imbued by FoC classification (or even merely the threat of it) has undermined, at least partially, the success of the “Second Registry” as well its popularity as the most appropriate policy tool for reversing the decline in developed world fleets. Instead, in most of the world’s major maritime adminstrations, fiscal policies have emerged as the major policy instrument for ensuring the continued survival of a national flag fleet. In particular, in Europe in the mid-1990s, several countries began to invoke tonnage tax proposals that, since implementation, are perceived to have been a major success in reversing the declining trend in flag fleets. The chapter by Brooks and Hodgson provides an exhaustive analysis and review of the alternative generic policy options for reversing the decline in a nation’s merchant fleet. In so doing, a completely understandable emphasis is placed upon the alternative fiscal regimes that may be applied to a nation’s shipping industry. By providing an analysis of the historical development of this aspect of Canadian shipping policy, the authors highlight the problems experienced in Canada that are common to all developed, traditional maritime nations. Rather uniquely, however, as long ago as 1985 the Canadian authorities opted for an approach which revolved around maintaining the management of overseas-flagged vessels and based on the concept of an “International Shipping Corporation (ISC).”
10
KEVIN CULLINANE
Such a policy implicitly recognises the potential contribution of an “International Maritime Cluster” (IMC) to a national economy but, as Brooks and Hodgson point out in their detailed analysis of the policy, it still may fail to provide some of the more subtle benefits that can accrue in the longer-term from fiscal policies (such as a tonnage tax) that promote a national flag fleet. Examples might include benefits that accrue from utilising seafaring skills and experience ashore, as well as from safety and environmental regulation etc. A particularly interesting aspect of the authors’ analysis lies with the distinction they draw between the policy context for a nation’s international, as opposed to its domestic, shipping industry and the implications that ensue from this. Brooks and Hodgson conclude that the ISC concept has achieved only marginal success in meeting the objectives that were set for it and that other approaches applied elsewhere (notably tonnage tax) have achieved significantly more. On this basis, they call for a comprehensive review of Canadian shipping policy. At the time of writing, however, it remains to be seen whether this and/or other representations to the Canadian government will exert any influence in stimulating a shift in policy.
5. SAFETY AND SECURITY The chapter by Jan Hoffman, Ricardo Sanchez and Wayne Talley is entitled “Determinants of Vessel Flag.” The content of this chapter is clearly very closely related to the previous one by Brooks and Hodgson. Although it too deals with the issue of ship registration, it does so primarily from the perspective of developing an empirical model of a ship operator’s decision on which flag will be selected as the registry of choice for a particular ship. In contrast to the previous chapter, where policy was analysed and ultimately brought into question, the policy context in this work is taken as a given and the objective of the model is to yield accurate predictions of the ship operator’s flagging decision under a range of conditions and environmental factors that may influence it. Although this is the avowed objective of the work, the true raison d’ˆetre for the analysis contained in this chapter is the fundamental relationship which is widely pre-supposed to exist between a ship’s flag and the level of safety and compliance with environmental regulations at sea. There has been a significant volume of research which has attempted to deduce “falsification” evidence (see Popper, 1934) against the supposition that open registry ships have the worst safety records (e.g. see Li & Wonham, 1999). Since there are numerous potential confounding variables (some of which are hinted at in this chapter), such an analysis is actually very difficult to do. In consequence, the results of previous research have not been totally conclusive. The balance certainly seems to fall, however, on the side of open registries faring relatively poorly in
Key Themes in Shipping Economics Research
11
their safety records compared to the flags of what are generally regarded as the “better” traditional maritime nations. Despite this, it should be recognised that the worst safety records of all can probably be justifiably attributed to some of the world’s major national (state) shipping companies. In their empirical work, Hoffman, Sanchez and Talley analyse a database relating to almost 48,000 of the world’s commercial vessels. They estimate two probit models (“basic” and “detailed” versions), with the probability of a particular ship being flagged out as the dependent variable in each and the set of independent variables varying in number and detail between the two models. Some of the results achieved are not what might be expected. For example, that: the probability of flagging out decreases with vessel age; the larger the container carrying capacity of a vessel the lower the probability of being flagged out; the probability of flagging out is lower if the ship was built in the country of the operator than otherwise etc. The authors succeed, however, in providing feasible, and largely convincing, explanations for some of these seeming anomalies. As the authors themselves acknowledge, one potially important problem with their analysis may be the unknown impact of muticollinearity among the independent (explanatory) variables. Further investigation of this aspect will certainly go a long way towards isolating the truly important influences on the flagging out decision. Such an analyis may be further facilitated by analysing a smaller sample with a larger size cut-off than was applied in this study. Since the fundamental distinction between the nature of ships used for international and for domestic shipping has a role to play in the authors’ explanations of some of their anomalous results, it would seem that the analysis of data relating to a more cognate sample would likely lead to more consistent and logical results that exhibit even greater explanatory power. What is particularly alarming in the findings of the analysis is that the probability of a ship being flagged out increases as the maritime safety record of the home country improves. Thus, the better a nation’s maritime safety record, the more likely it is that a ship operator from that nation will flag out and vice versa. Similarly, the more IMO conventions that a nation has ratified, the higher is the probability of flagging out. By analysing the relationship between socio-economic factors and flagging out, Hoffman, Sanchez and Talley conclude that flagging out is most definitely a developed economy phenomenon. Most poignantly, they suggest that “higher wages and labour standards may scare operators away from national registries.” This is a significant, though potentially problematic, conclusion to draw given the dual ambitions of most contemporary political establishments both to secure increases to GDP and to maintain and enhance the size of the national flag fleet. This set of findings prompts the authors to speculate as to the main beneficiaries of
12
KEVIN CULLINANE
the open registry system. It is certainly the case that the open registry system does allow the ship operators of the developed world to compete against developing economy competitors that possess inherently lower costs. However, the authors point out that it is also a source of revenue-generation for the developing nations where most of the open registries are based, it has promoted the employment of seafaring labour from developing nations and, ultimately, by reducing the cost of transport, has facilitated the international trade of developing nations. The authors seem to implicitly suggest a change in approach from the United Nations Conference on Trade and Development (UNCTAD) may be called for and that open registries should no longer be castigated as an undesirable feature of the modern shipping industry. Since the advent of the Second Register (discussed earlier) and the implementation of Port State Control, this appears to be an entirely sensible suggestion that warrants further discussion. The one caveat over any possible relaxation of attitudes towards open registries and their possible acceptance should be, however, a concommitant tightening up of safety and environmental regulation and its enforcement in order to avoid the worst excesses of the open registry system. Persuading the ITF that this is a policy worth pursuing, however, may prove to be a rather more difficult concession to obtain. It is almost certainly the case that research into safety and security at sea (and in ports) will burgeon over the next decade. A major part of this research effort will fall within the bailiwick of the maritime economist as difficult decisions need to be taken as to what risks are deemed acceptable and what risk management measures are economic to implement (see Li & Cullinane, 2003). It is important that the shipping economics research community is prepared to accept its responsibility to inform policy on the basis of the research it conducts in this field on the industry’s behalf.
6. A WIDER PERSPECTIVE: LOGISTICS AND SUPPLY CHAIN MANAGEMENT In recent years, there has been a veritable explosion of conceptual papers and research output that has positioned the liner shipping industry (in particular) as a contributing element within a logistical system or supply chain (e.g. Heaver, 2002; Notteboom & Winkelmans, 2001). This more macroscopic perspective on liner shipping specifically, and freight transport more generally, is having an influence on the nature and form of shipping economics and upon how research is conducted. This trend has been reinforced by the increasing importance of liner shipping’s contribution to the overall shipping industry. In addition, because liner shipping is perceived as fitting more comfortably into a wider industrial and logistical
Key Themes in Shipping Economics Research
13
picture than is bulk shipping, the market analyses that have characterized research into bulk shipping have not had such a major rˆole to play in the liner shipping context. Together, all these influences have stimulated a sea-change in the culture of maritime economic analysis. Different and more holistic approaches that adopt a wider perspective have been applied to the analysis of liner shipping than are applied to the bulk sector. As is exemplified in the final two chapters in the volume, the influence of this more holistic perspective is such that it is equally likely to have a bearing in port economics as within studies that relate exclusively to the liner shipping industry. Concentrating on what has traditionally been the “Cinderella” sector for the conduct of market analysis, my own contribution to this volume presents an overview of the global liner shipping market and a vision for its mediumterm future. In undertaking such a task, consideration must be given to the crucial macroeconomic influence of international trade patterns in fundamentally determining trends in shipping market characteristics. The chapter focuses on China’s accession to the World Trade Organisation (WTO) as a particularly profound example of how changes in international trade (however they might be brought about) can have a critical impact on the nature of the whole shipping industry. Even though it obviously exerts a much greater sphere of influence, the after-effects of this momentous political event are, however, considered exclusively from the perspective of their potential impact on the liner shipping sector. This chapter, as well as the last in this volume (by Ross Robinson), both exemplify the difficulty in analysing shipping markets in isolation of any discussion of the port sector that serves the ships that operate within the market. This is true of all shipping, but particularly so in the liner shipping sector, where ships are deployed to operate on networks that link ports as nodes and where the efficiency of ports and their connectivity to inland logistics systems are of paramount importance to the service that liner shipping companies can offer and the rewards that they can reap. In hypothesizing the potential impact of China’s WTO accession, therefore, it proves impossible to completely separate out the simultaneous impact that this phenomenon is likely to have on both the liner shipping and container handling sectors. This is especially the case given the close interdependence of the two sectors; a symbiotic relationship that manifests itself in: the service contracts that are negotiated between industry players from each of the two sectors; the increasing tendency towards the greater industrial concentration of both sectors and; the efforts of each of the two sectors to vertically integrate into the other’s traditional area of specialization and expertise. The chapter by Ross Robinson espouses what he refers to as a “Chain Systems” perspective on liner shipping strategy, the structure of liner networks and the competitive advantage of companies operating in the liner shipping milieu. A
14
KEVIN CULLINANE
detailed explanation of the conceptual context for the “Chain Systems” perspective reveals it as an all-embracing amalgam of the “five forces framework” due to Porter (1979), the resource-based view of competitive advantage that can be traced back to Penrose (1959) and the supply chain perspective of Cox (1997). Betraying their origins in industrial economics, the two former, more traditional, analytical frameworks emphasise the nature of the market and business environment in which an individual firm operates and the relationship (or strategic positioning) of the firm to that market or environment. Cox (1997) advocated that firms exist not only in markets, but also within supply chains and that, therefore, competitive advantage and “business success” (however that may be defined) are not only a function of a firm’s strategy and positioning with respect to competitors within a market, but also with respect to other players both within the supply chains in which the firm plays a part and within competing supply chains. While each of these conceptual frameworks may provide useful insights into the nature of the competitive advantage that an individual firm may enjoy, Robinson suggests that none in its own right provides a comprehensive explanation of the propensity for competitive advantage and consequent “business success.” An holistic approach that stresses the importance of incorporating a supply chain perspective can only be applied at a disaggregate level of analysis; where the behaviour of individual firms becomes the focus of study. While the analysis of an industry within which a firm operates may be undertaken at a reasonably macroscopic level, the need for a disaggregate approach becomes obvious when one considers the uniqueness of the supply chains within which that same individual firm has a role. Liner shipping is the ultimate example, in fact, of an industry that plays a pivotal role in the supply chains of the world’s manufacturing sector; it has no other role than to facilitate the flow of goods along supply chains, primarily within the global context. Having expounded the intricacies of the analytical framework he proposes, Robinson goes on to argue that competitive advantage in liner shipping and ultimate business success is achieved as the result of the power a firm exerts not only within its market, but also over the supply chains in which it plays a part. In the market, a liner shipping company will flex its muscles to secure, maintain and enhance control over the market (possibly as manifest in market share) and to exclude competitors from access to that market. In its supply chain activities, on the other hand, a liner shipping company will leverage its power in the supply chain to extract cost and quality improvements from its suppliers, while simultaneously increasing the dependence of its customers on the services that it offers and, thereby, extracting greater value from the customer base. In the first instance, power in the supply chain is derived from a firm’s ability to meet the needs and expectations of its customers;
Key Themes in Shipping Economics Research
15
i.e. from a liner shipping company’s ability to deliver the value which its customers are seeking. The natural corollary of adopting the analytical framework propounded by Robinson is that the development and implementation of corporate strategy in liner shipping is motivated and explained by the position of an individual company with respect to not only its market, but also the supply chains in which it is engaged. Strategy formulation is motivated by the attempt to achieve power in the market and the supply chain and to exploit that power by capturing value in both. In liner shipping, one of the major manifestations of strategy is the structure of the network offered. Quite correctly, Robinson expresses the view that network configuration, size and coverage are all of secondary importance to the achievement of strategic objectives and that, in fact, networks are structured with a view to securing value in both the market and supply chain systems. A graphic illustration of this thinking is provided by two in-depth case studies of the situation prevailing in Malaysia/Singapore in the mid-1990s (at a time when the Port of Tanjung Pelepas was not even a gleam in its designer’s eye) and in Hong Kong/South China in the early 2000s.
7. CONCLUSION While the richness and diversity of research in shipping economics precludes a totally comprehensive representation of the opportunities it offers, most of the main areas of interest are covered by contributions to this volume. It is clear that market analysis will continue to form the cornerstone of the shipping economics research agenda and that this is especially so for the bulk markets and the secondary markets that are related to it. With the further emergence of liner shipping as the most significant sector within the industry, there are signs that market analysis will find its way into container trade research. This is already happening for certain of the key mainline and niche trades. The influence of the logistics and supply chain perspectives will inevitably become more pervasive and this is likely to stimulate research that is more multi-disciplinary in nature and is characterised by more systemic analysis. It is already difficult to distinguish between shipping and port economics. Under the holistic logistics/supply chain perspective, the distinction is likely to become even further blurred.
REFERENCES Akatsuka, K., & Leggate, H. K. (2001). Perceptions of foreign exchange rate risk in the shipping industry. Maritime Policy and Management, 28(3), 235–249.
16
KEVIN CULLINANE
Beenstock, M., & Vergottis, A. (1993). Econometric modelling of world shipping. London: Chapman & Hall. Chen, Y.-S., & Wang, S.-T. (2004). The empirical evidence of the leverage effect on volatility in international bulk shipping market. Maritime Policy and Management, 31(2), 109–124. Cheng, P. C. (1979). Financial management in the shipping industry. Centreville, MD: Cornell Maritime Press. Cox, A. (1997). Business success. UK: Earlsgate Press. Cullinane, K. P. B., & Panayides, Ph. M. (2000). The use of capital budgeting techniques among UK-based ship operators. International Journal of Maritime Economics, 2(4), 313–330. Cullinane, K. P. B., & Robertshaw, M. (1996). The influence of qualitative factors in Isle of Man ship registration decisions. Maritime Policy & Management, 23(4), 321–336. Dickey, D. A., & Fuller, W. A. (1979). Distributions of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74, 427–431. Dickey, D. A., & Fuller, W. A. (1981). Likelihood ratio statistics for autoregressive time series with a unit root. Econometrica, 49, 1057–1072. Engle, R. F., & Granger, C. W. (1987). Cointegration and error-correction: representation, estimation, and testing. Econometrica, 55, 251–276. Euromoney (2004/2005). Shipping finance annual. London: Euromoney Books. Gardner, B. M., & Marlow, P. B. (1983). An international comparison of the fiscal treatment of shipping. Journal of Industrial Economics, 31(4), 397–415. Gardner, B. M., & Richardson, P. (1973/1974). Fiscal treatment of shipping. Journal of Industrial Economics, 22(2), 95–117. Goss, R. O. (1985). Social cost, transfer payments, and international competition in shipping. Maritime Policy and Management, 12(2), 135–143. Goss, R. O. (1993). The decline of British shipping: A case for action? A comment on “The decline of the UK merchant fleet: An assessment of government policies in recent years”. Maritime Policy & Management, 20(2), 93–100. Grammenos, C. Th. (1979). Bank finance for ship purchases. Bangor Occasional Papers in Economics, No. 16. University of Wales, Cardiff. Grammenos, C. T h., & Arkoulis, A. G. (2002). Macroeconomic factors and international shipping stock returns. International Journal of Maritime Economics, 4(1), 81–99. Grammenos, C. Th., & Marcoulis, S. N. (1996). Shipping initial public offerings: A cross-country collection. In: M. Levis (Ed.), Empirical Issues in Raising Equity Capital. Advances in Finance, Investment and Banking (Vol. 2). Amsterdam: Elsevier. Haralambides, H. E. (1993). Sensitivity analysis of risk in shipping finance. In: K. M. Gwilliam (Ed.), Current Issues in Maritime Economics. Dordrecht: Kluwer Academic. Heaver, T. D. (2002). The evolving roles of shipping lines in international logistics. International Journal of Maritime Economics, 4(3), 210–230. Johansen, S. (1988). Statistical analysis of cointegration vectors. Journal of Economic Dynamics and Control, 12, 231–254. Kavussanos, M. G., & Visvikis, I. D. (2004). Market interactions in returns and volatilities between spot and forward shipping freight markets. Journal of Banking and Finance, 28(8), 2015–2049. Koopmans, T. C. (1939). Tanker freight rates and tankship building, an analysis of cyclical fluctuations. Netherlands Economic Institute Report No. 27. Haarlem, Holland. Leggate, H. K. (1999). Norwegian shipping: Measuring foreign exchange risk. Maritime Policy and Management, 26(1), 81–91.
Key Themes in Shipping Economics Research
17
Leggate, H. K. (2000). A European perspective on bond finance for the maritime industry. Maritime Policy and Management, 27(4), 353–362. Li, K. X., & Cullinane, K. P. B. (2003). An economic approach to maritime risk management and safety regulation. Maritime Economics and Logistics, 5(3), 268–284. Li, K. X., & Wonham, J. (1999). Who is safe and who is at risk: A study of 20-year-record on accident total loss in different flags. Maritime Policy and Management, 26(2), 137–144. Marlow, P. B. (1991a). Shipping and investment incentives: A trilogy. Part 1: Investment incentives for industry. Maritime Policy and Management, 18(2), 123–138. Marlow, P. B. (1991b). Shipping and investment incentives: A trilogy. Part 2: Investment incentives for shipping. Maritime Policy and Management, 18(3), 201–216. Marlow, P. B. (1991c). Shipping and investment incentives: A trilogy. Part 3: The effectiveness of investment incentives for shipping; The UK experience, 1950–1987. Maritime Policy and Management, 18(4), 283–311. Marlow, P. B. (2002). Ships, flags and taxes. In: C. Th. Grammenos (Ed.), Handbook of Maritime Economics and Business (pp. 512–529). London: Lloyds of London Press. Morrell, P. S. (2002). Airline finance (2nd ed.). Ashgate: Aldershot. Notteboom, T. E., & Winkelmans, W. (2001). Structural changes in logistics: How will port authorities face the challenge? Maritime Policy and Management, 28(1), 71–89. Paine, F. (1990). The financing of ship acquisitions. Redhill: LRFairplay. Panayides, Ph. M., & Gong, X. (2002). The stock market reaction to merger and acquisition announcements in liner shipping. International Journal of Maritime Economics, 4(1), 55–80. Penrose, E. (1959). The theory of the growth of the firm. Oxford: Oxford University Press. Popper, K. R. (1934). Logik Der Forschung. Vienna: Springer. Porter, M. E. (1979). How competitive forces shape strategy. Harvard Business Review, 57(2), 137–145. Sjogren, H. (1999). Shipping as gambling: Governance mechanisms and the 1984 bankruptcy of Saleninvest. Scandinavian Economic History Review, 47(1), 48–64. Sletmo, G. K., & Holste, S. (1993). Shipping and the competitive advantage of nations: The role of international ship registers. Maritime Policy & Management, 20(3), 243–255. Slogett, J. E. (1999). Shipping finance (2nd ed.). Redhill: LRFairplay. Stephenson Harwood (1995). Shipping finance. London: Euromoney Books. Stokes, P. (1997). Ship finance: Credit expansion and the boom-bust cycle (2nd ed.). London: Wetherby Publishing. Tinbergen, J. (1931). Een Schiffbauzyclus. Weltirtschafliches Archiv, 34, 152–164. Tinbergen, J. (1934). Scheepsruimte en vrachten. De Nederlandsche Conjunctuur (March), 23–35. Veenstra, A. (1999). Quantitative analysis of shipping markets. Delft: Delft University Press.
2.
A SURVEY OF THE MODELLING OF DRY BULK AND TANKER MARKETS
D. R. Glen and B. T. Martin 1. INTRODUCTION The modern analysis of bulk shipping markets has been heavily influenced by two events in the past 15 years. First, the publication of Econometric Modelling of World Shipping, by Beenstock and Vergottis (1993b), provides an excellent survey of past econometric work in this area, and at the same time appears to be the last of its type. Their study is the most recent (to the author’s knowledge) fully specified structural econometric model of both the tanker and dry cargo freight markets. In that sense its publication marks a turning point in research in the area, as attention has shifted into issues raised by their study, using techniques that have been developed relatively recently. In general, most of what has been developed since has avoided the specification and estimation of complete structural models. Second, the revolution in econometric techniques generated by the development of cointegration analysis has changed the focus of researchers’ attention. Some of the most notable contributions to the literature in the past ten years have almost taken the competitive conditions of the bulk shipping markets as a given and focussed on the more specific aspects which had hitherto been neglected. In particular, modelling the behaviour of demand, trying to test assumptions about expectations, testing for conditional heteroscedasticity, examining seasonality, and trying to model the behaviour of ship prices, are all topics which will be explored
Shipping Economics Research in Transportation Economics, Volume 12, 19–64 Copyright © 2005 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0739-8859/doi:10.1016/S0739-8859(04)12002-7
19
20
D. R. GLEN AND B. T. MARTIN
later in this chapter. The one unifying theme which runs throughout these studies is the use of reduced form and VAR modelling – in other words, there has been an implicit rejection of the use of structural models of the type found in Beenstock and Vergottis (1993a, b).
2. MARKET MODELLING UNTIL BEENSTOCK AND VERGOTTIS 2.1. Early Econometric Modelling of Bulk Freight Markets1 An excellent survey of the major contributors to the econometric modelling of shipping markets is contained in Chapter 2 of Beenstock and Vergottis. Their review highlights the fact that in terms of modelling the freight rate and shipping fleet, little has changed since the seminal studies of Tinbergen (1931, 1934) and Koopmans (1939). Tinbergen’s (1934) study put forward a model of the dynamics of the freight rate market of the following form: F(t) = −rK(t)
(1)
DK(t) = Q(t − u)
(2)
Q(t − u) = lF(t − u)
(3)
. . . where F(t) is the freight rate, K(t) the fleet size (at time t), and Q represents orders, and (t − u) is the lag effect, with u a positive number and D the difference operator. Solving this simple differential equation yields a first order system defined as DK(t) = −␣K(t − u)
(4)
. . . which will generate cycles over time if the parameters that determine ␣ are of the correct value. In an earlier paper, Tinbergen (1931) had specified a two equation model which assumed demand as exogenous and equal to supply, which was determined by the fleet size, costs, which are proxied by bunker prices, and the freight rate. Beenstock and Vergottis (1993b, p. 73) show that the implied elasticities derived from the model are 0.94 for fleet size, −0.23 for bunker prices, and 0.59 for freight rates. In other words, fleet size is proportional to demand, which itself is inelastic with respect to freight rates and bunker costs. These original contributions were notable for first introducing the idea of a nonlinear short run supply curve. This device allowed the freight rate’s response to changes in demand to be contingent upon short term market conditions: when there was low levels of fleet utilisation, a demand change would not alter freight rate
A Survey of the Modelling of Dry Bulk and Tanker Markets
21
levels much, but when existing fleet utilisation was high, increases in transportation demand would have a disproportionate effect on rates. In other words, the freight rate supply elasticity was high at low utilisation levels, but very low at high utilisation levels. Whilst the work of Tinbergen and Koopmans formed the foundation of what might be called the classical econometric approach to modelling shipping markets, other approaches have also emerged. A notable contribution is to be found in the work of researchers at the Norwegian School of Economics. Norman and Wergeland (1981) developed Nortank, a simulation model of the freight market for large tankers. By using microdata on four different types of large tanker, the model tries to determine optimum speed on a representative round trip, assuming that each owner tries to maximise profits per trip, given fuel prices, freight rates, capacity utilisation and port time. The vessel speed range is limited, and this implies the possibility of corner solutions to the optimisation problem. Individual supply functions for each type can be discontinuous at the point at which the ship type moves into layup. This approach generates a kinked supply curve for each ship type. The interaction of an inelastic demand for tonne miles and the derived supply relationship thus determines market freight rates for this type of tanker (Beenstock & Vergottis, 1993b, p. 78). Whilst this approach to modelling the tanker market is interesting, it has not been widely adopted by others. The most popular approach has been to adopt the estimation of aggregate time series models, first pioneered by Tinbergen.
2.2. Other Aspects of Bulk Market Behaviour Both Koopmans and Tinbergen studied the dynamics of the spot freight rate in the context of the freight market itself. But both dry bulk and tanker markets are quite complex, with a range of contractual arrangements available for cargo transportation, ranging from spot (trip charter contracts), through consecutive voyage charters, via contracts of affreightment (contracts for moving cargo irrespective of specific ships employed), to time charter contracts (contracts under which the charterer takes over increasing amounts of responsibility in running and operating the vessel in exchange for a fixed monthly payment to the shipowner). The widespread existence of these contracts meant that researchers became interested in looking at the relationship between long term charter rates and market spot rates, as well as spot rate determination. In addition, researchers became interested in modelling the behaviour of ship prices, as these are but another aspect of the freight markets, being the price of the assets employed to produce the transport service itself.
22
D. R. GLEN AND B. T. MARTIN
Strandenes (1984) studied the relationship between time charter rates and second-hand ship prices. Efficient markets theory would predict that the present value of a time charter contract should be equal to the expected present value of the spot market profits (before fixed operating costs) over the duration of the contract, after adjusting for differences in risk. The measure of profits before costs is called the Time Charter Equivalent, (TCE). In her model, the time charter rate at time t of duration H tt is determined by two measures, one of the short term time charter equivalent, measured as the current value, and the expected long term time charter equivalent. The sum total of these two components was shown to be close to unity for three different time charter periods (less than 12 months, between 12 and 36 months, and over 36 months), for bulk carriers, medium size tankers, and large tankers. The fact that the values were close to unity suggested that there was no significant risk premium, at least for the estimation period 1967–1983. In the same study, Strandenes explored the relationship between ship prices and short and long term profits (note, not TCE measures). Again, the sum of the coefficients estimated should equal one, if the model was to be consistent with the efficient markets hypothesis. In effect her model separated out the ship price into two determinants, the present value of short run and the present value of expected long run profits. If the efficient markets hypothesis is correct, the value of second hand ship prices would be entirely explained by these two terms. Indeed, her empirical results suggested that for tankers the hypothesis is accepted. For Panamax bulk carriers, the hypothesis was rejected. Strandenes (1986) has also developed a model which integrates the separate models developed at Bergen for dry bulk, tankers, and shipbuilding and scrapping. Norship divided both tankers and dry cargo vessels into two size classes. It incorporates a switching mechanism in the form of combination carrier capacity, which altered their employment between dry and oil sectors in accordance with relative market profitability. Demand and supply equilibrium is imposed by matching the total demand for tanker transportation services to the supply, from the large and small tanker fleet and those combination carriers trading in oil. Similar conditions are imposed on the dry bulk sector. Time charter rates are modelled as in Strandenes (1984), as indeed, is the relationship between ship prices and earnings. Long term profitability L is measured by the difference between the expected long term charter equivalent and long term operating costs (OC). If capital markets are competitive, arbitrage would ensure that the newbuilding price of a ship would be valued at the expected present value of its future earning stream, which in this case is given by: Pn =
1 (L − OC) r+d
(5)
A Survey of the Modelling of Dry Bulk and Tanker Markets
23
. . . where r is the rate of interest and d a depreciation factor. This relation allows L to be determined in terms of the known values of the other terms in Eq. (5). Secondhand ship prices are then derived using a weighted average of short and long term profits, with a separate relation linking operating costs, long term profits, and long term time charter equivalent L . Scrapping supply is assumed to be proportional to the ratio of scrap prices to the ship market price for each ship type. On the other side of the market, the demand for new ships is assumed to be positively related to the expected future stream of profits and negatively related to the current newbuilding price. The first term, V, is defined as 1 V=k (H ␣ − OC) (6) r+d . . . where k varies according to the delivery lag on the newbuilding contract. The demand for new ship type i is given by V i ki D ni = ␣i (7) P ni On the shipbuilding side, the newbuilding price is assumed to equate to the marginal cost (net of subsidy). Marginal cost is assumed to be positively related to total construction relative to exogenous shipbuilding capacity. Norship is then used to generate simulation results given shifts in the value of exogenous variables. Beenstock and Vergottis (1993b, p. 84) report the results of one such experiment, when the price of bunker fuel is reduced by 36% in the second year of the simulation. Briefly, the fall in fuel prices leads to a permanent reduction in both tanker and dry cargo freight rates, with a greater reduction in tanker rates, reflecting their higher relative sea time. Combination carriers switch their employment more towards dry cargo, in response to the change in relative freight rates. Second hand prices fall, as does time charter equivalent earnings. This result is generated because the net effect of the fall in bunker prices on freight rates and profit is negative. (The reduction in operating costs is more than offset by the induced reduction in freight revenues since freight demand is very inelastic.) Hawdon (1978) develops and estimates an econometric model of the behaviour of annual average tanker spot rates for the period 1950–1973. The model has seven equations which explain the behaviour of the dry cargo and tanker spot markets. Two reduced form equations describe the behaviour of the dry cargo and tanker spot rates, in terms of demand volumes, factor input prices, and the spot rate in the other shipping market. No cross market substitutability was detected. The second part of the model consists of five equations determining the behaviour of orders, ship prices, deliveries and scrapping. These were found to
24
D. R. GLEN AND B. T. MARTIN
be dependent on current and past spot rates, existing fleet size, the price of steel, and trade volumes. Hawdon finds that changes in the overall volume of oil, bunker costs, and fleet size all contribute to the explanation of changes in the annual average spot rate. Neither labour costs nor average size were found to have an impact. The average size variable may be highly correlated with the trade volume measure, which may explain its insignificance. Having estimated the model using OLS and TSLS, Hawdon employs it to estimate the long run elasticities of the rate index to exogenous shifts in the volume of oil traded, and bunker costs. In the case of the former, the short run elasticity is 3.1, whilst the long run elasticity is −0.6. The response to bunker costs is similar in both time frames, with values of 1.7 and 1.9 respectively. 2.3. Charemza and Gronicki’s Disequilibrium Model Charemza and Gronicki (1981) present a very different perspective to the usual maintained hypothesis of market clearing price adjustment. Their model is notable for the fact that they permit demand and supply to differ in any given time period. Instead, they assume that the change in the market price is proportional to the level of excess demand computed in that period. The determining value of output is the “short side” of either demand or supply quantity in any given period (that is, the minimum of the two values for quantity supplied and quantity demanded). The model has a section describing the freight market, and a block defining shipbuilding and scrapping. The demand for freight services is driven by world trade, and is independent of freight rates. The supply of services is driven by fleet size and freight rates. Long term charter rates are modelled in terms of spot rates and the level of demand in relation to fleet size. The model is estimated in linear form (most are logarithmic) for the period 1960–1983. The results for this section imply that both the dry cargo and tanker spot markets are influenced by changes in the size of the fleet and in freight rates. Charemza and Gronicki’s results for long term time charters imply that they are positively affected by spot rates and the level of demand in relation to the fleet. The block describing shipbuilding and scrapping is not explained in detail. The interesting point to be made about this study is that there appears to have been very few published studies which have adopted a similar approach to the modelling of either tankers or dry cargo markets since. 2.4. Term Structure Models of Shipping Freight Rates Zannetos (1966) was one of the first scholars of the tanker market to point out the term structure like relationship between spot rates for ship hire and those offered
A Survey of the Modelling of Dry Bulk and Tanker Markets
25
for period chartering (time charters). Time charters for vessels usually run from one year to 10 years or more. At the extreme, the bareboat charter is in effect an instrument for financing a vessel, as the “owner” gives rights to the complete operational duties of running, crewing and insuring the vessel to the charterer. When the vessel is hired for the year, the owner pays the operating and capital costs, but not the voyage related costs. In exchange, a monthly hire fee is paid. Zannetos argued that the long run bareboat charter rate should in effect, be the long run marginal cost of providing the vessel. Shorter term hire rates should converge (with time period lengthening) either from above (in good markets), or below, (in bad markets) to this long run cost, which is of course the same for the ship irrespective of the type or market (spot or period), it is engaged in. This idea has been developed by several authors, noting the similarity between ship hire rates and short and long term interest rates, from whence the name was borrowed. If the shipping markets are efficient, owners should obtain the same net rate of return from their asset, irrespective of where it is employed. Given that the levels of risk borne by the owner and charterer varies across these arrangements, it might be argued that the time charter profit rate of a fixture of a given duration should equal the present value of the expected stream of earning generated if the vessel where otherwise employed in a sequence of spot transactions of the same overall duration. This forms the basis of the model developed in Glen et al. (1981), and used by Hale and Vanags (1989) and Veenstra (1999) amongst others. Glen et al. develop a continuous time model of the relation between time charter and spot rates and operating costs, which is then approximated by a discrete model relating the time charter rate to past values of the spot rate. The model is estimated using quarterly data derived from individual fixtures for tankers of a specific size range for the period 1970–1977. They use the model to try to test the Zannetos hypothesis of elastic price expectations, and argue that the results suggest that shipowners are risk averse, a result which is corroborated in findings reported in Strandenes (1984). Hale and Vanags (1989) explore the term structure hypothesis using rational expectations and market efficiency considerations as their twin assumptions. They derive a formal relationship between the period hire rate for ships, and the expected spot rate. They show that the spot equivalent of the period rate (i.e. adjusted for operating costs) should be equal to a weighted average of expected future spot rates for the equivalent period. Their model assumes that employment prospects are identical in the two states. They show that the one period change in the spot equivalent period rate is related to the “spread” between the lagged value of itself and the spot rate that prevailed, plus a term in the difference between the spot rate lagged one period and the present expectation of the spot period n periods hence. This second term is modelled in three different ways in their econometric work.
26
D. R. GLEN AND B. T. MARTIN
Hale and Vanags’ results were not encouraging for the term structure relationship between spot and time charter rates. Examining the dry bulk sector using monthly data for the period 1980–1986, they found that the coefficient on the spread was negative and insignificant, for two of the three ship sizes for dry bulk vessels. However, Veenstra (1999, p. 60) has shown that these results suffer from “omitted variable” bias, and his correction of this error dramatically altered the results (see below).
2.5. The Beenstock and Vergottis Model In the late 1980s and early 1990s Michael Beenstock and Andreas Vergottis published a series of papers in which they developed an integrated econometric model of the tanker and dry cargo markets (Beenstock & Vergottis, 1989a, b, 1993a). The work reported in these articles was further developed and combined with Beenstock’s study of the theory of ship prices (Beenstock, 1985) and published in book form (Beenstock & Vergottis, 1993b). The key feature of this work, in the authors’s opinion, is not in the econometrics; rather it is the seminal development of a coherent explanation of ship price behaviour, which is grounded in the application of the two basic hypotheses of rational expectations of freight prices, and market efficiency. The first assumption forces the predicted values of the freight rate which are generated by the model to be the expected values of the freight rate held by the market participants. This implies that the expectations are internally consistent with the predictions of the model. In other words, each agent acts as if the market expectation that they hold is generated by the estimated model of the shipping markets. Note that this implies all agents must hold the same expectations, if all agents act as if they know the true “model” which drives the market process. The second assumption, that of market efficiency, ensures that market prices clear excess demand or supply in any one period, and arbitrage will ensure that prices of assets will adjust to new market conditions as soon as they are known in the market. In Beenstock’s (1985) model, the time path of expected ship prices is derived, and it is shown that expected ship prices are a function of the past history of exogenous variables, and expectations of the future values of the same drivers. The difference relation has two roots, one of which is stable (lies inside the unit circle), the other of which is unstable, and represents the possibility of speculative bubbles. The rational expectations hypothesis assumes that the coefficient of the second root is zero, thus ensuring convergence. The theoretical model in the 1985 paper was then “simulated” with the use of numerical estimates of the various parameters, to show the response to anticipated and unanticipated shocks in world
A Survey of the Modelling of Dry Bulk and Tanker Markets
27
trade. This illustrated how the theory permitted prices to alter temporarily, in response to temporary shocks, but then returned to their initial values. In the case of a permanent response, exaggerated shifts in price occurred, generating cycle like behaviour in the ship price. Beenstock and Vergottis then published two papers outlining separate models, one for dry cargo (1989) and one for tankers (1989b). Both models use the same theory of ship price determination, and the same basic model relating freight rate behaviour to the newbuilding and second hand markets for ships. A later (1993a) paper developed a model relating the two sectors together, via the switching capacity of combination carriers, and the fact that shipyard capacity could either be employed in the production of dry cargo vessels, or in the production of tankers. This integration forms the basis of chapter 6 of the authors’ Econometric Modelling of World Shipping (1993b). At the conclusion of this chapter, they note that the spillover effects between the shipping markets, although measurable, are “relatively weak” (1993b, p. 228). After a thorough exposition of both discrete and continuous time theoretical models of the determination and behaviour of freight rate and ship prices, BV report on the estimation of their econometric models. The models for both dry cargo and tankers have common characteristics. The freight rate is determined by the proportional difference between quantity demanded q (in tonne miles), and the supply of ship services (measured by the fleet tonnage, K v ). Demand is taken as exogenous, and freight is a function of the balance between the exogenous demand and the active fleet. The active fleet is a proportion, 1 − , of the actual fleet, and is itself a function of freight rates, bunker prices, operating and lay up costs. The section analysing freight rates is then completed with the introduction of the relation between time charter rates, bunker prices, and expected freight rates. The BV theory of freight rate determination argues that the expected profitability of time chartering activity must equal the expected profitability of spot chartering over the same duration, plus a risk premium. This leads to the specification of the present relation, in which the present time charter rate is positively related to the present expectation of next period’s spot rate, and negatively related to the present expectation of next period’s bunker price. The (logarithmic) difference between these two coefficients is restricted to unity, because a 1% reduction in the expected profitability of trading in the spot market (i.e. the difference in the two terms above) must be translated into a 1% reduction in the expected time charter hire rate, which is the measure of period trading profitability. The BV models are estimated using annual data from various time periods. For the dry cargo market for example, the freight market and time charter rate segments are estimated using 3SLS with 24 observations, from 1962 to 1985 (Beenstock & Vergottis, 1993b, Table 4.2, p. 175). The shipbuilding sector utilises
28
D. R. GLEN AND B. T. MARTIN
data over 1952–1986, and the newbuilding price relation is estimated for the period 1960–1985. As Veenstra (1999, p. 38) points out, the empirical estimates generated by the Beenstock and Vergottis models are similar to earlier studies. For example, the inverse of the parameter on (q − K) is an implied estimate of the short run elasticity of supply with respect to the freight rate. The values are 0.24 and 0.31 are for dry bulk and tanker markets respectively. The models developed and presented in Beenstock and Vergottis (1993b) are a high water mark in the application of traditional econometric methods. They remain the most recent published work that develops a complete model of freight rate relations and an integrated model of the ship markets. It is a high water mark because the tide of empirical work has turned, and shifted in a new direction. This change has occurred for three reasons: first, the development of new econometric approaches which have focussed on the statistical properties of data; second, the use of different modelling techniques; and third; improvements in data availability have meant a shift away from the use of annual data to that of higher frequency, i.e. quarterly or monthly.
3. MODELLING MARKETS POST BEENSTOCK AND VERGOTTIS 3.1. Developments in Econometric Modelling The modelling of bulk shipping markets that has emerged during the 1990s can be characterised by three elements. First, the use of reduced form autoregressive models has replaced the estimation of structural ones. Second, and related to the first, has been the dissemination of new methods of econometric practice into shipping market applications, in the form of extensive use of stationarity testing and examining data for cointegration (common integration). Where data series are integrated of order one, long run relations between the levels of the variables can only exist if the vector which relates them is integrated of order zero. This is because an integrated series of order one is stationary after first differencing, whilst the series will have a trend when left undifferenced. Engle and Granger (1987) developed a method of establishing whether or not a long run relation between the variables existed. In their approach, a simple OLS regression would provide estimates of the long run relation between the variables. However, the t-ratios would be ignored, because of the non-stationarity of the series. Instead, the residuals from the original regression were saved, and then used as an error correction term in a regression on the first differences of the
A Survey of the Modelling of Dry Bulk and Tanker Markets
29
major series. This too, is estimated by OLS. Because the latter is now a regression on stationary variables, OLS is appropriate. This is because a cointegrated set of variables will be integrated of order zero (I(0)) if the original series are all I(1). Independently of Engle and Granger, Johansen (1988) published a seminal paper in which he developed a method for analysing the existence of cointegration in the context of a vector autoregressive (VAR) model. He showed that such a model, written in levels form, can be transformed into first differences. This transformed model would also contain a generalised form of Engle and Granger’s error correction term, containing the levels of the cointegrating (i.e. long run) relationship. The example below should clarify. Johansen’s general model for a vector autoregressive system of n variables is written as: Y t = + ␦t +
p=1
i Y t−I + Y t−p + e t
(8)
i=1
. . . where Yt is an (m × 1) vector of variables, with a possible common constant trend, , as well as a possible quadratic trend in the levels given by ␦t. The and matrices are linear combinations of the parameters of the original form of the relationship when expressed in levels form, which is given below as: (L)Y t = + e t
(9)
. . . where (L) is the lag operator of order p. The two equations differ only in that Eq. (8) includes a trend term for the differenced values of the variables; the matrix contains the components of the long run cointegration relations between the variables, if they exist. If the rank of , r(), is zero, no such relations exist, even though the first differences are stationary; if r() is of order m, then the variables are stationary in their levels form; if r() is less than m but greater than zero, then there are m − r cointegrating relations, which can be extracted from the coefficients of the matrix. The implication for modelling of these new approaches has been profound. The recent focus has been on re-examining the statistical properties of shipping market data, and then exploring the data set in terms of the existence or otherwise of cointegrating relations between the variables. Since the original VAR models make no distinction between endogenous and exogeneous variables, it is clear that the recent research agenda has shifted away from the estimation of large econometric models of shipping markets, and towards examining reduced form dynamic relations using the relatively new methodologies generated by the work of Engle and Granger, and Johansen. To the authors’ knowledge, there have been no published studies of the bulk shipping markets that have followed the direction indicated by Beenstock and Vergottis.
30
D. R. GLEN AND B. T. MARTIN
3.2. Econometric Research post Beenstock and Vergottis Since the “high-water mark” of Beenstock and Vergottis’ book, there has been a shift of orientation in the published studies of the tanker and dry bulk markets. For reasons outlined above, scholars gave much more attention to the statistical properties of the data sets they employed prior to model estimation. The newly developed tests of stationarity required larger data sets to be effective, and this requirement is reflected in the second noticeable trend, that of using quarterly or monthly data, in place of annual observations which (with one or two notable exceptions, had been a feature of past studies). The third shift in orientation was toward the use of vector autoregressive or reduced form models of rate behaviour. Finally, new models, developed for modelling financial data, were applied to the analysis of shipping markets. 3.2.1. VAR Modelling of Freight Market Relations Veenstra (1999) provides a good example of the convergence of some of these themes. He re-examines the term structure relationship between spot and time charter rates in the context of a VAR model which embodies restrictions imposed by the assumption that markets are efficient, and that there is a definite link between the time charter and voyage rates, as laid out for example, in Glen et al. (1981). The model and approach is as follows: If market efficiency rules, the shipowner should be indifferent between the choice of a series of spot charter fixtures which generate the same profit as a time charter hire of an equivalent duration, plus a (positive or negative) liquidity premium. In the context of a two variable system, Veenstra (1999, p. 202) expresses this model as: 1 i ␦ E t sp t+i + 1 − ␦n n
St =
(10)
i=1
. . . where St represents the “spread” or difference between the time charter and spot fixtures in time period t, and Et spt +i is the expected one period change in the spot rate, ␦ is the discount rate and is a positive or negative liquidity premium. In order to proceed, the unobservable expectations variable has to be replaced by observable variables. Campbell and Shiller show how this can be done by utilising the companion form of the VAR model. The companion form of a VAR model of order (i.e. any high order model) allows the reparamaterisation of the model into a first order difference equation which can be solved by repeated substitution. The VAR model of interest here is shown in Eq. (11) below. To derive the restrictions implied by the present value model, a number of steps are required. First they show that the spread variable can be defined as in Eq. (10)
A Survey of the Modelling of Dry Bulk and Tanker Markets
31
above. They then define a VAR model, involving the first difference in time charter rates (in the present case), and the spread variable. If spot and time charter rates are integrated of order one, but cointegrated, it follows that both of these variables will be stationary as defined, so the VAR model contains stationary variables as required. This VAR (1) model in is presented below for the first order case (in deviation from means form). spt ␣11 ␣12 spt−1 e1t = + (11) St ␣21 ␣22 St−1 e2t The present value restrictions are derived as follows. The system in Eq. (11) can be written as: z t = Az t−1 + vt
(12)
. . . where A is the companion form of the VAR matrix (in the special case of p = 1, the companion form and the original matrix are identical). For all i time periods the expected values of z given the available information set H, will be given by E(z t+i |Ht) = Ai z t
(13)
Thus there is a method available for eliminating the expectations terms in (10) as follows. The spread equation can be rewritten as:
g zt =
∞
␦i h A i z t
(14)
i=1
The g vector will be p + 1 by 1, with the unit value appearing in the p + 1th position (zeroes elsewhere), so that g z equals St . In order to pick out spt , vector h is defined with unity as the first element and zeroes elsewhere, i.e. h = [1, 0] in the first order case. As all elements of z are stationary, the infinite series in (14) converges, so that g may be rewritten as: g = h ␦A(I − ␦A)−1
(15)
Postmutiplying by (I − ␦A) yields: g (I − ␦A) = h dA
(16)
In the case of the first order system specified in Eq. (11), g = [0, 1]; h = [1, 0]; and the system can be written as: ␦␣11 ␦␣12 (1 − ␦␣11 ) (−␦␣12 ) (17) = [ 1 0] [ 0 1] (−␦␣21 ) (1 − ␦␣22 ) ␦␣21 ␦␣22
32
D. R. GLEN AND B. T. MARTIN
From which it follows that: −␦␣21 = ␦␣11 ,
or ␣21 = −␣11 , and 1 (1 − ␦␣22 ) = ␦␣12 , or ␣22 = − ␣12 ␦
(18)
The first restriction implied in equations (18) is that the sum of the coefficients on the spot rate variable in a first order VAR must equal unity. The second implies that the sum of the estimated VAR coefficients for the “spread” equation must equal the inverse of ␦, the constant discount rate applied for the present value calculations. Generalising the order of the VAR creates additional restrictions, and the companion matrix becomes more complex, but the essence is captured in the above relations (see Appendix for the second order model). In addition to the restrictions derived for the present value model, Campbell and Shiller (1987) constructed two tests of the amount of volatility “explained” by the model. They defined the “theoretical spread” as S t = E(S ∗t |Ht) = h ␦A(I − ␦A)−1 z t
(19)
. . . which of course is the value predicted from the VAR model when written in companion form. The theoretical spread is given by ∞ 1 i (20) ␦ Et X t+i − (X t+i−1 − (1 − ␦)x t+i−1 )|H t St = ␦ i=1
This is calculated from the model, and the ratio of the variances of the actual and theoretical spreads var(S)/var(S ), is expected to be unity, if the EMH and PV assumptions apply. Veenstra (1999) applies this approach to the spread of time charter and voyage rates for three size classes of dry bulk and tanker vessels, and argues that the results provide support for the “term structure” model of freight rate relations. In order to apply this approach however, he had to make the assumption that the time charter fixture duration was infinite, following Campbell and Shiller’s (1987) model. In addition, the discount rate had to be constant over time. In a subsequent paper they (Campbell & Shiller, 1988) develop a model which is appropriate for assets with finite lives, and also permitted a “time varying” discount rate. This modified version forms the theoretical basis of the study of Kavussanos and Alizadeh (2001). Using the same methodology as Veenstra, the authors’ develop a model of ship price behaviour in terms of the asset price, the operating profit of the vessel whilst trading, and the expected realisation gain or loss when the vessel is scrapped. In their model, there are two spreads, one between the ship price and the operating profit, the other between the ship scrap price and
A Survey of the Modelling of Dry Bulk and Tanker Markets
33
operating profit. The measure which is modelled in the system, along with the two spreads, is the series for the first difference in operating profit, discounted by the time varying discount rate. These three form the VAR model which is used to generate excess returns, which are then used as inputs in a further modelling exercise, designed to examine the behaviour of the rates of return in terms of their modelling of rate volatility itself. Wright (2003) uses the same techniques to test the rational expectations hypothesis by generating ex post estimates of the one year spot rate, which is the average of the present spot rate and the next eleven months spot rate, since the period rate is a one year time charter. He shows that the null hypothesis of strict proportionality between the spot rate and the constructed expectations set and the period rate cannot be rejected – he concludes that the rational expectations hypothesis is accepted as a long term relation in this market. Other researchers have also applied VAR techniques. Mirmiran (2002) developed VAR models as part of his analysis of the impact of the Common Agricultural Policy on the trade in grain. Because of data limitations, VAR models were developed for two dry cargo ship sizes. Using the resulting relationships, he was able to examine, by simulation, the potential impact that would have been felt in these markets if two competing proposals (those of the USA and the European Union), designed to eliminate the export surplus in grain created by the CAP, had been implemented. The models were estimated separately, given the empirical evidence that ship sizes were segmented. OLS could be used to estimate the VAR models in these cases. 3.2.2. Time Varying Volatility – ARCH and GARCH Models A second theme that emerges from a review of the more recent literature is the application of models which relax the standard econometric assumption of constant volatility, as expressed in terms of the normally distributed error term with constant variance over time. Two such models are the autocorrelated conditional heteroscedasticity (ARCH) or generalised autocorrelated conditional heteroscedasticity (GARCH), first developed by Engle (1982), and extended by Bollerslev (1986). These models allow both the conditional mean and variances of the data set to be modelled simultaneously, allowing some of the variance to be modelled as well as the means of the target series. This permits the variance to change over time, thus relaxing a key element of the assumptions of the classical regression model. The original model (ARCH) was developed to allow for the persistence of periods of greater and lesser volatility of the data. GARCH models, generalised this approach, so that the modelling of the conditional variance can become as sophisticated as the modelling of the conditional mean of the data. An important contributor to this field has been Kavussanos (1996a, b, 1997). Using monthly data for the period 1973:1 to 1992:12, Kavussanos (1996a) models
34
D. R. GLEN AND B. T. MARTIN
the behaviour of the average value of spot and time charter rates for three sizes of dry cargo vessel, and the volatility of the rate over time, using a model of the following form: yt = xt b + t ; t ∼ N(0, ht ) p q ht = ␣0 + i=1 ␣1 2t−1 + i=1 iht−1
(21)
. . . where x t is a vector of independent variables, yt the dependent variable, t the error term with expected mean of zero, and non-constant normally distributed variance given by ht , whose behaviour over time is modelled in the second relation in (21). In the formulation above, the error term is a function of the p squared values of past random errors, and q past values of the conditional variances ht . The model is estimated using maximum likelihood. In Kavussanos’ estimations, a variety of specifications were found to be optimal, ranging from ARCH (1) for the aggregate freight rate, a GARCH(1,2) for time charters, and GARCH (1,1) for the three size categories of dry cargo. The set of variables employed to explain the conditional mean varied slightly, but for the spot rates, they included lagged values of the relevant spot rate, bunker prices, industrial production, and the stock of the world dry bulk fleet. For time charter modelling, the significant independent variables were lagged time charter and spot rates, with no empirical role found for bunker prices, in contrast to those expected from the application of Beenstock and Vergottis’ model in a “reduced form.” Kavussanos tested for stationarity of the data series employed, using DF and ADF tests. All series were I(1) except for the measure of the fleet, which was I(0). As is well known, if the I(1) variables cointegrate, their linear combination can be I(0); presumably in these models, the linear combination implied is between those variables excluding the dry cargo fleet which, then forming an I(0) vector, can be regressed together with the fleet measure. The implication is that the fleet measure may not cointegrate with the rest of the model, which appears to be counterintuitive. There is no discussion of this in the article. Kavussanos (1996b) applied the same approach to tanker market behaviour, applying the ARCH approach to three ship sizes (Aframax, Suezmax & VLCC). Here the focus is on the behaviour of ship prices rather than on freight rates. However, he does discuss the possibility of ARIMA or ARIMA-X modelling of freight rates as an adjunct to the prime focus, which is on the modelling of risk. Data is again monthly, from 1980 to 1993, and prices are adjusted by the USA CPI to create a series in real terms. He finds, among other things, that the real oil price affects the variance of both Suezmax and VLCC ship prices, and permits more efficient modelling of both the conditional mean and variances of VLCC ship prices.
A Survey of the Modelling of Dry Bulk and Tanker Markets
35
Turning his attention to the dry cargo sector once again, Kavussanos (1997) explores the issue of the existence or otherwise of stochastic seasonal effects on second hand ship prices. The innovation in this paper is to examine for the significance of seasonal factors. He finds that there are no significant seasonal unit roots, other than at “zero frequency” (Kavussanos, 1997, p. 435) which implies that the series are I(1) in annual terms. These first differenced series again form the basis of an ARCH model of means and variances of second hand prices. He finds that time charter rates help explain the behaviour of second hand prices, whilst variations in the LIBOR rate affect the conditional variance, for all three size classes. There is no attempt to explain why different structural variables (originally identified by Beenstock and Vergottis in terms of the conditional means) should enter the different components of the ARCH model in different ways. Kavussanos’ own work has itself led to other researchers applying the technique. Glen and Martin (1998) revisited the tanker market, using freight rates derived for specific routes and a given ship size. They confirmed Kavussanos’ findings that a GARCH specification provided superior results to that which did not jointly model the variability of the freight rate, in the context of data specific to a given trading route, rather than using a generalised index for a given ship type across all routes. 3.2.3. Other Recent Contributions A number of other contributions have also been made to the literature, which have focused on one particular aspect of bulk market behaviour. Wright (1999) used cointegration techniques to study the “degree of integration” of the three tanker sizes classes represented by the Lloyds Shipping Economist data, namely, 30,000 dwt, 130,000 dwt, and 250,000 dwt. Wright argued that complete integration of the market segments required that they cointegrate, which he establishes for the three spot rate series together with the one year time charter rates for the 30,00 dwt tankers. He found three cointegration vectors between the four variables, implying long run integration. He then tests the hypothesis that the restricted cointegrating relationship is pairwise between the smallest size spot rate index and the other three rate indices. This is tested against the alternative hypothesis of a spread plus a constant. The restriction is accepted. He concludes that the wet bulk market is highly integrated. Note that this view of the market is focussed entirely on the mean values of the rates, and ignores any consideration of the risk differences that might exist, as has been noted in the review of the work of Glen (1990) and Kavussanos (1996a, b). Berg-Andreassen (1997) has explored alternative means of modelling expectations formation in the context of time charter rate determination. Five different alternative formulations are developed, namely, the “Zannetos hypothesis” (Zannetos, 1966), the Koyck Lag hypothesis (derived from adaptive
36
D. R. GLEN AND B. T. MARTIN
expectations), the rational expectations hypothesis, and the “conventional wisdom hypothesis.” As Veenstra (1999, p. 60) points out, the author’s tests of the relative performance of these models is flawed because of an inappropriate application of modern techniques. However, there is merit in trying to answer this specific question, given the evidence noted above, that the ‘dominant’ model of rational expectations appears to be rejected when tested in an appropriate framework (Kavussanos & Alizadeh, 2001). Although not primarily concerned with freight rates per se, Tsolakis et al. (2002) attempt to make a comparison of the relative efficiency of VAR and structural modelling when trying to forecast the shipping business cycle. Their paper is interesting in that it is the first to make such a comparison of alternative modelling approaches, rather than selecting a procedure a priori. They present mixed results with regard to the superiority of VAR over a structural model – in some cases, a two period VAR outperformed the structural model, in others not. This paper is notable for the use of annual data, which limits the degrees of freedom quite substantially when using the VAR methodology. Recently Tvedt (2003) has challenged the overwhelming consensus that has emerged from the above literature, namely the first difference stationarity of most shipping time series. Whilst accepting that the data for dry bulk markets is first difference stationary, he argues that when the data is transformed by conversion to yen from dollars, it becomes stationary in levels form. He justifies this transformation by claiming that yen is the more appropriate currency, because the Japanese economy is a major driver of dry bulk shipping markets. He says nothing about the stationarity properties of the yen dollar exchange rate, however, which would be of interest. Essentially Tvedt provides a multiplicative transformation of the original series, call it x, the dollar value of the series in question. If this is multiplied by the yen/$ rate, (z), the yen value of the series (y) is obtained, i.e: y =x·z
(22)
A log transformation generates a linear combination: ln y = ln x + ln z
(23)
But a linear combination of two I(1)series will only be I(0) if they are cointegrated and are both I(1). If one series is I(1) and the other is I(0), the new series will be I(1) so Tvedt’s finding is valid only if the exchange rate series is also I(1). This is not discussed in the article. Even if it were the case, the case presented for the transformation appears to be very weak; despite Japan’s importance, there are many possible currencies that might be candidates for such transformations, such as those of China and South Korea.
A Survey of the Modelling of Dry Bulk and Tanker Markets
37
Two other aspects that emerges from this review of the literature is the limited numbers of studies that examine the relations between trade flows and shipping markets. Apart from Mirmiran (2002), two notable exceptions stand out. Kavussanos’ (1996c) modelling of disaggregate bilateral dry cargo trade flows, and Veenstra’s (1999) study of the cointegration relations that exist between the specific routes for the principal dry cargo and oil trades. Given the reliance of the literature on a demand elasticity of zero, this is not surprising. The second point of interest is the lack of any studies at a micro level. For example, Glen and Martin (2002) have tried to explore the implications within the VLCC sector of the existence of a tanker pool, in terms of its effects on rates and earnings.
4. EMPIRICAL WORK The review of the literature presented in Sections 2 and 3 reveal the marked change in data use, methodology and modelling strategy that has evolved over the past decade. Higher frequency data, testing for stationarity, examining series for cointegration, the use of VAR modelling techniques, have all figured heavily. In addition, applying ARCH and GARCH to model the means and variances of the data series are all recent developments that have contributed to the present state of knowledge on the modelling of shipping freight markets and the associated asset markets for ships themselves.
4.1. Re-estimation of the Veenstra VAR Model The first step taken in the empirical work reported in this section is to revisit Veenstra’s study (Veenstra, 1999) of the relationship between spot and time charter rates for tankers and dry bulk vessels. The Lloyd’s Shipping Economist data used in that study was extended to its maximum; in 2003 the data on time charter voyage equivalents ceased to be published, thus providing a time series from October 1980 until February 2003, 269 observations. The published data series are for both spot rates, and time charter rates for three different size classes of vessels, and for both oil and dry cargo. For tankers, the relevant sizes are 30,000 dwt (Handy), 115/130,000 dwt (Suez) and 250,000 dwt (VLCC). Tanker rate data is available in both dollars per ton of cargo and in Worldscale, a standard tanker industry index of freight rates. For dry cargo, the relevant data series are for 30,000 dwt (Handy), 55,000 dwt (Panamax) and 120,000 dwt (CapeSize). The time charter data2 is available in both Worldscale and dollars per ton of cargo for tankers, and in dollars per ton of cargo for dry bulk vessels (dry cargo). Data for the smallest size dry
38
D. R. GLEN AND B. T. MARTIN
cargo ship is spliced from two series, and the series for the 130,000 dwt tanker had also to be constructed, following Veenstra (1999). All other series continued on a comparable basis until February 2003, when the data was no longer published. It should be noted that the time charter data is expressed in dollars per ton of cargo, which is an estimate, derived from the market hire rate of the time charter, expressed in dollars per month for vessel hire, which is then converted to a rate expressed in terms of cargo delivery costs. This calculation requires assumptions about the employment of the vessel on a specific route, as well as making assumptions about the cost of fuel consumed on those voyages, during the given hire period. The use of this measure of time charter rates means that both spot and time charter series are comparable, which is important given that the spread variable is the difference between the time charter and spot voyage rates, so they should be in the same units of measurement. The question to be addressed is: Is the restricted VAR model estimated by Veenstra still appropriate for the extended data set? The answer appears to be yes. In order to estimate this model, a number of stages have to be completed. First, the data series employed in the model have to be examined for their statistical properties; in particular, their stationarity. In the VAR model expressed in Eq. (11), the first differenced series for spot rates and their lagged values have to be shown to be stationary. In addition, the measure of the spread between time charter and spot voyage charter rates must also be stationary. Note that the latter is a linear combination of two series measured in “level” form (i.e. undifferenced), in contrast to the differenced spot rate data. Once the stationarity properties of the data have been established, the appropriate order of the VAR can be determined. That is, should the model be estimated using first differences only (a VAR of order 1), or should the second difference be included? This question can be addressed by the use of an information criterion, which balances the increased explanatory power generated by the addition of extra independent variables against a “cost” of greater complexity. Once the order of the model has been determined, estimation can then take place. Using the updated data set, the model developed by Veenstra was estimated for both the original data set (1980:10 to 1996:02) and the extended one (1980:10 to 2003:02). The series were examined for stationarity, using the Augmented DickeyFuller (ADF) test statistic. This has the formula given in Eq. (24) sp t = + t + ␥sp t−1 +
p−1
j sp −j + t
(24)
j=1
where the first difference of the spot rate, spt , is regressed against a constant (), a time trend (t), the one period lagged value of the level of spot rates (␥spt −1 ), and
A Survey of the Modelling of Dry Bulk and Tanker Markets
39
higher order lagged values of the first differences of the spot rate ( j sp −j ). The trend term is included if it is known or assumed that the relevant data series has a deterministic linear trend, whilst the past lagged values are included to ensure that the error term is white noise. The critical terms are the values of  and ␥. Under the null hypothesis, the original series should have no “unit root,” which implies that the value of ␥ is expected to be zero. Thus the Augmented Dickey-Fuller test is in this version, a joint test of the null hypothesis that the first difference of a series has no trend, nor any influence from lagged values of the level of the series (Greene, 2000, p. 784). In practice, ADF statistics are computed for the level and first difference of the relevant series. If the values are smaller than the reported critical values (because of the negative sign), the null of a unit root is accepted, and the series as tested is non-stationary. The distribution of the test statistic is not straightforward, as it depends upon the model assumed in the null hypothesis. The results presented in this chapter are derived using Eviews 4.0 (Eviews, 2002), which uses the Mackinnon (1991, 1996) critical values to determine whether or not the series is stationarity. Assuming that the necessary stationarity properties of the relevant series are satisfactory, the next stage of the VAR procedure is to determine the appropriate number of lagged variables that is optimal. The standard tests employed for this is the Schwartz Criterion (SC) or the Akaike Information Criterion (AIC). The former is the one selected here. The SC measures the value of the AIC measured for lag of order p, plus a term that increases in the lag order and reduces with the numbers of observations. The appropriate order is the one that minimises the value of the SC as its value is negative in most computer software programmes (Greene, 2003, p. 565). Finally, with the appropriate order of VAR determined, the model can be estimated. Tables 1 and 2 present the results of the ADF tests for stationarity of the series to be employed in the VAR model. Table 1 presents the results for the three dry cargo size segments, and Table 2 repeats the exercise for the Tanker segment (using dollar per ton of cargo data). The series are presented in ascending order of size, so that the first three rows in Table 1 represent the spot, time charter and spread series for the Handy (HNDY) size dry cargo segment. Recall that the “spread” is the difference between the time charter and spot rates. Similar results are reported for Panamax (PNMX) and Cape size (CAPE) size segments. The reported values from the test clearly indicate all dry cargo rate series are non-stationary in level form, but stationary at the 1% critical value in first difference form. The “spread” variables are stationary in level form, although the Panamax spread is only stationary if the 5% critical value is used. The results for tanker data in Table 2 indicate that the levels of the data on time charter equivalents (measured in current Worldscale) and spot rates (measured in
40
D. R. GLEN AND B. T. MARTIN
Table 1. ADF Test Results for Stationarity of LSE Data – Dry Cargo. Variable
Levels
1st Difference
Critical Values
Value
Stationary
Value
Stationary
1%
5%
HNDYSP HNDYTC SPRDHNDY
−3.52 −2.35 −5.74
No No Yes
−9.23 −7.74 n/a
Yes Yes
−4.01 −4.01 −4.01
−3.44 −3.44 −3.44
PNMXSP PNMXTC SPRDPNMX
−2.83 −2.83 −3.80
No No Yes
−11.19 −11.70 n/a
Yes Yes
−3.99 −4.00 −4.00
−3.43 −3.43 −3.43
CAPESP CAPETC SPRDCAPE
−2.33 −3.08 −6.06
No No Yes
−10.82 −10.82 n/a
Yes Yes
−3.99 −3.99 −3.99
−3.43 −3.43 −3.43
Source: Derived by the authors from LSE data. Sample: 1980:10 – 2003:02, except for Handy sized, which is 1980:10 – 1994:02. All variables in log form. Null is of trend and intercept. Mackinnon critical values from Eviews 4.0.
Table 2. ADF Test Results for Stationarity of LSE Data – Tankers. Variable
No Trend W Trend No Trend No Trend No Trend W Trend No Trend W Trend No Trend W Trend
Levels
1st Difference
Critical Values
Value
Stationary
Value
Stationary
1%
HNDYSP HNDYTC SPRDHNDY
−5.12 −4.53 −6.38
Yes Yes Yes
−13.02 n/a n/a
Yes
−4.00 −4.00 −4.00
−3.43 −3.43 −3.43
SUEZSP
−2.19
No
−13.80
Yes
−4.00
−3.43
SUEZTC SPRD130K
−3.05 −3.65
No Yes
−11.94 n/a
Yes
−3.46 −3.46
−2.87 −2.87
VLCCSP
−2.82 −4.51 −2.13 −4.25 −4.52 −4.61
No Yes No Yes Yes Yes
−12.96 −12.98 −12.98 −12.89 n/a n/a
Yes Yes Yes Yes
−3.46 −4.00 −4.00 −4.00 −3.46 −3.46
−2.87 −3.43 −3.43 −3.43 −2.87 −2.87
VLCCTC SPRDVLCC
5%
Source: Derived by Authors from LSE data. Sample: 1980:10 – 2003:02. All variables in log form. Null is of trend and intercept, except where indicated. Mackinnon critical values from Eviews 4.0.
A Survey of the Modelling of Dry Bulk and Tanker Markets
41
Table 3. Schwartz Criterion Values for VAR Model. Order
1 2 3 4 5
Dry Bulk
Tanker
120000
55000
30000
250000
130000
30000
−10.08 −9.98 −9.91 −9.82 −9.74
−10.33 −10.22 −10.13 −10.07 −9.96
−11.01 −10.97 −10.84 −10.74 −10.65
−8.64 −8.75 −8.72 −8.71 −8.32
−8.94 −8.77 −8.71 −8.62 −8.52
−9.04 −8.94 −8.90 −8.79 −8.73
Source: Derived by the authors from LSE data – 1980:10–1996:12 except 30,000 dwt (1980:10–1994:02).
current dollars per cargo ton delivered) are all non-stationary at the 1% level. It should be noted that some tanker series depend crucially on the maintained null hypothesis for the test to generate non-stationarity in the levels of the spot and time charter rate series. The spread series are clearly stationary in level form, thus allowing valid regressions of the spread and first differences of the spot rates, since they would all be stationary, and classical regression models are appropriate. Tables 3 and 4 present the results of testing for the appropriate order of VAR to be employed in the model, using the Schwartz criterion. These were calculated by estimating the models for spot and spread relations for each of the size classes, both dry cargo and tankers, for Veenstra’s original sample (Table 3), and for the extended data set (Table 4), whilst permitting the number of lagged dependent variables (order) to be up to five months. The values reported in Table 3 are similar to, and with the same structure as, Veenstra’s (Table 7.11, p. 200), implying an appropriate order of 1 or 2. Table 4 presents the results of repeating the analysis for the full sample. The results follow the same structure as in the original sample, except that the preferred order is always one in every case. This implies that there is a sound empirical basis for Table 4. Schwartz Criterion Values for Full Sample VAR Model. Order
1 2 3 4 5
Dry Bulk
Tanker
120000
55000
30000
250000
130000
30000
−9.70 −9.62 −9.56 −9.47 −9.38
−10.01 −9.96 −9.88 −9.84 −9.77
n/a n/a n/a n/a n/a
−8.16 −8.13 −8.05 −7.99 −7.91
−8.56 −8.51 −8.43 −8.37 −8.31
−8.70 −8.62 −8.57 −8.50 −8.45
Source: Derived by the authors from LSE data – 1980:10 – 2003:02.
42
D. R. GLEN AND B. T. MARTIN
using a first order VAR model of the spot rate and the spread relationship, since all series are stationary in their appropriate form, and the appropriate lag order is one. The preliminary analysis of the data suggests that a first order VAR is an appropriate way to model the data, with one of the two series being the “spread” variable, the other the first difference in the spot rate. All variables are measured in logarithms. Estimation of the Spread model can now be carried out. There are slight differences in the approach compared to Veenstra, however, to estimate a VAR with the present value restrictions in place, as discussed earlier. The technique of Seemingly Unrelated Regression (SUR) was employed. This permits both within and cross equation restrictions, and the use of Wald tests to examine the validity of those restrictions. Recall that, under the term structure hypothesis in a VAR model, the sum of the coefficients of the spread equation should equal the inverse of the discount rate, and the sum of the coefficients of the first difference of the spot rate should sum to zero. Accordingly, SUR estimation was employed on an unrestricted model which in essence is the VAR, relating the variables as follows. D(log(x)) = c(1) + c(2) × d(log(x(−1)) + c(3) × SpreadX(−1) + error(1) SpreadX = c(4) + c(5) × d(log(x(−1)) + c(6) × SpreadX(−1) + error(2) (25) The above system is a first order VAR with no restrictions on the parameters. The results are presented in Tables 5 and 6. They are similar, but not identical to, those reported for restricted VAR estimation in Tables 7.14 and 7.15 of Veenstra (1999, pp. 204, 205). A Wald test on the following joint restrictions:−c(1) + c(2) + c(3) = 0; c(3) + c(4) + c(5) = 1.003, can then be conducted. The value of 1.003 being imposed by Veenstra as an approximation to the assumed discount rate of 0.997 (Veenstra, 1999, p. 204). Table 5. Unrestricted VAR Model of Term Structure – Tankers. 30,000
Spot rate Spread Constant R-bar squared
130,000
250,000
Spot Rate
Spread
Spot Rate
Spread
Spot Rate
Spread
0.05 (0.06) 0.08 (0.03) −0.008 (0.01) 0.02
0.0006 (0.07) 0.85 (0.03) 0.015 (0.01) 0.73
−0.15 (0.06) 0.07 (0.03) 0.00 (0.01) 0.05
0.18 (0.07) 0.88 (0.03) 0.00 (0.01) 0.73
−0.02 (0.06) 0.16 (0.04) −0.01 (0.01) 0.07
0.08 (0.07) 0.81 (0.04) 0.02 (0.01) 0.63
Source: Computed by authors, Sample: 1980:10 2003:02. All variables in logs.The Spot Rate is measured as the first difference of the log value, as required by the model.
A Survey of the Modelling of Dry Bulk and Tanker Markets
43
Table 6. Unrestricted VAR Model of Term Structure – Dry Cargo. 30,000
Spot rate Spread Constant R-bar squared
55,000
120,000
Spot Rate
Spread
Spot Rate
Spread
Spot Rate
Spread
0.12 (0.05) 0.11 (0.05) −0.02 (0.01) 0.04
0.05 (0.08) 0.83 (0.05) 0.02 (0.01) 0.71
−0.004 (0.06) 0.03 (0.02) −0.001 (0.006) 0.01
0.12 (0.06) 0.93 (0.02) 0.001 (0.006) 0.85
0.06 (0.06) 0.13 (0.05) −0.02 (0.01) 0.03
−0.05 (0.06) 0.71 (0.05) 0.05 (0.01) 0.52
Source: Computed by authors, Sample: 1980:10 2003:02, except 30,000 dwt. which is 1980:10–1994:02. All variables in logs. The Spot Rate is measured as the first difference of the log values, as required by the model.
Inspection of the coefficient values in Table 5 shows that the coefficients do not exactly comply with the restrictions needed for the model to be consistent with the spread relation and present value restrictions. For example, the sum of the coefficients on the differenced spot rate equation for 30,000 dwt tankers is 0.122 when it should be equal to zero, and the sum of the coefficients in the spread relation is 0.8656, not 1.003 as suggested by Veenstra. For the 30,000 dwt dry cargo size in Table 6, the corresponding sums are 0.21 and 0.90 respectively. This approach is not valid however, because a joint test on both restrictions is required. A Wald test was computed to achieve this. Essentially, the model is re-estimated subject to the two joint restrictions on the coefficients laid out above, and the difference in the Log-likelihood values for the unrestricted and restricted models are computed. If this difference is significant at conventional critical values, the restrictions are rejected (Greene, 2003, p. 96). The results of the Wald Tests for the restrictions on the full sample sets of data are provided in Table 7. The data series for the dry cargo Handy size could not be extended beyond Veenstra’s original sample, so there are no results to report for that segment. It is interesting to note that the restrictions are accepted for the 55,000 Table 7. Wald Tests of the Joint Restriction on the VAR System. Variable
Dry Cargo 2
Handy Panamax CapeSize
n/a 1.49 59.3
Tanker Worldscale
df
Prob
Size
2
2 2
0.47 0.00
30 130 250
3.52 0.839 1.989
Source: Computed by authors, Sample: 1980:10 – 2003:02.
df 2 2 2
Tanker $/ton
Prob
2
df
Prob
0.17 0.66 0.37
2.86 22.74 369.60
2 2 2
0.24 0.00 0
44
D. R. GLEN AND B. T. MARTIN
Table 8. Correlations and Variance Ratios from VAR Models. Variable
Handy Panamax CapeSize
Dry Cargo
Tanker
Variance Ratio
Correlation
Size
Variance Ratio
Correlation
n/a 1.18 1.87
n/a 0.92 0.71
30 130 250
1.31 1.36 1.52
0.86 0.86 0.80
Source: Computed by authors, Sample: 1980:10 – 2003:02.
dwt (Panamax) dry cargo, and for all the tanker sizes, when measured in current Worldscale. However, when the models are estimated using current dollars per ton of cargo, also published in Lloyds Shipping Economist, the results are much worse, with only the smallest tanker size model accepting the restrictions (the results are not reported here, but are available from the authors, on request). Veenstra also reported two “informal tests” of the spread model, which are the correlation between the variance of the spread and that forecast from the spread model, and the ratio of the forecast to the actual variance of the spread. In ideal conditions, both of these values should be unity. Table 8 presents the results of the tests on the extended sample. The Capesize dry cargo variation is quite a long way from unity, at 1.87. Indeed, all the ratios exceed unity, implying that the observed variability of the spread values within the sample period are consistently larger than that generated by the predictions of the model, and the correlations are high, but not close to unity. The above results provide confirmatory evidence that the VAR approach can generate a reasonable model of the dynamics of the spread between time charter and spot rate behaviour over time, with no exogenous variables being considered, although the informal test results for the extended sample are weaker than those reported in Veenstra (1999, p. 204). The models can be used to generate the one period forward static forecasts of the spread. The results for the three tanker sizes are shown in Fig. 1. It would appear that despite the rejection of the parameter restrictions imposed on the VAR in certain size classes of ship, the use of the spread relationship between time charter and spot is a significant factor in explaining the behaviour of the spot rate as well as the behaviour of the spread. It has been demonstrated that the spread is also a stationary series in levels form. Can this relationship be used to improve the modelling of market rates in a “reduced form” context? 4.2. Modelling Spot and Time Charter Rates Given that the main thrust of this chapter is the modelling of the rates for dry cargo and tanker segments, can the use of the spread relationship employed in the VAR
A Survey of the Modelling of Dry Bulk and Tanker Markets
45
Fig. 1. Estimated Spreads from the Veenstra VAR Model (1980:10 – 2003:02).
approach above be incorporated into a model of the spot rate itself? One way of incorporating it is to view the pairwise relationship of time charter and spot rates as forming a cointegrating vector, which makes sense from the term structure viewpoint. Such a relation can be used to model the first differences of the spot rate: – as long as the model consists of elements which have the same order of integration, some factors may affect the short term dynamic of the spot rate without being in the cointegrating relation. When a pair of variables are cointegrated, a linear combination of the two series may be stationary (implying a constant long term relationship between them), even if both series are non-stationary on their own. Thus, if two series are I(1), that is, they need to be differenced once to generate a stationary time series, then if a linear combination of the two series exists which is itself stationary, the pair of variables are said to be cointegrated (Greene, 2003, p. 650). Furthermore, if two time series are non-stationary, but have a cointegrating vector, then by the Granger representation theorem (Greene, 2003, p. 654), there
46
D. R. GLEN AND B. T. MARTIN
also exists an “error correction mechanism” (ECM) that adjusts the current rate of change of the dependent variable (in our case, the spot rates) to differences between the lagged values of the difference between the levels of the two variables (in our case, this is the “spread,” the difference between the time charter and voyage spot rates). This approach would be consonant with the use of an “error correction mechanism” (ECM) type model, with the ECM being created by the time charter and spot spread relationship. This approach models the first difference in the voyage spot rates as a function of variables that affect the rate of change of the variable in the short term, but which do not necessarily alter the long run values of the level. Indeed in the “stationary state” of such models, all first differenced values are set to zero (implying zero growth) which leaves just the terms in the ‘error correction mechanism’, which generate the implied long run relationship between the set of variables contained therein. Alternatively, some variables may cointegrate with the spread variables, implying that some long term relationship is possible, even if the level of integration of each variable on its own is different. What variables should be considered? Kavussanos (1996a, b) modelled the monthly spot rate against a number of factors, derived from the reduced form version of the Beenstock and Vergottis’ model of competitive bulk markets. These variables included a proxy for demand growth (industrial production), and bunker prices (as a proxy for costs). The proposed “reduced form” model of spot rate behaviour is given below. +/−
− + sp t = f(S + t−1 , Z t−i , K t−i , BP t−i )
(26)
. . . where spt is the first difference in the spot rate, S, the one period lagged value of the difference between the time charter and spot rate measured in real $ per ton of cargo equivalents, Z a vector of exogenous demand variables, K the existing relevant fleet, and BP the bunker price. This simple model is the basis of attempting to forecast the short run behaviour of the relevant spot rate on a monthly basis, whilst keeping the link between long run spot and time charter rates. It could be argued that the time charter rate is a proxy for the long run marginal cost of providing a tanker service, from which the short run rate can depart in the short term, but not in the long term. This somewhat unusual model forms the basis of the results reported below. 4.3. Data, Stationarity and Cointegration Tests Data for the fleet variables was collected from two sources: dry cargo and one set of tanker data was obtained from Lloyd’s Shipping Economist, for three different size
A Survey of the Modelling of Dry Bulk and Tanker Markets
47
Fig. 2. Alternative Modelling of Tanker Stock Data. Interpolation Equation TOTK = 314.14878 + 0.91404899 × (@TREND(1980 : 01)) − 0.088993354 × (@TREND(1980 : 01)2 ) + 0.0012728098 × (@TREND(1980 : 01)3 ) − 7.2014086e − 06 × (@TREND(1980 : 01)4 ) + 1.5828519e − 08 × (@TREND(1980 : 01)5 ) − 3.1555964e − 14 × (@TREND(1980 : 01)7 ) R-bar squared = 0.96.
classes. The Lloyd’s data for tankers alters its size categories for the two smaller classes, at January 1995. The tanker fleet data used in this study was constructed as follows. E A Gibson provided tanker deadweight tonnage figures on an annual basis from 1980 to 1995. From 1995 on quarterly figures were available. The data were for the same five size ranges from 1980 until 2002. The missing observations were interpolated in two different ways. First, a higher order polynomial in time was used to fit a relation to the existing data over the period. The resulting model was used to generate the missing monthly observations. The R-square for this model was 96.1%. This approach was rejected in favour of a second. Monthly data for the entire tanker fleet was available from Lloyd’s Shipping Economist. Simple interpolation between the actual Gibson data generated the missing observations, which were then used to construct market shares of the five sizes available over the time period. These market shares were then applied to the Lloyds’ Shipping Economist data, thus generating monthly estimates for each sector without the need for interpolation. This approach has the merit of generating data based on actual monthly figures. Visual inspection of Fig. 2a–c suggests that the interpolation method keeps some of the “erratic” nature of the data, as well as preserving the true figures that are available. The generated data appears very smooth, as one would expect. However, it does assume that the Gibson and Lloyds data is consistent. The Gibson estimate of the tanker fleet is considerably smaller than that generated by Lloyds at the beginning of the sample, but closer at the end (see Fig. 3). The overall pattern of
48
D. R. GLEN AND B. T. MARTIN
Fig. 3. Comparison of Lloyds Shipping Economist and Gibson’s Tanker Fleet Data.
the data is the same however, as can be seen in the relevant figures. The ADF test results presented in Table 9 are generated using the Gibson data. As noted by Veenstra, LSE stopped publishing data on the smallest size classes in 1994, so the modelling of this size class is for the period 1980:10 to 1994:02. All monetary series were measured in dollars per ton of cargo, both for tankers and dry cargo. This permitted the deflation of the monthly series by the application of the U.S. monthly producer price index, thus converting spot, time charter, oil prices and bunker prices to real values. In principle this could be done for tanker data when expressed in Worldscale, providing the Worldscale Flat values are known. The application of a common deflator to all the series that were measured in nominal dollar values appeared to be a more convenient way forward.3 Tables 9 and 10 presents the results of ADF tests on a number of variables that might be expected to form a long run cointegrating relationship with bulk freight rates. The variables chosen were: the real values of the spot and time charter rates for 25,000 and 30,000 dwt vessels (Handy size) together with LSE fleet data for 10 < 40,000 dwt vessels. The sample period is limited to February 1994 for this data. The Panamax size (40 < 80,000 dwt) fleet information was used with real spot and time charter rates for 55,000 dwt vessels, and rates for 120,000 dwt vessels
Dry Cargo
Levels 1979:03 – 1995:08
Fleet variables Fleet 25K 50K
10 <40 40 < 80 80 +
−1.75 No −1.25 No 1979:03 – 2002:12 −0.64 No −1.76 No 1979:10 – 2003:9 −3.39 No −1.98 No −2.77 No
Tankers 10 < 50 50 < 100
First Differences Lags
1995:08 – 2002:12
1 1
−26.86
1979:03 – 1995:08
−6.86 −7.51
Yes Yes
1 1
−11.49 −8.72
3 2 6
1980:02 – 2002:10 −2.14 No −2.28 No
1 1
100 < 200 200 < 320
−1.77 −1.89
No No
1 1
320+
−1.92
No
1
TotalTkr
−1.38
No
1
Capesize TotalFleet
1995:08 – 2002:12
−4.89
Critical Values
−4.01
−3.43
Yes Yes
−3.99 −3.99
−3.43 −3.43
−6.05 −6.45 −3.94
Yes Yes Yes
−3.99 −3.99 −3.46
−3.43 −3.43 −2.87
−4.03 −3.50 −11.55 −7.26 −3.87 −13.11 −1.72 −11.69 −2.94 −10.19
Yes Yes at 5% Yes Yes Yes at 5%
−3.99 −4.00 −4.00 −4.00
−3.43 −3.43 −3.43 −3.43
−3.46
−2.87
−3.99
−3.43
2nd diff
A Survey of the Modelling of Dry Bulk and Tanker Markets
Table 9. ADF Test Results on Fleet and Demand Variables.
2nd diff No Yes No 2nd diff
49
50
Table 9. (Continued ) Dry Cargo
Levels 1979:03 – 1995:08
Exogenous variables Demand Industrial production Tanker demand
−2.89 −2.81
Source: Computed by authors.
No No
Lags
1995:08 – 2002:12
1979:03 – 1995:08
1995:08 – 2002:12
Critical Values
1
−10.33
Yes
−4.00
−3.43
1
−9.20
Yes
−4.01
−3.44
1
−8.65 −10.61
Yes Yes
−4.02 −4.01
−3.44 −3.44
1
−12.53 −12.76
Yes Yes
−3.46 −3.99
−2.87 −3.43
D. R. GLEN AND B. T. MARTIN
Nominal oil price Real oil price Costs No trend Bunker prices Real bunker prices
1981:09 2003:06 −1.65 No 1984:10 2000:03 −1.85 No 1990:03 2003:07 −3.13 No −3.56 No 1980:02 2003:11
First Differences
A Survey of the Modelling of Dry Bulk and Tanker Markets
51
Table 10. ADF Test Results for Real Dry Cargo Price Series. Variable
Level
First Difference
RLSPHNDY RLTCHNDY RSPRDHNDY
−3.68 −2.74 −3.67
Borderline No Borderline
−9.30 −7.94 −10.15
Yes
RLSPPX RLTCPX RSPRDPX
−3.02 −3.08 −3.80
No No Yes
−11.44 −12.10
Yes Yes
RLSPCS RLTCCS RSPRDCS
−2.46 −3.34 −6.06
No No Yes
−11.05 −11.16
Yes Yes
Source: Computed by authors.
were matched to data for the fleet of vessels of 80,000 dwt and above. The values for the test for the various series in Table 9 indicate that they are first difference stationary in all cases, although some series reject the null of stationarity in level form with values that are borderline (e.g. 200 < 320,000 dwt tanker fleet, which is only stationary in first difference form at the 5% level of significance, since the critical value at 1% is −4.00 and the ADF test score is −3.87). The results in Table 10 are also broadly in accordance with first difference stationarity for the “real” dry cargo price data series, but again, some series may possibly be stationary in level form.4 Having considered the empirical evidence on the stationarity properties of the data series that is to be employed in the dry cargo models, the issue of their cointegration is now considered. This is examined by the use of the Johansen procedure, which utilises a VAR methodology to develop appropriate tests for the identification of the set of cointegrating vectors that may exist between the variables (Greene, 2003, p. 656). The existence of a cointegrating vector is interpreted as statistical evidence for the presence of a long term stable relation between the variables identified. When only two variables are considered, only one such cointegrating vector is possible. When there are several, the maximum number of such vectors is one less than the number of variables. From an economic perspective, it is ideal that a unique vector exists, which can be interpreted as the long run equation defining the relation between the variables. When there are more than one, a straightforward economic interpretation of the statistical relations becomes problematic. Table 11 presents the results of testing for the number of cointegrating vectors for the three dry cargo sizes as well as the relation between the real time charter and spot rates for Capesize vessels on their own. Inspection of the table reveals
52
D. R. GLEN AND B. T. MARTIN
Table 11. Cointegration Test Results for Dry Bulk Models – 1981:07 2003:02.
RLTCHNDY RLSPHNDY LINDPRN LHNDYE LNBKRP LDRY40E LINDPRN RLBKRP RLSPPX RLTCPX LDRY110 RLSPCS RLBKRP RLTCCS LINDPRN
Series: RLSPCS RLTCCS
Hypothesized No. of CE(s)
Eigenvalue
Trace Statistic
5% Critical Value
1% Critical Value
No. of CI Vectors
None**
0.258
89.807
68.52
76.07
At most 1
0.148
45.066
47.21
54.46
1
None**
0.150
83.512
68.52
76.07
1
At most 1
0.092
41.439
47.21
54.46
None**
0.243
136.129
68.52
76.07
At most 1** At most 2* At most 3* At most 4*
0.111 0.069 0.048 0.016
64.987 34.836 16.614 4.071
47.21 29.68 15.41 3.76
54.46 35.65 20.04 6.65
4
None**
0.144
40.293
12.53
16.31
1
At most 1
0.000
0.004
3.84
6.51
Source: Computed by authors. ∗ Result is significant at the 5% critical value. ∗∗ Result is significant at the 1% critical value.
that for the first two vessel sizes, there is only one cointegrating vector under the null hypothesis of no constant in the cointegrating equation, and a linear trend in the data. The trace statistic of 89.807 exceeds the 1% critical value, so the null hypothesis of no cointegrating vector is rejected. The next row shows the sequential test which now assumes that there is at most one cointegrating vector between the time charter (RLTCHNDY) and spot rates (RLSPHHNDY) for Handy size dry cargo vessels, as well as the level of industrial production (LINDPRN), the Handy size fleet (LHNDYE), and bunker prices (LNBKRP). The trace test statistic is now 45.066, which is less than the 5% critical value of 47.21, so the null hypothesis of no more than one cointegrating vector is accepted. However, there appears to be at
A Survey of the Modelling of Dry Bulk and Tanker Markets
53
Table 12. ADF Tests for Real Prices, Tankers. Level
Stationary
RLSP30 RLTC30 RSPRD30
−0.36 0.16 −4.28
No No Yes
−13.67 −13.73
Yes Yes
RLSP130 RLTC130 RSPRD130
−0.10 0.39 −3.57
No No Yes
−13.69 −11.95
Yes Yes
RLSP250 RLTC250 RSPRD250
−0.49 0.18 −4.84
No No Yes
−13.70 −12.78
Yes Yes
No trend, no intercept Sig at 1% level
First Difference
−2.57
−1.94
Source: Computed by authors, Sample 1980:12 – 2003:2.
least 4 cointegrating vectors identified for the Capesize sector, and under certain assumptions of the null, 5! If five actually exist, the series are not I(1), and can be modelled in level form, but this contradicts the evidence from the ADF tests on individual variable stationarity. This issue can be partly resolved by noting that there is one unique cointegrating vector between the spot and time charter rate series for this data, and thus the short run model my be unstable but the long run relationship is still valid. It is concluded that the cointegration tests permit the estimation of the proposed model set out in Eq. (25). A similar set of results for the tanker segments are now presented. The tanker fleet data has already been described in Table 9; stationarity properties of the series for real spot and time charter rates for the three size classes, together with the results of the ADF test for real oil prices and real bunker prices, are shown in Table 12. Visual inspection of the time series suggested that there was no trend in the data, and the ADF test null hypothesis was adjusted accordingly. It should be noted that in its most general form (i.e. with a trend and with an intercept), the ADF test results imply that all the real spot and time charter series are stationary in level form, so there is room for debate here. The results of the cointegration tests for the tanker segments are presented in Table 13. As with the Dry cargo sector, two of the sectors had unique cointegration vectors in at least one combination, but the third (the largest size) did not. The maximum number is two, at the 1% level of significance. Whilst these results are not overwhelmingly supportive of a fundamental long term relationship between the hypothesised variables of the model, it is sufficient to proceed with estimation. The models were estimated in first difference form,
54
D. R. GLEN AND B. T. MARTIN
Table 13. Cointegration Tests for Tanker Variables. Hypothesized No. of CE(s)
Eigenvalue
Trace Statistic
No. of CI Vectors
None**
0.22
149.81
At most 1** At most 2*
0.20 0.15
99.99 56.06
None**
0.19
118.16
At most 1
0.14
62.46
None**
0.22
107.74
At most 1 At most 2*
0.16 0.09
58.76 24.41
2
RSPRD30 INDPRN LHNDYE RLBKRP
None** At most 1
0.17 0.07
69.99 23.09
1
RLSP130 RLTC130 LSUEZE INDPRN RLBKRP RLOILP
None**
0.20
At most 1
0.14
59.95
None**
0.12
73.77
At most 1
0.09
41.19
None**
0.20
RLTC30 RLSP30 INDPRN LHNDYE RLBKRP RLOILP
RLTC30 RLSP30 INDPRN LHNDYE RLBKRP RSPRD30 INDPRN LHNDYE RLBKRP RLOILP
RLSP130 RLTC130 LSUEZE INDPRN RLBKRP RSPRD130 LSUEZE INDPRN RLBKRP RLOILP RSPRD130 LSUEZE INDPRN RLBKRP RLSP250 RLTC250 LVLCCE INDPRN RLBKRP RLOILP
RLSP250 RLTC250 LVLCCE INDPRN RLBKRP
RSPRD250 LVLCCE INDPRN RLBKRP RLOILP
2
2
103.5 1
1
88.99
At most 1
0.13
45.10
1
None** At most 1
0.09 0.06
52.00 28.83
1
None**
0.25
154.51
At most 1** At most 2
.0.20 0.13
None**
0.15
104.48
At most 1** At most 2
0.10 0.07
62.52 35.83
None**
0.24
118.66
At most 1** At most 2
0.16 0.12
64.68 31.59
98.45 54.27
2
2
2
A Survey of the Modelling of Dry Bulk and Tanker Markets
55
Table 13. (Continued )
RSPRD250 LVLCCE INDPRN RLBKRP
Hypothesized No. of CE(s)
Eigenvalue
None** At most 1** At most 2
0.15 0.09 0.06
Trace Statistic
82.16 42.23 19.34
No. of CI Vectors
2
Source: Computed by authors, Criteria set at 1% significance, Sample period – With Oil Price 1987:05 – 2003:02, 194 obs., Sample period – Without Oil price 1981:08 – 2003:02 254 obs. ∗ Result is significant at the 5% critical value. ∗∗ Result is significant at the 1% critical value.
except for the spread variable, which has been established to be stationary. This procedure appears to be valid on the basis of the results established in the previous tables.
4.4. Estimation Tables 14 and 15 present the Ordinary Least Squares (OLS) estimation results of the model specified above. The results in Table 14 are broadly consistent with the ideas expressed in the formal model, although there are a few issues. The “spread” relation is statistically significant in all three of the size classes for the dry cargo sector, although it is “borderline” for the smallest size class. The short run effects of industrial production growth on the growth of the spot rate for the smallest size is also a little odd, with the overall effect appearing to be negative. The other coefficients have the expected signs, although some have quite large standard errors. The four statistical tests reported at the foot of the table give reasonable results. The Q and Q-squared statistics, and the Breusch-Godfrey (Lagrange Multiplier) test, are different means for assessing the degree of autocorrelation remaining in the residuals; with one case excepted, the models appear to be satisfactory. The standard errors are White heteroscedasticity consistent estimates. The Dry cargo models all emphatically fail the Jarque-Bera test for the normality of the distribution of the residuals. This is consistent with the results of many earlier studies on monthly data for these markets (e.g. Kavussanos, 1996a, 1997). The tanker results (Table 15) again are broadly consistent with the a priori expectations, except for the effect of the growth in the real oil price on spot rate growth in the 250,000 dwt tanker market, which appears to be negative, in contrast to that found for the 130,000 dwt vessels. No link at all could be found for the
56
D. R. GLEN AND B. T. MARTIN
Table 14. Estimation Results for Dry Cargo. Variable C RSPRD(−1) D(LINDPRN(−1)) D(LINDPRN(−4)) D(LDRYFLT(−1)) D(LDRY110(−9))
Handy Coeff 0.01 (0.01) 0.12 (0.05) 1.77 (0.88) −2.23 (1.01) −4.71 (1.89) –
D(RLBKRP(−1))
–
D(RLBKRP(−3))
–
@SEAS(4) @SEAS(6) @SEAS(7) R2 Adjusted R2 S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat Q (12 lags) Probability Q2 (12 lags) Probability Reset test (2 fitted) Probability Breusch-Godfrey (2 lags) Probability Jaques-Bera Probability Sample
Panamax Coeff 0.02 (0.01) 0.04 (0.03) – – −1.45 (1.05) – 0.09 (0.05) –
Capesize Coeff −0.01 (0.01) 0.11 (0.05) 2.36 (0.82) – – −1.19 (0.40) – 0.09 (0.06) –
−0.08 (0.02) −0.07 (0.03) −0.06 (0.03)
−0.12 (0.02) −0.06 (0.02)
−0.08 (0.02) −0.06 (0.02)
0.25 0.21 0.08 0.79 175.61 1.98
0.17 0.16 0.09 1.87 272.55 2.22
0.14 0.11 0.10 2.54 230.01 1.95
12.81 0.31 21.58 0.04 1.32 0.27 0.01 0.99 23.19 0.00
4.59 0.10 5.35 0.95 0.02 0.98 2.11 0.12 63.35 0.00
5.55 0.94 6.90 0.86 1.12 0.33 0.24 0.79 12.07 0.00
1981:07 2003:02
1981:09 2003:02
1981:12 1994:02
–
Source: Computed by authors. Dependent Variable: First Difference of real spot rate, Bracketed terms are White Heteroscedestacity Consistent Standard Errors.
A Survey of the Modelling of Dry Bulk and Tanker Markets
57
Table 15. Estimation Results for Tankers. Variable C RSPRD(−1) D(LINDPRN(−3))
30,000 dwt 0.03 (0.02) 0.16 (0.05) 2.03 (1.63)
130,000 dwt 0.01 (0.01) 0.10 (0.04) –
D(LINDPRN(−4)) D(LINDPRN(−5)) D(LHNDYE(−3)) D(LSUEZE(−3)) D(RLBKRP(−1)) D(RLBKRP(−5))
– −9.29 (6.84) – 0.19 (0.11) –
D(RLOILP(−2))
–
D(RLOILP(−4))
–
D(RLSP130(−1))
–
SEAS(3) SEAS(4) SEAS(5) SEAS(8)
−0.08 (0.04) −0.07 (0.04) −0.09 (0.03) −0.08 (0.04)
2.10 (1.02) – −2.81 (1.69) 0.06 (0.07) – – 0.19 (0.10) −0.11 (0.09)
R2 Adjusted R2 S.E. of regression Sum squared resid Log likelihood Durbin-Watson stat
0.33 (0.02) 0.18 0.14 0.14 3.34 117.48 2.15
−0.01 (0.01) 0.19 (0.05) – 3.51 (1.39) – – −1.19 (0.40) – 0.20 (0.09) −0.27 (0.14) – –
–
–
–
–
–
–
−0.06 (0.03)
SEAS(9) Dummy variable
250,000 dwt
0.40 (0.04) 0.17 0.16 0.09 1.87 272.55 2.22
– −0.12 (0.04) 0.20 (0.03) 0.19 0.16 0.16 4.85 86.19 1.94
58
D. R. GLEN AND B. T. MARTIN
Table 15. (Continued ) Variable Q (12 lags) Probability Q2 (12 lags) Probability Reset test (2 fitted) Probability Breusch-Godfrey (2 lags) Probability Jaques-Bera Probability Sample
30,000 dwt
130,000 dwt
250,000 dwt
8.4 0.75 17.11 0.15 2.62 0.76 1.84 0.62 19.41 0.00
11.09 0.53 12.16 0.43 3.93 0.14 3.33 0.96 0.39 0.00
4.38 0.98 25.48 0.13 0.91 0.41 0.49 0.61 4.35 0.11
1987:02 2003:02
1987:10 2003:01
1986:04 2003:02
Source: Computed by authors. Dependent Variable: First Difference of real spot rate, Bracketed terms are White Heteroscedestacity Consistent Standard Errors.
smallest size, which is to be expected since the data is for clean trades, and there is less likelihood of a link between crude oil price growth and rates for cargoes of clean products.5 The dummy variable was incorporated to account for the effects of the IranIraq wars (1979:10–1980:12), the first Kuwait war (1990:08–1991:04), and the recent USA and U.K. invasion of Iraq (October 2002 on). It is interesting to note the different seasonal patterns between the smallest tanker size, which appears to have a 7% lower growth rate in the months of March to May, compared to that of the 250,000 dwt tanker, whose seasonal demand was only significantly lower in the month of September over the sample period. It is also worth noting that no “Erika” or “Prestige” effect could be detected in the data that was used in this model. One would need data on ships by age as well as over time to examine the possibility that the market reaction to these events has triggered a premium’ for new tankers relative to old. The data set used here does not permit such an analysis. As in the case of the Dry Cargo results, the “diagnostic tests” were reasonably satisfactory, except for the Jarque-Bera test results, which again indicate nonnormality in the distribution of the residuals. In order to try to reduce this problem, and to deal with potential time varying heteroscedasticity, the above models were also estimated using GARCH. This had the effect of eliminating non-normality in only one of the three tanker size classes, and none of the estimated dry cargo ones, results which appear to differ from those reported in Kavussanos (1996a, b). The use of GARCH had little material effect on the equation for the mean, which is the focus of this study, so are not reported here.
A Survey of the Modelling of Dry Bulk and Tanker Markets
59
4.5. Within Sample Forecast Performance Given that the models estimated above are an attempt to use the “spread” idea to explain the behaviour of the spot rates, an examination of their performance is required. Accordingly, a “na¨ıve” model was constructed for each segment, and the in sample static one period forecast performances of the “spread” and the na¨ıve models were compared. The na¨ıve model was specified as X(t) = c + cX(t − 1) + dummies + error
(27)
In other words a simple model using the one period lagged value of the relevant dependent variable as the only explanatory variable. The results of this exercise are presented in Table 16. Four criteria are used – Root mean squared error (RMSE), Mean Absolute Error (MAE), mean percentage error (MA%E), and the Theil Inequality coefficient. In almost every case the estimated model performed better than the na¨ıve model, and there were no cases when the na¨ıve model outperformed the estimated model across all four criteria. It should be noted that the degree of improvement in many cases is very slight, although the tanker sectors appeared to generate the largest improvements in performance.
Table 16. Forecast Performance of OLS models and Na¨ıve Model. Dry Cargo
Tankers
Model
Na¨ıve
Model
Handy RMSE MAE MA%E Theil Ineq
0.073 0.054 3.193 0.022
0.081 0.057 3.513 0.024
0.132 0.096 3.581 0.024
Panamax RMSE MAE MA%E Theil Ineq
0.085 0.064 3.629 0.023
0.086 0.064 3.586 0.023
0.105 0.079 3.086 0.020
Capesize RMSE MAE MA%E Theil Ineq
0.099 0.077 2.552 0.016
0.105 0.079 2.615 0.017
0.157 0.122 4.655 0.029
Na¨ıve 30,000 0.137 0.101 3.784 0.025 130,000 0.130 0.092 3.631 0.025 250,000
Source: Computed by authors.
0.171 0.125 4.735 0.032
60
D. R. GLEN AND B. T. MARTIN
4.6. Concluding Comments to Section 4 This section has presented the results of two related research efforts. First, the spread model, based on a VAR representation of spot rates and developed by Veenstra (1999) was estimated with updated information. The new data allowed the sample to be extended to February 2003, adding nearly 100 observations to the data set. It was shown that the VAR model appeared to work extremely well with the new data set, generating very good one period static forecasts over the sample period. The results for the tanker sector were presented graphically in Fig. 2a–c. Second, a pragmatic model, employing the “spread relationship” between time charter and spot rates was then developed, with the aim of improving the modelling of the spot rates for dry cargo and tanker ship sizes. Data for time charter rate equivalents in $ per ton of cargo were used, thus linking the model to the data employed in the VAR study. However, one important difference is the attempt to use real dollars per ton, rather than nominal dollars or current Worldscale. It was shown that a number of variables could be added to improve the “fit,” which incorporated the spread component that was in effect the estimating equation for the first difference of the spot rate inside the original VAR model. The forecasting performance of the estimated models were compared to that of a na¨ıve model, and shown to be marginally better than the “na¨ıve” alternative in most cases. There are a number of issues that are worth noting for future research. First, the ADF testing of the data on spot and time charter rates does not appear to be totally convincing. A statistical case can be made out for both non-stationarity and for stationarity, depending upon the maintained null hypothesis. This is not very helpful when determining whether or not to model in levels or in first differences. Secondly, the model presented in this study can be criticised for not being directly derived from a structural form. Perhaps the time is right to review the heavy reliance on stationarity tests and permit economic theory a greater role in the modelling process. And thirdly, it would be interesting to be able to take a model of spot rate behaviour and see how it fares when used to model annual, quarterly, and monthly data. Is there a model that makes sense of all three frequencies? There is an interesting research challenge.
NOTES 1. This section draws heavily on Chapter 2 of Beenstock and Vergottis (1993b). 2. This terminology is non-standard, but follows the presentation of data in Lloyd’s Shipping Economist. The industry convention is to refer to time charter rates as a hire rate expressed in dollars per day. This is because the voyages of time chartered vessels are determined by the charterer, not the owner.
A Survey of the Modelling of Dry Bulk and Tanker Markets
61
3. An anonymous referee made the point that the use of either the U.S. consumer or producer price indices were both subject to the criticism that global inflation in the shipping industry is not well captured by the use of a one country index. They suggested that an index of operating costs might be more appropriate as a deflator. The real question is what is the deflator for? If it is to allow for the changing purchasing power of the dollar, which is widely regarded as the numeraire of shipping, then the use of a dollar deflator is justifiable. The use of a cost deflator would also run into the objection that cost structures depend on key assumptions about the company’s operations, and might be just as problematic to construct. Ideally perhaps, a Purchasing Power Parity index might be more appropriate. 4. This is an appropriate point to note that the acceptance of the null of non-stationarity using the ADF approach is not as powerful a test as it might appear, as these tests cannot separate “nearly integrated” time series from truly non-stationary ones very efficiently. Bear this in mind when considering the results. We are grateful to the referee for this point. 5. The authors are grateful to the comments of an anonymous referee, who pointed out that it is possible that the a priori relationship between the crude oil price and spot tanker rates could be ambiguous. This is because there are two possible mechanisms. First, a rise in the oil price is caused by a rise in oil demand, which also generates an increase in the demand for oil transportation. This generates a positive association. Second, a rise in the oil price might be caused by a reduction in the supply of oil, which implies a fall in the demand for oil transportation services and an expected fall in the spot price. Thus, both a positive and a negative relation could be justified. Which is the stronger of the two is determined only by experience.
ACKNOWLEDGMENTS Special thanks Ms. Z. Woodhouse (of E. A. Gibson Ltd.) for providing a large amount of data on the Tanker Markets. Dr. P. Rogers of Galbraiths Limited provided useful data on the dry bulk markets. We would also like to thank Dr. Glen’s colleagues at London Metropolitan University, particularly Dr.’s M. Rowlinson and R. Mirmiran, for providing one of the authors with the resources to complete this work. Last but emphatically not least, we would like to acknowledge the forbearance of the editor, Professor Kevin Cullinane, of Newcastle University, who has waited patiently for this contribution to his book.
REFERENCES Beenstock, M. (1985). A theory of ship prices. Maritime Policy and Management, 12, 31–40. Beenstock, M., & Vergottis, A. (1989a). An econometric model of the world market for dry cargo freight and shipping. Applied Economics, 21, 339–356. Beenstock, M., & Vergottis, A. (1989b). An econometric model of the world tanker market. Journal of Transport Economics and Policy, 23, 263–280.
62
D. R. GLEN AND B. T. MARTIN
Beenstock, M., & Vergottis, A. (1993a). The interdependence between the dry cargo and tanker markets. Transportation and Logistics Review, 29, 3–38. Beenstock, M., & Vergottis, A. (1993b). Econometric modelling of world shipping. London: Chapman & Hall. Berg-Andreassen, J. A. (1997). The relationship between period and spot rates in international maritime markets. Maritime Policy and Management, 24, 335–350. Bollerslev, T. (1986). Generalised autoregressive conditional heteroskedasticity. Journal of Econometrics, 31, 307–327. Campbell, J. Y., & Shiller, R. J. (1987). Cointegration and test of present value models. Journal of Political Economy, 95, 1062–1088. Campbell, J. Y., & Shiller, R. J. (1988). Stock prices, earnings and expected dividends. Journal of Finance, 43(3), 661–676. Charemza, W., & Gronicki, M. (1981). An econometric model of world shipping and shipbuilding. Maritime Policy and Management, 8, 21–30. Engle, R. F. (1982). Autoregressive conditional heteroskedasticity with estimates of the variance of inflation of UK inflation. Econometrica, 50, 987–1007. Engle, R. F., & Granger, C. W. (1987). Cointegration and error-correction: Representation, estimation, and testing. Econometrica, 55, 251–276. Glen, D. (1990). The emergence of differentiation in the oil tanker market 1970–1978. Maritime Policy and Management, 17, 289–312. Glen, D., & Martin, B. (1998). Conditional modelling of tanker market risk using route specific freight rates. Maritime Policy and Management, 25, 117–128. Glen, D., & Martin, B. (2002). Do tanker pools influence market rates? The case of tankers. International Paper presented at IAME Annual Conference, Panama City, Panama, November 15. Glen, D., Owen, M., & van der Meer, R. (1981). Spot and time charter rates for tankers 1970–1977. Journal of Transport Economics and Policy, 45–58. Greene, W. (2003). Econometric analysis. London: Prentice-Hall. Hale, C., & Vanags, A. (1989). Spot and period rates in the dry bulk market: Some tests for the period 1980–1986. Journal of Transport Economics & Policy, 23(3), 281–291. Hawdon, D. (1978). Tanker rates in the short and long run. Applied Economics, 10, 203–217. Johansen, S. (1988). Statistical analysis of cointegration vectors. Journal of Economic Dynamics and Control, 12, 231–254. Kavussanos, M. (1996a). Comparison of volatility in the dry-cargo ship-sector. Journal of Transport Economics and Policy, 30, 67–82. Kavussanos, M. (1996b). Price risk modelling of different size vessels in the tanker industry. Logistics and Transportation Review, 32, 161–176. Kavussanos, M. (1996c). Highly disaggregate models of seaborne trade: An empirical model for bilateral dry-cargo trade flows. Maritime Policy and Management, 23, 27–43. Kavussanos, M. (1997). The dynamics of time-varying volatilities in different size second-hand ship prices of the dry-cargo sector. Applied Economics, 29, 433–443. Kavussanos, M., & Alizadeh, A. (2001). Efficient pricing of ships in the dry bulk sector. In: Conference Proceedings, IAME Annual Conference 2001 (pp. 1005–1040). Hong Kong: Hong Kong Polytechnic University. Koopmans, T. C. (1939). Tanker freight rates and tankship building: An analysis of cyclical fluctuations. F. De Erven, N. V. Bohn (Eds). Netherlands Economic Institute Report No. 27. Haarlem. Mackinnon, J. (1991). Critical values for cointegration tests. In: R. F. Engle & C. W. Granger (Eds), Long Run Relationship: Readings in Cointegration (Chap. 13). Oxford: Oxford University Press.
A Survey of the Modelling of Dry Bulk and Tanker Markets
63
Mackinnon, J. (1996). Numerical distribution functions for unit root and cointegration tests. Journal of Applied Econometrics, 11, 601–618. Mirmiran, R. (2002). The impact of the EU’s common agricultural policy on the structure of demand for shipping transport of grain. Unpublished Ph.D. Thesis, London Guildhall University, London. Norman, V., & Wergeland, T. (1981). Nortank: A simulation model of the freight market for large tankers. Report No. 4, Norwegian School of Economics and Business Administration, Bergen, Norway. Quantitative Micro Software (2002). Eviews 4.0. Irvine, CA, USA: Quantitative Micro Software. Strandenes, S. R. (1984). Price determination in the time charter and second hand markets. Discussion Paper 0584, Norwegian School of Economics and Business Administration, Bergen, Norway. Strandenes, S. R. (1986). NORSHIP – A simulation model for bulk shipping markets. Norwegian School of Economics and Business Administration, Centre for Applied Research, World Market Prospects Report No. 6, Bergen, Norway. Tinbergen, J. (1931). Ein Schiffbauzyclus? Weltwirtcshaftliches Archiv, 34, 152–164. Tinbergen, J. (1934). Scheepsruimte en vrachten. De Nederlandsche Conjunctuur (March), 23–35. Tsolakis, S. D., Cridland, C., & Haralambides, H. (2002). Econometric Modelling of Second-hand Ship Prices, 34. Paper presented at the Annual Conference of IAME, Panama City, Panama, November 15. Tvedt, J. (2003). A new perspective on price dynamics of the dry bulk market. Maritime Policy & Management, 30(3), 221–230. Veenstra, A. (1999). Quantitative analysis of shipping markets. Delft: Delft University Press. Wright, G. (1999). Long run freight rate relationships and market integration in the wet bulk carrier shipping sector. International Journal of Transport Economics, 26(3), 439–446. Wright, G. (2003). Rational expectations in the wet bulk shipping market. International Journal of Transport Economics, 29(3), 309–318. Zannetos, Z. (1966). The theory of oil tank shipping rates. Cambridge: MIT Press.
64
D. R. GLEN AND B. T. MARTIN
APPENDIX Derivation of Present Value Restrictions in a Second Order VAR p
Modifying the notation of the text, so that ␣ij represents the relevant coefficient for the pth order matrix, the amended VAR is as follows: ⎛ ⎞ ⎛ 1 ⎞ ⎛ ⎞ ⎞ ⎛ spt ␣11 ␣211 ␣112 ␣212 e1t spt−1 ⎜ sp ⎟ ⎜ 1 ⎟ ⎜ ⎟ ⎟ ⎜ 0 0 0 ⎟ t−1 ⎟ ⎜ ⎜ ⎜ spt−2 ⎟ ⎜ 0 ⎟ + (A.1) · · · ⎜ ⎟=⎜ 1 ⎟ ⎜ ⎟ ⎟ ⎜ ⎝ St ⎠ ⎝ ␣21 ␣221 ␣122 ␣222 ⎠ ⎝ St−1 ⎠ ⎝ e2t ⎠ St−1
0
0
1
0
St−2
0
where the 4 × 4 matrix above is the companion form of the VAR(2) model. Applying the restrictions given in (15) in the text yields ⎞ ⎛ 1 − ␦␣111 −␦␣211 −␦␣112 −␦␣212 ⎜ −␦ ␦ 0 0 ⎟ ⎟ ⎜ [ 0 0 1 0] ⎜ 2 1 2 ⎟ ⎝ −␦␣121 −␦␣21 1 − ␦␣22 −␦␣22 ⎠ 0
⎛ = [1
0
−␦
0
0
␣111
⎜ 1 ⎜ 0]␦ ⎜ 1 ⎝ ␣21 0
0
␣211
␣112
0
0
␣221
␣122
0
1
␣212
⎞
0 ⎟ ⎟ ⎟ ␣222 ⎠ 0
The four restrictions are: −␦␣121 = ␦␣111 −␦␣221 = ␦␣211 1 − ␦␣122 = ␦␣112 −␦␣222 = ␦␣212 The present value restrictions thus apply to each of the pth order matrices in the original VAR.
3.
ECONOMETRIC MODELLING OF NEWBUILDING AND SECONDHAND SHIP PRICES
H. E. Haralambides, S. D. Tsolakis and C. Cridland 1. INTRODUCTION: THE MARKET FOR NEW AND SECONDHAND VESSELS One could hardly find a shipping paper that does not start by stressing that shipping is a very volatile, unpredictable and risky market; that, at the same time, shipping is the bloodline of trade, growth and welfare and that, therefore, its fortunes are inextricably linked to those motivations that drive people and nations to trade more with each other. Undoubtedly, all the above is true but, in addition, the intrinsic characteristics of the shipping markets themselves, such as its shipbuilding cycles and speculative investments in secondhand ships, also come into play to compound the industry’s notorious volatility. This chapter deals with the factors determining the prices of ships; new or old. Shipping is one of the few industries having a separate and active market where the main assets themselves (ships) are traded. The price of a ship, like that of every other capital asset, depends on the ship’s expected future profitability or, in other words, on the investor’s expectations regarding future developments in the markets he operates. Prices, particularly those of secondhand ships, thus correlate strongly with freight rates and, together with them, fluctuate widely. The timing of the investment is therefore the single most important factor of business success. Shipping Economics Research in Transportation Economics, Volume 12, 65–105 Copyright © 2005 Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0739-8859/doi:10.1016/S0739-8859(04)12003-9
65
66
H. E. HARALAMBIDES ET AL.
Volatility in secondhand ship prices coupled with long delivery times of new ships gives rise to considerable speculation (asset play). The yearly volume of secondhand ships changing hands is indeed significant. Transactions in secondhand ships play an important economic role in the shipping industry: They give shipowners and other investors the opportunity to buy and sell ships directly, thus allowing easy entry and exit to the freight market. This is a major condition for market competitiveness. Instances of low freight rates usually coincide with low vessel values but, despite the fact that this is bad news for owners of existing tonnage, it provides opportunities for new investors to buy in at a low cost. Shipbuilding too is a market whose variables (demand, supply and prices) are subject to distinct cyclical fluctuations. Such volatility has repeatedly led to collapses in newbuilding prices, as well as to disturbances in production, and, consequently, to severe financial problems, even bankruptcy, for shipyards and shipowners. In addition to market expectations, the price of new ships depends on shipbuilding costs and shipyard capacity. New and secondhand ship prices also correlate among themselves. Some would even argue that new and old ships are substitute commodities: the first more technologically advanced, but also more expensive to acquire and with long delivery times, the second usually cheaper and in immediate delivery. In the face of a burgeoning demand and tight shipyard capacity, secondhand ships would thus sell at a premium. On the contrary, in a depressed and over-supplied market, secondhand ship prices would tend to converge to the ships’ scrap values while newbuilding prices could still keep close to shipbuilding costs. Knowing the factors and the way and extent they impact ship prices could afford investors precious insights for the timing of their investments. So, how do new and secondhand ship prices relate to each other and to freight rates? Do exchange and interest rate fluctuations also play a role? Are shipyards competitive, and do newbuilding prices reflect costs of production? Does yard capacity, often created through subsidies and national industrial policies, impact newbuilding and secondhand ship prices? And if it does, is this good for the long term profitability of the shipping industry? These are some of the questions this chapter attempts to address, starting from a review of the work on the subject carried out by earlier researchers. Following this, a Theoretical Error Correction model is presented and estimated for both new and secondhand ship prices. Its forecasting performance is finally compared with that of an Atheoretical Autoregressive (AR) model for all ship types under investigation. Conclusions are presented in the end highlighting the importance of market expectations, shipbuilding costs and shipyard capacity in the determination of ship prices.
Econometric Modelling of Newbuilding and Secondhand Ship Prices
67
2. LITERATURE REVIEW 2.1. Newbuilding Prices The first researcher to analyse the cyclicality of the shipbuilding market was Tinbergen (1931). Tinbergen supported that the condition of the shipbuilding market is very much dependent on the amount of freight offered for shipping. The freight rate is then in its turn dependent on the shipping tonnage present in the market. This leads to an endogenous shipbuilding market cycle, which is caused by the time lag between the demand for shipping capacity and the actual availability of this capacity. In addition to this, he comments that there is also evidence of exogenous disruptions causing the cycle to act unpredictably at different periods of time. Tinbergen takes a supply-demand approach to analysing the newbuilding market based on the cobweb theorem. In other words, Tinbergen describes a model where supply adjusts to price with a specific time lag. More specifically, low total tonnage leads to high freight rates. Ships ordered during a prosperous market period are delivered about one year later, thus increasing the total amount of tonnage. The cobweb theorem approach is subsequently followed by Koopmans (1939). For Koopmans, the shipbuilding market is influenced by expectations concerning the degree of equilibrium between the transportation capacity of the world fleet and the aggregate demand for its services. The main reason behind this reliance on expectations is the time lag between ordering and delivery of new tonnage, since the market situation is shaped according to orders placed several years earlier. As part of his Tankship Building Model, Hawdon (1978) estimates a tanker newbuilding price equation by including both cost related (steel price) and asset pricing (freight rates) variables in his approach. For Hawdon, the price per dwt of new tankers is assumed to be linearly related to freight rates (current and lagged), the size of the fleet, the average size of tankers and the price of steel. He estimates a linear relationship by Ordinary Least Squares (OLS) and 2 Stage-Least Squares (2SLS). Hawdon finds that current levels of freight rates have a significant impact on newbuilding prices, while lagged freight rates are non-significant. Steel prices, as indicators of shipbuilding costs, are also found to be highly significant. He also finds a statistically significant negative coefficient for fleet size. This result reflects the depressing influence of overcapacity on ship prices. Finally “average ship size” is not shown to have a significant effect on newbuilding prices. The latter is a most remarkable result, probably due to model misspecification, for one would expect economies of scale in ship construction to have a perceptible negative impact
68
H. E. HARALAMBIDES ET AL.
on ship prices in competitive markets. Hawdon’s research is not presented along simple demand/supply terms and, as a result, circular causality questions can be raised; for instance, it is not fleet oversupply that directly impacts prices but rather the effect of oversupply on orderbook, which in its turn distorts the supply/demand equilibrium for newbuildings and, then, prices. By taking an asset pricing approach, Beenstock (1985) observes that secondhand prices are flexible whereas newbuilding prices are relatively sticky. He follows this by implying that newbuilding prices adjust to secondhand prices over time. This position however is open to criticism since the shipbuilding industry is supply- and cost driven, whereas secondhand vessels are market driven. Newbuilding prices cannot possibly adjust to something that is so volatile and speculative. By the same token, no country would adjust its shipbuilding capacity, involving a lot of heavy investment and sunk costs, to speculative movement of prices. After all, the drivers of newbuilding prices, as discussed later in the chapter, are competition among shipyards, excess shipbuilding capacity and subsidies (i.e. industrial sectoral policies). These are the important variables in a supply driven industry like shipbuilding, with the rest being of secondary importance. Beenstock and Vergottis (1989) distinguish between newbuilding and secondhand markets and adopt an asset pricing modelling approach. At the time of their yard contracting, ships will typically sell at prices that can differ from those of identical existing new ships by a larger or smaller amount. The main reason for this difference in price, they claim, stems from the fact that a new ship is immediately available to trade, while a contracted newbuilding only becomes available after the construction period has lapsed. Because new contracting is for forward delivery, the market for these ships should resemble a forward market. Prices arranged should reflect market expectations, at the time of contracting, regarding the value of new ships at the time of delivery. This conclusion is partly true today, since, on many occasions, in some countries, it is national policy, i.e. subsidies or aggressive pricing to capture market share, that is reflected in newbuilding prices rather than market expectations. Their newbuilding model is: Ln NB = E(lnPt+1 ) + k k = k1 + k2 . . . where: NB = price of new-built ships k = premium
Econometric Modelling of Newbuilding and Secondhand Ship Prices
69
Beenstock and Vergottis assume here that the prices of new existing ships move in line with expected prices in the secondhand market. The parameter k, representing a number of factors (k1 > 0), is a premium that reflects the fact that the new-built ships may embody superior technology. The case where k2 < 0 represents a risk premium, required to attract investors into taking forward positions in tankers. This means that the demand side of the newbuilding market is one of the factors determining newbuilding prices. The Beenstock and Vergottis model of newbuilding prices is a case of an asset pricing model. They assume that newbuilding and secondhand ships are practically perfect substitutes. However, although it seems plausible that demand for newbuildings has a high price elasticity, because secondhand prices are a close substitute, it is unlikely that newbuilding and secondhand vessels are perfect substitutes. This is due to several reasons: (1) (2) (3) (4) (5) (6)
Secondhand and new ships are available in different time frames; Different trading conditions apply as a result of timing; Different costs apply as a result of timing; Different risks apply as a result of timing; Secondhand ships have shorter trading lives than the newly-built ones; There may be technological advantages of one over the other.
Jin (1993) takes a supply and demand approach to analysing the newbuilding tanker market. However, her analysis also incorporates elements of the cost-based approach, such as shipyard labour costs. The study provides a framework to analyse quantitatively the relationship among market factors, such as newbuilding prices and orders, and other exogenous factors, such as technological change in the shipbuilding industry and changes in shipping distances. The study also identifies the impacts of these factors on tanker supply and demand. For Jin, shipbuilding is both a capital- and labour intensive industry, in terms of absolute uses of factor inputs. There are high fixed costs associated with shipyard operations. Ship construction creates a large number of jobs, which generate important regional and national economic impacts. Shipbuilding capacity is represented by the quantity of capital and labour in the industry. Shipbuilding capacity is the most important factor determining the supply of new tankers, and higher capacity is associated with greater supply. Technological change is another important factor influencing new tanker supply. According to Jin, shipbuilding costs should be a decisive factor in a supply function but, in the short run, costs may be less influential than capacity. Furthermore, it is difficult to measure world shipbuilding costs, since there is great variation in labour costs among different suppliers; costs of shipbuilding
70
H. E. HARALAMBIDES ET AL.
materials and energy in different countries are also different. For this reason Jin uses the average number of employees in the Japanese shipbuilding industry at t-1 as a cost indicator. This is not a very convincing argument since there are many reasons for labour force numbers to fluctuate. Therefore, it is labour costs that one should be interested in rather than number of workers. The inverse supply function can be specified as: NB = f s (Orderbook, Cap, EMPt−1 Steel, T i, e 2i ) . . . where for observation i: NB = The price of new tankers; Orderbook = The quantity of tanker newbuilding supplied which is equal to the quantity of tanker newbuilding demanded; Cap = Shipbuilding capacity; Steel = The steel price; EMPt−1 = Average number of employees in the Japanese shipbuilding industry at t-1; Ti = Technological change; e2i = A stochastic error term. Both “orderbook” – a demand variable – and “capacity” – a supply variable – have the right signs and are statistically significant. Apparently, a high orderbook signifies a “tight” market and high prices, while excess shipyard capacity ought to have a depressing effect on ship prices. The price of steel – a cost and supply variable – is found statistically insignificant and it is consequently dropped from the estimations. This is an important finding, showing that the international shipbuilding market may be driven by subsidies and other sectoral policies rather than competitive pressures. Having said this, the model is able, on the other hand, to establish a significant negative impact of “technology”, and thus productivity, on ship prices. Again, these contradictory results can be attributed to model misspecification, due to the lack of an underlying integrated economic model able to capture causality relationships among variables. By combining the asset pricing model with the cost-based model approach, Volk (1994) attempts to explain shipping and shipbuilding market cycles through developments in freight rates, shipping innovation, psychological and speculative factors in shippers’ behaviour, and a limited endogenous influence of replacement orders. Summing up, we could say that the major determinants of newbuilding prices, considered by the above authors, are:
Shipbuilding costs; Shipyard capacity; Vessel orderbook; Freight rates; Secondhand prices.
Econometric Modelling of Newbuilding and Secondhand Ship Prices
71
2.2. Secondhand Ship Prices With the exception of Charemza and Gronicki (1981) who report equations in which ship prices adjust to freight and activity rates, Beenstock (1985) is the first one paying attention to ship markets and the determination of ship prices. Beenstock argues that simple supply-demand analysis is not appropriate for ship prices since a ship is a capital asset of considerable longevity. Thus, based on portfolio theory, Beenstock comes up with the following specification: F × PSt = f PS Wt
E t t+1 E t PSt+1 , , it PSt PSt
. . . where: PS: the secondhand price; W: the world wealth; F: the fleet size (bulk carrier or tanker depending on the circumstances) Et t+1 : expected ship earnings for the coming year, equal to timecharter minus operating costs; Et PSt+1 : expected secondhand price for next year; i: the interest rate. According to this, the share of ships in total world wealth varies directly with the expected return on ships as capital assets and is inversely related to alternative investments. This capital asset equation is based on the assumption that the share of ships in total world wealth behaves as part of a well-diversified portfolio consisting of all world wealth. In other words, for given rates of return and wealth, the demand for ships varies inversely with the price of ships since the relative return on ships falls as price rises and because of wealth effects induced by relative ship price changes. Beenstock and Vergottis (1989a, b, 1992, 1993) follow this approach in all their subsequent research. A capital asset allocation model is meant to calculate the optimal shares in assets of a portfolio, given certain fixed expected return and risk for all assets. However, Beenstock and Vergottis do not give any arguments why this model can also be applied to calculate ship prices, or the reason it can be applied to stock prices or the price of any asset. Besides that, world wealth is something very illusive and it is certainly not equivalent to world GDP, which Beenstock and Vergottis use in their model. In addition, in the same chapter, Beenstock (1985) assumes that new and secondhand ship prices are perfectly correlated, thus new and secondhand ships are the same asset, only differing in age. As argued elsewhere in this chapter, this assumption is open to criticism. In his subsequent work together with Vergottis (1989, 1992, 1993), he introduces an additional dynamic element into the model, i.e. the newbuilding market. Strandenes (1984, 1986) regards secondhand values as a weighted average of short and long-term profits. In Strandenes (1986), the secondhand market is integrated with the newbuilding market. Secondhand prices depend on expectations
72
H. E. HARALAMBIDES ET AL.
concerning future developments in other shipping markets. As a result, a present value equation is estimated including the ship’s value, its expected cashflow in each period, and the number of expected trading days. Individual cash flows are then substituted by average expected earnings per year. An interesting point is the assumption of infinite economic life and the inclusion of a depreciation factor. The assumption of infinite economic life however is not realistic, as ships have a finite life and a substantial terminal value. Kavussanos and Alizadeh (2002) show that if this is not taken into account results are different. Expected earnings are subsequently expressed as a weighted sum of current and future expected long-term earnings. Finally, by using the real depreciation factor to correct for the ship’s age, a general formula, for any ship price regardless of age, is obtained. Recent studies by Kavussanos, Kavussanos et al. and Veenstra employ atheoretical models (Vector Autoregressive (VAR), Autoregressive Conditional Heteroscedasticity (ARCH), and Autoregressive (Integrated) Moving Average (AR(I)MA) models) for the estimation of secondhand ship prices. By using the cointegration methodology, Veenstra (1999) establishes that secondhand ship prices for various ship sizes in bulk markets are stationary in first differences, thus permitting the search for long-run cointegrating vectors between them. The variables chosen for examination include the secondhand price, a timecharter rate, as well as newbuilding and scrap prices. Veenstra distinguishes between replacement and speculative sales. He analyses data on two different ship ages: 5-year-old ships, representing replacement sales, and 10-year-old ships, representing speculative sales. Kavussanos (1996, 1997) examines the dynamics of volatilities in the dry-bulk and tanker markets. By employing time-series modelling – atheoretical ARCH models – he finds that prices of small vessels are less volatile than those of larger ones and the nature of this volatility varies across sizes. Glen and Martin (1998) make a similar study on tanker market risk. Their results are in line with those of Kavussanos, despite differences in data, sample period and modelling technique. In another study, Glen (1997) examines the dynamic behaviour of secondhand prices of tankers and dry cargo vessels over various time periods. He aims to determine whether or not the market for such assets is efficient. He extends and re-analyses the results of an earlier study by Hale and Vanags (1992) by employing the Johanssen method of testing for cointegration. He concludes that the existence of cointegration does not necessarily imply market inefficiency, if the factors that create the common trends are stochastic in nature. Therefore he argues that the evidence put forward in his paper is consistent with market efficiency in the long run.
Econometric Modelling of Newbuilding and Secondhand Ship Prices
73
3. METHODOLOGY: MODEL SPECIFICATION, DATA COLLECTION AND ANALYSIS 3.1. Variable Identification Both the analysis of the newbuilding and the secondhand market can be represented in terms of a supply and demand framework. Demand for new ships can be expressed as a function of timecharter rates (expressing expectations), secondhand ship prices (price of substitutes), and of course new ships’ own price, NB. QD NB = f(fr, SH, NB) By the same token, the supply of new vessels can be seen a function of the orderbook as a percentage of the fleet (used as a proxy for shipyard capacity changes), shipbuilding costs, exchange rate fluctuations and newbuilding prices: Q SNB = f(O/F, NB, Xrate, C) S Since, at equilibrium, Q D NB = Q NB , newbuilding prices can be expressed as: ⎛ ⎞ +/−
⎜+ + O + − ⎟ NB = f ⎝fr, SH, , C, Xrate⎠ S In the same way, demand for secondhand ships can be expressed as a function of timecharter rates, newbuilding prices, secondhand ship prices and the cost of capital. QD SH = f(fr, SH, NB, LIBOR) Supply of secondhand vessels is here assumed to be a function of the orderbook as a percentage of the fleet and secondhand prices: Q SSH = f(O/F, SH) S Since Q D SH = Q SH , secondhand ship prices can be expressed as: +
+/−
− O , LIBOR) F . . . where: NB = newbuilding price; SH = secondhand price of a vessel; fr = the vessel’s average timecharter rate per day for the year; O/F = orderbook as a percentage of the total fleet (including combined carriers) used as an indicator of shipyard capacity; C = Shipbuilding costs expressed as the price of rolling steel +
SH = f(fr, NB,
74
H. E. HARALAMBIDES ET AL.
plates in Japan; Xrate = The USD/Yen or USD/Won exchange rate; LIBOR = Interest Rate (London Inter-Bank Offered Rate). The expected sign is given above each variable. 3.1.1. Variables Used in Both Models 3.1.1.1. The timecharter rate. According to our models, both newbuilding and secondhand prices are a function of the vessel’s expected revenue. This is expressed as the average timecharter equivalent rate per day. The reason for this is that timecharter rates denote shipowners’ and charterers’ expectations of things to come. Therefore, it is assumed that the higher the timecharter rate, the higher a ship’s future profitability and, as a result, the higher its value. Timecharter rates determine sector dwt demand, as well as sector orderbook, since shipowners will be eager to build more ships of the size or type yielding the highest returns. Therefore, shipping dwt demand determines which type of vessel yards will build. An overall fall in dwt demand within a specific sector or segment will prompt a shipbuilder to switch production to another sector or segment. For the large yards that are able to switch their production from one sector/segment to another, their order of preference, in descending order, is: Containerships LNG carriers Tankers Bulk Carriers and, always, the larger the better. Therefore, for most large shipbuilders, the bulk carrier is the vessel of last resort. The building of bulk carriers in large shipyards implies a softer demand for the preferred vessel types and a lowering of newbuilding prices to stimulate ordering in a sector where vessel earnings are restricted by low cargo values. In other words, the newbuilding market for different ship types and sizes is not homogeneous and different ship types and sizes exhibit different characteristics with respect to the determinants of newbuilding prices. Furthermore, according to the view firstly expressed by Beenstock (1985), new and secondhand ships are also capital assets. This means that they compete with other investments in terms of profitability. The higher the return on investment in shipping, the more money investors will be willing to pour in the market and as a result the higher the demand for new and secondhand ships. In principle, this argument is correct but there are exceptions: Norwegian K/S schemes for example have served to illustrate that there have been times when secondhand prices have appeared to move outside of the primary influences of earnings (and newbuilding prices). At the height of their popularity, there were many K/S schemes in the market, to buy tonnage, and this pushed the price up for secondhand vessels even
Econometric Modelling of Newbuilding and Secondhand Ship Prices
75
though vessel earnings were low. In our model, speculative demand for secondhand ships can be captured by the cost of capital variable, LIBOR, expressing the “availability of finance,” particularly when this is combined with attractive tax positions of individual investors. Apart from denoting expectations, the timecharter rate is also a much better market indicator than the spot rate, which is route specific, and more importantly in the tanker market, the Worldscale. There are several reasons for this. A timecharter rate is by definition the result of an owner and a charterer agreeing to a given hire, over a future period of time, and it can be assumed that this, in some way, reflects the parties’ market expectations in the period ahead. Timecharter rates are less volatile than spot rates and therefore do not reflect the highs and lows of the spot market. Shipowners also like to project their income in terms of being net of voyage costs and in this the timecharter rate is appropriate whereas the spot rate only provides a gross income position. Spot rates also reflect only a snapshot in time, give no indication of forward expectations and are notoriously difficult to forecast by nature of their volatility. For all of these reasons, spot rates are seldom a driver to newbuilding ordering and it would be unwise by owners to place too much faith in them. By contrast, timecharter rates display greater stability and are easier to understand. It is often a condition of ordering that owners cover their initial forward position with a timecharter. For Worldscale in particular, it is a common mistake by many researchers (Beenstock & Vergottis, 1989, 1993; Hawdon, 1978), to use it as an income indicator over a period of time. Worldscale is a cost based schedule that is re-calculated on an annual basis for a full cargo for the standard vessel based upon a round voyage from loading port to discharging port and return. This simply means that when changes occur in bunker prices, port dues or the exchange rate of the currency of the States included in this route, WS100 for this year, as Fig. 1 shows, will be different in dollar terms than WS100 for the same route the previous year. Furthermore, while in Worldscale terms one year may look better than another, a different story may appear if the Worldscale flat rate (WS100) is adjusted for increases in voyage costs. In this case, as Fig. 2 shows, while the WS rate may be higher, the actual dollar per tonne rate may be lower, thus depicting worse rather than improving market conditions. Consequently, discrepancies will occur between the WS and the actual rates obtained by the vessels over the years thus distorting the final results. 3.1.1.2. Newbuilding and secondhand prices (for the secondhand and the newbuilding model respectively). Secondhand and new ships are substitutes since an increase for example in the price of secondhand ship prices will lead to an
76
H. E. HARALAMBIDES ET AL.
Fig. 1. Comparison of Original WS with WS 2001.
increase in the demand for new ships. Moreover, since it is easy for shipowners to switch from secondhand ships to newbuildings, the demand is more elastic thus making these goods close substitutes. In other words, a freight rate increase will increase demand for ships with an immediate positive effect on secondhand ship values. Ceteris paribus, this will make shipowners more eager to order new ships thus pushing newbuilding prices up as well. As a result, a positive sign is expected for these variables. 3.1.1.3. Shipyard capacity and orderbook. Shipyard capacity is an important determinant of shipbuilding output. Shipbuilding is an increasing returns to scale industry and, as such, minimum efficient size often requires significant excess capacity; something often accentuated by shipbuilding subsidies, due to
Fig. 2. VLCC Rates in Worldscale and US Dollars per Tonne.
Econometric Modelling of Newbuilding and Secondhand Ship Prices
77
this industry’s significant economic impacts. In the presence of excess capacity, competition among shipyards holds prices down. Statistical discrepancies and lack of long enough time-series data have not allowed the use of this variable in the present model. Instead, O/F has been used as a proxy. The idea here is that a high O/F signifies a tighter shipbuilding market and thus a positive relationship is expected to exist between this variable and shipbuilding prices. In the case of secondhand ships, a high O/F may have either a positive or a negative impact on secondhand ship prices. A large order book signifies both a prosperous market, as well as longer delivery times. For both reasons, owners of existing tonnage would, ceteris paribus, be reluctant to sell and this would drive secondhand ship prices up. At the same time, a high O/F may affect negatively expectations of future profitability (considerable tonnage entering the market in the near future, thus depressing rates) and this could exercise downward pressures on secondhand ship prices. 3.1.2. Cost Related Variables Used in the Newbuilding Model 3.1.2.1. Shipbuilding costs. Despite subsidies and industrial policies in a number of shipbuilding countries, shipyards do compete with each other and, as a result, shipbuilding costs should bear, one way or the other, upon newbuilding prices, at least in the long run. Shipbuilding costs, and benchmarking among different countries, are difficult to measure, particularly over a length of time suitable for econometric analysis. As a result, and in contrast with Jin’s approach to use labour costs as a proxy of shipbuilding costs, the price of steel plates in Japan is used here as a newbuilding cost indicator. Steel plates account for approximately 30% of newbuilding prices and their fluctuations could provide a reliable proxy for total shipbuilding costs. 3.1.2.2. Exchange rates. Exchange rates can help shipbuilding nations to become more competitive by offering lower prices for new ships. Since most shipbuilding costs are incurred in local currency but the final product is quoted in dollars, a devaluation of the currencies of the major shipbuilding nations against the U.S. dollar could have a negative effect on newbuilding prices. An example of this could be found in Korean yards whose competitiveness may be traced, among others, to the weakness of their domestic currency. Figure 3 shows the price of a new VLCC built in Korea quoted in both U.S. dollars and Korean Won. After the depreciation of the Won against the U.S. dollar in 1999, the Koreans were quoting prices in Won that were much higher than the 1997 levels. However, in U.S. dollar terms they were quoting a lower price that made them more competitive
78
H. E. HARALAMBIDES ET AL.
Fig. 3. Newbuilding Prices in Korean Won and $US. Source: Drewry Shipping Consultants Ltd.
and kept newbuilding prices low. As a result, a negative sign is expected for this variable. 3.1.3. Capital Cost Related Variables Used in the Secondhand Model The last variable included in the secondhand model is the cost of capital. The reason that this variable is not included in the newbuilding model as well is that there is always an abundance of sources and levels of capital to finance a newbuilding project, especially government subsidies and shipyard credit schemes that are normally unavailable to secondhand investors. Furthermore, the fact that investors opting for newbuildings rather than secondhand ships are in principle more “liquid” than those investing in secondhand tonnage means that this variable does not affect the determination of new ship prices. It is fairly obvious that, ceteris paribus, high interest rates will impact negatively the demand (and thus price) for secondhand ships. On the one hand, their acquisition becomes more expensive, while, on the other hand, rising interest rates signify a growing world economy in which bankers may have plenty of alternative opportunities of lending their resources. As a result, in rising markets, bankers may exercise stricter due diligence, ship inspections, and similar endeavours to manage risk, that may make overall credit availability scarcer and more expensive to shipowners. The opposite however could well be the case, and bankers could by
Econometric Modelling of Newbuilding and Secondhand Ship Prices
79
all means rush in an investment bonanza of rising markets and expectations. This is a good place to point out again these authors’ conviction, following Tinbergen, that the only “meaningful econometrics” are those of well-specified structural models able to capture significant causality relationships among economic variables.
3.2. Data Collection and Analysis Financial data on the three-month London Inter-Bank Offered rate (3-month LIBOR), were obtained from Datastream International for the years 1960–2001. Steel Prices for roll-plates in Japan were obtained from World Steel Dynamics. Shipping related data was collected from the annual reviews and the monthly reports of the major shipbroking houses including Clarksons, Fearnley’s and SSY. The oldest review found was that of 1965 by Fearnley’s including data dating back to 1960. The first step was to distinguish between ship types. For the bulk carrier market, three ship types were selected: Handy (15–49,999 dwt), Panamax (50–79,999 dwt) and Capesize (80,000+) bulk carriers. Tankers were classified into Handy (15–49,999 dwt), Panamax (50–79,999 dwt), Aframax (80–120,000 dwt), Suezmax (120–199,999 dwt), and Very Large Crude Carriers (200,000+). Based on data provided by shipbrokers, for every ship type, a specific vessel was taken as a benchmark: 30,000 dwt geared, 65,000 dwt and 120,000 dwt ships for bulk carriers and 30,000 dwt, 70,000 dwt, 105,000 dwt, 130,000 dwt and 280,000 dwt ships for the tanker market, all built in Japan. The exception was the tanker market where, from 1985 onwards, Korean built vessels were used. The reason was that Korean shipyards became more competitive price-wise due to massive investment, efficiency and quality improvement as well as due to the currency situation, which clearly disfavoured the Japanese Yen. As a result, prices quoted by Korean shipyards have become the industry’s benchmark. The reason for the above disaggregation was to investigate whether and to what extent the variables influencing secondhand vessel prices affect differently the various ship sizes/segments. Beenstock (1985), Beenstock and Vergottis (1993) have developed models based on aggregate data. Glen (1990) however argues that the traditional assumption that the oil tanker market is homogeneous is no longer valid and it has been replaced by route and size differentiation. According to Glen (1990), size differentiation emerged because of the limited flexibility of the largest oil tankers, due to increased supply and lower levels of port capacity growth thus creating severe constraints on port availability and hence route flexibility.
80
H. E. HARALAMBIDES ET AL.
Consequently, large vessels became riskier assets to own. This is also supported by the results reported by Kavussanos (1996, 1997). These results require analysis of both the bulk carrier and the tanker market at a disaggregate level. Furthermore, this approach can become a useful tool for the shipowners to base their investment decisions, since they will be able to know which factors, and to what extent, affect each different ship size. Finally, a disaggregate approach can also be an interesting tool for brokers, who will be able to assess the price of a secondhand vessel more accurately and then discount or increase that price according to the specifications of the ship that is marketed for sale. Due to the number of variables and the problems experienced with data collection, the use of annual data was decided. This way fewer data discrepancies occurred and results were more reliable, particularly for the earlier years where data was scarce and unreliable. Another issue was the starting date of the observations. The 1960s were a time when new types of larger, purpose-built ships were coming on stream on a regular basis to satisfy the increasing demand for sea transport services deriving from the impressive increase in trade volumes. Thus the late 1960s saw the introduction of VLCCs and Cape size bulk carriers. Despite the fact that data existed for some ship types from mid-1950s, it was decided that the starting point of this analysis would be 1968. This year was chosen because, for the first time, the world fleet included vessels of all the different types analysed in this chapter. 3.2.1. The Combination Carrier Effect Often reported separately from bulk carriers and tankers, combination carriers were developed in order to exploit trade imbalances and demand fluctuations. These vessels primarily offered the potential to maximise laden/ballast ratio and/or switch trades through their greater cargo carrying capabilities. A secondary advantage was the ability to re-position into other markets more easily than straight bulk vessels through carrying backhaul cargoes. They reached their peak in the 1970s but their operational problems and charterers’ prejudice did not make them a success with shipowners who steadily abandoned this concept. The result was an ever decreasing fleet, mainly consisting of old vessels since over the last years only a handful of this type of vessels has been ordered. Nevertheless, these ships pose some pressure, far less significant today than in the 1970s, on the supply side of both the tanker and bulk carrier market. To cope with this issue, an investigation into the trade patterns of these ships, their split between the oil and the dry bulk trades, was necessary. Fearnleys and Jacobs (SSY) Reviews from 1965 up to 2001 provided the necessary data. The combined carrier fleet and orderbook were divided between dry bulk and oil trades and distributed according to ship sizes based on data from Drewry. This way, a clearer picture of the real size of the fleet employed in both dry
Econometric Modelling of Newbuilding and Secondhand Ship Prices
81
and oil trades was obtained and more accurate estimations of the fleet size could be made.
3.2.2. Problems Experienced with Data Data compilation proved to be the most time consuming and difficult task. This was not due to a lack of information sources but to the inconsistency and quality of the data itself. One of the major problems incurred during this research was the nonmatching of the size class categories. Reported ship sizes change over time. Even when long data series were obtained from a single source, year on year changes with respect to size could be observed. This problem was more intense in the earlier years where not all class size categories were as well developed and distinguished as today. For example, in the early and mid-1960s, large tankers used to be described as anything above 80,000 dwt, something that is certainly not the case today. Such things can create significant problems with the data series, particularly when one benchmarks against today’s size distributions. Take for example newbuilding prices. Until 1975, Fearnleys was reporting newbuilding values based on the prices quoted by European shipyards. From 1976 onwards, prices quoted by Japanese shipyards became the benchmark. As a result, a price reduction of 10% compared to the previous year in the price of a newbuilding built in Japan may look as a 60% reduction if the price is compared to the 1975 price quoted by a European shipyard for the same ship. Therefore, in some cases, adjustments had to be made by comparing Fearnleys quotations with those of other shipbroking houses such as Clarksons, and SSY. Another problem is vessel classification. Some vessels that some shipbroking houses regard as Aframaxes may be classified as Suezmaxes by others, while some people distinguish between handy, handymax and superhandymax vessels, with the latter being in the same size category as some Panamax vessels. This creates significant discrepancies amongst the fleet data reported by various shipbroking houses. For example, some may report in their orderbook figures only the deals that go through, whereas others may also include unexercised options. Furthermore, another problem, particularly with tanker vessels, was the inclusion of other ship types in fleet statistics. Some, for example, may report tanker fleet figures including chemical carriers, or others include combined carriers in the tanker orderbook. The situation regarding fleet related data became even worse after 1995. Nearly all shipbroking houses use Lloyd’s Register to source fleet information. However, in 1995 Lloyds Register changed the ship type classification for the World Merchant Fleet. This meant that some ships previously regarded as bulk carriers are now classified under general cargo vessels. Consequently, fleet statistics pre and after 1995 are even more difficult to compare.
82
H. E. HARALAMBIDES ET AL.
3.2.3. Data Problems Experienced in the Newbuilding Model 3.2.3.1. Problems associated with quoted newbuilding prices. As a general rule, the main problem with public information is that it is based on reported prices. These may be subject to a margin of error, depending in part on the source they derive from. In addition, they seldom, if ever, reveal crucial elements such as payment terms. In most business areas, there is likely to be a discount for cash or prompt payment. It may be that, typically, payments are staged (contract signing, keel laying, launching, delivery). Some deals have payments which are endloaded, in which case a higher reported price would seem a reasonable expectation. Payment terms may have implications for issues such as refund guarantees. Even for vessels of similar outline specifications, there are valid reasons for price differentials between and within yards. Some of these are listed below (Drewry, 2001): Reported prices may not be accurate; The vessel may or may not be of a standard design, and the degree of changes from the standard specifications demanded by the customer may vary; It may be a single ship order, or part of a series, or the order may include future options; The vessels may have differing levels of “bought in” equipment such as engines or containment systems; Existing customers may be offered a more competitive price as some yards place customers relations above the level of profit on each individual contract; Differing financing terms may alter the final cost. 3.2.3.2. The shipyard capacity measurement problem. Shipyard capacity was proven to be the most difficult in terms of data availability. Shipbuilding capacity is difficult to measure and different sources quote different figures. This makes it difficult to obtain long series of reliable, widely accepted data. The two major sources of information in this area are the Organisation for Economic Cooperation and Development (OECD) and the Association of Western European Shipbuilders (AWES). In many cases, in OECD reports, annual output seems to exceed actual capacity, something that cannot happen in reality, while AWES statistics do not make a clear division between available capacity and annual output. Table 1 compares the OECD shipyard capacity figures, in CGT (000s), with the delivered ship output from AWES statistics. In Japan and the EU, output has substantially exceeded capacity, whereas no capacity figures are reported for Korea and the U.S. Furthermore, AWES provides statistics on shipyard capacity from 1975 onwards. These examples raise a significant query over the accuracy and reliability of the above information as a measure of merchant shipbuilding capacity, and hence
Econometric Modelling of Newbuilding and Secondhand Ship Prices
83
Table 1. OECD Shipbuilding Capacity and Output. 1997
EU Japan Korea USA
1998
Capacity
Output
Capacity
Output
3190 5600 – –
3225 6298 3983 134
3230 5600 – –
3586 6834 3656 360.4
Source: Capacity OECD, Output AWES.
the supply/demand imbalance in shipbuilding. As a result, it was decided to exclude this variable from the model and instead use O/F as a proxy for shipyard capacity. 3.2.4. Transformation into Logarithms Time series were transformed into natural logarithms. Several reasons can be mentioned to justify this. First, in certain circumstances, taking logarithms may stabilise a non-stationary variance. Also, an exponential trend in time series becomes linear after transformation, thus making it easier to analyse it in more detail. Finally, parameters in linear structural models with variables in logarithms can be interpreted as elasticities. 3.2.5. Dummy Variables During model estimation, some observations could be considered as outliers. To overcome the problem of these observations having a large impact on the estimation results, dummy variables were included in the models. For all tanker segments dummy variables were used for the years 1971, 1973 and 1979 to make up for the effects on the freight market, and subsequently on the secondhand market, of unforeseen political events. These are the Tap line closure along with the restrictions on Libya oil production by the new regime in 1971, the Yom Kippur War followed by the first oil crisis in 1973 and the subsequent second oil crisis in 1979. For Bulk carriers dummies were used for 1985 and 1986. These were the years that the ships ordered by Sanko in 1983 were delivered. Sanko, a Japanese company, was facing tremendous financial problems due to the slump in the tanker market in the late 1970s and the sharp market fall in the bulk sector in 1982. Instead of looking for a restructuring plan it went on to order secretly 125 handy size bulk carriers. Such behaviour was not justified by market conditions, since the market was showing no signs of recovery but it was due to the company’s precarious economic situation and its desire to seek ways to avoid bankruptcy. Furthermore, during these years many banks decided to foreclose on many of the
84
H. E. HARALAMBIDES ET AL.
loans they had provided to shipowners in previous years. This caused a record number of ship auctions at very low prices, which had a distorting effect on the value of secondhand vessels.
3.3. Model Specification 3.3.1. Testing for a Unit Root The stationarity (or not) of a series can strongly influence its behaviour and properties. For example, persistence of shocks will be infinite for nonstationary series. Also, the use of non-stationary series can lead to spurious regressions. If two variables are trending over time, a regression of one on the other could have a high R2 even if the two are totally unrelated. Furthermore, if the variables in a regression model are not stationary, the standard assumptions for asymptotic analysis are not valid. In other words, the usual “t-ratios” will not follow a t-distribution and thus hypothesis testing of regression coefficients cannot be carried out. The statistical methodologies used to test for stationarity were initiated by Dickey and Fuller (1979) and further improved by Dickey and Fuller (1981). To test for a unit root in shipping related variables, the Augmented Dickey Fuller Test is applied here. To perform this test, the number of lagged first difference terms to add to the test regression need to be specified. The usual advice is to include lags sufficient to remove any serial correlation in the residuals. The number of lags is determined by the order of the AR model one assumes to be valid. An idea about this can be obtained by looking at the Auto Correlation (AC) and Partial Auto correlation Functions (PAC). A look at these functions for all variables used in this chapter (not reported here) indicates that the AR is of order 1. As a result, and in addition to the fact that we use annual data, we choose to add 1 lagged first difference term to the test regression. Secondly, we have to decide whether to include other exogenous variables in the test regression. There is the choice of including a constant, a constant and a linear time trend, or neither in the test regression. The choice here is important since the asymptotic distribution of the t-statistic under the null hypothesis depends on our assumptions regarding these deterministic terms. In this case we chose to run the test with both a constant and a linear trend since the other two cases are just special cases of this more general specification. In case there is no trend in the variable, the t-statistic for trend will be statistically insignificant. Based on the above specification, the Augmented Dickey Fuller test for every variable was performed. The results are reported in Tables 2 and 3.
Econometric Modelling of Newbuilding and Secondhand Ship Prices
85
Table 2. ADF Test Results for Stationarity (Bulk Carriers). Ship Type/Variable
Levels
Stationarity 1st Difference Stationarity
1% −4.2826
−2.910586 −2.334160
No No
−5.158053 −6.464367
Yes Yes
−2.088419
No
−4.084219
Yes
−2.314715
No
−4.591673
Yes
Handy LnNB Lnfr Ln(O/F) LnSH
−3.169618 −3.816648 −3.625152 −3.433941
Yes Yes Yes No
−5.636254 −5.462005 −5.186416 −4.653240
Yes Yes Yes Yes
Panamax LnNB Lnfr Ln(O/F) LnSH
−3.191532 −4.241085 −2.898470 −3.251515
No Yes No No
−4.057055 −5.551264 −4.794752 −4.498089
Yes Yes Yes Yes
Capesize LnNB Lnfr Ln(O/F) LnSH
−3.223442 −4.124810 −1.882862 −3.371316
No Yes No No
−5.073918 −5.843474 −7.666681 −5.001132
Yes Yes Yes Yes
LnLIBOR LnC (Shipbuilding Costs) LnXrate (Exchange Rate, yen) LnWon (Exchange Rate won)
MacKinnon Critical Values 5% −3.5614
10% −3.2138
Note: ADF test models contain an intercept and no trend. The null hypothesis is that the series is non-stationary. This hypothesis is rejected if the statistics are larger in absolute values than the critical values. The test equation includes a trend and an intercept. Number of lagged first difference terms added to the regression: One (1).
The results in Tables 2 and 3 indicate that log-levels of most variables are nonstationary, while their log-first differences are stationary. This suggests that these variables are in fact integrated of first order, I(1). Exceptions are the timecharter variable for all bulk carriers, as well as handy bulk carrier newbuilding and O/F variables, which are stationary I(0). The properties of variables with different degrees of integration are markedly different. Autocorrelations of an I(0) series decline rapidly as the lag increases, while those of an I(1) series decline slowly, if at all. Innovations of I(1) series affect all subsequent values, so they have infinite memories, while I(0) series give smaller weight to events in the more distant past. Generally speaking, conventional results and tests in the classical normal regression model are valid only if all variables are I(0). If the variables are I(1) or higher or a mix of I(0) and I(1), the distributional theory is different and the usual
86
H. E. HARALAMBIDES ET AL.
Table 3. ADF Test Results for Stationarity (Tankers). Ship Type/ Variable
Levels
Stationarity
1st Difference
Stationarity
MacKinnon Critical Values
1% −4.2826 Handysize LnNB Lnfr Ln(O/F) LnSH
−2.000447 −3.368731 −2.786261 −2.619826
No No No No
−4.779069 −4.408206 −3.767966 −4.339763
Yes Yes Yes Yes
Panamax LnNB Lnfr Ln(O/F) LnSH
−3.018837 −3.493548 −3.007487 −2.709682
No No No No
−4.496352 −4.929942 −5.381145 −3.980877
Yes Yes Yes Yes
Aframax LnNB Lnfr Ln(O/F) LnSH
−2.513246 −3.025137 −2.904352 −3.003548
No No No No
−3.603109 −5.096917 −4.367984 −4.318656
Yes Yes Yes Yes
Suezmax LnNB Lnfr Ln(O/F) LnSH
−2.992179 −3.132850 −1.849302 −2.252434
No No No No
−5.686306 −4.506962 −4.953553 −4.477216
Yes Yes Yes Yes
VLCC LnNB Lnfr Ln(O/F) LnSH
−2.936825 −2.317703 −3.022314 −2.523191
No No No No
−3.570359 −4.881463 −4.894199 −4.612988
Yes Yes Yes Yes
5% −3.5614
10% −3.2138
Note: ADF test models contain an intercept and no trend. The null hypothesis is that the series is non-stationary. This hypothesis is rejected if the statistics are larger in absolute values than the critical values. The test equation includes a trend and an intercept. Number of lagged first difference terms added to the regression: One (1).
test statistics are no longer valid. This means that no inferences can be made from such models. From the analysis of the shipping related variables above it is shown that they are a mixture of I(0) and I(1). If a regression includes such a mixture, as is the case in Beenstock and Vergottis’ work, no inferences can be made from the results. In order however not to lose the whole framework of regression-based statistical inference, we need to deal with the unit root so that standard asymptotics do apply again. One way to deal with the unit root problem is to filter the series so that the unit root is filtered out. This can be done by a class of models using a combination of first differenced and lagged levels of cointegrated variables. This is known as Error Correction Model or Equilibrium Correction Model. Provided that the
Econometric Modelling of Newbuilding and Secondhand Ship Prices
87
variables constituting the error term are cointegrated, then the error term will be I(0) even though the constituents are I(1). It is thus valid to use Ordinary Least Squares (OLS) and standard procedures for statistical inference. An error correction model takes the following form (Brooks, 2001): y t = 1 x t + 2 (y t−1 − ␥x t−1 ) + u t where yt −1 − ␥xt −1 is known as the error correction term. The variables in lagged levels representing this error correction term must be cointegrated in the long run. This way, an error correction model is able to capture the effects the independent variables have on the dependent one both in the short (less than a year) and the long run (more than a year). Therefore, in order to specify such a model, cointegration tests have to be performed so that cointegrating relations are found. 3.3.2. Testing for Cointegration A group of non-stationary time series is cointegrated if there is a linear combination of them that is stationary; that is the combination does not have a stochastic trend. The linear combination is called the cointegrating equation. Its normal interpretation is as a long-run equilibrium relationship. This allows us to test whether the variables we have chosen to include in our model exhibit such relationships and, consequently, can be used together in the model. Taking the Johansen (1991) approach, we find that for all ship types there are at most four relations in our models, which makes us confident with estimating them. For every ship type, the hypothesis that there is no cointegration among our variables is rejected at the 5% level for both shipping markets (a full set of results for this analysis can be obtained from the authors). Based on the above, different combinations of cointegrated variables were made according to the particular characteristics of every ship segment.
4. OUTPUT OF ESTIMATIONS AND TEST RESULTS To demonstrate whether a model or equation is rightly selected, the estimation results in the empirical analysis are presented together with a number of statistics. The output of estimations and test results contain the following statistics: Parameter values, standard error of estimate and corresponding t-values; Coefficient of determination (R2 ), Adjusted R2 , Durbin-Watson statistic (DW), and S. E. of regression. Results were estimated with Ordinary Least Squares (OLS), using the statistical package EViews 4.0.
88
H. E. HARALAMBIDES ET AL.
4.1. Diagnostic Tests In order to test whether inferences can be made from the model, the following tests were performed: White and ARCH tests for heteroscedasticity; Ljung-Box Q Statistic and Breusch-Godfrey tests for autocorrelation; The results are summarised in Table 4 for newbuilding and Table 5 for secondhand ship prices. Neither autocorrelation nor Heteroscedasticity exist. The OLS estimator is thus BLUE (Best Linear Unbiased Estimator) and consequently statistical inferences can be made from it.
4.2. Model Results The results for all ship types, as well as reports on the standard errors, t-statistics, R2 , adjusted R2 and the Durbin Watson statistic are summarised in Tables 6 and 7 for newbuilding and secondhand vessels respectively. In the newbuilding market, in all models, R2 is high and ranges from 0.74 (VLCCs) to 0.92 (Suezmax tankers) with the rest ranging between 0.80 and 0.83. The fit of all equations is good as indicated by the SE of regressions, ranging from 0.065 to 0.09. The same apply more or less to the secondhand market where, in all models, R2 is high and ranges from 0.74 (VLCCs) to 0.95 (Panamax tankers) with the rest ranging between 0.75 and 0.90. The fit of all equations is good as indicated by the SE of regressions ranging from 0.05 to 0.27. 4.2.1. Newbuilding Price Results: Tankers Shipbuilding costs were found to have a significant effect for all ship types both in the long and the short run. For example, in the short run, a 10% increase in shipbuilding costs will make newbuilding prices rise by 3.5% in the Handy sector, 5.2% in the Suezmax and 4.3% in the VLCC sector respectively. This was expected since shipbuilders usually filter any increase in shipbuilding costs into the final product, with a positive impact on the vessel’s price. For Suezmaxes in particular, this important influence of shipbuilding costs on newbuilding prices can be attributed to the special nature of such vessels. It should be understood that Suezmax tankers are not natural products for shipyards to build, as the demand for them is relatively limited. This is because the footprint of a Suezmax in a building dock is not much smaller than a VLCC and this offers shipbuilders little opportunity to maximise output (better to build a VLCC or two aframaxes than one Suezmax).
Ship Type
Correlogram Correlogram Q Statistic Squared (1 Lag) Residuals (1 Lag)
Serial Correlation LM Test (2 Lags)
ARCH LM Test (1 Lag)
White Heteroscedasticity Test (No Cross Terms)
F Statistic
Obs × R2
Q Statistic
Obs × R2
F Statistic
Obs × R2
Handy size bulk carrier
0.2426 (0.622)
1.9006 (0.168)
0.781390 (0.470631)
2.216435 (0.330147)
1.711335 (0.201087)
1.727420 (0.188741)
0.717432 (0.750591)
19.23374 (0.570148)
Panamax bulk carrier
0.00005 (0.999)
0.1506 (0.696)
0.000262 (0.999739)
0.000797 (0.999602)
0.130371 (0.720664)
0.138739 (0.709538)
1.091025 (0.456667)
21.275 (0.381108)
Cape size bulk carrier
0.0832 (0.773)
0.5062 (0.477)
0.35495 (0.965185)
0.107812 (0.947521)
0.452745 (0.506361)
0.476529 (0.489999)
0.521479 (0.893942)
19.19617 (0.689781)
Handy tanker
0.3185 (0.573)
0.1115 (0.738)
0.453243 (0.643025)
1.417454 (0.492270)
0.090337 (0.766234)
0.097213 (0.7552)
2.924185 (0.118161)
25.98073 (0.25251)
Panamax tanker
0.7923 (0.373)
0.7274 (0.394)
0.735229 (0.491912)
2.054624 (0.357968)
0.664726 (0.422030)
0.696810 (0.403858)
1.536093 (0.258405)
23.20274 (0.278950)
Aframax tanker
0.7832 (0.376)
0.1367 (0.712)
0.630205 (0.543829)
1.963214 (0.374708)
0.119120 (0.732665)
0.127382 (0.721162)
0.607255 (0.820391)
20.98507 (0.581999)
Suezmax tanker
0.14 (0.706)
1.3674 (0.242)
0.681836 (0.518279)
2.112726 (0.347718)
1.190146 (0.284940)
1.224337 (0.268511)
0.336145 (0.973285)
16.89129 (0.814552)
VLCC
0.0620 (0.803)
2.2510 (0.134)
0.044835 (0.956267)
0.152165 (0.926740)
2.055124 (0.163612)
2.051086 (0.152098)
0.932875 (0.598497)
23.51923 (0.430816)
Econometric Modelling of Newbuilding and Secondhand Ship Prices
Table 4. Diagnostic Tests Newbuilding Prices (p-Values in Parentheses).
89
90
Table 5. Diagnostic Tests Secondhand Ship Prices (p-Values in Parentheses). Ship Type
Correlogram Correlogram Q Statistic Squared (1 Lag) Residuals (1 Lag)
Serial Correlation LM Test (2 Lags)
ARCH LM Test (1 Lag)
White Heteroscedasticity Test (No Cross Terms)
F Statistic
Obs × R2
F Statistic
Obs × R2
F Statistic
Obs × R2
0.0046 (0.946)
0.2956 (0.587)
0.795994 (0.464915)
2.211914 (0.330894)
0.249190 (0.621689)
0.265201 (0.606569)
0.999066 (0.493361)
13.44134 (0.414328)
Panamax bulk carrier
0.0649 (0.799)
0.2446 (0.621)
0.063727 (0.938450)
0.189972 (0.909386)
0.203408 (0.655585)
0.216842 (0.641457)
1.532178 (0.215615)
18.64333 (0.230354)
Cape size bulk carrier
0.0005 (0.981)
0.9625 (0.327)
0.568132 (0.576419)
1.662572 (0.435489)
0.915287 (0.347876)
0.953597 (0.328805)
1.036122 (0.495028)
17.86024 (0.397709)
Handy tanker
0.1720 (0.678)
2.9409 (0.086)
0.305421 (0.740556)
0.919013 (0.631595)
2.955675 (0.097937)
2.854634 (0.091111)
1.002426 (0.531027)
19.71784 (0.411733)
Panamax tanker
0.4579 (0.499)
2.7722 (0.096)
1.098071 (0.393211)
4.789421 (0.091199)
2.477668 (0.127565)
2.436109 (0.118570)
2.374978 (0.092319)
24.17779 (0.189421)
Aframax tanker
0.4222 (0.516)
0.7316 (0.392)
2.018679 (0.159013)
5.038853 (0.080506)
0.744014 (0.395975)
0.777696 (0.377847)
2.919298 (0.307011)
24.80702 (0.130325)
Suezmax tanker
0.0227 (0.88)
0.6488 (0.421)
1.417491 (0.268111)
4.082050 (0.129896)
0.576974 (0.454082)
0.606747 (0.436015)
2.212747 (0.099358)
24.23545 (0.187275)
VLCC
0.2210 (0.638)
1.0072 (0.316)
1.303796 (0.295942)
3.796065 (0.149863)
0.869009 (0.359491)
0.904275 (0.341638)
1.345220 (0.322490)
21.56336 (0.306523)
H. E. HARALAMBIDES ET AL.
Handy size bulk carrier
Econometric Modelling of Newbuilding and Secondhand Ship Prices
91
Table 6. Regression Results OLS Newbuilding Prices (Series in Logarithms). Coefficient
Std. Error
t-Statistic
Handy size bulk carrier C(1) D(LnSH) D(Lnfr) D(LnO/F) D(Xrate) D(C) DUMMY LnNB(−1) LnC(−1) R2 S. E. of regression Durbin-Watson stat
−2.808639 0.021486 0.446005 −0.040593 −0.260451 −0.044245 −0.206927 −0.834990 0.251178 0.832633 0.094357 1.827539
0.604254 0.094128 0.100715 0.040932 0.163778 0.129424 0.092127 0.131455 0.095638 Adjusted R2 Sum squared resid
−4.648109 0.228261 4.428381 −0.991697 −1.590267 −0.341864 −2.246102 −6.351907 2.626342 0.774418 0.204775
Capesize bulk carrier C(1) D(LnSH) D(Lnfr) D(LnO/F) D(Xrate) D(C) DUMMY LnNB(−1) Lnfr(−1) R2 S. E. of regression Durbin-Watson stat
−0.522854 0.212045 0.095844 −0.002659 0.049841 0.225067 −0.263857 −0.401815 0.524845 0.803482 0.089946 2.020580
0.642140 0.104588 0.096035 0.061606 0.170594 0.092136 0.046083 0.141933 0.160177 Adjusted R2 Sum squared resid
−0.814236 2.027426 0.998013 −0.043160 0.292159 2.442787 −5.725656 −2.831020 3.276659 0.735129 0.186077
Panamax bulk carrier C(1) D(LnSH) D(Lnfr) D(LnO/F) D(Xrate) D(C) LnNB(−1) LnC(−1) LnSH(−1) R2 S. E. of regression Durbin-Watson stat
−0.184712 0.284819 0.064559 −0.020365 −0.120946 0.171702 −0.777740 0.221273 0.487117 0.794473 0.094050 1.911060
0.644875 0.095207 0.117818 0.037597 0.161740 0.137659 0.186959 0.103414 0.115334 Adjusted R2 Sum squared resid
−0.286431 2.991587 0.547953 −0.541673 −0.747781 1.247298 −4.159938 2.139683 4.223532 0.722986 0.203444
Handy size tanker C(1) D(LnSH)
−1.856030 0.368502
1.131004 0.080124
−1.641046 4.599132
92
H. E. HARALAMBIDES ET AL.
Table 6. (Continued ) Coefficient
Std. Error
t-Statistic
D(Lnfr) D(LnO/F) D(Xrate) D(C) LnNB(−1) LnSH(−1) Lnfr(−1) R2 S. E. of regression Durbin-Watson stat
0.116709 0.139870 −0.257478 0.354045 −0.593587 0.598007 0.515879 0.809683 0.094510 2.166437
0.064895 0.044514 0.145204 0.095495 0.142780 0.154623 0.222441 Adjusted R2 Sum squared resid
1.798416 3.142196 −1.773214 3.707486 −4.157349 3.867529 2.319176 0.729549 0.169710
Panamax tankers C(1) D(LnSH) D(Lnfr) D(LnO/F) D(Xrate) D(C) LnNB(−1) LnSH(−1) R2 S. E. of regression Durbin-Watson stat
0.331578 0.624289 0.002889 0.007224 0.075520 0.275474 −0.397795 0.813074 0.818590 0.084724 1.534792
0.422290 0.110170 0.060050 0.036660 0.135318 0.103914 0.174050 0.203357 Adjusted R2 Sum squared resid
0.785190 5.666613 0.048104 0.197053 0.558094 2.650982 −2.285522 3.998264 0.760868 0.157920
Suezmax tankers C(1) D(LnSH) D(Lnfr) D(LnO/F) D(C) D(Xrate) LnNB(−1) LNC(−1) Lnfr(−1) DUMMY R2 S. E. of regression Durbin-Watson stat
−3.449089 0.140219 0.017473 −0.072858 0.521903 0.146040 −0.634332 1.091599 0.282129 0.261546 0.914862 0.065387 2.122181
0.427146 0.061916 0.043303 0.029660 0.112924 0.106720 0.077873 0.101810 0.040285 0.043830 Adjusted R2 Sum squared resid
−8.074734 2.264645 0.403512 −2.456453 4.621711 1.368436 −8.145773 10.72192 7.003325 5.967258 0.876549 0.085510
Aframax tankers C(1) D(LnSH) D(Lnfr) D(LnO/F) D(Xrate)
−0.333502 0.296938 −0.044777 −0.100502 −0.000544
0.425629 0.078684 0.066989 0.035774 0.107772
−0.783551 3.773803 −0.668416 −2.809388 −0.005048
Econometric Modelling of Newbuilding and Secondhand Ship Prices
93
Table 6. (Continued ) Coefficient
Std. Error
t-Statistic
D(C) C(7) LnNB(−1) LnSH(−1) LNC(−1) R2 S. E. of regression Durbin-Watson stat
0.077936 0.184053 −0.569411 0.517911 0.423981 0.804544 0.067945 2.301108
0.127140 0.049295 0.098593 0.075431 0.156109 Adjusted R2 Sum squared resid
0.612996 3.733671 −5.775393 6.866032 2.715935 0.716588 0.092330
VLCC tankers C(1) D(LnSH) D(Lnfr) D(LnO/F) D(Xrate) D(C) DUMMY LnNB(−1) Lnfr(−1) LNC(−1) R2 S. E. of regression Durbin-Watson stat
−2.581998 0.006097 0.028831 −0.095021 0.184547 0.430361 0.139694 −0.614874 0.140489 1.180380 0.743925 0.095327 2.068940
0.583819 0.056031 0.078318 0.047174 0.155258 0.145115 0.067486 0.112858 0.055780 0.155027 Adjusted R2 Sum squared resid
−4.422598 0.108818 0.368131 −2.014270 1.188644 2.965661 2.069959 −5.448218 2.518628 7.614031 0.622627 0.172659
Consequently, Suezmax prices have been closely pegged to VLCCs, albeit with a discount not proportional to size. In other words, Suezmaxes are rarely bargain vessels. Newbuilding prices of Suezmaxes could therefore be more closely tied to shipbuilding costs than other tanker sizes. Orderbook fluctuations, used as a proxy for the effects of shipyard capacity on newbuilding prices, were found statistically significant for all tankers except Panamaxes. The results indicate the important role shipyard capacity plays in the determination of newbuilding prices. Exchange rate fluctuations were found not to have an impact on newbuilding prices. However, the fact that shipbuilding costs were found to effect newbuilding prices significantly, along with the fact that most of shipbuilding materials are purchased by the shipyard in local currency but quoted to the buyer in dollars, shows that exchange rate fluctuations still have an important, although less direct and visible at first sight effect on the determination of newbuilding vessel prices. Freight rate fluctuations were found not to affect newbuilding prices in the short run but in the cases of VLCCs, Suezmax and Handy tankers they had a long run effect.
94
H. E. HARALAMBIDES ET AL.
Table 7. OLS Regression Results for Secondhand Ship Prices (Series in Logarithms). Coefficient
Std. Error
t-Statistic
Handy size bulk carrier C(1) D(Lnfr) D(LnNB) D(LnLIBOR) D(LnO/F) DUMMY LnSH(−1) LnLIBOR(−1) R2 S. E. of regression Durbin-Watson stat
−4.383875 0.808928 −0.337381 −0.064260 −0.118613 −0.574985 −0.586247 −0.199993 0.897470 0.103370 1.957645
1.052590 0.161227 0.240349 0.052922 0.091695 0.086129 0.075222 0.092625 Adjusted R2 Sum squared resid
−4.164846 5.017311 −1.403715 −1.214236 −1.293556 −6.675859 −7.793610 −2.159171 0.864847 0.235076
Capesize Bulk Carrier C(1) D(Lnfr) D(LnNB) D(LnLIBOR) D(LnO/F) DUMMY LnSH(−1) LnLIBOR(−1) R2 S. E. of regression Durbin-Watson stat
−2.766969 0.633842 0.569612 −0.451611 −0.125750 −0.358353 −0.737501 −0.533302 0.859445 0.167717 1.857334
0.855025 0.109445 0.240580 0.177127 0.077173 0.197331 0.121821 0.125016 Adjusted R2 Sum squared resid
−3.236127 5.791444 2.367659 −2.549646 −1.629459 −1.816001 −6.053974 −4.265862 0.810251 0.562577
Panamax bulk carrier C(1) D(Lnfr) D(LnNB) D(LnLIBOR) D(LnO/F) DUMMY LnSH(−1) LnLIBOR(−1) R2 S. E. of regression Durbin-Watson stat
−2.411518 0.432454 0.573533 −0.093507 0.051508 −0.230754 −1.031324 −0.227950 0.814031 0.167608 1.893576
0.879041 0.139062 0.201378 0.156492 0.063284 0.107801 0.128787 0.090016 Adjusted R2 Sum squared resid
−2.743353 3.109796 2.848047 −0.597517 0.813912 −2.140561 −8.007962 −2.532328 0.754860 0.618034
Handy size tanker C(1) D(Lnfr) D(LnNB) D(LnLIBOR)
0.013861 0.521381 0.591285 −0.032265
0.267601 0.148215 0.226860 0.127800
0.051798 3.517745 2.606383 −0.252465
Econometric Modelling of Newbuilding and Secondhand Ship Prices
95
Table 7. (Continued ) Coefficient
Std. Error
t-Statistic
D(LnO/F) DUMMY LnSH(−1) LnNB(−1) R2 S. E. of regression Durbin-Watson stat
0.016001 −0.309288 −0.731054 0.841021 0.767409 0.134678 2.015567
0.081606 0.123306 0.197649 0.115496 Adjusted R2 Sum squared resid
0.196081 −2.508300 −3.698758 7.281828 0.686003 0.362766
Panamax tankers C(1) D(Lnfr) D(LnNB) D(LnLIBOR) D(LnO/F) DUMMY LnSH(−1) LnNB(−1) LnO/F(−1) Lnfr(−1) R2 S. E. of regression Durbin-Watson stat
−1.160923 0.141960 0.795719 0.115854 0.002469 −0.114689 −0.602795 0.793316 −0.303451 0.312472 0.948807 0.054035 1.854913
0.252129 0.036129 0.096020 0.048667 0.030804 0.056577 0.118270 0.127931 0.065935 0.088355 Adjusted R2 Sum squared resid
−4.604475 3.929295 8.286981 2.380560 0.080145 −2.027143 −5.096783 6.201142 −4.602282 3.536549 0.924557 0.055476
Suezmax tankers C(1) D(Lnfr) D(LnNB) D(LnLIBOR) D(LnO/F) DUMMY LnSH(−1) LnLIBOR(−1) LnNB(−1) LnO/F(−1) R2 S. E. of regression Durbin-Watson stat
0.845594 0.302893 0.845547 −0.251877 0.242642 −0.106457 −0.598061 −0.620879 0.721973 −0.140932 0.809332 0.159511 2.051026
0.613104 0.100394 0.224933 0.157085 0.071348 0.079235 0.143935 0.246056 0.210248 0.065640 Adjusted R2 Sum squared resid
1.379201 3.017048 3.759098 −1.603441 3.400801 −1.343554 −4.155066 −2.523328 3.433908 −2.147035 0.723532 0.508877
Aframax tankers C(1) D(Lnfr) D(LnNB) D(LnLIBOR) D(LnO/F) LnSH(−1)
−5.540092 0.574838 0.484290 0.067353 0.133124 −1.129018
1.485586 0.194434 0.373448 0.078066 0.084527 0.187182
−3.729230 2.956466 1.296809 0.862765 1.574930 −6.031655
96
H. E. HARALAMBIDES ET AL.
Table 7. (Continued ) Coefficient
Std. Error
t-Statistic
Lnfr(−1) LnNB(−1) R2 S. E. of regression Durbin-Watson stat
0.581171 0.710068 0.817413 0.144165 1.623224
0.152742 0.202969 Adjusted R2 Sum squared resid
3.804914 3.498409 0.759317 0.457240
VLCC tankers C(1) D(Lnfr) D(LnNB) D(LnLIBOR) D(LnO/F) DUMMY LnSH(−1) LnNB(−1) LnO/F(−1) Lnfr(−1) R2 S. E. of regression Durbin-Watson stat
−6.132013 0.727271 1.978095 0.103148 0.127926 −0.330624 −1.174948 0.920653 −0.216395 0.450924 0.739839 0.274820 2.140796
1.964714 0.240097 0.649283 0.247917 0.133934 0.156736 0.219023 0.207527 0.065320 0.225092 Adjusted R2 Sum squared resid
−3.121071 3.029072 3.046587 0.416060 0.955144 −2.109439 −5.364502 4.436313 −3.312852 2.003293 0.622767 1.510525
The dummy variables were found to be significant in large tankers. However, the dummy variable exercises higher influence in the Suezmax and VLCC segments. The reason is that the dummies included in the model are related to political events affecting the Middle East in particular, where most loadings for these types of ships take place. 4.2.2. Newbuilding Price Results: Bulk Carriers For bulk carriers, like tankers, shipbuilding costs were found to be statistically significant for all ship types. However, the orderbook as a percentage of the fleet and the exchange rate fluctuations were found to have no significant effect on the price of a new vessels. Freight rate fluctuations seem to affect Handy newbuilding prices in the short run and Cape size ones in the long run. For example, a 10% increase in freight rates for Handy Bulk carriers will make handy newbuilding prices increase by approximately 4.5%. However, newbuilding prices for Panamax bulk carriers do not seem to be affected by this variable either in the short or the long run. 4.2.3. Overall Results: Newbuildings Overall, shipbuilding costs were found to have the most significant and extensive effect on the determination of newbuilding prices for all ship types. This result
Econometric Modelling of Newbuilding and Secondhand Ship Prices
97
is in line with what was expected since, traditionally, shipbuilding has shifted to countries with a comparative cost advantage. Exception to this were the handy size bulk carriers where freight rate fluctuations were seen to have the most pervasive effect on newbuilding prices in the long-run. As mentioned above, being cheaper to build, the construction of a bulk carrier is a shipyard’s last resort in its strive to maximise revenue. New contract prices for dry bulk carriers may therefore be driven by the demand and price of alternative vessels like tankers. The timing of the dry bulk carrier order is more crucial than in tankers and it often takes place when demand for, and prices of, new tankers have fallen. The O/F variable was found statistically significant for tankers but not for bulk carriers. This result merits further investigation. In principle, in a “tight” shipbuilding market, the price of the less preferred types of ships (bulk carriers) should be relatively higher due to high opportunity costs. The answer could be in the anti-cyclical investment behaviour of bulk carrier owners, in no small measure attributable to their asset-play strategies, vis a vis tanker owners who are generally more trade driven. Exchange rate fluctuations were found to be insignificant for all ship types, both in the long and the short run. Normally, as argued above, the weakening of a shipbuilding country’s currency ought to lead to lower dollar prices of new ships for export. At the same time shipbuilding, particularly in Korea and China nowadays, has a very high import content (raw materials, machinery and equipment) and a weak currency would make such imports more expensive. The overall effect could thus only be established through the construction of a trade model, something outside the scope of this chapter. At this point it would suffice to say that the impact of exchange rates on newbuilding prices is captured by other variables in the model such as shipbuilding costs and shipyard capacity. Timecharter rates were found to be statistically significant in the determination of newbuilding prices in handy and Cape Size bulk carriers and in Handy, Suezmax and VLCC tankers. The fact that timecharter rates were not found to be significant for Panamax bulk carriers and Panamax and Aframax tankers may be attributed to the forward nature of newbuilding ordering. This implies that demand for new vessels reflects an anticipated future trading environment rather than present market conditions. There have been many instances however where increased ordering has been witnessed in anticipation of improved future market conditions that were not realised. There have also been times when rising freight rates have not prompted new ordering. This may be the reason why, in the aforementioned markets, newbuilding prices do not seem to be driven by either freight rates or timecharter rates. As far as secondhand prices are concerned, our results indicate that they have a significant effect on the newbuilding price of all ship types except VLCCs and
98
H. E. HARALAMBIDES ET AL.
Handy size bulk carriers. This means that, apart from being predominantly cost driven, newbuilding prices, for some ship types, may also be driven, to a certain extent, by asset pricing and speculation. 4.2.4. Secondhand Price Results: Tankers For secondhand ship prices, timecharter rates were found statistically significant in all market segments in the short run. The effect of a 10% increase in timecharter rates ranges from a 2% (Panamax tankers) to a 7.5% (VLCCs) increase in the price of secondhand vessels, with the rest lying between 3.5 and 6%. Newbuilding prices were found significant in all market segments both in the long- and the short run. Short run effects on secondhand values range from an increase of 6% (Handy tankers) to almost 20% (VLCCs), with the rest of the secondhand values rising between 8 and 10%, for a 10% increase in newbuilding prices. The only exception was the Aframax tanker market. The Aframax tanker is regarded as the workhorse of the industry and as such it relies too much on timecharter rates. Therefore, newbuilding price changes are only significant in the long run. Also, changes in O/F are significant in the long- rather than the short run, with negative effects on the secondhand values of VLCCs, Suezmax and Panamax tankers respectively. This tends to confirm the frequently observed “nervousness” of shipowners in front of an inflated orderbook and its impact on future freight rates and profitability. The cost of capital was found to be insignificant in all tanker segments apart from the Suezmax market. Finally, the dummy variable included to investigate the effects of the oil crises in 1973 and 1979, and the Tap Line closure in 1971, was found to have significant explanatory power in all markets except Aframax tankers. 4.2.5. Secondhand Price Results: Bulk Carriers Newbuilding prices and timecharter rates are statistically significant in the short run for all types of ships. The only exception is the handy bulk carrier market where newbuilding prices were found to be statistically insignificant. The reason may be that there are not many newbuildings entering the market or on order. Furthermore, handy is the industry’s workhorse and relies heavily on timecharter rates (a 10% increase in timecharter rates would increase secondhand prices by 8%). While, in “handies”, timecharter rates have a higher effect than newbuilding prices, in “capes” they have the same effect (a 5% increase in secondhand values from a 10% increase in either of them). In Panamax bulk carriers, it is the newbuilding variable that has the greatest effect.
Econometric Modelling of Newbuilding and Secondhand Ship Prices
99
In the long run, an increase in the cost of capital will have negative effects for all vessel segments. Orderbook as a percentage of the fleet, O/F, was found statistically insignificant for all vessel types. The dummy used for the Sanko deal (1985, 1986) was found significant for Handies and Panamaxes. 4.2.6. Overall Results: Secondhand Newbuilding prices and timecharter rates have the greatest effect of all variables on the determination of secondhand values, in most cases both in the short- and in the long run. This shows that secondhand ship prices are primarily market driven whereas newbuilding prices are mainly cost driven. The coefficients of the NB variable are higher than those of the fr one. A reason behind this may be asset play. A special case is the Suezmax tanker segment. According to our findings, a 10% increase in the newbuilding price of such a tanker will make a five-year old vessel’s price rise by about 8.5%. This can be attributed to the special nature of such vessels. It should be remembered that Suezmax tankers are not mainstream products for shipyards to build, as the demand for them is relatively limited. This is because the footprint of a Suezmax is not much smaller than a VLCC in a building dock and offers shipbuilders little opportunity to maximise output. Consequently, Suezmax prices have been closely pegged to VLCCs with a discount but not proportional to size. In other words, Suezmaxes are rarely bargain vessels. Secondhand prices of Suezmaxes could therefore be more closely tied to newbuilding prices than other tanker sizes. In the above circumstances secondhand values for bulkers have become closely related to the cost of building a new ship and owners have placed greater emphasis on trading ships as commodities (asset play) than their counterparts in the tanker market. The cost of capital seems to be significant only for bulk carrier owners, and this only in the long run. The only exception is the Suezmax segment due to its particular characteristics described above. What can be implied from this is that shipowners operating in the tanker sector have more capital than their dry bulk colleagues. Therefore, an increase in the cost of borrowing may not affect their investment decisions as much as it would in the case of bulk carrier owners. This argument can be further strengthened by the fact that cashflow and revenues are significantly higher in the tanker market than in the bulk carrier one, as well as by the fact that, traditionally, the world’s largest and richest shipowners have mostly been active in the tanker sector. After all, for many shipping enterprises, particularly for the Greek ones, entering the tanker market from the bulk carrier one is regarded as a step forward, and a sign of maturity and success.
100
H. E. HARALAMBIDES ET AL.
Finally, O/F has a negative effect on secondhand prices only in the long run and only in large (Suezmax, VLCCs) and medium size (Panamax) tankers. This may be due to the fact that an already existing large orderbook may make tanker owners reluctant to invest in ships, particularly as expensive as large tankers, since this may be an indication of future oversupply. As a result, demand for secondhand ships falls and so do their prices. The exception is the Aframax segment where, as has already been mentioned, owners place most of their decision to invest in secondhand ships on timecharter fluctuations. Overall, with respect to bulk carriers, the following observations may add up to the above findings: The values of the dry cargo commodities carried by bulkers are far lower than those carried by tankers, such as crude oil and petroleum products. This tends to put greater pressure on dry bulk carriers to provide value for money transportation given that transportation costs form a higher proportion of the landed commodity prices in bulkers than in tankers. This tends to limit the upside potential of dry bulk carrier freight rates more than tankers. Dry bulk shipping is therefore more cost driven than revenue driven, compared to tanker shipping. The performance of the dry bulk carriers, both in terms of absolute earnings and return on investment, has traditionally been much lower than that of tankers. This is one of the reasons why combined carriers have on average traded predominantly in oil rather than dry cargo (sometimes the ratio has reached 90% wet 10% dry). Given the above, dry bulk carrier owners have tended to be far more readily attracted to low newbuilding prices than tanker owners, as the cost of the vessel is more crucial than return on investment. Thus, dry bulk carrier orders often react much quicker to lower contract prices, especially at times when orderbooks are much lower than available shipbuilding capacity. Being cheaper to build, bulk carriers are not the first choice of shipyards where both tankers and bulkers are constructed. This is because a shipyard needs to maximise value from available space. The demand and price of alternative vessels, like tankers, may therefore drive new contract prices of dry bulk carriers. The timing of the dry bulk carrier order is more crucial than in tankers and it often takes place when demand for new tankers has fallen, together with shipbuilding prices.
5. FORECASTS Based on the models estimated above, forecasts were made for all ship types and for the period 1999–2001. In addition to this, an atheoretical Autoregressive (AR)
Handy B/C
Panamax B/C
Cape B/C
Handy Tanker
Panamax Tanker
Aframax Tanker
Suezmax Tanker
VLCC Tanker
SEM Absolute errors Percentage errors Mean squared errors
0.025605 0.958693 0.028030
0.073022 2.388894 0.074553
0.003727 0.107708 0.003789
0.099145 3.020109 0.122884
0.149928 4.284879 0.169511
0.095854 2.606928 0.095881
0.089577 2.277221 0.102784
0.062812 1.444907 0.077671
AR Absolute errors Percentage errors Mean squared errors
0.035573 1.331820 0.038764
0.049134 1.618908 0.059389
0.069687 2.001372 0.085654
0.066711 1.975353 0.088875
0.072636 2.077007 0.073338
0.045158 1.228126 0.045198
0.047529 1.200931 0.065574
0.068900 1.582485 0.090157
Econometric Modelling of Newbuilding and Secondhand Ship Prices
Table 8. ECM-AR Forecast Comparison for 2000–2001 Newbuilding Prices.
101
102
Table 9. SEM-AR Forecast Comparison for 2000–2001 Second-Hand Ship Prices. Panamax B/C
Cape B/C
Handy Tanker
Panamax Tanker
Aframax Tanker
Suezmax Tanker
VLCC Tanker
SEM Absolute errors Percentage errors Mean squared errors
0.166690 6.756877 0.185686
0.082117 3.025542 0.083020
0.149612 4.711866 0.152600
0.219726 7.710057 0.255889
0.187533 5.500556 0.224987
0.061618 1.755294 0.064628
0.245304 6.491360 0.245511
0.170031 4.078291 0.170868
AR Absolute errors Percentage errors Mean squared errors
0.123028 5.007458 0.125500
0.047807 1.803812 0.061935
0.080333 2.093071 0.067665
0.123384 4.273113 0.128692
0.268577 7.944779 0.283153
0.363906 10.22644 0.375982
0.283709 7.440459 0.299382
0.317071 7.586094 0.323485
H. E. HARALAMBIDES ET AL.
Handy B/C
Econometric Modelling of Newbuilding and Secondhand Ship Prices
103
model was estimated for all ship types and forecasts were made for the same period. The Bayesian Information criterion was used in all cases in order to determine the lag of the model, which in this case was 1. Therefore, our Auto Regressive model has the following form: X t+1 = + X t + t+1 Where Xt = P(t) = lnSH(t) or lnNB(t) depending on the model under investigation, is the mean, (t) denotes the shock at time t, and is the autoregressive coefficient. Tables 5 and 8 compare the forecasting performance of the two different methods for newbuilding and secondhand ship prices respectively. The absolute, percentage, and root mean squared errors are used to compare the performance of the forecasts of the two models. For newbuilding prices, AR forecasts outperform those of ECM in five out of eight cases namely for the Panamax bulk carriers, as well as for the Handy, Panamax, Aframax and Suezmax tankers. The error correction model outperforms AR estimates for Handy and Cape size bulk carriers and VLCC tankers. For secondhand ship prices, it can be seen that AR forecasts outperform those of the Error Correction Model (ECM) in all three bulk carrier segments as well as in the Handy Tankers sector. The ECM outperforms AR estimates for Panamax, Aframax, Suezmax and VLCC tankers. These results imply that the theoretical Error Correction Model is still to be preferred if one wants to achieve the classical objectives of Econometric Business Cycle Research (EBCR) simultaneously, which, according to Tinbergen (1959), are to describe and forecast cycles and to evaluate policies. However, if not all goals have to be met with a single vehicle, other methods might serve the purpose equally well or even better as is the case with the Auto Regressive models whose forecasts outperform those of ECM in most cases (Table 9).
6. CONCLUSIONS Despite remarkable past efforts to model the behaviour of new and secondhand ship markets, few studies have so far attempted to develop a model based on economic theory. It is hoped that the models presented in this chapter have thus contributed in this lack in literature. Newbuilding prices and timecharter rates have the greatest effect of all variables on the determination of secondhand ship prices, in most cases both in the shortand long run. The cost of capital is only significant for bulk carrier owners. The only exception is the Suezmax segment due to its particular characteristics. The
104
H. E. HARALAMBIDES ET AL.
Orderbook (as a percentage of the fleet) has a negative effect on the price of secondhand vessels only in the long run and only in large and Panamax tankers. For newbuilding prices, shipbuilding costs are found to have the most significant effect for all ship types. Timecharter rates have an effect only on a few ship segments. This is in line with theory that newbuilding prices are cost driven, rather than market driven, as secondhand ship prices are. It is also found that actual exchange rates do not affect shipbuilding prices, but cost variations, due to exchange rate fluctuations, do. Orderbook (as a percentage of the fleet), used as a proxy for shipyard capacity, is found significant only for tankers, indicating that shipyards’ expansion policy is aimed at high value ships like tankers rather than bulk carriers. Finally, newbuilding prices for some ship types may be driven, to a certain extent, by asset pricing and speculation.
REFERENCES Beenstock, M. (1985). A theory of ship prices. Maritime Policy and Management, 12, 215–225. Beenstock, M., & Vergottis, A. (1989a). An econometric model of the world shipping market for dry cargo, freight and shipping. Applied Economics, 21, 339–356. Beenstock, M., & Vergottis, A. (1989b). An econometric model of the world tanker market. Journal of Transport Economics and Policy, 23, 263–280. Beenstock, M., & Vergottis, A. (1992). The interdependence between the dry cargo and tanker markets. Logistics and Transportation Review, 29(1), 3–38. Beenstock, M., & Vergottis, A. (1993). Econometric modeling of world shipping. London: Chapman & Hall. Brooks, C. (2001). Introductory econometrics for finance. Cambridge, UK: Cambridge University Press. Charemza, W., & Gronicki, M. (1981). An econometric model of world shipping and shipbuilding. Maritime Policy and Management, 8, 21–30. Dickey, D. A., & Fuller, W. A. (1979). Distribution of estimators for time series regressions with a unit root. Journal of the American Statistical Association, 74, 427–431. Dickey, D. A., & Fuller, W. A. (1981). Likelihood ratio statistics for autoregressive time series with a unit root. Econometrica, 49(4), 1057–1072. Drewry (2001). The European and worldwide shipbuilding market: An economic analysis on the comparative strengths and weaknesses of EU and Korean shipyards. London. EViews (2002). EViews 4.0 users’ guide. Irvine California, USA. Glen, D. R. (1990). The emergence of differentiation in the oil tanker market, 1970–1978. Maritime Policy and Management, 17(4), 289–312. Glen, D. R. (1997). The market for second-hand ships: Further results on efficiency using cointegration analysis. Maritime Policy and Management, 24, 245–260. Glen, D. R., & Martin, B. T. (1998). Conditional modelling of tanker market risk using route specific freight rates. Maritime Policy and Management, 25, 117–128. Hale, C., & Vanangs, A. (1992). The market for second hand ships: Some results on efficiency using cointegration. Maritime Policy and Management, 19(19), 31–40.
Econometric Modelling of Newbuilding and Secondhand Ship Prices
105
Hawdon, D. (1978). Tanker freight rates in the short and long run. Applied Economics, 10, 203–217. Jin, D. (1993). Supply and demand of new oil tankers. Maritime Policy and Management. Johansen, S. (1991). Estimation and hypothesis testing of cointegration vectors in gaussian vector autoregression models. Econometrica, 59, 1551–1580. Kavussanos, M. G. (1996). Price risk modelling of different size vessels in the tanker industry. Logistics and Transportation Review, 32, 161–176. Kavussanos, M. G. (1997). The dynamics of time-varying volatilities in different size second-hand ship prices of the dry-cargo sector. Applied Economics, 29, 433–443. Kavussanos, M. G., & Alizadeh, A. M. (2002). The expectations hypothesis of the term structure and risk premiums in dry bulk shipping freight markets. Journal of Transport Economics and Policy, 36(2), 267–304. Koopmans, T. C. (1939). Tanker freight rates and tankship building, an analysis of cyclical fluctuations. Netherlands Economic Institute Report No. 27. Haarlem, Holland. Strandenes, S. R. (1984). Price determination in the timecharter and second hand markets. Discussion Chapter 0584. Norwegian School of Economics and Business Administration, Bergen, Norway. Strandenes, S. R. (1986). Norship: A simulation model of markets in bulk shipping. Discussion Chapter 11. Norwegian School of Economics and Business Administration, Bergen, Norway. Tinbergen, J. (1931). Een Schiffbauzyclus? Weltirtschafliches Archiv, 34, 152–164. Tinbergen, J. (1959). Selected chapters. L. H. Klassen (Ed.), North Holland, Amsterdam. Veenstra, A. W. (1999). Quantitative analysis of shipping markets. T99/3, TRAIL Thesis Series. The Netherlands: Delft University Press. Volk, B. (1994). The shipbuilding cycle – A phenomenon explained. Bremen: Institute of Shipping Economics and Logistics.
4.
CROSS-INDUSTRY COMPARISONS OF THE BEHAVIOUR OF STOCK RETURNS IN SHIPPING, TRANSPORTATION AND OTHER INDUSTRIES
Manolis G. Kavussanos and Stelios N. Marcoulis 1. INTRODUCTION AND AIM OF THE PAPER The aim of this paper is to review the empirical work in the area of shipping finance, which deals with the companies in the shipping sector that have taken the decision to resort to the public capital markets in order to finance their activities. More specifically, the paper is concerned with the performance of listed companies in stock exchanges around the world. The pricing of stocks in the financial markets is a result of the collective action of investors analysing the stocks and taking action while pursuing profit making opportunities. If markets work efficiently, the characteristics of stocks as determined in these markets should reflect past, present and future prospects of the stocks and their sector, and must be the best indicator upon which to base decisions. To put things in perspective, the history of the shipping industry is inextricably linked with the world economy and its economic and technological development. As Adam Smith (1776) notes, “shipping is one of the major catalysts of economic development. . . . shipping is a cheap source of transport which can open up wider Shipping Economics Research in Transportation Economics, Volume 12, 107–142 Copyright © 2005 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0739-8859/doi:10.1016/S0739-8859(04)12004-0
107
108
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
markets to specialisation, offering shipment of even the most everyday products at prices far below those that can be achieved by any other means.” Over 95% of world trade in volume terms moves by sea. Over the past years specialization of activities has taken place in shipping transportation itself. Ship design and ship building technology, investment in infrastructure such as ports and port equipment, logistics and warehousing have allowed for that. In discussing the structure of the industry, one should be concerned with the way business is organised to achieve the efficient transport of different kinds of cargo. The major division within the cargo carrying part of the industry is between bulk and liner shipping. The former specialises in the transport of large cargo parcels which can be carried on a one ship one cargo basis. Thus, tramp ships travel on any sea-lane around the world in search of such cargoes to transport. Liner ships, on the other hand, specialise in the transportation of small cargo parcels, usually carried in containers. These ships provide regular services on specific sea-lanes around the world. Other segments of the water transportation industry include passenger vessels, ferries, tugs and other ancillary services, among others. The bulk, liner and other segments of the water transportation industry have totally different approaches as far as the type of organisations involved and the respective shipping policies are concerned. For example, liner operators need to organise the transport of many different parcels (usually in containers in our days) and need a large shore-based staff, capable of dealing with shippers, handling documentation and planning the ship loading. Hence, due to their high overheads and also the need to maintain a regular service even when a full payload of cargo is not available, the liner business is vulnerable to uneconomic price competition from other shipowners operating on the same trade routes. Contrary to that, the bulk shipping companies operate under the principle of “one ship – one cargo” and hence handle fewer, but much larger cargoes. Therefore, their operations do not require a large shore-based activity. Nevertheless, the few decisions that need to be made by a bulk shipping business are crucial and require the attention of the owner and/or vice-president. Large companies, shipping substantial quantities of bulk materials, often run their own shipping fleets to handle a proportion, or all, of their transport requirements. For example, according to Jacobs (1986), in 1984 the major oil companies collectively owned around 40% of the whole world tanker fleet. However, it should be stressed that industrial conglomerates do not necessarily become shipowners just to optimise the shipping operation. It is also to ensure that basic transport requirements are met at a predictable cost without the need to resort to the charter market.1 Another sector of the industry is cruise shipping. This sector is linked very much to the tourist and leisure industry. Despite utilizing a ship, the service provided is
Cross-Industry Comparisons of the Behaviour of Stock Returns
109
entirely separate from that of the cargo carrying sector; it is leisure and tourism provided on board a ship. The ferry subsector of the shipping industry is also affected by tourism and the movement of passengers, but also by the movement of cargo. Despite its seasonal nature, part of the demand for its services emanate from quite a steady movement of cargo carrying lorries in certain routes. The economic forces driving the demand and supply factors are distinct then. Other shipping linked sectors involve yards and offshore companies, and these form important parts of the total – the world shipping industry. In our days, with the development of the concept of door to door service for products, specialized agents are involved in providing such logistical services. The final customer can stop worrying about arranging transportation by several modes of transport, the handling in between and the security issues involved before getting his hands to the final product. Of course this involves the question of considering other modes of transport in the supply chain process. Again, the large sums of money involved in investment and infrastructure suggest that perhaps the performance of listed companies in other modes of transport are relevant. This is discussed later on as well from the point of view of the investor, who is not necessarily involved in the physical operation of the transportation service, but is interested in investing in portfolios with transportation stocks. A major question then throughout the years has been how to finance investments in this highly capital-intensive industry. Ships cost millions and such large sums need careful investment decisions. Methods of financing have varied over time and place, as well as with the corporate structure of the company requiring funds to invest in shipping. Thus, while traditional borrowing from banks2 has always been prominent in the industry, charter backed finance3 has been very popular in the post second world war period. This has been followed by asset backed finance in the 1980s (e.g. ship funds), and since the 1990s – a lot of interest has been placed in drawing funds from the public. The latter may be materialised either by borrowing through bonds, or by offering part-ownership to the public through shares in the company. This paper concentrates on this last form of finance of the industry. In particular, it concentrates on alternative valuation models of shares of water transportation companies. It examines single index models under which the market index alone is assumed to be driving market returns, and then extends this to multi index models as sets of micro economic and macroeconomic factors are considered as possible additional factors determining security returns. Security valuation models are also considered for stocks of other transportation industries and some other industries. International comparisons are also made by considering valuation issues through the formation of industries at the global level. At this global level, distinction is made in the shipping industry of a number of subsectors, and the characteristics
110
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
of these subsectors are compared between themselves. This investigation enables comparisons of risk-return trade-offs between industries and subsectors of the international shipping industry, a process which takes place by investors when taking practical investment decisions. The availability of listed companies in stock exchanges for each industry allows investors, by observing share price formation, to get a view of the market value of companies in each industry through the interplay of demand and supply forces driven by experts dealing in these stocks. Despite the significance of the water and other transportation sectors and the use of industry indices in investment decisions, there has been limited work in attempting to compare the risk-return performances of these industries: (1) in a capital markets context; and (2) in a comparative cross-industry context using asset pricing models. Equally limited attempts have been made to uncover factors, other than the market, that may influence returns in these industries. This might be due to the fact that since the seventies the sectors which provided superior returns to investors were, according to Jones (1993), the finance sector and, after the crash of 1987, the industrial sector. The exception to this lack of published work in the area is a series of studies by Kavussanos and Marcoulis (1997a, b, c, 1998, 2000a, b), Kavussanos et al. (2002) and Kavussanos et al. (2003). Nevertheless, such industry analysis would be revealing for investors in the water and other transportation industries, portfolio managers and corporate financiers since it would enhance investment decisions, possibly induce investors to place a different share of their investment funds in these industries and also shed some further light as to what drives values in these industries. The significant amount of research produced in recent years on the concept of efficient markets has shaped, to a large extent, the way academics and practitioners think about the stock selection process. More specifically, there are two well-known approaches to analysing and selecting stocks, fundamental and technical analysis. Traditionally, the former approach has occupied the majority of resources devoted to the analysis of common stocks and it is concerned with the valuation of stocks according to fundamentals such as gearing ratios, book to market ratios, size and many more. The other approach mentioned, technical analysis, deals with the search for identifiable and recurring stock price patterns and attempts to exploit market inefficiencies. The research mentioned above deals with fundamental analysis, the study of a stock’s value using basic data at microeconomic and macroeconomic level. Fundamental analysis is based on the concept that any stock has an intrinsic value which is a multidimensional function of the general state of the economy (e.g. industrial production, inflation, oil prices etc.), the market, the structure of the industry the company operates in, and the company’s fundamental microeconomic factors (e.g. capital structure, book-to-market ratio, market capitalization, etc.).
Cross-Industry Comparisons of the Behaviour of Stock Returns
111
This paper is organised as follows. The next section discusses the decision process of the investor when analysing and selecting stocks, and the determinants of stock returns that enter his decision function. Section 3 discusses the decision process by investors when analysing and selecting stocks at the micro and macro economic level. Section 4 extends further this process, outlining why industry analysis makes sense and that it is part of the investors decision rule. Section 5 describes the industry classification systems that may be used to define industrial sectors, such as the Standard Industrial Classification (SIC) system, the Bloomberg and the Morgan Stanley classifications. Section 6 presents the major findings of the research in shipping economics and transportation industries. The final section concludes and discusses the possibilities that this research presents for applications by industry practitioners.
2. THE DECISION PROCESS OF THE INVESTOR IN ANALYSING AND SELECTING STOCKS It is important to understand the way investors analyse and select stocks during their investment decision process. More specifically, it is important to understand that investors perform industry analysis during their investment decision process. This in turn places the question of how the universe of companies can be classified into industries. Kavussanos and Marcoulis (2001), for instance, classify stocks into industries based on the SIC (Standard Industrial Classification) index. An equally important question is the identification of possible factors that drive returns for these stocks. To turn to the last question first, Kavussanos and Marcoulis (2001) deal with fundamental analysis, the study of a stock’s value using basic data at microeconomic and macroeconomic level. Fundamental analysis is based on the concept that any stock has an intrinsic value which is a multidimensional function of the general state of the economy (e.g. industrial production, inflation, oil prices etc.), the market, the structure of the industry the company operates in, and the company’s fundamental microeconomic factors (e.g. capital structure, book-tomarket ratio etc.). It is important to start any analysis dealing with stocks by assessing the state of the economy which explicitly influences investors’ everyday investment decisions. For example, if a recession is likely, or under way, stock prices will be heavily affected (they are likely to drop) at certain times during the contraction. Conversely, if a strong economic expansion is under way, stock prices will be heavily affected (they are likely to rise), again at particular times during the expansion. For instance, in 1997 when the world economy was performing
112
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
exceptionally well, unprecedented rises were fuelled in stock markets all round the world. Having completed an analysis of the economy an investment manager usually performs industry analysis. King (1966) was the first to present evidence that there is an industry factor affecting the variability in stock returns. Individual industries tend to respond to general market movements, but the degree of response can vary significantly due to the fact that industries undergo significant changes over both short and long periods of time. Furthermore, different industries are affected, to different degrees, by economic recessions and expansions. For example, the heavy goods industries will be severely affected in a recession (e.g. the auto and steel industries in the recession of 1981–1982). On the other hand, consumer goods industries might be much less affected during such a contractionary period. Of course the opposite occurs when the economy is growing where income may be spent on investment or consumption goods. According to Begg, Fischer and Dornbusch (1987), when an increase in investment occurs, it raises income by a large amount (the investment multiplier), which in turn may produce an increase in demand for the product (the income accelerator) generating demand for more investment goods, so that the economic system expands rapidly. Eventually, labour and capital become fully utilised and the expansion is sharply halted, throwing the whole process into reverse. During a severe inflationary period, such as the late seventies and early eighties, regulated industries, such as utilities, were severely hurt by their inability to pass along all price increases. Finally, from time to time, there appear to emerge new “hot” industries which enjoy spectacular growth. Examples that come to mind include genetic engineering and synthetic fuels and more recently dot.coms. Once the investment manager has performed economy, market and industry analysis, he has to shift his emphasis on to company analysis. Security analysts focus on a number of factors which are important in analysing a company. These can be divided into two wide categories, qualitative factors and quantitative factors. Qualitative factors focus mainly on the managerial capacity of the company and its future prospects, which are fundamental to its success. Quantitative factors focus mainly on past and present income statements and the balance sheets of a company and include variables such as earnings, price-to-earnings multiples, dividend yields, capital structure, book-to-market ratios, size etc.
3. THE DETERMINANTS OF STOCK RETURNS – GENERAL LITERATURE SURVEY Experience has shown that stock analysis and selection procedure is not an easy task and as such it is not surprising that a substantial part of the financial literature
Cross-Industry Comparisons of the Behaviour of Stock Returns
113
has dealt with the issue i.e. the determinants of stock returns at the microeconomic, company level, and at the macroeconomic level. Numerous academics have followed the example of King (1966) who is believed to be the first to study the determinants of stock returns. His study, utilising statistical methodologies of that time concluded that stock price changes can be expressed in terms of a market, an industry and a company effect. Effectively, what King proposed was that stock prices are shaped and determined by developments at the macroeconomic level, which in turn affect industries and the stock market in general, and by developments at the microeconomic level which affect the company’s fundamentals, hence its value. King’s findings were extremely important and were going to be the basis for a substantial amount of academic research which was to follow. It is one of the aims of this part of the paper to very briefly review the voluminous literature regarding the determinants of stock prices. One of King’s findings, that of the market effect, was to be presented in a more formal way by Sharpe (1964) and Lintner (1965) and was to shape the way academics and practitioners perceived asset returns for a long time to come. The reference point is the Capital Asset Pricing Model (CAPM) which expresses the stock returns of any company as a linear function of just one factor, the return on the market portfolio of assets. The CAPM splits asset risk into two components, market or systematic risk, representing that portion of asset risk related to the riskiness of the market, and residual or non-systematic risk, which is unrelated to market movements. These ideas are summarised in Eq. (1). R jt = R ft + j (R Mt − R ft ) + jt
(1)
. . . where Rjt , Rft , RMt are the returns of company j, the risk-free rate and the market return, respectively, measured over time, t, j is a measure of market risk and jt is the error term which captures residual risk. However, during the late seventies and the early eighties, the discipline of finance and financial economics evolved and as computers became more powerful, a number of theoretical and practical criticisms regarding the validity of the CAPM arose. For example, Roll (1977) criticised the CAPM on the grounds that the composition, let alone the return, of the true market portfolio is not known to the researcher; what became widely known as the Roll Critique of the CAPM. Others, such as Banz (1981), Basu (1977, 1983), Reinganum (1981), Lakonishok and Shapiro (1986), and Fama and French (1992), among others, observed that small firms appear to consistently earn higher returns than big firms, the well known “size effect.” The economic rationale behind the “size effect” has been a puzzle and attempts have been made to explain it by arguing that: (1) returns for small firms have been computed in a way that biases them upwards; (2) the risks on small firms have been understated; and (3) size may act as a proxy for some other economic influence.
114
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
Stattman (1980), Rosenberg, Reid and Lanstein (1985), Fama and French (1992), among others, found that there is a positive relationship between the average stock returns of U.S. stocks and their ratio of book value of equity, BE, to market value of equity, ME. A similar relationship is documented for Japanese stocks by Chan, Hamao and Lakonishok (1991). Companies with a high BE/ME ratio are believed to be “value” stocks, while companies with a low BE/ME ratio are believed to be “growth” stocks. Generally speaking, “growth” stocks are stocks exhibiting rapid increases in earnings and that is why their market value of equity, reflecting their hypothetical excellent prospects, may be significantly higher than their book value of equity. On the other hand, “value” stocks are stocks whose market price seems to be low relative to their net worth. The empirical evidence suggests that “value” stocks seem to outperform “growth” stocks. However, “value” and “growth” stocks may also be categorised by their earnings to price (E/P) ratio, the inverse of the well-known and widely used P/E (Price Earnings) ratio. Relatively low values of this ratio characterise “growth” stocks, while relatively high values characterise “value” stocks. Therefore, an interesting question that arises is whether there is any relationship between stock returns and their E/P ratios. Ball (1978), Reinganum (1981) and Basu (1983) argue that E/P ratios help to explain the cross-section of average returns on U.S. stocks. Another variable which has been the subject of empirical tests regarding the explanation of the cross-section of stock returns is leverage. Bhandari (1988) defined leverage as (Book Value of Total Assets minus Book Value of Equity)/(Market Value of Equity) and found that stock returns are positively related to market leverage. Fama and French (1992) included leverage, among a number of other variables, in an attempt to explain the cross-section of U.S. stock returns and, in line with Bhandari (1988), found a positive relationship between market leverage, defined as the ratio of Total Assets/Market Value of Equity (A/ME) and stock returns. Furthermore, they defined book leverage as the ratio of Total Assets/Book Value of Equity (A/BE) and found a negative relationship between this ratio and stock returns. An excellent discussion of the above determinants or “anomalies” of stock returns can be found in the seminal paper of Fama and French (1992), who employ a multifactor model to estimate the influence of the aforementioned factors on the cross-section of U.S. stock returns. The discussion so far has focused on a number of determinants of stock returns which, with the noticeable exception of the market, share a common characteristic, they all affect the company at the microeconomic level. An equally voluminous amount of literature exists with regards to macroeconomic factors which are also believed to have a role to play in the determination of stock returns. As in the case
Cross-Industry Comparisons of the Behaviour of Stock Returns
115
of the microeconomic factors, it is not possible to review all the papers dealing with macroeconomic factors and stock returns. However, an attempt will be made to review a number of the most important ones so as to obtain an insight of the role of the macroeconomy in the explanation of stock returns. The effect of macroeconomic factors on stock returns can be thought to be a consequence of the pricing of stocks, as the stream of discounted expected future cash flows from holding a security. According to Damodaran (1994), the general stock valuation model is of the form presented in Eq. (2). Price of Security =
t=∞ t=1
DPSt (1 + r)t
(2)
. . . where DPSt denote expected dividends per share and r is the required rate of return on stocks. It is clear from Eq. (2) that any macroeconomic factors which affect either the expectations of future cash flows to the investor (dividends per share) and/or the rate used to discount them will indirectly, though effectively, influence stock returns. Chen, Roll and Ross (1986) were among the first to specify and test a set of economic factors which, based on economic theory and intuition, should affect stock returns either through future cash flows or through the discount rate. They utilised the following factors: inflation; the term structure of interest rates; risk premia and industrial production and found them to be significant in explaining stock returns. Although Chen, Roll and Ross (1986) can by no means claim that they have found the full set of variables for asset pricing they have most certainly made an important step in the right direction. Their model is of course the base for the well known Arbitrage Pricing Theory (APT). Their work has been continued in a series of papers by Burmeister and Wall (1986) and Burmeister and McElroy (1987, 1988) who developed a multifactor model and found that five macroeconomic factors, similar, though not identical, to those employed by Chen, Roll and Ross (1986) are statistically significant in explaining stock returns. The factors they utilised were: default risk; the term structure of interest rates; inflation; sales as a proxy for the profits of the economy and; the market. In the spirit of the above work, other researchers have studied the explanatory power of macroeconomic variables over stock returns in various stock exchanges across the world. For example, among others, Poon and Taylor (1991) applied the ideas discussed above to the U.K., Martinez and Rubio (1989) applied them to the Spanish stock market, Hamao (1988) applied them to Japan and Wasserfallen (1989) applied them to a number of European countries. Most of the above studies
116
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
find evidence in favour of the hypothesis that a set of macroeconomic factors has explanatory power over stock returns. Chen and Jordan (1993), in a more recent study featuring the U.S. market, employed a set of factors similar to the one used by Chen, Roll and Ross (1986) but grouped the companies in their sample according to their industry classification, the significance of which is something which is discussed in great extent in Chap. 1 of the book of Kavussanos and Marcoulis (2001). Chen and Jordan (1993) also find evidence in favour of macroeconomic factors being significant in the pricing of stocks. More specifically, they document that, apart from the market returns, changes in oil prices and inflation are possible sources of risk. However, as in the case of microeconomic factors, the study of macroeconomic factors and stock returns is not confined only to academics. Salomon Brothers (better known today as Salomon Smith Barney) have developed a macroeconomic model in the spirit of the macroeconomic models we have been discussing above (Sorensen et al., 1989). This model uses seven variables to explain stock returns. They are: economic growth, measured by changes in industrial production; the business cycle, measured by the difference between corporate bonds and U.S. Treasuries; long-term interest rates, measured by yield changes in the 10-year Treasuries; short-term interest rates, measured by yield changes in the 1-month Treasuries; inflation shock, measured by changes in the consumer price index; the U.S. dollar, measured by changes in a 15 country currency basket Vs the dollar and; a market proxy. Salomon have been using their multifactor model for some time now and they claim that using monthly data, this model explains about 40% of the fluctuations in the returns of a sample of 1,000 stocks. Definitely then, models of this type are very promising. In addition to the above, following their invention of the APT model, Roll and Ross created the Roll and Ross Asset Management Corporation to translate their theory to practice. They begin by stating the systematic sources of risk that they believe are relevant to capital markets. These, according to the Roll and Ross Asset Management Corporation are: the business cycle; interest rates; investor confidence; short-term inflation and; long-term inflationary expectations (Sharpe et al., 1995). Roll and Ross quantify these factors by designating certain macroeconomic variables as proxies. For example, the business cycle factor is represented by changes in the industrial production index. At the heart of the Roll and Ross approach lies the assumption that each source of risk is subjected to volatility and is entitled to some return. Moreover, Roll and Ross assume that individual securities and portfolios of securities possess different sensitivities to each source of risk. With these in mind, Roll and Ross attempt to construct investment portfolios which offer the most attractive risk-reward profile for the investor.
Cross-Industry Comparisons of the Behaviour of Stock Returns
117
Another institution of international prominence which is dealing with multifactor models from a more practical angle is BARRA International. Their professionals utilise both fundamental and macroeconomic factors to explain stock returns. For example, in 1993 BARRA International published a research paper (Engerman, 1993) whose aim was to compare a multifactor model utilising fundamental factors to another multifactor model utilising macroeconomic factors in the U.S. For the record, the fundamental factors were, the market; the stock’s price momentum; size; trading activity; growth; the earnings-to-price ratio and; the book value-to-market value ratio. The macroeconomic factors were, interest rates; bond spreads; industrial growth; inflation; oil prices; gold price and; dollar’s value. Their results indicate that both the microeconomic, fundamental factors and the macroeconomic factors have a role to play in the explanation of stock returns. They, moreover, claim that microeconomic factors are more effective than macroeconomic factors in accounting for the cross-sectional variation in stock returns. In line with the above ideas, BARRA International is currently providing multifactor models for several countries including the U.S., Japan, the U.K., Canada, France, Switzerland and many more. Reference to the BARRA models will be made again later in this paper, when the importance of industry classification is discussed. Industry classification is indeed a prominent feature of the majority of BARRA’s models.
4. THE IMPORTANCE OF THE INDUSTRY EFFECT Apart from King (1966), who as mentioned above was the first to argue that changes in stock prices can be explained by an industry effect as well as a market and a company effect, others, such as Arditti (1967) and Nerlove (1968) find that industry differences are highly significant in explaining cross-sectional differences in stock price returns in the U.S. stock market. Similar results are also produced in the industry focused analyses of Saunders and Yourougou (1990), Ferson and Harvey (1991) and Isimbabi (1994), among others. Hence, an investor whose aim is to build a diversified portfolio could well be interested in these cross-industry differences since he could utilise them to build a portfolio diversified across industries of different risk-return characteristics. The importance of the industry effect is also emphasised in a number of other studies focusing on the U.S. stock market. For example, Sorensen and Burke (1986) studied the relative price performance of several industry groups and concluded that an investment strategy based on buying and holding the best performing industry groups may enhance returns. Kane and Unal (1988) and Neuberger (1991) focused on the risk return characteristics of the U.S. banking sector. They find banks to be greatly underpriced
118
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
and exhibiting systematic risk that is lower than that of the “average” company. They also find that this systematic risk is increasing over time and that stocks of larger banks exhibit higher risk than small ones. Gyourko and Keim (1993) focused on another “popular” U.S. industry, real estate, and found that the average monthly returns of the industry are just under 1% with the associated standard deviation being over 7%. They then compared these figures with the corresponding figures of the S&P 500 index and found that the average monthly returns of the real estate industry are higher, but as might be expected are accompanied by higher standard deviation. Furthermore, the industry effect has not only been studied by academics. Prominent practitioners have also spent a lot of research time and resources to formulate models which encompass this effect. For example, BARRA International who was introduced above, very often estimates multifactor models for different industry groupings. One of their most widely used multiple-factor models, the BARRA E3 U.S. model (Kahn, 1994), utilises microeconomic factors and estimates a model for each of 55 industry groups. Furthermore, their Canadian (Roy, 1992) and Japanese models (Rosenberg, 1991), both released in 1992, also utilise microeconomic and macroeconomic factors to estimate models across several industrial groupings. By considering some real world examples Kavussanos and Marcoulis (2001) in Chap. 1 of their book argue that there is scope for undertaking industry analysis and that it pays to direct investment resources on to some industries at the expense of others at different points in time. From the evidence observed, the authors go on to argue that it would have been beneficial for an investor to have performed careful industry analysis at any point in time, from the post-war period to date. Towards this end, Jones (1993) is cited, where a study spanning over the 48 year period 1941–1989, indicates that from 1941 to 1973 the computer and business equipment industry did extremely well (145 times what it was in 1941) and the electronics industry also did well, rising to almost 69 times its 1941 level. On the other hand, during the same 32 year period, the lead and zinc industry was less than twice its original level, and the sugar and textile apparel industries were only three times their original levels. Examples from the eighties cited in the same study, also indicate tremendous differences among industries’ performances as measured by the Standard and Poor’s Stock Price Industrial Indexes. For example, during the period 1982–1989, the drugs industry almost quadrupled, the broadcast media industry rose to almost six times its 1982 level while on the other hand the chemicals industry declined by almost 30%. However, the most dramatic example during the period mentioned is that produced by the market crash of October 1987 when the Dow Jones Industrials index lost 23.2%. Industries such as toys, machine tools, leisure time, gold, and
Cross-Industry Comparisons of the Behaviour of Stock Returns
119
offshore drilling, among others, lost around 40% of their value while others such as electric companies and telephone companies lost less than 10%. Jones (1998) clearly demonstrates that the losses suffered by investors varied widely according to the particular industries held.
5. THE ISSUE OF THE CLASSIFICATION OF COMPANIES INTO INDUSTRIES A practical question with industry analysis, concerns the classification of companies into industrial sectors. The problem is not new. Kavussanos and Marcoulis (2001) utilise the SIC (Standard Industrial Classification) codes, to classify the U.S. companies used in their research. The sectors used were transportation industries, such as water transportation, air transportation, rail transportation and trucks and non transportation industries such as electricity, petroleum refining, gas and real estate. The rationale for selecting the aforementioned industries is that the air transportation, rail transportation and trucks industries have been chosen due to the fact that they are transportation industries. Hence, they might be argued to compete, in one way or another, with the water transportation industry in the investor’s stock selection decision. The petroleum refining, electricity and gas industries were chosen due to their stable growth nature which is directly opposite to the very cyclical nature of the water transportation industry. Finally, the real estate industry was included due to its cyclical nature which, is a characteristic of the water transportation industry as well. For exact SIC definitions for each of the industries see Kavussanos and Marcoulis (2001), Appendix A of Chap. 1. The SIC system which is based on Census data classifies companies into industries according to their final product or service. SIC codes have 11 letter divisions, A to K and each division consists of several major industry groups, designated by a two digit numerical code. The major industry groups, within each division, are further subdivided into three, four, and five-digit SIC codes to provide even more detailed classifications. The larger the number of digits in the SIC system, the more specific the breakdown. For instance, Division E defines transportation. This consists of 10 major groups as per 2-digit number; 40, 41, . . . , 49. Major group 44 corresponds to water transportation which consists of six further 3 digit groupings. Grouping 441 refers to deep-sea foreign transportation of freight, which only has one four-digit category under it. Other SIC codes for the other industry groups examined in the above book include 40 (Railroad Transport) and 45 (Air Transportation).
120
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
SIC codes have significantly contributed towards bringing order to the industry classification problem by providing a consistent basis for describing industries and companies. Analysts using SIC codes can focus on economic activity in as broad, or as specific, a manner as desired. For example, industry groupings in a number of studies dealing with industry stock return comparisons in the U.S. (e.g. Boudoukh et al., 1994; Eun & Resnick, 1992; Ferson & Harvey, 1991; Isimbabi, 1994), among others, have been formulated according to their SIC codes. Typically, the above studies have used 2–4 digit industry groupings. Despite the fact that the SIC system of industry classification is the easiest to use and the most consistent system available, it is not the only one in actual use. Standard and Poor’s Corporation, as from the end of 1982, provide weekly stock indices on approximately 100 industry groupings, many of which go back 30 or even 40 years. The Value Line Investment Survey covers roughly 1,700 companies, divided into approximately 90 industries. Another useful industry classification system is that of The Media General Financial Weekly which divides stocks into 60 industries. Given its many advantages and its wide use among academics and practitioners the SIC classification system was used by Kavussanos and Marcoulis (2001). Nonetheless, like any classification system, it faces certain limitations, the most important of which is the classification of multiproduct, diversified companies under more than one heading. When performing cross industry comparison, in order to avoid distorting the results during comparisons between the selected industries, companies that were found to be listed under more than one industry groups were omitted. This resulted in industry groupings which were relatively homogeneous. Morgan Stanley Capital International (MSCI) also have an industrial classification system. They publish data on monthly price indices (with 1970: 1 = 100) for 38 International Industries. The MSCI price indices are value-weighted and aim for 60% coverage of the total market capitalisation for each market. Companies in the indices replicate the industry composition of each local market. The chosen list of stocks, formed from the share prices4 of approximately 1600 securities in 22 countries, includes a representative sample of large, medium, and small capitalisation companies from each local market, taking into account the stocks’ liquidity. Furthermore, stocks with restricted float or cross-ownership are avoided. Kavussanos et al. (2002) use this data set to perform industry analysis at the international level using global portfolios. In yet another study, Kavussanos et al. (2003) focus on the international maritime sector and its subsectors. The answer to the question of classification of shipping companies into its subsectors is not readily available. Yet, such industry classification can enhance further the understanding of the risk return trade offs
Cross-Industry Comparisons of the Behaviour of Stock Returns
121
Table 1. Maritime Industry Sectors. Sector
Description
Bulk Container Cruise Drilling Ferry Offshore Shipping
Dry bulk, older type General cargo ships, excluding OBO LOLO and some ROLOs with large container section Cruise ships Rig owners and operators Passenger ferries including ROPAX Supply boats and anchor handlers Companies with 90% or more of revenue derived from shipping or shipping related activities but which could not be classified into any other sector Oil Tanker, excluding chemical and gas Tankers as well as FSPO. OBOs were included when operated as oil Tankers Shipyards excluding Rig yards Companies with between 60% and 90% of revenue derived from shipping or shipping related activities – the balance being derived from elsewhere All of the above sectors
Tanker Yard Diversified All
Note: OBO – Oil Bulk Ore; LOLO – Lift On Lift Off; ROLO – Roll On Lift Off; ROPAX – Roll On Passenger.
within the shipping industry. Kavussanos et al. (2003) identify every possible maritime company listed continuously in any stock exchange in the world over the 3-year period July 1996 and July 1999, and classify it under pre-defined sub-sectors of the industry (Table 1). This, sampling of companies across stock exchanges (rather than focusing on companies listed in one exchange), gives the largest possible cross-sectional sample of maritime companies in each sub-sector, and at the same time a sufficient length of time series data for returns (36 monthly observations) to enable estimation and inferences. The starting point was the Maritime Transport and Energy list of traded shares appearing in the Financial World page of the Lloyds List, which in turn is based on the Bloomberg classification list. This was supplemented by any other public companies known to be involved in shipping or shipping related industries but not listed there. In order to classify companies into sectors, a short questionnaire was sent to 250 of these companies in July 1999 asking them to classify the percentage of their core business activity in a number of predefined sectors. This information was supplemented by consulting their annual reports for 1998 and 1995. There was an approximately 20% response to the initial questionnaire, with a further 30% replying after a reminder letter, which was sent four weeks later. Financial information for companies which did not reply was obtained from the Wright Investors’ Service web page (http://www.wisi.com), from the Fairplay Online Directory (http://www.wsdonline.com) and from individual company web pages.
122
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
In order to make inferences for each sector which reflect the risk/return profile of operating in the specific sector, companies whose economic activity in shipping or shipping related activities was less than 60% were considered overly diversified and were discarded from the sample.5 Companies for which there was no information on revenue, companies involved in mergers, acquisitions and/or changes in their core business during the sample timeframe, and companies for which stock data could not be found on DataStream were excluded from the analysis. To account for the possibility that different degrees of diversification have varying effects on the risk-return profiles of sectors, the companies that remained were classified and analysed according to whether 60, 75 and 90% of their core business activity was in the same sector. Specialised companies operating only in one sector were straightforward to classify. Companies whose core activity was over 90% in more than one sector of shipping but for which no detailed breakdown of the percentages attributable to each sector were available were classified in a general category called “shipping.” Companies with diverse core business that included over 10% of activities not shipping or shipping related were classified as “diversified.” The sectors “Reefer,” “Gas,” “Chemical Tankers,” “Brokers” and “Ports” had to be abandoned due to too few listed companies belonging to them. In total 108 companies made up the final sample used for analysis. For analysis, the return on stocks must be defined, as well as other data on relevant micro macro economic variables and the market. In the above studies monthly stock price and dividend yield (in percentage form) data for each share were collected from DataStream International Service. Logarithmic monthly returns for company i at time t, Rit , are calculated in percentage form using the equation: (P it + (P it × DYit /1200)) R it = 100 × ln P it−1 Where Pit and Pit −1 are the stock prices of company i at time t and t−1, respectively, and DYit is the annualised dividend yield paid by company i at time t. In calculating the CAPM regression and multifactor regressions, a question of what is the relevant market is always raised. Because the sample includes companies listed on stock exchanges in different countries the Morgan Stanley Capital International (MSCI) All Country World Index was used for analysis in the studies by Kavussanos et al. (2002, 2003). Also, given the practice of evaluating industry specific funds by benchmarking on sectoral indices, the MSCI International Shipping Index was also used for analysis. The MSCI All Country World Index is calculated as a market capitalisation weighted average of equity returns in 51 countries (23 developed and 28 emerging),
Cross-Industry Comparisons of the Behaviour of Stock Returns
123
and is quoted in gross form inclusive of dividends. The MSCI Shipping Index is one of the 38 industry indices produced by Morgan Stanley. Companies are classified based on their principal economic activity as determined by the breakdown of earnings. If no detailed earnings data are available then breakdown of sales data are used. In defining industries MSCI attempts to construct homogeneous groups which are expected to react similarly to economic and political trends and events. A further practical question relates to how the risk free interest rate is defined. The U.S. three-month Treasury bill rate is generally considered as a measure of the global risk free rate of interest, RFt . It is instructive to view summary statistics for average equity returns by maritime sector and for the returns on the MSCI world and shipping indices for the period July 1996 to July 1999, as estimated in Kavussanos et al. (2003). These are presented in Table 2. Yet another question is how to construct empirically the macro and micro economic variables used as explanatory variables in a multifactor model, such Table 2. Summary Statistics of Mean Monthly Returns of Each Sector by Classification Criteria; July 1996–July 1999. Sector
Classification Criteria 90%
75%
60%
Mean
SD
No
Mean
SD
No
Mean
SD
No
Bulk Container Cruise Drilling Ferry Offshore Shipping Tanker Yard Diversified All
−2.18 −0.85 3.04 0.32 −0.05 0.17 −1.68 −2.53 0.60 −0.50 −0.92
1.22 2.90 1.32 1.12 2.73 1.67 3.79 2.50 0.46 1.57 2.85
6 7 3 7 11 7 34 12 4 17 108
−1.79 −0.92 3.04 0.32 −0.47 0.17 −1.77 −2.53 0.23 −0.12 −0.91
1.54 2.45 1.32 1.12 2.60 1.54 3.90 2.50 1.10 1.72 2.86
7 9 3 8 15 8 30 12 6 10 108
−1.88 −0.92 2.93 0.33 −0.14 0.25 −2.01 −2.46 0.23 N/A −0.91
1.45 2.45 1.10 1.05 2.61 1.38 3.76 2.41 1.00 N/A 2.86
8 9 4 9 17 10 30 13 7 N/A 107
MSCI-All MSCI-Sh
Mean 1.42 0.28
SD 4.50 6.34
Skew −1.61 4.19
Kurt 4.19 3.65
Notes: (1) SD = Standard Deviation, No = Number of companies classified under each sub-sector, Skew = Coefficient of Skewness, Kurt = Coefficient of Kurtosis, MSCI-All and MSC-Sh are the Morgan Stanley All Country World Index, and the Shipping Index, respectively, (2) Under the 60% criterion, the Diversified sector only contained one company (Wilh Wilhelmsen) and this sector was therefore dropped.
124
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
as that estimated in Kavussanos et al. (2002) and Kavussanos and Marcoulis (2000b). It is suggested that only the unanticipated part of the macro economic series is relevant, as the anticipated part will be incorporated instantly into prices in efficient markets. ARIMA models are used to filter out the anticipated part of the macro variables, with the unanticipated part being used for estimation in multifactor models. For details see these last two papers.
6. MAJOR FINDINGS OF THE RESEARCH It should be mentioned at this point that the reason Kavussanos and Marcoulis (1997a, b, c, 1998, 2000a, b, 2001) focused on the U.S. water transportation industry is that the U.S. had the largest number of companies in this sector, relatively sufficient for meaningful analysis, compared to any other single country. Even for the U.S. the listing of water transportation companies did not have such a long history, with most companies entering the stock exchange to raise funds in the 1980s and 1990s. It is worth noting that currently, the vast majority of U.S. based water transportation companies still remain privately owned. The primary reason, given by both investor groups and investment bankers for this, is the industry’s perceived high level of risk. This risk stems from the fact that on the one hand the water transportation industry is a predominantly capital intensive industry which requires huge investment outlays, while on the other hand it is subject to cyclicality which more often than not is beyond the industry’s control. This cyclicality is a result of the industry serving the world economy through the transportation of world trade. It is a fact of life that the world economy goes through cycles and along with it world trade goes through cycles as well. Consequently, the water transportation industry is also subject to cycles whose amplitudes are a function of those of the world economy and the demand and supply situation in the water transportation market. Given the above risk profile of the water transportation industry, the prime motive of the studies have been to examine, and compare, this industry’s stock exchange risk perception over time and across industries of similar and different risk profiles. In Kavussanos et al. (2002, 2003) the analysis was carried in a global setting. This goes along with the notion of investors operating in a transnational goods and capital markets and forming global portfolios of industries across countries. It is a practice widely followed by big investment houses. At the same time, such an analysis enables the increase in the sample of companies listed in each industry to a respectable level. As a consequence results are more reliable. Furthermore it allows the distinction of subsectors within the shipping industry, as in Kavussanos
Cross-Industry Comparisons of the Behaviour of Stock Returns
125
et al. (2003) – see Tables 1 and 2 and discussion in the previous section, something which would be impossible if restricted to a single country setting. Of course, as more companies decide to become listed in stock exchanges, results can be even more reliable in future analysis. To start with the findings of Kavussanos and Marcoulis (2001), Appendix A of this paper presents them, in tabular form, so that the reader can obtain a full picture of the major findings of the research. Moreover, Appendix B of the paper presents a summary of the empirical evidence regarding the micro and macroeconomic factors employed in the book. Other studies similar to Kavussanos and Marcoulis are tabulated, which have also attempted to identify the determinants of stock returns by using sets of either microeconomic or macroeconomic factors. Appendix B.1 presents empirical evidence in the finance literature regarding the macroeconomic factors employed in the book and compares it with the results of the book, while Appendix B.2 repeats the exercise for the microeconomic factors employed in the book.6 As can be seen in Appendices B.1 and B.2, there are similarities as well as differences among the results of the book and the general literature. Most differences show up with respect to the macroeconomic factors which, incidentally, tend to also exhibit differences across other empirical studies. However, when comparing the results, one should keep in mind that the majority of published work utilises a wide sample of companies, which have used portfolios as a basis for asset pricing rather than industry classification as in our study.
6.1. Results from the CAPM Kavussanos and Marcoulis (1997a, 1998) and Chap. 4 in Kavussanos and Marcoulis (2001) deal with the Risk and Return of U.S. Water Transportation Stocks over time and over Bull and Bear market conditions and its results do indicate a number of interesting aspects regarding the behaviour of water transportation company stocks during the period 1985–1994. Firstly, the average beta of shipping companies was estimated to be numerically lower but statistically not different from the beta of the “average” company (unity). This result of the systematic, non-diversifiable, risk of the water transportation industry being not different from the average market risk could make the shipping industry an attractive candidate for potential investors. Furthermore, another attractive characteristic of the beta coefficient of the water transportation industry, from the investor’s point of view, is the fact that it appears to be stable over time. Secondly, water transportation companies appear, with some notable exceptions, not to be underpriced over the full ten year period examined. Sub-period
126
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
estimations of the CAPM coefficients suggest that underpricing did not occur during the first five year period as well as during the second five year period. Thirdly, there appears to be a “size” effect in the shipping industry for the period 1984–1989 in the sense that smaller shipping companies tend to exhibit higher returns. These higher returns, as might be expected, were found to be accompanied both by higher total and systematic risk. However, this “size” effect disappears in the period 1990–1994 possibly due to the shifting of several “small” caps to the medium and large groupings. The chapter also examines whether underpricing, as measured by alpha, or the systematic risk, as measured by beta, of the water transportation companies included in the sample examined changed over bull and bear market conditions during the ten year period examined. It is found that alpha, not beta, tends to be mostly affected by upward or downward market movements. Therefore, investors considering including shipping stocks in their portfolios need not worry about possible changes in the systematic risk of these stocks during changing market conditions. Another aim of the chapter is to examine the risk-return relationship of the water transportation industry along another dimension, that of comparing the systematic risk of companies belonging to this industry to the systematic risk of companies belonging to other related and non-related industries. In this context, the beta of the water transportation industry is compared to the beta of the following seven industries: rail transportation; air transportation; trucks; electricity; petroleum refining; gas and; real estate. To achieve that, the Capital Asset Pricing Model is employed. Results reveal some further interesting characteristics of the stock returns of the water transportation industry in the U.S. during the period 1984–1995. The beta of the water transportation industry is significantly lower than the beta of the rail transportation industry and the beta of the real estate industry. It is statistically similar to the betas of the five other industries employed in the chapter. Looking at the findings of the chapter, one might conclude that stocks belonging in the water transportation industry do not appear to possess any risk characteristic the investment community is not aware of. The industry average beta of 0.92 seems to be in line with the average company’s beta, unity, and the average explanatory power of the regressions of around 23% is also typical in these kind of estimations. Furthermore, tests suggesting that the industry beta does not appear to change over time, despite the cyclical nature of the underlying industry, can only be “good” news for the industry’s stock market perception. Moreover, numerically speaking, the beta of the water transportation industry is the lowest of all transportation industries’ betas.
Cross-Industry Comparisons of the Behaviour of Stock Returns
127
6.2. Microeconomic (Company specific) Factors as Determinants of Equity Returns Kavussanos and Marcoulis (1997b) and Chap. 5 in Kavussanos and Marcoulis (2001) undertake a comparative analysis of the stock market perception of the riskreturn relationship of U.S listed water transportation stocks in relation to stocks belonging to the other transportation and non-transportation sectors mentioned above over the same time period. The significant difference between this analysis and the one in Kavussanos and Marcoulis (1997a) is that this one is carried out in a multidimensional risk environment. More specifically, apart from relating crosssectional differences in the returns of companies belonging to different sectors to the market, the model used in this chapter relates those differences to a number of fundamental, company specific, factors, which according to intuition and academic research are believed to influence stock returns. These factors are: the market value of equity (size), the book-to-market value of equity ratio; the earnings-to-price ratio; the asset-to-market value of equity ratio; and the asset-to-book value of equity. The methodology used by the authors to estimate the above relationship for each industry is the Seemingly Unrelated Regression Model (SUR) due to it possessing two significant advantages over the classic Ordinary Least Squares Model (OLS). The first is that the sensitivities of each company’s returns to the market (betas) are estimated simultaneously across companies together with the impact of the fundamental variables and the alphas also allowing the imposition of cross-equation restrictions on the parameters. The second is that the SUR, in contrast to more classic methodologies utilised in the past in similar studies, adjusts for the cross-sectional correlation in the residual returns across companies thus leading to parameter estimates which are more efficient than those given by OLS models, the gain being proportional to the correlation between disturbances from the different equations. This advantage is particularly important in studies of this nature since companies grouped according to their industry classification are likely to exhibit residual returns’ correlation. The results of the study indicate that there appear to be factors, from the microeconomic, company specific environment, which, in addition to the market which remains the driving force behind returns, tend to influence the returns of the water transportation industry and the other industries. It should be noted, nonetheless, that the significance of the fundamental variables appears to vary across sectors and over time. The book-to-market value, the asset-to-market and the asset-to-book value of equity ratios, and the market value of equity are significant in some industries but not in others while the earnings-to-price ratio has no role to play in any industry’s returns. Generally speaking, the coefficients of the fundamental
128
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
variables come out with the expected sign, save the positive “size effect” in the petroleum industry. The returns of the water transportation industry appear to be influenced by only one factor, the asset-to-book value of equity ratio. The relationship is a negative one and suggests that as shipping companies become more highly geared, in terms of book leverage, their stock market performance deteriorates. As far as systematic risk (beta) is concerned and in line with the findings of Chap. 4, the water transportation industry, along with three other nontransportation industries, exhibits lower than average systematic risk for the whole period studied as well as for the two subperiods. Furthermore, the beta of the water transportation industry, along with that of six other industries, does not vary from subperiod one to subperiod two while only the real estate industry exhibits a lower beta than the water transportation industry. Finally, as in Chap. 4 in Kavussanos and Marcoulis (2001) the alpha of the water transportation industry along with the alphas of all other industries analysed are positive thus suggesting that these industries have been underpriced over the time horizon studied.
6.3. Macroeconomic (Economy Wide) Factors as Determinants of Equity Returns Kavussanos and Marcoulis (2000a) and Chap. 6 of Kavussanos and Marcoulis (2001) is utilizing the traditional one-factor market model, augmented to include a number of other economic factors believed to influence security returns. However, in this chapter, the factors used are macroeconomic, as opposed to microeconomic. Hence, cross-sectional differences in the returns of the companies in each industry are related to the stock market and the following set of macroeconomic factors: industrial production, the term structure of interest rates, oil prices, consumption, and inflation. The selection of this set of macroeconomic factors was driven both by intuition, since the aforementioned factors affect both the future cash flows and riskiness of a company, as well as due to their popularity among academics (they have been widely used in previous studies). Assuming efficient markets, only the innovations or unanticipated changes in the above macroeconomic variables should influence stock returns. Hence ARIMA models were used to filter out the anticipated component of series with “memory.” Following that, Multiple Least Squares (MLSQ) regression methods were used to estimate the relationship of the unanticipated changes in the above factors to the stock returns of each industry over the period 1985–1995.
Cross-Industry Comparisons of the Behaviour of Stock Returns
129
The results of this chapter, like the one preceding it, show that there are factors other than the market which influence the returns of the water transportation industry and other industries, thus justifying the use of multifactor models instead of the traditional one-factor, market model. More specifically, in line with the findings of Chaps 4 and 5, the beta of the water transportation industry is found to be lower than the “average” beta of unity and it is also found to be among the lowest in the industries analysed. Moreover, the alpha of the water transportation industry, along with the alphas of the other industries analysed are significantly higher than zero thus implying that these industries have, on average, been underpriced over the period 1985–1995. Regarding the macroeconomic factors, the authors find that their effect varies across industries. This is probably the most interesting and important finding of the study, that different industries tend to react differently to different economic shocks. The investment manager could utilise this finding, and by examining the sensitivities of industry stock returns to the macroeconomy, make better investment decisions. As far as the returns of the water transportation industry are concerned, they appear to be influenced by two macroeconomic factors, monthly industrial production and oil prices. The former exerts a negative effect which suggests that increases in monthly industrial production are accompanied by dropping returns in the industry while the latter indicates a positive relationship which suggests that the returns of the water transportation industry are an increasing function of increases in oil prices.
6.4. Microeconomic and Macroeconomic Factors – A Unified Approach Kavussanos and Marcoulis (2000b) and the seventh chapter in Kavussanos and Marcoulis (2001) relates cross-sectional differences in the returns of the companies in each of the industries mentioned in the previous chapters to the set of the microeconomic factors utilised in Kavussanos and Marcoulis (1997b) and the set of the macroeconomic factors utilised in Kavussanos and Marcoulis (2000a) simultaneously over the period 1985–1995. The chapter recognises that both micro and macro economic factors may be determinants of stock returns across industries and attempts to uncover the determinants of each industry’s stock returns in a more general setting where both sets of factors are included. This practice is supported not only by academics (Chen et al., 1986; Fama & French, 1992 among others) but also by practitioners (BARRA, amongst others). Results from a Seemingly Unrelated Regression Model (SUR) regarding market betas indicate that the market, as expected, has a significant role to play in
130
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
explaining the returns of all industries. The beta of the water transportation industry is, as in the previous chapters, the lowest among the transportation industries and the third lowest, ranking behind electricity and gas, of all the industries analysed. Further inferences regarding market betas suggest that the market beta of the water transportation industry, along with the market betas of the electricity, gas and real estate industries, are significantly lower than the “average” market beta which is of course one. Moreover, the alpha of the water transportation industry, along with the alphas of the other industries analysed are significantly higher than zero thus implying that these industries have, on average, been underpriced over the period 1985–1995. As far as the economic factors are concerned, the returns of the water transportation industry were found to be positively related to oil prices and market value of equity and negatively related to monthly industrial production and the total assets-to-book value of equity ratio. The sensitivities of each industry’s returns to the set of microeconomic and macroeconomic factors, as expected, varies across industries. All factors, except the price to earnings ratio, appear to be priced in one industry or another. The estimation of this general model incorporating both microeconomic and macroeconomic factors for each industry sheds some light regarding differences both in the structure and sensitivities of each industry’s stock returns to the set of factors employed. A useful point that comes out of this analysis is that the stock returns of the water transportation industry, for example, are positively affected by oil prices and the market value of equity and negatively affected by monthly industrial production and the asset-to-book value of equity. The aforementioned factors comprise a different set when compared to any other industry under analysis and hence the investment manager, by picking, or not picking, this industry, may expose, or not expose, his portfolio to the specific set of economic factors. The same holds of course for any other industry. Furthermore, the industry analyst can also compare the direction and magnitude of the sensitivities of the different factors employed in the analysis to the returns of each industry. For example, an unanticipated change in oil prices affects the gas industry much more than the water transportation industry or the negative effect of market leverage is more profound in the electricity industry than in the water transportation industry. Finally, it should be noted that the sign of the microeconomic and macroeconomic factors utilised is not always in line with the majority of the existing literature, thus pointing out that empirical results regarding the direction of the determinants of industry stock returns may differ, in some cases, to the direction of the determinants of the full universe of stocks. This is another important finding of this chapter and certainly an area which could provide interesting research possibilities for the industry analyst.
Cross-Industry Comparisons of the Behaviour of Stock Returns
131
6.5. Macroeconomic Factors and International Industry Returns Given the increasing degree of integration in the capital markets internationally, an interesting question is the identification of factors affecting the risk return profile across industries at the global level. This is done in Kavussanos et al. (2002) and in Chap. 8 of the Kavussanos and Marcoulis (2001) book. The objective of the paper was to present evidence, for the first time, of the ability and the usefulness of world macroeoconomic news in explaining the variability of global industry returns. The monthly risk variables employed in the study are: the excess return on a world equity market portfolio, fluctuations in global exchange rates, global measures of inflation, industrial production growth and credit risk. OLS regressions were used to estimate the relationship between unanticipated changes in the above factors and the excess returns of a set of 38 international industries as compiled by Morgan Stanley Capital International (MSCI). Among the factors considered, the return of the world market portfolio affects significantly all the 38 international industries under analysis. Moreover, it is by far the most important factor in explaining the variation in international industry returns. Inclusion of macroeconomic factors marginally increases the explanatory power of the model. Several significant relationships are detected with respect to the remaining factors that do not, generally, exhibit a consistent pattern in the way in which they affect returns of global industries. The long run impact a factor may have, can be positive on the returns of a particular industry, and negative or insignificant on the returns of another, depending on industry specific characteristics. This finding is also consistent with evidence presented in chapters 6 and 7 of the Kavussanos and Marcoulis (2001) book. The practical implications of this study are important for portfolio managers. The industry integration or segmentation in the world economy, makes any evidence on the sources of risk that may affect stock returns across industries at the international level essential in adopting an optimal strategy for global investing. The industrial classification of a given asset becomes crucial, as certain global industries develop to be homogeneous, and capital markets are becoming increasingly integrated. The significant relationships between global macroeconomic factors and international industry stock returns detected in this paper, are useful to the investor who can exploit these relationships in order to increase his diversification capacity or speculate by timing his investment. 6.6. International Industry Returns for Subsectors of the Shipping Industry The paper by Kavussanos et al. (2003) gives yet another dimension to the analysis, as explained before. It compares the behaviour of shipping and shipping related
132
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
company stock returns to reveal whether systematic risk differs from the average in the market and across sub-sectors of the maritime industry. Following an extensive collection of information through postal questionnaire survey, 108 publicly listed shipping and shipping related companies, across stock exchanges of the world, are classified by sector according to their core business activity – see earlier discussion for this. The Capital Asset Pricing Model (CAPM) is employed for the period 1996 to 1999 to model stock returns and measure sector s (systematic risk). During the 1996–1999 period analysed, when the shipping industry was not doing particularly well, companies in sectors were broadly overpriced, and average returns seemed to be negative. Market ’s for all the stocks in the industry appeared to be significantly lower than the market beta. The Drilling and the Offshore sectors were significantly higher than one, however all other average sector ’s appeared to be either equal or lower than the market average. The sectors that appeared to have ’s which were significantly lower than the market are the Shipping, Tanker, Ferry, and also Bulk and Containers mostly. It seems then that the maritime industry stocks do not carry above average market risk, at the international setting. In comparing the ’s amongst sectors it seems that the Drilling and Offshore sectors have the same proportion of systemic risk in them. The  values of these sectors do not differ significantly from each other but are significantly different from all the other sector  values except for Cruise. However, the Cruise sector , whilst not significantly different from Drilling and Offshore, is not significantly different from any other sector. There is no significant difference in the  values of the remaining sectors: Bulk; Container; Ferry; Shipping; Tanker; Yard; Diversified and All. When regressed against the MSCI Shipping Index, the Drilling and Offshore sectors again appear to have the same degree of market risk in them. There is no significant difference in the  values of all other remaining sectors. Finally, as more companies in the industry become public the scope for increased sample sizes for each sub-sector of the maritime industry on which to base inferences will also increase. Perhaps a further study when more data is available and also when market conditions are different (on the upturn) may add to the body of knowledge established in this paper.
7. USEFULNESS OF THE FINDINGS Having presented the review on the state of research regarding listed companies in the shipping and other related sectors and subsectors of shipping, some comments
Cross-Industry Comparisons of the Behaviour of Stock Returns
133
are in order regarding the usefulness of these results. The focus is on two major categories of investors including portfolio managers and corporate financiers. As far as the former category is concerned, traditionally, investors and portfolio managers’ strategies, regarding stock selection, are perceived as the choice of the proper mix of stocks in order to maximise returns subject to their risk profile. In order to achieve that however, they would need to identify what “features” really matter in a stock’s or an industrial sector’s performance. The research on shipping and transportation sectors answers this question by expressing the returns of the industries analysed as a linear combination of each industry’s returns’ sensitivity to a number of microeconomic and macroeconomic factors times the risk premium on this factor. As might be recalled, every industrial sector analysed has its own pattern of sensitivities to the different microeconomic and macroeconomic factors employed. This might be used by the architect of the portfolio’s investment strategy to determine the most desirable exposure to each risk factor. Altering the mix of industries included in the portfolio will certainly affect the amount, and type, of risk exposure to each factor studied. For example, suppose the portfolio manager wishes to move away from any unanticipated change in the term structure risk (an unanticipated widening or narrowing of the long vs short -term interest rates) since he believes that there will be some turbulence in the future regarding this factor. Utilising the multifactor model, for example he could exclude the air transportation, trucks, electricity and petroleum refining industries from his portfolio. Alternatively, the portfolio manager could employ the above model or any other model presented, to analyse the sensitivities of the factors employed to the returns of each industry. Using each industry relevant equation, the investment manager can substitute his expectations of each microeconomic and macroeconomic factor employed in order to arrive at the expected returns of each industry. Then, according to the confidence that he may be able to place in his expectations, he can decide upon the proportion of stocks that belong to each industry that he will include in his portfolio. Furthermore, by comparing industry specific equations, the portfolio manager can diversify, more effectively, his risk with respect to the factors employed. As far as the second category is concerned, investment bankers, the usefulness of the results centres around the concept of the cost of capital or discount rate, which is a critical factor used by corporate financiers in several projects which have discounted cash-flow valuation as their backbone (such as capital budgeting and the valuation of privately and publicly owned companies). Despite the fact that there is no consensus among practitioners regarding the right model to use for estimating the cost of capital, traditionally, most applications have been employing the capital asset pricing model (CAPM) mainly due to its simplicity.
134
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
NOTES 1. The charter market occupies a central place in the structure of the shipping industry. Both shippers and shipping companies may prefer to hire a vessel rather than buy it outright. This may be due to the high capital cost of the vessel or the need to either cover cyclical peaks in demand for shipping capacity or replace ships out of service. 2. When borrowing from banks owners offer equity as part financing in a mortgage. 3. In charter backed finance shipowners borrow against the security of a long time charter agreement they have with a charterer. 4. Share prices included in the indices were adjusted for any rights issues, stock dividends and/or splits. 5. Accordingly, some companies known to have major shipping or shipping related interests were excluded because they were too diversified elsewhere. 6. For details of estimation methods and full set of results see Kavussanos and Marcoulis (1997a, b, c, 1998, 2000a, b, 2001), and particularly Kavussanos and Marcoulis (2001).
REFERENCES Arditti, F. (1967). Risk and the required rate of return. Journal of Finance, 22, 19–36. Ball, R. (1978). Anomalies in relationships between securities’ yields and yield – surrogates. Journal of Financial Economics, 6, 103–126. Banz, R. (1981). The relationship between returns and market value of common stocks. Journal of Financial Economics, 9, 3–18. Basu, S. (1977). Investment performance of common stocks in relation to the price-earnings ratios: A test of the efficient market hypothesis. Journal of Finance, 32, 663–682. Basu, S. (1983). The relationship between earnings, yield, market value and the return for NYSE common stocks: Further evidence. Journal of Financial Economics, 12, 1. Bhandari, L. (1988). Debt-equity ratio and expected common stock returns: Empirical evidence. Journal of Finance, 43, 507–529. Boudoukh, J., Richardson, M., & Whitelaw, R. (1994). Industry returns and the Fisher effect. Journal of Finance, 49(5), 1595–1615. Burmeister, E., & McElroy, M. (1987). APT and multifactor asset pricing models with measures and unobserved factors: Theoretical and econometric issues. Discussion paper, Department of Economics University of Virginia and Duke University. Burmeister, E., & McElroy, M. (1988). Joint estimation of factor sensitivities and risk premia for the APT. Journal of Finance, 43(3), 721–733. Burmeister, E., & Wall, K. (1986). The Arbitrage pricing theory and macroeconomic factor measures. The Financial Review (February). Chen, J., & Jordan, D. (1993). Some empirical tests in the APT: Macrovariables vs. derived factors. Journal of Banking and Finance, 17, 65–89. Chan, L., Hamao, Y., & Lakonishok, J. (1991). Fundamentals and stock returns in Japan. Journal of Finance, 46, 1739–1789. Chen, N., Roll, R., & Ross, S. (1986). Economic forces and the stock market. Journal of Business, 59, 383–403. Damodaran, A. (1994). Damodaran on valuation – Securities analysis for investment and corporate finance. New York: Wiley.
Cross-Industry Comparisons of the Behaviour of Stock Returns
135
Engerman, M. (1993). Using fundamental and economic factors to explain stock returns. BARRA Newsletter (Fall), 1993. Eun, C., & Resnick, B. (1992). Forecasting the correlation structure of share prices: A test of new models. Journal of Banking and Finance, 16, 643–656. Fama, E., & French, K. (1992). The cross-section of expected stock returns. Journal of Finance, 47, 427–465. Ferson, W., & Harvey, C. (1991). The variation of economic risk premiums. Journal of Political Economy, 99, 385–415. Gyourko, J., & Keim, D. (1993). Risk and return in real estate: Evidence from a real estate index. Financial Analysts Journal (September–October), 39–46. Hamao, Y. (1988). An empirical examination of the arbitrage pricing theory. Japan and the World Economy, 1, 45–61. Isimbabi, M. (1994). The stock market perception of industry risk and the separation of banking and commerce. Journal of Banking and Finance, 18, 325–349. Jones, C. (1993). Investments: Analysis and management (4th ed). New York: Wiley. The Jones Act Reform Coalition Report (1998). Published by the Jones Act Reform Coalition. Kahn, R. (1994). The E3 Project. BARRA Newsletter (Summer). Kane, E., & Unal, H. (1988). Change in market assessment of deposit-institution riskiness. Journal of Financial Services Research, 1, 207–229. Kavussanos, M. G., Arkoulis, A., & Marcoulis, S. (2002). Macroeconomic factors and international industry returns. Applied Financial Economics, 12, 923–931. Kavussanos, M. G., Juell-Skielse, A., & Forrest, M. (2003). International comparison of market risks across shipping related industries. Maritime Policy and Management, 30(2), 107–122. Kavussanos, M. G., & Marcoulis, S. (1997a). Risk and return of U.S. water transportation stocks over time and over bull and bear market conditions. Maritime Policy and Management, 24(2), 145–158. Kavussanos, M. G., & Marcoulis, S. (1997b). The stock market perception of industry risk and microeconomic factors: The case of the U.S. water transportation industry vs. other industries. Transportation Research, E33(2), 147–158. Kavussanos, M. G., & Marcoulis, S. (1997c). Risk, return and investment decisions. Lloyd’s Shipping Economist, Capital for Shipping, 8–11. Kavussanos, M. G., & Marcoulis, S. (1998). Beta comparisons across industries – A water transportation industry perspective. Maritime Policy and Management, 25(2), 175–184. Kavussanos, M. G., & Marcoulis, S. (2000a). The stock market perception of industry risk and macroeconomic factors: The case of the U.S. water and other transportation stocks. International Journal of Maritime Economics, 2(3), 235–256. Kavussanos, M. G., & Marcoulis, S. (2000b). The stock market perception of industry risk through the utilisation of a general multifactor model. International Journal of Transport Economics, XXVII(1), 77–98. Kavussanos, M. G., & Marcoulis, S. (2001). A market analysis of risk and return in the water and other transportation industries. Boston, MA: Kluwer. King, B. (1966). Market and industry factors in stock price behaviour. Journal of Business, 39, 139–190. Lakonishok, J., & Shapiro, A. (1986). Systematic risk, total risk and size as determinants of stock market returns. Journal of Business Finance, 10(1), 115–132. Lintner, J. (1965). Security prices, risk, and maximal gains from diversification. Journal of Finance, 587–615. Martinez, A., & Rubio, J. (1989). Arbitrage pricing with macroeconomic variables: An empirical investigation using Spanish data. Working Chapter, Universidad del Pais Vasco.
136
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
Nerlove, M. (1968). Factors affecting differences among rates of return on investment in individual common stocks. Review of Economics and Statistics, 50, 312–331. Neuberger, J. (1991). Risk and return in banking: Evidence from bank stock returns. Federal Reserve Bank of San Francisco Economic Review, 18–39. Poon, S., & Taylor, S. J. (1991). Macroeconomic factors and the UK stock market. Journal of Business, Finance and Accounting, 18, 619–636. Reinganum, M. (1981). Mis-specification of capital asset pricing: Empirical anomalies based on earnings yields and market values. Journal of Financial Economics, 9, 19–46. Roll, R. (1977). A critique of the asset pricing theory’s tests (Part I): On past and potential testability of the theory. Journal of Financial Economics, 4, 129–176. Rosenberg, B., Reid, K. R., & Lanstein, R. (1985). Persuasive evidence of market inefficiency. Journal of Portfolio Management, 11, 9–17. Rosenberg, J. (1991). The new Japanese equity model. BARRA Newsletter (November/December). Roy, V. (1992). BARRA releases new Canadian model. BARRA Newsletter (March/April). Saunders, A., & Yourougou, P. (1990). Are banks special? The separation of banking and commerce. Journal of Economics and Business, 42, 171–182. Sharpe, W. (1964). Capital asset prices: A theory of market equilibrium under conditions of risk. Journal of Finance (September), 425–442. Sharpe, W., Alexander, G., & Bailey, J. (1995). Investments (5th ed). Prentice-Hall International Editions. Sorensen, E., Salomon, R., Davenport, C., & Fiore, M. (1989). Risk analysis: The effect of key macroeconomic and market factors on portfolio returns. Salomon Brothers. Stattman, D. (1980). Book values and stock returns. The Chicago MBA: A journal of selected chapters, 4, 25–45. Wasserfallen, W. (1989). Macroeconomic news and the stock market – Evidence from Europe. Journal of Banking and Finance, 13, 613–626.
Cross-Industry Comparisons of the Behaviour of Stock Returns
137
APPENDIX A Summary of Major Findings in Kavussanos and Marcoulis (2001) Chapter 4: Risk and return of U.S. water transportation stocks over time and over bull and bear market conditions. Period Covered: January 1985 – December 1994 Methodology: CAPM Major Findings:
Industry Average Beta: 0.9199 (= 1) Industry Average Alpha: 0.00218 (> 0) Parameters exhibit stability over time No constant “size effect” over time Shift of alpha, but not beta, over bull and bear market conditions
Note: Figures in parenthesis indicate statistical equality or non equality to the number in the parenthesis Beta comparisons across industries – a water transportation industry perspective. Period Covered: July 1984 – June 1995 Methodology: CAPM Major Findings: Industry CAPM Parameters Industry Water transportation Air transportation Rail transportation Trucks Electricity Gas Petroleum refining Real estate
Alpha
Beta
0.0352 (> 0) 0.0124 (> 0) 0.0150 (> 0) 0.0206 (> 0) 0.0668 (> 0) 0.0447 (> 0) 0.0039 (> 0) 0.0260 (> 0)
0.9411 (< 1) 0.9748 (= 1) 1.0155 (= 1) 0.9676 (= 1) 0.9465 (< 1) 0.9581 (< 1) 0.9838 (= 1) 0.6933 (< 1)
Note: Figures in parenthesis indicate statistical equality or non equality to the number in the parenthesis.
138
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
The beta of the water transportation industry is significantly lower than the beta of the rail transportation industry and significantly higher than the beta of the real estate industry.
Chapter 5: The stock market perception of industry risk and microeconomic factors: The case of the U.S. water transportation industry vs. other transport industries. Period Covered: July 1984 – June 1995 Methodology: Multifactor Model employing fundamental microeconomic factors Major Findings: Industry Multifactor Model Parameters Industry Water transportation Air transportation Rail transportation Trucks Electricity Gas Petroleum refining Real estate
Alpha
Beta
0.0420 (> 0) 0.0030 (> 0) 0.0080 (> 0) 0.0210 (> 0) 0.0770 (> 0) 0.0650 (> 0) 0.0230 (> 0) 0.0280 (> 0)
0.9410 (< 1) 0.9760 (= 1) 1.0110 (= 1) 0.9680 (= 1) 0.9420 (< 1) 0.9520 (< 1) 0.9760 (= 1) 0.6890 (< 1)
ME B/M A/ME A/BE E/P *
+ − +
− −
+ +
+
−
Note 1: Figures in parenthesis indicate statistical equality or non equality to the number in the parenthesis. Note 2: Where the sign is positive, this means that there is a positive relationship between that factor and returns. Where the sign is negative, the opposite holds. The magnitude of each factor is discussed in detail in the relevant chapter. Where no sign appears, the coefficient is statistically insignificant. ∗ ME, B/M, A/ME, A/BE and E/P correspond to market value of equity, book to market, total assets to market equity, total assets to book equity and earnings to price ratios respectively.
The beta of the water transportation industry is significantly lower than the beta of the rail transportation industry and significantly higher than the beta of the real estate industry. The only industry beta which exhibits significant temporal variability is that of the petroleum refining industry. In the water transportation and the other six industries no significant temporal change has occurred.
Cross-Industry Comparisons of the Behaviour of Stock Returns
139
Chapter 6: The stock market perception of industry risk and macroeconomic factors: The case of the U.S. water and other transportation stocks. Period Covered: July 1985 – June 1995 Methodology: Multifactor Model employing macroeconomic factors Major Findings: Industry Multifactor Model Parameters Industry Water transportation Air transportation Rail transportation Trucks Electricity Gas Petroleum refining Real estate
Alpha
Beta
0.0346 (> 0) 0.0103 (> 0) 0.0289 (> 0) 0.0176 (> 0) 0.0642 (> 0) 0.0426 (> 0) 0.0320 (> 0) 0.0348 (> 0)
0.9449 (< 1) 0.9538 (= 1) 0.9875 (= 1) 0.9698 (= 1) 0.9248 (< 1) 0.9579 (< 1) 0.9741 (= 1) 0.7543 (< 1)
MIP UTS UOG UCG UI* −
−
+ + − +
+ − + + −
− −
Note 1: Figures in parenthesis indicate statistical equality or non equality to the number in the parenthesis. Note 2: Where the sign is positive, this means that there is a positive relationship between that factor and returns. Where the sign is negative, the opposite holds. The magnitude of each factor is discussed in detail in the relevant chapter. Where no sign appears, the coefficient is statistically insignificant. ∗ MIP, UTS, UOG, UCG, UI correspond to unanticipated changes in monthly industrial production, the term structure, oil prices, consumption and inflation respectively.
The beta of the water transportation industry is not significantly different to the beta of any other transportation or non – transportation industry.
140
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
Chapter 7: The stock market perception of industry risk through the utilisation of a General Multifactor Model. Period Covered: July 1985 – June 1995 Methodology: Multifactor Model employing both micro and macroeconomic factors Major Findings: Industry Multifactor Model Parameters Industry Water transportation Air transportation Rail transportation Trucks Electricity Gas Petroleum refining Real estate Industry Water transportation Air transportation Rail transportation Trucks Electricity Gas Petroleum refining Real estate
Alpha
Beta
0.0334 (> 0) 0.0206 (> 0) 0.0289 (> 0) 0.0228 (> 0) 0.0710 (> 0) 0.0603 (> 0) 0.0197 (> 0) 0.0348 (> 0)
0.9438 (< 1) 0.9471 (= 1) 0.9878 (= 1) 0.9593 (= 1) 0.9264 (< 1) 0.9580 (< 1) 0.9676 (= 1) 0.7543 (< 1)
MIP UTS −
−
+ + − +
ME
−
UOG + − − + + −
UCG
UI
− −
B/M A/ME A/BE E/P + − − + +
+
−
Note 1: Figures in parenthesis indicate statistical equality or non equality to the number in the parenthesis. Note 2: Where the sign is positive, this means that there is a positive relationship between that factor and returns. Where the sign is negative, the opposite holds. The magnitude of each factor is discussed in detail in the relevant chapter. Where no sign appears, the coefficient is statistically insignificant.
The beta of the water transportation industry is not significantly different to the beta of any other transportation or non – transportation industry.
Empirical Evidence Regarding the Macroeconomic Factors Employed in Kavussanos and Marcoulis (2001). Factor Employed
Study
Findings of Study
Findings of this Book
Applicable Industry
Monthly growth of industrial production
Poon and Taylor (1991) Bong – Soo Lee (1992) Chen, Roll and Ross (1986) Pearce and Roley (1985) Chen and Jordan (1993)
− ve effect + ve effect + ve effect n.s. n.s.
− ve effect − ve effect
Water transportation Electricity
Unanticipated changes in term structure
Chen, Roll and Ross (1986) Poon and Taylor (1991) Chen and Jordan (1993)
+ ve effect n.s. n.s.
+ve effect +ve effect +ve effect − ve effect
Air transportation Trucks Petroleum refining Electricity
Unanticipated changes in oil prices
Chen, Roll and Ross (1986) Chen and Jordan (1993)
+ ve effect − ve effect
+ve effect +ve effect +ve effect − ve effect − ve effect
Water transportation Gas Petroleum refining Trucks Real estate
Unanticipated changes in consumption
Rubinstein (1976) Lucas (1978) Breeden (1980) Wasserfallen (1989)
+ ve effect + ve effect + ve effect − ve effect
− ve effect − ve effect
Rail transportation Gas
Unanticipated inflation
Chen, Roll and Ross (1986) Hamao (1988) Wasserfallen (1989) Poon and Taylor (1991) Martinez and Rubio (1989)
− ve effect − ve effect − ve effect n.s. n.s.
n.s.
Cross-Industry Comparisons of the Behaviour of Stock Returns
APPENDIX B
All industries
141
142
APPENDIX C Empirical Evidence Regarding the Microeconomic Factors Employed in Kavussanos and Marcoulis (2001). Study
Findings of Study
Findings of this Book
Applicable Industry
Market value of equity
Banz (1981) Basu (1977, 1983) Reinganum (1981) Lakonishok and Shapiro (1986) Fama and French (1992) Stattman (1980) Rosenberg et al. (1985) Fama and French (1992) Chen et al. (1991) Fama and French (1992) Bhandari (1988) Fama and French (1992)
− ve effect − ve effect − ve effect − ve effect − ve effect + ve effect + ve effect + ve effect + ve effect + ve effect + ve effect − ve effect
+ ve effect + ve effect − ve effect
Water transportation Petroleum refining Gas
+ve effect +ve effect
Electricity Real estate
Ball (1978) Reinganum (1981) Basu (1983) Fama and French (1992)
+ ve effect + ve effect + ve effect n.s.
+ve effect +ve effect − ve effect − ve effect − ve effect n.s.
Rail transportation Petroleum refining Water transportation Air transportation Petroleum refining All industries
Book to market value of equity
Asset to market value of equity Asset to book value of equity
Earnings to price
MANOLIS G. KAVUSSANOS AND STELIOS N. MARCOULIS
Factor Employed
5.
THE FISCAL TREATMENT OF SHIPPING: A CANADIAN PERSPECTIVE ON SHIPPING POLICY
Mary R. Brooks and J. Richard Hodgson 1. INTRODUCTION Since the mid-1990s, attention has been focused by most of the leading maritime countries, particularly those in Europe and more recently the U.S., on the increasing dominance of open registry vessels in international shipping at the expense of the closed registry fleets of traditional maritime nations. The trend has been observed with increasing discomfort by many governments as they have suffered a steady erosion of shipping knowledge and expertise. A number of European countries have attempted to mitigate this trend by implementing initiatives intended to encourage the positioning of effective ownership, management, and to some extent registration, of ships in their respective countries, and to stimulate recruitment and training. Initiatives, including fiscal relief for ship owners through a tonnage tax regime, simplification of registration procedures and tax relief for seafarers have been implemented. The concerns that these steps have been designed to address are not new to Canada. In the early 1970s, Canada initiated a series of studies to decide upon Canada’s international shipping policy and to evaluate the measures that needed to be introduced to implement that policy. A review of these studies is briefly
Shipping Economics Research in Transportation Economics, Volume 12, 143–171 Copyright © 2005 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0739-8859/doi:10.1016/S0739-8859(04)12005-2
143
144
MARY R. BROOKS AND J. RICHARD HODGSON
undertaken in the next section. Following the Report of the Task Force on Deep-Sea Shipping (Transport Canada, 1985), Canada adjusted its tax regime as it applied to international shipping activities; and the government encouraged the establishment of several International Maritime Centres (IMCs). The approach Canada took in its treatment of international shipping was unique. Now that more than a decade has passed since the 1991 tax changes, this paper will explore the “Canadian way” in light of more recent legislative changes by its OECD partners. Many have opted to implement a tonnage tax regime, together with other adjustments, as a means of beating back the growing incidence of national ship owners flagging out to open registries. The purpose of this paper is to examine the models that have been adopted by developed countries to address the issue of the fiscal treatment of international shipping in a globalized financial world. To round out the topic, the paper will also briefly discuss the extension of this fiscal treatment to domestic shipping.
2. SHIPPING POLICY IN CANADA TO 1985 Canada’s participation levels in the shipping industry have been both volatile and cyclical. At the end of World War I, the government established the Canadian Government Merchant Marine Limited, which, by the early 1920s, was operating over 60 ships on a world-wide basis. This pre-eminent position was not to last, and the company was forced to close in 1936. By the end of the Second World War, which too provided a significant stimulus to Canadian shipbuilding, the publiclyowned Park Steamship Company Limited operated 150 Canadian-built, Canadianregistered, Canadian-manned ships – the fourth largest merchant fleet in the world. A post-war decline in the fortunes of Canadian ship owners also occurred after World War II. Changing technology, growing disparities in wage levels between developed and developing countries and the emergence of low-cost flags of convenience meant that, by 1969, Canadian ocean-going merchant tonnage totaled no more than 70,000 GRT, with only four vessels in this fleet exceeding 1,000 GRT! This situation sparked a series of studies into the form that Canada’s international shipping policy should take, and, more important for this paper, what the Canadian government might do in terms of fiscal measures to support that policy on the international shipping side. Hedlin Menzies & Associates Ltd. (1970) was the first in a series of studies that would lead to the development of Canadian international shipping policy as it exists today. It examined “the economic feasibility of developing a modern Canadian flag deep-sea merchant fleet.” The report argued that trading interests and ship-owning interests were not always convergent, and started the long debate as to what policy would be in the best interests of the Canadian public.
The Fiscal Treatment of Shipping
145
It was followed by Darling (1974, p. i), a study with a focus on examining the “needs, scope and implementation of an international shipping policy.” Darling argued that Canadian shipping policy was poorly conceptualized, and that a reorientation was needed before any policy could be effective. Specifically, a clear distinction needed to be made between Canada’s core interests and its advantages where shipping was concerned; while a national flag fleet might have represented the latter, its absence was no threat to the former, which was trade-based (Darling, 1974, pp. 2, 3). Darling concluded that Canada should base its policy on core interests, not advantages; given the dynamic nature of the international environment, Canada needed an improved policy and organizational structure to cope with the resulting challenges. The authors regard the Darling report as a thoughtful and far-sighted study; it articulated well the conceptual, policy, and organizational difficulties challenging Canada’s approach to international shipping at the time. These two studies were followed by two more that examined the economic benefits and commercial requirements of a Canadian flag fleet (Alcan Shipping Services Ltd, 1977; Department of Finance, 1978). The former, a consulting report, was forced to make a large number of assumptions, leading to questions about the reliability of its findings. The latter, a report of a government department to an industry advisory body (the Shipping Advisory Board), had a broader mandate, including repeating the analysis of the Alcan study (Department of Finance, 1978, p. iv). The Department of Finance report concluded that Canadian flag services could not be operated profitably, and even in the event that foreign cargo preference laws did result in potentially higher costs for Canadian shippers, the introduction of Canadian-flag shipping would not represent an “appropriate policy response.” The Department of Finance considered that Canadian trading interests had little to gain – and possibly much to lose – by the establishment of a Canadian-registered fleet. The Department of Finance report, as a government document, was the first in a series of documents whereby the Canadian government publicly struggled with the issue of international shipping policy and its support. Transport Canada (1979, pp. 39, 40) concluded that the significant changes in international shipping at that time implied that Canada could no longer continue to rely automatically on market forces to ensure the availability of low-cost shipping. The declared purpose of Transport Canada (1980) was “to set out the main elements of a deep-sea shipping policy for Canada, taking into account recent and anticipated developments in this country and abroad.” This policy document signaled to the nation that the government viewed the public interest as more focused on trade-protection and growth than on the maintenance of a shipping industry. Transport Canada’s focus for the next few years shifted to examining the merits of adoption of a defensive deep-sea shipping policy. In the end, Transport Canada (1983) concluded that defensive legislation was inappropriate as only a small share
146
MARY R. BROOKS AND J. RICHARD HODGSON
of Canada’s trade was affected by the discriminatory practices of other nations and Canada’s vital interests did not seem to be imperiled. However, the report recognized that, although the impact of foreign interventions appeared to be limited at the time, there was no guarantee that this would remain the case. This led to the formation of the Task Force on Deep-Sea Shipping. Prior to the formation of the Deep-Sea Task Force, Brooks and Marlow (1983) examined Canadian fiscal policy as it applied to shipping for the federal government. Based on the methodology used in fiscal studies for the U.K. government (discussed later) and reported by Gardner and Marlow (1983), the study evaluated the existing Canadian fiscal climate against that provided in West Germany, Japan, the United Kingdom and the United States. It also examined four variants to the existing Canadian fiscal regime. The paper concluded that, if the government wanted to improve the fiscal climate to encourage registration of deep-sea shipping in Canada, it would need to improve the fiscal regime for foreign-built ships, and to standardize safety requirements to international levels so that the capital costs of safety investments were not higher for Canadian owners than for foreign owners. The Brooks and Marlow (1983) findings were rendered moot by the government’s decision shortly thereafter both to establish the DeepSea Task Force in April 1984 and to enter into a new Canada-U.S. income tax treaty in August of 1984 (Lambe, 1985, p. 123). According to Lambe (1985), many Canadian and U.S. ship owners felt that they could not remain internationally competitive if ships were registered at home. Both countries therefore granted limited tax exemptions for nationals operating vessels through non-resident or foreign corporations. A Canadian company could be declared a resident, even if it had no offices in Canada, if its Board of Directors opted to meet in Canada. If it chose to structure its business such that nonresident status was maintained, income tax became payable on the dividends when remitted to Canada. The exemption was granted if the country where the foreign corporation resided granted similar relief to a Canadian resident. Under the new treaty, Canada and the U.S. allowed reciprocity in this arrangement. A Canadian owned international shipping business, as long as it was domiciled in the U.S. and conducted its international shipping business from the U.S. and its Board met outside Canada could rely on its exemption from income tax as a non-resident shipping company; the reverse was only partly true for U.S. firms as the basis of the exemption was the use of a foreign flag and the earnings had to be derived from a foreign corporation (the activity did not need to be international shipping). In other words, “U.S. domestic law does not enjoy the same harmony with the treaty as does Canadian domestic tax law (Lambe, 1985, p. 127).” The advantage did not extend to domestic shipping activities in either country and Lambe concluded that ship owners would continue to adopt flags of convenience. The new tax treaty,
The Fiscal Treatment of Shipping
147
therefore, did not really address the Canadian desire to protect its international trading interests through the development of maritime expertise in Canada. The Task Force on Deep-Sea Shipping was mandated to “evaluate changing conditions in the international shipping market and the possible need for measures to encourage the expansion of the Canadian deep-sea fleet (Transport Canada, 1985, p. 3).” It was also the last substantive examination of deep-sea shipping policy undertaken by the government. (Some might argue otherwise, citing the most recent Canada Marine Act Review Panel for example, but the authors maintain that all studies since this report have been related to either competition policy as it applies to liner shipping or the provision of infrastructure, not of shipping services.) The Task Force concluded that it was vital that Canada have a strong base of shipping expertise in order to ensure continued smooth flow of exports. The focus of such expertise would be one of being astute buyers of maritime transportation services, this astuteness being derived from industry knowledge. Such industry knowledge would also be garnered through the creation of a financial environment that was more attractive to shipping companies. On the one hand, the Task Force (Transport Canada, 1985) expressed strong reservations regarding adoption of a policy directed at encouraging a Canadian flag deep-sea fleet by direct operating subsidies. On the other, it provided strong support for Canadian involvement in the management of foreign flag shipping, as the means by which Canada could encourage and strengthen its expertise and interests in international shipping. In the latter case, it recommended important fiscal adjustments that included the creation of “International Shipping Corporations” (ISCs) as the proposed mechanism for achieving these fiscal objectives. The ISC structure that evolved, while somewhat different from that envisaged by the Task Force, nevertheless aspired to the same goals. It is in place today and is the subject of the next section of the paper. In summary, Canada concluded that the costs of developing a Canadian flag fleet, mainly via subsidies in one form or another, were seen to outweigh what were considered comparatively modest benefits in the form of balance of payments contributions, employment creation, stimulus to employment in auxiliary services, and strategic defence considerations. However, Canada needed to ensure the continuance of free and fair competition through careful monitoring of trade practices on Canada’s import and export routes. In order to achieve this objective, it was imperative that Canada maintain expert and practical knowledge and experience in the wide range of commercial activities relating to ocean shipping, of which ship ownership was but one element. Since the development of the requisite commercial environment could not be achieved through the operation of Canadian-flagged ships, then the establishment in Canada of the effective ownership, management and control of foreign flag ships would provide the means.
148
MARY R. BROOKS AND J. RICHARD HODGSON
It was to this end that the Task Force report recommended legislation to support the establishment of ISCs.
3. THE ISC APPROACH The fiscal treatment of shipping in Canada, prior to the ISC solution proposed by the Deep-Sea Task Force (Transport Canada, 1985), was very restrictive. The Canadian government applied a “mind and management” test that resulted in Canadian corporations being taxed on their world-wide income. For an industry where most companies pay little or no tax, this was a major barrier to establishing an international shipping business in Canada. The Task Force therefore recommended that Canadian tax law be modified in order to create a financial environment that was more attractive to shipping companies, through the implementation of legislation that allowed for the creation of International Shipping Corporations. It was further recommended that an ISC be majority owned by Canadian residents, and be managed in Canada. The ISC would be permitted to own ships registered under any flag. As well, if the ISC earned at least 90% of its income from international shipping, it would not be taxed in Canada on its income. It was believed that this would act as a clear signal that successful shipping companies could operate from within Canada. Canada’s Department of Finance response to the Deep-Sea Task Force recommendation was relatively swift; it took the position that to grant an exemption from Canadian income tax for a business enterprise resident in Canada would set an unacceptable precedent for the Department. Furthermore, the Department was engaged in discussions on corporate tax restructuring and corporate tax reform, and there was concern about the use of ISCs to avoid taxes. Clarkson Gordon (as reported by Jonathan Seymour & Associates, 1988, Appendix II) undertook an assessment of the ISC model in the context of competing tax regimes and in particular that of the U.S. Their resulting Memorandum to the Department recommended that incorporation in Canada not be required and that the ISC only needed to have a majority of Canadian resident directors. To qualify as an ISC, the recommendation was that 75% of salaries must be paid to Canadian residents. From a tax perspective, the ISC issue became ensnared in a larger debate on tax reform and it was to be some time before resolution on that front would be forthcoming. At the same time as these discussions with the Department of Finance were taking place, there were efforts being made within Transport Canada to establish a Canadian position on the United Nations Convention on Conditions for the Registration of Ships, 1986. Until the details of the convention were finalized, any
The Fiscal Treatment of Shipping
149
policy decisions on the treatment of Canadian flag vessels were also effectively precluded. In 1988, Transport Canada examined the continuing development in Europe and elsewhere of alternative fiscal arrangements for the shipping industry in a small study of the second registry option (Abbott, 1988) but the report did not make recommendations to the Government about whether establishing a second registry, as had been done in Norway, would best meet Canada’s trade and transport objectives. The report provided additional information for the ongoing debate and, despite the absence of any recommendation to do so, a second registry was given serious consideration. Meanwhile, on the west coast of Canada, progress was seen as too slow. In 1988, the Vancouver-based Transportation Task Force of the Asia-Pacific Initiative (a joint federal-provincial task force) hired Jonathan Seymour & Associates to undertake a study with the primary objective of creating . . . an internationally competitive business environment suitable for the ownership, management and control of ships engaged in international trade; one which will permit Vancouver to achieve its natural potential as a major international centre for maritime commerce (Jonathan Seymour & Associates, 1988, p. 3).
The study concluded that the primary constraint on Vancouver becoming an international shipping centre was the tax structure applicable to the industry in Canada. Over the course of the next three years, the debate led to eventual amendments to Canada’s Income Tax Act (Bill C-18), which received royal assent on December 17, 1991. These amendments provided for the possibility that a company incorporated in another country, but located in Canada, where all or substantially all of its activities were primarily involved with the movement of international traffic and whose revenues were derived totally or at least substantially from the operation of ships in transporting international traffic would be considered, for tax purposes, not to be a resident of Canada, but rather to be a resident of its country of incorporation. (This was a quite elegant, if complex, solution to the problem the Department of Finance faced with respect to establishing a precedent.) This provision did not preclude, however, the possibility that a corporation might also be a resident in a country other than Canada or the country of its incorporation. In 1998, Section 250(6) of the Canada Shipping Act was modified to recognize the frequent use of holding companies and single-purpose/single-ship corporations by international shipping companies. Following these changes, where an international company controlled ships through a wholly-owned corporation which itself would qualify as a non-resident, it could be treated as if it were operating the ships themselves in order to establish non-residency for tax purposes.
150
MARY R. BROOKS AND J. RICHARD HODGSON
Fig. 1. The International Shipping Corporation. Source: Appendix 2 from Seymour, J. (1998, p. 184). It is used with permission.
The relationship between the two critical sections of the Income Tax Act (and therefore the determination of who qualifies as an ISC) is best explained by Fig. 1. There were also changes made to the customs tariff following representation by the International Maritime Centre Vancouver (1994). ISCs were able to purchase vessels offshore and, as long as these vessels remained in the deep-sea trades, their ISC owners would not be required to pay the customs duty applicable to vessels operating in the Canadian coasting trade. There were no changes made to the fiscal
The Fiscal Treatment of Shipping
151
treatment of domestic shipping at the time. Canadians therefore had to be satisfied with the shore-based maritime services jobs created by the ISC legislation and the limited number of sea-going jobs reserved for Canadians by protecting domestic shipping for Canadian flag vessels. It is particularly interesting to note that the majority of vessels using the ISC legislation today are in the tanker and dry bulk trades. Cliff Pratten, an economist at Cambridge University, in a recent study commissioned by the Baltic Exchange (Anonymous, 2002a, p. 10), noted that “Ship owners who pay taxes are not competitive in the tanker and dry cargo trades and, in the long run, would be forced out of the business.” Through the ISC legislation, Canada had made a move to secure participation in the trade, and shore-based jobs in this sector of shipping, well before the European Community examined options for its Member States. Today, there are those who are ardent supporters of the ISC concept, but also others who are frustrated that the changes did not go further and provide conditions that would enable such capacity to be located in Canada, and ultimately under Canadian registry. This paper will return to the evaluation of this approach for international shipping in the discussion. In summary, the ISC legislation has gone some way towards achieving the policy objectives set for it, although the degree of achievement is viewed by some to be relatively modest. Its supporters may legitimately argue that the majority of the estimated 500 jobs, as identified by the authors through discussion with those in the industry, would not have been created in Canada without the legislation.
4. THE FISCAL TREATMENT OF CANADIAN DOMESTIC SHIPPING Not all countries treat international and domestic shipping similarly for tax purposes. Canada has protectionist cabotage policies when it comes to domestic shipping; traffic between two Canadian ports must be carried in a Canadian flagged ship on which all applicable duty has been paid. While it is possible for foreign flag vessels to enter the Canadian domestic trades under waiver (granted when no suitable Canadian flag ship can be found after a diligent search has been conducted), domestic shipping is effectively segregated from foreign flag shipping and taxed in line with Canadian domestic corporations. Canadian flag vessels must have crews with Canadian certificates, and such certification is available only to Canadian citizens and permanent residents. When adjustments to this requirement were proposed in 1998 as part of the ISC amendments, there was no appetite within the unions or within the walls of Canadian industry association offices to stimulate long-term opportunities for sea-going
152
MARY R. BROOKS AND J. RICHARD HODGSON
jobs by liberalizing the crewing requirements. Many have argued that Canada’s domestic shipping must be protected to preserve sea-going jobs for Canadians, jobs otherwise threatened by lower cost sources of labour. (As U.S. crews were more expensive, the jobs would not have been lost to American flag vessels.) One of the issues, originally raised by Brooks and Marlow (1983) as influencing the choice of Canadian-flag deep sea shipping, continues today under the fiscal climate facing Canadian domestic shipping. Canadian vessel safety standards are often viewed by Canadian operators as unique, onerous, and involving costs sufficient to make the Canadian flag unattractive in relation to other flags that are held to less prescriptive international safety standards. Canadian vessel standards are thus regarded as a deterrent to the development of Canadian flag shipping. The tax climate for Canadian domestic shipping is sufficiently onerous that it effectively renders it uncompetitive with foreign flag shipping. The Income Tax Act does not treat domestic shipping as a special case and Canadian shipping companies, as for any Canadian manufacturing enterprise, are subjected to the full extent of corporate income tax. Furthermore, the capital cost allowance on Canadian vessels is only 33.3%, 16.67% in the year of acquisition, much lower rates than are found in other jurisdictions. Most damaging of all is the 25% customs tariff on all foreign-built vessels imported for use in Canadian trades, a point made throughout the years by numerous studies on the issue. Finally, unlike many other jurisdictions, Canada does not allow any tax-free reserves; the only exception to this is a reserve for the expenses of the vessel survey required every four years. Only U.S. costs are higher, and Canada’s cabotage rules, therefore, are necessary to protect domestic trades from foreign flag incursion. Perhaps the most important aspect to draw forward into the discussion later in this paper is that Canada taxes all domestic companies on their worldwide income. This is certainly quite different from the situation in other countries. For example, Hong Kong and Singapore treat international income as exempt from the base on which domestic corporate tax is calculated. The critical issue for Canada moving forward will be whether to treat domestic shipping more like international shipping or to maintain the existing distinct separation in treatment between the two. This issue will enter into deliberations if Canada contemplates a move towards an alternate fiscal regime, like the tonnage tax approach currently used in a number of European countries. The next section will explore the tonnage tax option.
5. FISCAL TREATMENT IN EUROPE: THE TONNAGE TAX APPROACH As the oversight role of the European Commission in maritime transport evolved after the signing of the Treaty of Rome, the Commission became increasingly
The Fiscal Treatment of Shipping
153
concerned with the decline in the traditional shipping registers of its Member States. This manifested itself in various ways since 1986. The first step to establish a substantive European policy for shipping was the 1986 package of measures liberalizing the internal (intra-Community) market, revising conditions for foreign flag operators calling at EC ports, and addressing EC competition policy. Although the full scope of proposed cabotage reforms was not finally completed until 1999, substantial progress had been made by 1996, when the European Commission released its watershed policy paper entitled Towards a New Maritime Policy (European Commission, 1996). The revised shipping policy objective was stated as: “to ensure freedom of access to shipping markets across the world for safe and environmentally friendly ships, preferably registered in EC Member States with Community nationals employed on board” (European Commission, 1996, p. 1.1). The document presented a two-fold strategy: to ensure safety and fair competition in international open markets, and to establish a Community framework for enhancing shipping competitiveness. This framework encompassed policy support for training and employment (to develop and safeguard maritime expertise), enhancement of research and development, and the provision of state aid. Guidance on state aid was first provided in guidelines promulgated in 1989, and subsequently revised in 1997. The State Aid Guidelines (European Commission, 1997) endorsed the longterm aim as being to ensure freedom of access to shipping markets across the world for safe and environmentally friendly ships. While accepting this free market principle, the EC also reaffirmed its preference that, as far as possible, such ships should be registered in EC Member States with Community nationals employed on board (European Commission, 1997, p. 1.1). The principal objectives were declared to be to safeguard EC employment (both on board and ashore), to preserve maritime know-how and skills in the Community, and to improve safety (European Commission, 1997, p. 2.2). The guidelines recognized, however, that it was relatively expensive to operate EC registered ships with EC seafarers on board, and that this situation inhibited the pursuit of high quality operations and undermined the goal of safe, efficient, environmentally friendly transport. It was these circumstances that led to the fundamental decision taken by the European Commission to create conditions that allowed fair competition with flags of convenience. The fiscal costs, mainly corporate taxation and wage-related liabilities, were concluded to be the critical and distorting factors (European Commission, 1997, p. 1.4). The approach was to reduce contribution rates on social protection and income tax for EC seafarers employed on EC-registered vessels. With regard to fiscal treatment, the Guidelines endorsed a number of measures on the basis that they were shown to safeguard high quality employment. These included accelerated depreciation and the right to reserve profits made on the sale
154
MARY R. BROOKS AND J. RICHARD HODGSON
of ships for a number of years on a tax-free basis provided that they were eventually re-invested in the acquisition of ships. More notably, the guidelines also endorsed the replacement of the normal corporate tax system by a tonnage tax. Tonnage tax is an alternative method of calculating corporate tax profits by reference to the net tonnage of the ship operated. It ignores actual profit and, rather, computes a notional profit on the basis of the number and size of ships operated and taxes this notional profit. Therefore, tonnage tax profit replaces both the taxadjusted commercial profit (or loss) on a shipping trade and the chargeable gains (or losses) made on tonnage tax assets. The tax rate is set so that notional profits, and hence actual corporate tax paid, are minimal. Other profits of a tonnage tax company are taxable in the normal way. This approach is attractive not only because of the low level of taxation, but also because it may be accurately predicted and easily calculated. Hence, to quote Lord Alexander (HM Treasury Release, 1999, Para. 27): The mechanism seems to be an ingenious device for obtaining virtual tax exemption compatible with international tax treaty obligations.
The State Aid Guidelines (European Commission, 1997) made clear that, since their objective was to promote the competitiveness of the EC fleets, fiscal alleviation schemes should require a link with a Community flag as a rule. However, the guidelines accepted that such schemes might also exceptionally be approved where they applied to the entire fleet operated by a ship owning company established within a Member State’s territory. Approval depended upon it being demonstrated that the strategic and commercial management of all ships concerned was effectively carried out within the territory, and that this activity contributed substantially to economic activity and employment within the Community. Furthermore, in such instances, the aid had to be necessary to promote the repatriation of the strategic and commercial management of all ships concerned to the EU and the beneficiaries of such schemes had to be liable to payment of corporate tax in the Community. In such cases, the Commission would also require the provision of regular reports, and would closely monitor the situation to ensure it did not distort competition between Member States (European Commission, 1997, p. 3.1). Also, such schemes were intended to benefit only the shipping sector, and so the fiscal advantages offered had to be restricted to shipping activities by suitable “ring-fencing” provisions. What has been the experience to date with the tonnage tax? Although Greece has had a tonnage tax regime in place since 1938, the past comparatively high level of that tax meant it did not attract much attention for nearly five decades. In this respect, the first European State to adopt, as an option, a tonnage tax pitched at a level to compete with flag of convenience tax rates was the Netherlands in
The Fiscal Treatment of Shipping
155
1996. The Dutch government was thus the first to conclude that the decline in national shipping, maritime skills and maritime-related economic activity could not be halted by its current measures, which subsidized ship owners. The new approach relied instead on creating an attractive climate in which shipping was seen as the nucleus of the country’s maritime business cluster. With an optional tonnage tax as the central element of the Dutch policy package (reflecting the new emphasis on Dutch ownership rather than the Dutch flagged fleet), labour cost distortions were addressed by flag-related wage subsidies coupled with increased flexibility in manning rules (U.K. Department of Environment, Transport and the Regions, 1998). The dominant view in the Netherlands now is that the new initiatives have been an outstanding success. Despite the fact that ships did not need to fly the Dutch flag in order to qualify, the Dutch flag fleet grew by around 20% during the first year of application. Indeed, the demand for Dutch seafarers became so great that it created a significant shortage, to the point where the rules requiring Dutch ships to have a Dutch master had to be changed to authorize masters from other EU Member States (The Transportation Institute, 2003). The results have been viewed as so successful that the Netherlands is now exploring ways of extending the provisions to other sectors of the maritime industry. Very shortly thereafter, Norway introduced a tonnage tax regime modeled on the system introduced by the Netherlands. This move followed adoption of, and added support to, its two register concept, the Norwegian Ordinary Ship Register and the Norwegian International Ship Register, which had been in place since 1987. The introduction of the tonnage tax resulted in a significant growth in the Norwegian owned and operated fleet, aided by the fact that seafarers were entitled to allowances on their pre-tax income, and by the introduction of a “netwage” system. Since that time, however, there has been significant turbulence in Norwegian policy-making regarding fiscal and labour-related costs for the shipping industry, possibly simply because the success of the measures raised questions as to whether they were overly generous. This has resulted in adjustment to aid levels, either up or down, by successive governments. At present there exists an uneasy compromise, and the expectation of possible transfers of Norwegian-flagged vessels to the more attractive Danish register, for example (The Transportation Institute, 2003). Germany introduced an optional tonnage tax system in early 1999, again modeled on the Netherlands’ approach. Provisions, approved by the EC as meeting its Guidelines, were also introduced that permitted retention of a percentage of income tax, provided support in reducing non-wage labour costs, and stimulated expanded maritime training. Germany was followed in early 2000 by the United Kingdom.
156
MARY R. BROOKS AND J. RICHARD HODGSON
The U.K. move towards a tonnage tax regime had its genesis in the increasing concern about the decline of the U.K. shipping industry. In particular, there was considerable concern about the ability of the industry to withstand the changing demographics of the U.K. seafaring community, and the estimated labour contraction that was projected in studies by the University of Wales (1996) and McConville et al. (1997). The publication of British Shipping – Charting a New Course by the U.K. Department of Environment, Transport and the Regions (1998) was a critical step in the development of the fiscal reform package eventually adopted. The report stressed the importance of U.K. interest and involvement in shipping, not only from an economic perspective, but also in relation to safety and security considerations, and protection of the environment. It recognized that the support that had been provided, and the deregulation initiatives that had been taken to date, had not been effective in reversing the decline of the maritime sector. Taking the position that continued acquiescence to the steady erosion of the U.K.’s core maritime capability was not a tenable long-term policy, the paper concluded that a more interventionist policy was required. It proposed following the successful initiatives taken by certain European countries; that these should form the foundation for a new integrated maritime policy. The paper set out a comprehensive strategy focused on four broad objectives: increasing skills, encouraging employment, increasing the U.K.’s attractiveness to shipping enterprises, and gaining safety and environmental benefits. While initially the policy did not commit to fiscal reform, in 1999, the government initiated a comprehensive examination of the tonnage tax concept by Lord Alexander of Weedon. Lord Alexander (HM Treasury Release, 1999, Para. 97) concluded that: [The] Government’s policy for shipping cannot be achieved without creating a tonnage tax regime to ensure a fiscal environment which is, and is perceived to be, user-friendly . . .. The existing structure is unsatisfactory. It acknowledges the principle of subsidy, provides substantial backing to the industry, yields comparatively little tax, but is not commercially attractive to companies or investors. . . . Whilst the economic case alone is not wholly convincing, and the success of the policy in a highly competitive market is far from assured, the future of our shipping industry is now at a critical point. Without attempting the tonnage tax, further decline seems inevitable and may soon reach a point where it will become irreversible.
Tonnage tax legislation was enacted in July 2000, and Lloyd’s Shipping Economist (Anonymous, 2002a) has attributed to this initiative the subsequent 89% growth in the U.K.-registered fleet. Following on the heels of the U.K., a number of other European States have now opted to adopt a tonnage tax regime coupled with various labour-related aid measures. Ireland, Finland, Denmark and Spain have all now introduced such
The Fiscal Treatment of Shipping
157
measures, while France and Belgium have announced plans to do so, and are currently seeking EC approval for their proposals. France has also agreed to set up a new French International Register to replace the uncompetitive Kerguelen flag (Fairplay, 2003a). While mainstream support for a tonnage tax regime is clearly strong in Europe, there is skepticism in some quarters. Policy turbulence in Norway and Finland suggest differences of view as to what the optimum balance should be. Sweden, although it already has in place an attractive “net wage” system, decided in 2003 not to proceed with the introduction of a tonnage tax regime (Fairplay, 2003b). In Italy, no decision has been taken to introduce a tonnage tax regime, and, as a result, it is forecast that there will likely be a shift of Italian shipping to other European Member States where more favourable fiscal terms exist (The Transportation Institute, 2003). Non-EU countries have followed European developments with interest. In January 2003, the Japanese government sent a team to Europe to study the tonnage tax concept after strong representation by the Japanese Shipowners Association States (The Transportation Institute, 2003). As well, New Zealand and Australia are examining the concept. Of particular note is the interest shown by Canada’s largest trading partner, the United States. This interest took on more substantive meaning in December 2001 when draft legislation, The Merchant Marine Cost Parity Act (HR3262), was introduced for consideration by the Transportation and Infrastructure Committee. It received strong support from several key players including the Maritime Administrator (Schubert, 2002). The Bill was sidelined by U.S. security concerns and did not survive, but there is an expectation that it may be re-introduced. In conclusion, a number of European countries have adopted a tonnage tax approach rather than following the Canadian ISC model. The approach affords a level playing field between domestic and international shipping, which the ISC does not, and raises the question about whether Canada should continue on its existing path or harmonize with the European model. Before that question is discussed, the next section will explore the academic literature on the fiscal treatment of shipping in order to provide an external view of the two approaches.
6. LITERATURE ON THE FISCAL TREATMENT OF SHIPPING From the academic perspective, the debate in shipping over the existence of flags of convenience and open registries has been a long one. Shipping is probably the most international of industries, with highly mobile assets and registry rules
158
MARY R. BROOKS AND J. RICHARD HODGSON
that usually allow ship owners to choose a flag for their vessels on grounds of the most beneficial tax and labour regime. The use of a flag of registry (and its accompanying fiscal climate) is dependent upon whether the benefits to the firm are expected to be realized by the ship owner. Thus, the setting of national fiscal policy for shipping presupposes the strategic responses of firms in the industry to that policy. If the policy does not adequately reflect the needs of the nation’s ship owners (and operators), it will fail to meet its objectives, be they to attract owners to the flag or to encourage free and fair competition in global trade in shipping services. Thus the literature on the fiscal treatment of shipping tends to be contained within the broader literature on flags of registry and the choice of flag. There has been no shortage of academic work examining the fiscal treatment of shipping. Goss (1967) introduced the concept of analyzing a country’s fiscal climate by looking at its impact on a ship owner’s investment decision. He believed that the ship owner’s investment response varied, depending on whether the owner was able to make full use of any tax credits (the “full tax” situation) or was in a situation where tax credits were not at all useful (the “no tax” situation). His seminal research led to an exploration by Gardner and Richardson (1973/1974), which solidified the foundation for the research that followed. They added a third scenario to the model, that of the new entrant. Companies are in a full tax position when they earn sufficient profits to be able to take full and immediate advantage of any tax allowances offered by the fiscal regime. The no tax position implies that a company has insufficient profits and a sufficiently high level of accumulated allowances to generate no tax liabilities for the foreseeable future. Because most new businesses are faced with start-up costs, the new entrant position is one where the company is unlikely to be able to take immediate or full advantage of the tax incentives and so the utility of the fiscal incentive is not positive over the short-term. However, that utility improves as the company becomes more firmly established; it may take a number of years before the company finds itself in a profitable position, no longer moving tax credits forward from prior years’ losses or allowances. It is then able to take immediate rather than delayed advantage of incentives or credits offered to reward the behaviours desired by the government. These three situations encompass the full range of circumstances a company is likely to face. As some governments do offer “tax holidays” or similar incentives, the utility of the situation is an important consideration in examining the uptake of tax opportunities. (A full explanation of the approach to national fiscal policy analysis from the ship owner’s point of view can be found in Marlow (1991a, b, c)). Work done by Gardner, Goss and Marlow (1982, p. 11) noted that fiscal regimes that “seek to stimulate investment through a system of tax allowances
The Fiscal Treatment of Shipping
159
and accelerated depreciation are highly discriminatory” and “create high barriers to entry for new firms.” Furthermore, because the approach they employed used regression analysis, they noted that the methodology only identifies the best package, without attempting to isolate the impacts of the various individual components of the package. Their belief was that a focus on the individual elements of a package was not desirable, and that the objective of government policy should be to harmonize the whole of the fiscal and ship financing regime. Gardner and Marlow (1983) extended the three-scenario model by undertaking international comparisons, setting the stage for others to examine the fiscal environments of particular countries and to draw conclusions for governments about the most effective fiscal treatment for their shipping industry or aspirations. The model allowed infinite variation in depreciation rates, treatment of salvage (or scrap) value, commencement of incentives, availability of tax-free reserves, and so on. Included in these further research efforts was the Brooks and Marlow (1983) study undertaken for the Canadian government. While these foundations for the analysis of fiscal policy were built prior to the debate in the 1990s about the nature of the fiscal reform that developed countries should undertake, they did not impede the continued interest by academics and governments in the topic. For example, Lee (1996) examined the Korean flag fiscal environment in comparison with that provided by flags of convenience and concluded that Korean flag shipping was at a serious fiscal disadvantage. The combination of national tax (corporate tax, income tax, capital gains and withholding tax on bareboat charter income), customs duties (on both capital and repair and maintenance) and a wide variety of local taxes resulted in a punitive fiscal environment for Korean flag shipping. If a Korean shipping company’s fiscal burden was indexed to 100, a Japanese operator’s fiscal burden would be indexed at 42.9 (less than half of the Korean burden), and under the Liberian flag at a mere 6.0. Furthermore, Lee noted that the fiscal climate for shipping was more punitive than that facing the airline industry in Korea. Perhaps the most important aspect of all these studies is their focus on the single element of fiscal policy in isolation from other factors in the flag choice decision. It is also critical to examine the broader flag choice literature here. Tolofari et al. (1986) examined the flagging out issue and the impact on shipping costs of phasing out open registries. Presenting a translog framework, they explored the open registry issue from an econometric perspective. Their examination looked at manning, repair and maintenance costs, lubricating oil and stores, insurance and administration; it does not appear that tax expenditures were included. Most important, though, the findings illustrated flag differences arising from non-fiscal cost components. The study concluded that phasing out flags of registry would result in greater costs to be borne by the industry (and, ipso facto, by international
160
MARY R. BROOKS AND J. RICHARD HODGSON
trade interests), and that these costs would have to be weighed by governments against “the less quantifiable benefits of greater national jurisdiction over shipping” (Tolofari et al., 1986, p. 424). If the undefined element of administrative costs did not include income tax costs, then the gap between costs incurred by vessels operating under open registry and those flagging in traditional registries would be even greater. At the time of much of the research noted above, there was a distinct divergence in the effects of the fiscal treatment of international shipping between most OECD countries and flags of convenience. This has now changed, and today most of the world’s international shipping operates in a fiscal climate with little or no tax on profits. Few countries remain that have not adopted some form of fiscal relief or witnessed the flight of owners to more “convenient” flags. The current academic literature on the fiscal treatment of shipping has, therefore, primarily focused on assessing the implementation of tonnage tax, and academic criticism of the tonnage tax approach appears generally muted at this time. Brownrigg et al. (2001, p. 219) noted the potential drawbacks; “it could involve some cost to the Exchequer, encourage tax avoidance, or distort competition between shipping and other modes or U.K. industries.” To those who might be concerned that a zero tax environment may cause serious investment distortions, Knudsen (1997, p. 45) pointed out that while taxation of profits is the norm in other industries, zero tax is the rule in global shipping. In such circumstances, he observed that zero tax does not systematically lead to super profits, and therefore is no more attractive to investors than other options. Throughout the academic debate on alternative approaches to shipping taxation runs the issue as to whether or not the shipping industry warrants special treatment. Selkou and Roe (2002) in a Comment on the U.K. initiative, argued that the industry was not a special case and, therefore, the initiative was no more than a subsidy. The views expressed by Selkou and Roe are weakened by the absence of any analysis of the European maritime aid guidelines (European Commission, 1997) in which the rationale for treating shipping as a special case is set out. The special case argument has been a long-standing one for both shipping and airline economists; there is no consensus on the issue, and, one might argue, consensus is unlikely in the immediate future. Without agreement on whether shipping constitutes a special case meriting an alternate approach to fiscal treatment, the debate within the academic community will continue. Fiscal policy attracts the attention of governments without a clear understanding of its role in the strategic decision-making of ship owners trading internationally. Governments tend to focus on encouraging ship owners to flag in the country, or retaining owners contemplating flagging out. Both Cullinane and Robertshaw (1996) and Bergantino and Marlow (1998) examined the issue of fiscal policy in
The Fiscal Treatment of Shipping
161
the larger context of flag choice. The former examined the qualitative factors for the attractiveness of the Isle of Man registry to U.K. ship owners, while the latter conducted an empirical examination of the determinants of choice of flag. Both concluded that fiscal policy is only one element in a much broader decision-making process undertaken by ship owners. Cullinane and Robertshaw (1996) found that cost reduction was not the only motivator present in flag choice decisions. In fact, more than one-half of the companies they interviewed were cognizant that they were choosing a more expensive flag option, but other factors such as reduced potential for port state inspections, perceived vessel market value on sale, savings on insurance premiums, perceived quality, corporate culture and national pride played an important role in choosing the Isle of Man or U.K. registration over flag of convenience alternatives. Inertia was also a factor. These were confirmed as critical elements in a Lloyd’s Shipping Economist (Anonymous, 2002b) article on the topic, which added deregistration costs and flag switching costs to the equation. Bergantino and Marlow’s (1998) evidence is even more compelling. Their research provides empirical evidence that the sector in which the vessel operates is a critical input to the flag choice decision. General cargo vessels are more likely to be flagged out than tankers, and deep-sea vessels more likely than short-sea vessels. Overall, across the entire sample of 167 U.K. ship owners studied, fiscal reasons fell behind crew costs, control, availability of skilled labour and compliance costs in importance to ship owners. This may be traced to the simple fact that the majority of companies (65%) studied were in the no tax category and 60% expected to stay there for at least five years; for these companies the fiscal climate is simply less relevant. The study provided more recent empirical proof that the decision contains many qualitative factors, reinforcing the findings of Cullinane and Robertshaw (1996) and the earlier work of Goss (1985). The vessel owner’s preoccupation with crew costs, first and foremost, was explained by Goss (1985) and confirmed as a key element by Tolofari et al. (1986). Goss (1985, p. 135) noted: When the profits are lower, the crew’s wages represent a higher proportion of the freight earned; even though the absolute size of wages may be the same, a ship operator’s attention is thus more readily focused on them. Second, . . . when revenue levels are determined in something approaching perfect competition, as in the short-term charter markets, crew costs are one of the few items on which a ship operator’s attention can be concentrated with any hope of success.
Goss (1985, p. 143) concluded “the expansion of FOC fleets may depend on fairly complex questions of labour costs rather than upon corporate taxation.”
162
MARY R. BROOKS AND J. RICHARD HODGSON
7. DISCUSSION The ISC approach to the taxation of shipping has been in place for more than a decade. While it took a few years to refine its implementation through subsequent amendments to the Income Tax Act and supporting legislation, the existing treatment has now been in effect for a period of five years. With the recent adoption of a tonnage tax approach throughout much of Europe, and its consideration by the U.S., it is time for Canadians to re-examine the concept in the context of its pro-trade stance and the actions of its principal trading partners. Should Canada retain its existing fiscal policy or move to harmonize with its European colleagues? First, it needs to be emphasized that the shift to a tonnage tax concept (or equivalent near-zero tax regime) is not just a fiscal policy being pursued by one or two developed maritime nations to improve their competitive position relative to their developed country neighbours. It is, rather, a significant, multinational initiative that is directed at fundamentally reshaping the industrial environment of maritime transport, and one that has stemmed the tide of owners migrating towards greater use of open registry shipping. What are the considerations that are driving this fundamental shift? In evaluating the reasons behind the adoption of a tonnage tax regime, it is important to recognize that it is the related policy objectives that are driving the initiative. More particularly, as observed earlier, many if not most Member States of the EU have accepted that, in order to preserve maritime know-how and skills and strengthen safety and environmental protection, they must reduce the difference in operating costs between ships that are owned and managed in the concerned State (whether national of foreign flag) and its third flag competition including flags of open registry States. They have also recognized that this would require fiscal and other mechanisms designed to remove differentials in corporate tax and crew costs. The European Community (1997) guidelines on State aid to maritime transport make clear that a competitive position in international shipping is only achievable if differences in fiscal costs (both corporation tax and wage related liabilities) are removed. In this respect, and as we have seen, the guidelines emphasize that maritime transport presents a special case. While this might be seen as a radical step by States like Canada, whose transportation policy is built upon the concept of “modal neutrality” in policy treatment, Canadian policy-makers can draw comfort from the fact that this policy stance is becoming increasingly the international norm. It is justified on the basis that, unlike virtually all other commercial activities, the principal capital assets in shipping are highly mobile. In another global industry with mobile assets, the airline industry, national ownership requirements are an essential element specified in the Chicago Convention, which governs air service
The Fiscal Treatment of Shipping
163
bilaterals and effectively prevents a situation similar to shipping from developing. Furthermore, since the cost of capital varies little worldwide, and there is little or no difference in technology available, only corporate taxation and wage related liabilities remain as the critical factors that decide competitiveness in this unique industry. These are recognized as critical inputs by investors who are more likely to invest where the absence of corporate tax or payroll liabilities flow to the bottom line. Knudsen (1997, p. 45) observed: In the maritime countries that stick to conventional taxation of their shipping industry there is serious unrest in the industry. It seems impossible to compete for equity capital against players who need not include tax on profits as a cost element. The costs of joining or not joining the zero tax club must be considered by governments.
There is another important point that needs to be made in any review of a tonnage tax regime, namely that it is not generally adopted to achieve economic objectives. Its adoption in the U.K., as previously discussed, was clearly not based on economic grounds but was a political decision. Some academic studies examining the tonnage tax have also looked at it as a broader concept, more than just a fiscal reform measure. Brownrigg (1999) emphasized this point, observing that “the success of a shipping policy – whatever that may encompass – does not stem from its fiscal dimension alone, but from the full package of proposals within the policy.” This reinforces the point made much earlier by Gardner, Goss and Marlow (1982, p. 11). Thus, the justification for this fiscal regime is related more to achievement of a range of shipping and trade policy objectives than to direct economic benefit. The public interest of the country is deemed to be best served when sufficient maritime knowledge and experience is available to protect and nurture a State’s trading interests, and to ensure that its trading activities are conducted in an efficient, safe and environmentally sensitive manner. In this respect, it is time for countries to examine the issue using a triple bottom line approach, i.e. one that incorporates social and environmental as well as economic benefits and costs. There are particular issues in relation to future policy in Canada. The introduction of the ISC concept and the subsequent, more widespread adoption of a tonnage tax option by a number of other developed countries raise two important questions: (1) can the Canadian ISC initiative be regarded as a success, and (2) whether or not it can, would the tonnage tax option now be a better choice? Firstly, has the ISC approach been successful in meeting Canada’s policy objectives? The Deep-Sea Task Force (Transport Canada, 1985) called for the creation of shore-based jobs in the maritime services industry in support of
164
MARY R. BROOKS AND J. RICHARD HODGSON
Canadian trading interests. There is very little in the way of substantive data available that can shed a clear light on the degree of success of the ISC initiative in relation to this objective, and opinions as to its success depend in large part on who participates in the discussion. For the companies that relocated to Vancouver, the benefits of the Vancouver location – political stability (important to many in Hong Kong prior to 1997), inexpensive office space, friendly immigration policies – are cited as the essential ingredients attracting them to run international deep-sea shipping operations from Canada. Some of the original companies have now gone, and some new ones have arrived. While various statistics are offered as to the contribution that these ISCs have made, these statistics are difficult to verify; the device of non-resident off-shore corporations makes for a less than fully transparent result. With the assistance of Vancouver proponents, the authors identified 17 companies operating as ISCs in Canada today, and Sletmo (2002) counted about 25, both more than the result desired by the Asia-Pacific Initiative. (The initial Asia-Pacific Initiative goals were 6–12 companies controlling 200 ships and employing 800 people directly.) On the vessel side, the authors were able to attribute management of 247 vessels and 36 new buildings to 13 of the ISC companies. Again, in the light of this scale of activity, the outcome could be argued as a success. The employment estimates, however, fall short; only 500 direct jobs could be attributed to 13 of the ISCs. While the Asia-Pacific Initiative (Jonathan Seymour & Associates, 1988) had projected that the employment numbers would be higher, certainly the ambitions for the number of vessels under management and companies were achieved. The number of jobs created by the ISC initiative is greater than would have been the case without the legislation but is still quite modest in comparison to more recent European initiatives. As there was very little tax being paid by shipping companies in Canada, the opportunity cost of the ISC tax amendments and subsequent legislation to the Canadian taxpayer was minimal. There is the comfort that ISC employees pay Canadian income taxes and, furthermore, create indirect jobs held by Canadians who also pay Canadian income taxes. This would not have happened without the changes made to the legislation. Overall though, it can be concluded that if the success of the initiative is measured against the objectives of the Deep-Sea Task Force, it would be regarded as modest. In considering the merits of alternative fiscal mechanisms for Canada, and more particularly whether a tonnage tax regime would serve national shipping interests better, there are arguments for and against this proposition. On the one hand, it might be argued that a tonnage tax approach is easy to estimate and is less complex in its qualification requirements. In addition, there is no need to incorporate offshore. On the other hand, the ISC ensures full tax avoidance while
The Fiscal Treatment of Shipping
165
tonnage tax still involves a tax payment, albeit in most cases a very modest one. Perhaps of more pertinence is the fact that Canada has the ISC tax regime in place, and any move to a tonnage tax regime would require wrenching changes to Canada’s Income Tax Act and other related legislation. In this respect, if the only question was whether there should be adjustments in the nature of the corporate tax regime applying to shipping, then almost certainly the answer would be “no.” This is not, however, the case. Both approaches had the objective of stimulating national involvement and expertise in shipping by enabling nationally owned and controlled vessels to be internationally competitive. In many ways, the broad intent of the tonnage tax initiative mirrors Canada’s policy aims at the time. As the first to move, Canada adopted the ISC concept but others did not follow. There are a number of fundamental differences in the policy considerations behind the decision. Thus, in examining the most appropriate fiscal treatment that a State might choose for its national shipping in the future, it is instructive to look more deeply at some of these important differences. First is the policy perspective on safety and environmental protection. While national security (with a focus more on military mobility than terrorism at that time) was frequently addressed in the various shipping policy analyses undertaken by Canada, little consideration was given to the potential safety and environmental risks inherent in the expanding use of open registry vessels. Perhaps this may be attributed to the fact that Canada, indeed the whole of North America, did not fully appreciate the massive scale of potential environmental disasters until the Exxon Valdez incident off Alaska in 1989. On the other hand, Europe’s concern was raised repeatedly as a result of the Torrey Canyon (1967), the Amoco Cadiz (1979) and subsequently the Braer (1993). Europe has continued to be preoccupied with reducing its exposure to the risks inherent in the operation of poor quality ships around its shores. The recent sinking of the Erica and the Prestige, and the extensive damage that their cargoes caused to the coastlines of France and Spain, as well as the threat posed by the tanker Castor, even though disaster was averted in this latter case, have served to provide impetus to policy initiatives that could serve to lessen the presence of such ships in European waters. With the heightened concern over terrorism, illegal immigration, piracy and drugs, Europe has also viewed increased involvement in ship management by Member States as an attractive way to reduce potential risks in relation to maritime security. This concern provided a strong impetus for Europe to pursue enhanced Port State Control (including the Paris Memorandum and the French-led “Equasis” initiative) as well as for other initiatives such as the International Safety Management (ISM) Code and the new STCW Convention with its associated “white list”. Canada has, of course, also strongly supported these initiatives, but Europe has chosen to go further and make safety and environmental protection important
166
MARY R. BROOKS AND J. RICHARD HODGSON
goals in its initiatives designed to make EC registered ships competitive with flag of convenience ships. Canada has shown little interest in seeking cost parity with flag of convenience ships as a means of enhancing safety or protecting the environment. Concerns regarding safety, security and environmental risk are in no way abating. With the current escalation in terrorism, piracy and the like, along with heightened environmental discomfort with such issues as ballast water control, antifouling paint, greenhouse gases, and single hulled tankers, there continues to be substantial cause for concern and vigilance. While undoubtedly the enforcement of standards through such mechanisms as Port State Control has gone some way to enhancing safety and environmental standards, there are limitations to the extent to which those standards can be assured through the regulatory process. For example, the Memorandum of Understanding on Port State Control 1982 requires only 25% of ships to be inspected. The greater the degree to which world trade is moved in ships where ownership and control resides in and is under the scrutiny of responsible maritime States, the lower the risk that cargoes may be carried in ships of suspect quality. In this respect, the idea that shipping policy, including fiscal reform, should be mobilized to achieve safety and environmental objectives clearly merits heightened attention. Another aspect that perhaps received less attention in Canada than it should have related to the importance to be attached to seafaring experience and competencies in implementing the objective of promoting shipping expertise in both the public and private sectors. Canadian policy studies certainly recognized the value of establishing and maintaining shore-based institutions in such fields as ship acquisition and financing, ship management, ship chartering and brokerage, shipping agencies, ship chandlery, freight forwarding, and marine insurance. Less attention was apparently devoted to the need for public sector expertise in such areas as marine safety, ship inspection and marine accident investigation, but this too is required. Canada appears to have paid little attention to this requirement for substantive sea-going experience. It is now generally agreed that shipping knowledge and expertise, whether shore based or seagoing, is enormously important to a State in ensuring the safe and efficient functioning of shipping services, whether provided by State-flagged or foreign-flagged ships. Additionally, and in response to the recognized need to encourage the development and maintenance of maritime know-how and expertise, there is a need to provide expanded opportunities for a career in the shipping field, including ensuring the availability of the necessary training and advancement opportunities, both at sea and ashore. This again constituted an important difference in the respective policy perspectives of Canada and its European counterparts, where, in the case of the latter, sea training received considerable attention. It
The Fiscal Treatment of Shipping
167
seems clear that in future these considerations must also become an important dimension of the national shipping policy of a responsible State. Another important area of possible comparison is the degree of synergy between international and domestic shipping policy. European countries have sought, in their policy frameworks, to achieve synergies between international and domestic shipping policies, particularly in rendering it more feasible for ships predominantly operating in international trade, to engage in certain coasting trade activities as and when opportunities arise. On the other hand, Canada has chosen, or possibly been compelled as a result of the Canada United States Trade Agreement, to maintain a strong division between domestic and international regimes. At issue, therefore, is the degree to which it is appropriate to keep a separation between domestic and international shipping activities. Clearly the more limited the market to which the use of capital assets such as ships is confined, the higher the potential for less than optimum use of those assets. Despite this, Canada has effectively constructed a regulatory wall between its domestic and international shipping operations. On the one hand, cabotage laws require that a ship be Canadian registered and applicable duty paid in order to engage in domestic marine transportation or other activities. On the other, Canada’s current tax environment and labour laws effectively preclude use of Canadian registered, duty paid ships in international trade. This contrasts with the European approach where restrictions on access to maritime cabotage have been significantly relaxed among its membership, and there is no duty requirement. Instead international competitiveness is achieved for those ships that have the capacity and opportunity to participate in both sectors by extending tonnage tax privileges to domestic movements. Through the careful construction of a “ring fence” to ensure that no distortions are introduced in relation to competition between modes, the European approach would seem to offer advantages in ensuring improved synergies between international and domestic movements, and in optimizing the overall effective use of valuable shipping assets. For Canada this may seem a radical step, but it is not as radical as it might appear. Indeed Canada, with the U.S., trails much of the world in reducing cabotage protectionist measures. As Goss points out, the European Union and Mercosur have significantly reduced cabotage restrictions in Europe and South America respectively. He goes on to suggest that: it would be good to see a thorough study of the effects of opening the domestic trades of Canada and the USA to international competition. There can be no doubt that this would lead to lower costs and freight rates, to the great benefit of consumers in both countries (Goss, 2002, p. 5).
The assurance that trading activities are conducted efficiently in a safe environment is a policy conclusion reached by Canada in 1985, but it was addressed in a quite
168
MARY R. BROOKS AND J. RICHARD HODGSON
unique way. The Canadian approach, considered leading edge in its day, looked at the fiscal program in isolation, making little or no linkage to safety or environmental objectives, or any adjustments to accommodate the needs for Canadian sea-going labour. It also maintained a separation between domestic policy and deep-sea shipping policy, making a special case for international shipping only. In light of subsequent changes by Canada’s major trading partners towards the tonnage tax approach, Canada clearly needs to evaluate whether the current approach remains the best choice for the future. Finally, there is no assurance that altering the fiscal regime will have any impact on an operator’s choice of the national flag, if attraction of carriers to the flag is intended. Many of the States opting for a tonnage tax approach had a longer tradition of shipping participation and a more active industry at the time of implementation than Canada. Thus, if Canada were to adopt a similar regime that only matches what exists elsewhere, little or no change may result in the short term. This does not, however, mean that the concept will not be attractive over the longer term, and gradual, longer-term adoption could prove beneficial.
8. CONCLUSIONS There are currently three models for developed country governments to select, in relation to the fiscal environment they choose to offer the industry. First, a country can make no adjustment to its domestic tax regime in its application to shipping, and thus effectively choose to abide by market forces and have its international trade transported on vessels of other States. Second, a country can adopt an approach, similar to Canada’s, one that is based on ship management activities taking place in the trading State but by a company using the artificial construction of non-resident status coupled with offshore incorporation. In this case, the country will likely choose to make a clear separation between international and domestic shipping. Third, a country can adopt a national tonnage tax regime, with or without its second register variant. The outcome of all three approaches will be competitive, zero- or near-zero tax fiscal regimes applicable to the vessels that transport its international cargoes. However, while the last approach may be seen as the most fiscally radical in its special treatment of shipping, it provides the industry with a competitive footing and the most substantive policy benefits in terms of environmental control and shore-based net benefits. According to Marlow (2002), it is the indirect impacts, e.g. loss of the maritime skill base for on-shore jobs and the multiplier benefits to on-shore maritime sectors, that are key considerations in the Canadian case, citing that U.K. shipping has a larger multiplier effect than manufacturing, distribution
The Fiscal Treatment of Shipping
169
or agriculture. The critical issue then becomes one of identifying the net benefits for the country contemplating fiscal reform. Regardless of choice, it is clear that deep-sea trade will continue to be carried predominantly by ship owners operating vessels in a zero, or near-zero, tax regime. The issue for States, therefore, effectively boils down to whether they wish to encourage full asset utilization opportunities for national flag ships through the opportunity to participate in both domestic and international trades, and potentially gain greater control of the environmental factors. Whatever the approach, fiscal reform cannot be contemplated in isolation as it appears to have been in the past. It is only one component in an overall policy package.
ACKNOWLEDGMENTS The assistance of Iain Grant (Dalhousie MMM ’03 candidate), Neal Hewitt (Dalhousie MBA/LLB ’05 candidate) and Philip (A. J.) Nichols (MBA ’03 candidate), in undertaking the background research for this paper, is much appreciated.
REFERENCES Abbott, R. (1988). Second vessel registries: Options selected by European nations (TP 9418E). Ottawa: Transport Canada Economic Research. Alcan Shipping Services Ltd. (1977). Shipping options for Canadian international deep-sea trade. Montreal: Alcan Shipping Services Ltd. Anonymous (2002a). London’s taxing question. Lloyd’s Shipping Economist (November, 7–10). Anonymous (2002b). Shipowner’s flag flutter. Lloyd’s Shipping Economist (November, 11–14). Bergantino, A., & Marlow, P. B. (1998). Factors influencing the choice of flag: Empirical evidence. Maritime Policy and Management, 25(2), 157–174. Brooks, M. R., & Marlow, P. B. (1983). Canadian tax policy and the shipping industry: An evaluation. Unpublished. Brownrigg, M. (1999, September). Tonnage tax, the UKs choice. Speech to UK Chamber of Shipping, Barbican Centre. Brownrigg, M., Dawe, G., Mann, M., & Weston, P. (2001). Developments in U.K. shipping: The tonnage tax. Maritime Policy and Management, 28(3), 213–223. Cullinane, K. P. B., & Robertshaw, M. (1996). The influence of qualitative factors in Isle of Man ship registration decisions. Maritime Policy and Management, 23(4), 321–337. Darling, H. J. (1974). The elements of an international shipping policy for Canada. Ottawa: Transport Canada Marine, August. Department of Finance (1978). An economic analysis of Canadian deep-sea shipping options (executive summary). Working Paper prepared for the Shipping Advisory Board, September. European Commission (1996). Towards a new maritime policy (March).
170
MARY R. BROOKS AND J. RICHARD HODGSON
European Commission (1997). Community guidelines on state aid to maritime transport 97/C 205/05 5.07.1997. Fairplay International Shipping Weekly (2003a). France to replace Kerguelen Flag (8 May). Fairplay International Shipping Weekly (2003b). Stockholm says no to tonnage tax (23 June). Gardner, B. M., Goss, R. O., & Marlow, P. B. (1982). Ship finance and fiscal policy. Paper read to The Nautical Institute, 16 December, Cardiff, Wales: University of Wales Institute of Science and Technology. Gardner, B. M., & Marlow, P. B. (1983). An international comparison of the fiscal treatment of shipping. Journal of Industrial Economics, 31(4), 397–415. Gardner, B. M., & Richardson, P. (1973/1974). Fiscal treatment of shipping. Journal of Industrial Economics, 22(2), 95–117. Goss, R. O. (1985). Social cost, transfer payments, and international competition in shipping. Maritime Policy and Management, 12(2), 135–143. Goss, R. O. (2002). The future of maritime economics. In: C. Th. Grammenos (Ed.), Handbook of Maritime Economics and Business (pp. 3–8). London: Lloyds of London Press. Hedlin Menzies & Associates Ltd. (1970). Summary of Canadian merchant marine: Analysis of economic potential. Winnipeg: Hedlin Menzies & Associates Ltd. HM Treasury Release (1999). Independent inquiry into a tonnage tax – A report by the Lord Alexander of Weedon QC. London (July). International Maritime Centre Vancouver (1994). International shipping corporations opportunities and obstacles, prepared for the Department of Foreign Affairs and International Trade, March. Jonathan Seymour & Associates (1988). Vancouver: An international centre for maritime commerce final report. Vancouver: Asia Pacific Initiative Transportation Task Force (December 22). Knudsen, K. (1997). The economics of zero taxation of the World shipping industry. Maritime Policy and Management, 24(1), 45–54. Lambe, H. B. (1985, January–February). The Canadian and U.S. reciprocal shipping tax exemptions. Canadian Tax Journal (pp. 118–133). Lee, T-W. (1996). Flagging options for the future: A turning point in Korean shipping policy? Maritime Policy and Management, 23(2), 177–186. Marlow, P. B. (1991a). Shipping and investment incentives: A trilogy. Part 1: Investment incentives for industry. Maritime Policy and Management, 18(2), 123–138. Marlow, P. B. (1991b). Shipping and investment incentives: A trilogy. Part 2: Investment incentives for shipping. Maritime Policy and Management, 18(3), 201–216. Marlow, P. B. (1991c). Shipping and investment incentives: A trilogy. Part 3: The effectiveness of investment incentives for shipping The U.K. experience 1950–1987. Maritime Policy and Management, 18(4), 283–311. Marlow, P. B. (2002). Ships, flags and taxes. In: C. Th. Grammenos (Ed.), Handbook of Maritime Economics and Business (pp. 512–529). London: Lloyds of London Press. McConville, J., Glen, D. R., & Dowden, J. (1997). U.K. seafarers – An analysis. London: London Guildhall University. Selkou, E., & Roe, M. (2002). Comment: U.K. tonnage tax: Subsidy or special case? Maritime Policy and Management, 29(4), 393–404. Seymour, J. (1998). A time for change: An internationally competitive Canadian flag? Pragmatic reform of the Canada Shipping Act is the answer. Proceedings, Edmonton: Canadian Transportation Research Forum (pp. 170–184). Schubert, W. G. (2002). Remarks to the Maersk master and chief engineer conference. Portsmouth, Virginia, September 17. (www.marad.dot.gov/Headlines/speeches/Maersk conference.htm.)
The Fiscal Treatment of Shipping
171
Sletmo, G. K. (2002). The rise and fall of national shipping policies. In: C. Th. Grammenos (Ed.), Handbook of Maritime Economics and Business (pp. 471–494). London: LLP, 2002. The Transportation Institute (2003). Maritime aid policies of select countries. Washington, DC. http://www.trans-inst.org accessed May 2003. Tolofari, S. R., Button, K. J., & Pitfield, D. E. (1986). Shipping costs and the controversy over open registry. Journal of Industrial Economics, 34(4), 409–427. Transport Canada (1979). Background paper on deep sea shipping, TP 2347. Ottawa: Transport Canada, Shipping Policy Branch. Transport Canada (1980). A shipping policy for Canada, TP1676. Ottawa: Transport Canada, Marine Division. Transport Canada (1983). Examination of a defensive deep-sea shipping strategy, TP4918E. Ottawa: Transport Canada. Transport Canada (1985). Task force on deep-sea shipping, report to the Minister of Transport, TP 6347E. Ottawa: Transport Canada. U.K. Department of Environment, Transport and the Regions (1998). British shipping charting a new course. December. University of Wales (1996). The UK’s requirements for people with experience of working at sea. Cardiff.
6.
DETERMINANTS OF VESSEL FLAG
Jan Hoffmann, Ricardo J. Sanchez and Wayne K. Talley 1. INTRODUCTION For a vessel operator, the choice of the flag is one of the main business decisions. For the countries that register vessels under their flag, this operation implies responsibilities as well as income. For international organizations such as the International Maritime Organization (IMO), the United Nations Conference on Trade and Development (UNCTAD), the Organization for Economic Cooperation and Development (OECD), the International Labour Organization (ILO), the Food and Agricultural Organization (FAO), and also national governments who have an interest in cleaner oceans and safer shipping, the compliance with international environmental, safety and labour regulations is of high priority – and this compliance is assumed to be closely related to a vessel’s flag. Most recently, for security reasons, interest in the linkages between flags and those who actually control the ships has re-emerged, although OECD (2003a) finds that in practice it is not so much the open registries themselves that enable reclusive owners to remain anonymous, but the corporate instruments and structures that are freely available internationally. In order to increase transparency about the linkage between the flag states and those who actually control the ships, UNCTAD (1986) developed a “Convention on the Registration of Ships,” which, however, never entered into force. Ample literature exists about comparisons between different flags concerning their compliance with international standards. OECD (2003b, p. 8), for example, highlights that the “principal responsibility for complying with the IMO’s
Shipping Economics Research in Transportation Economics, Volume 12, 173–219 Copyright © 2005 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0739-8859/doi:10.1016/S0739-8859(04)12006-4
173
174
JAN HOFFMANN ET AL.
regulatory framework has always remained with Flag States. These states traditionally exercise direct control over national fleets and their crews that tended to be nationals of those states. However, the development of ‘open’ registries – where non-national shipowners could register their ships in national registries with a sometimes-tenuous link to the flag – saw the direct ship-Flag State-national crew link weakened. The development of open registries and the international sourcing of crews has offered cost savings to owners and new employment opportunities for seafarers around the world. However, this shift of registries has rendered the control of the quality of world fleets and their crews more problematic. Most Flag States carry out their regulatory responsibilities either directly or through intermediary Class Societies. However, a certain number of states have sought to reduce their expenditures related to the administration of their fleet and/or have sought to develop their registry solely as an income-generating venture. These and other smaller states simply do not have the budgets and/or administration necessary to ensure that their fleets continue to meet IMO requirements. Class Societies have played an increasingly more important role in ensuring the safety, seaworthiness and quality of these national registries. Yet, it is commonly recognised that stiff competition in the classification/certification market has led to the emergence of certain Class Societies willing to cut corners in order to gain or retain clients.” The OECD thus assumes a link between the flag of a vessel, its classifications society, and the likelihood of being non-seaworthy. For the specific case of bulk carrier accidents investigated by the U.S. Coast Guard, Talley (2002) analyses the non-seaworthy risks, and does not find them to be related to the ship’s flag. Roberts and Marlow (2002) on the other hand found that the risk of foundering was related to the ship’s flag of registration. Li and Wonham (1999) state that openregistry country ships tend to be substandard, although the safety records of some open-registry countries are quite acceptable. Similarly, Alderton and Winchester (2002) conclude that, on average, vessels registered in open registries do have higher ratings, but big differences persist, and the countries with the worst safety records are small national, but not open, registries. UNCTAD (1994) finds that open registry ships were involved in a higher proportion of General Average cases than would correspond to their share of the world fleet. With regard to classification societies, Talley (1999) concludes that there is evidence of variance in the safety performance of different classification societies. For those countries that provide registration services, many of which are relatively small developing countries, the income generated from this business is important for their national economies (Thanopoulou, 1995). For Panama, for example, just the direct income from the registry is around 60 million USD per year. If, as is assumed, lower registration fees imply a saving for the shipping
Determinants of Vessel Flag
175
company, this has a bearing on freight rates, which in turn benefit global trade, although – as some claim – possibly at the cost of more accidents and pollution (see for example Thanopoulou, 1998). This leads us to two related questions: (a) why are some flags safer than others?; and (b) what are the determinants of a vessel’s flag? Although the paper only directly investigates the latter question, the two questions are related when the determinants of a vessel’s flag are also determinants of its safety record. For example, if a given flag attracts more dry bulk than liquid bulk vessels, then it may have a worse average safety record than an alternative flag with more liquid bulk vessels in its registry, because the latter vessel type is historically less likely to sink than dry bulk ships. Hence, the first flag may appear worse in casualty statistics, although it may have the same quality controls as the second. Of interest in this context is also the impact of IMO conventions, the safety record of the operator country’s national register, and possibly the income per capita. Will the ratification of many IMO conventions scare away vessel operators into foreign registers? May a relatively bad safety record be an indicator of a lax national safety regime and encourage operators to stay in the national register? Are high wages, which are related to the national GDP per capita, a measurable burden which may increase the likelihood to choose foreign flags? The remainder of this paper will look at quantifiable determinants of a vessel’s flag. It practically covers the world fleet of commercial vessels. What is to be explained is the probability that a given vessel – or rather its operator – chooses or not a foreign flag. To “choose a foreign flag” is meant to stand for a situation where the vessel’s operator’s country of domicile is different from the vessel’s flag state or where a second register is used. In a topically related paper, Bergantino and Marlow’s (1999) have analysed the out-flagging of U.K. vessels. In our paper, we will later further distinguish between specific major Latin American and Caribbean open registers. We will use data about the known commercial fleet of vessels of 300 Gross Tons (GT) and above, based on information provided by LRFairplay1 and analyse it with regard to possible determinants of the choice of flag. The available data allows us to consider binary or quantifiable information concerning the age, size, country of build, vessel types, the classification society, and various characteristics of the operator’s country of domicile.
2. THE WORLD’S COMMERCIAL FLEET Figures 1 and 2 depict the size of fleets according to the country of domicile of the vessel operator.
176
JAN HOFFMANN ET AL.
Fig. 1. Top 30 Operator Countries, Number of Vessels, January 2003.
Lloyds Register Fairplay provided data for 47,740 commercial vessels, including ships on order. Of the delivered vessels, 45.9% use a foreign flag. Larger and cargo vessels are more likely to use a foreign flag than smaller and miscellaneous or passenger vessels. As a consequence, the proportion of foreign flagged tonnage is higher than is the case for the number of vessels: Only 34.6% of the world’s GT use the national flag, i.e. almost two out of three gross tonnes are registered under a foreign flag. Big differences exist between operator countries. In terms of vessel numbers, Japan is the most important operator country. Regarding GT, the Greek operators control the biggest proportion of the world’s tonnage. Of the major 30 operator countries, Iran (96%) and India (88%) have the highest proportion of nationally flagged GT, whereas operators based in Monaco (1%), Switzerland (7%) and Belgium (9%) are least likely to use their national flag. Table A1, in the Annex, contains information about open and national registries for vessels above 300 GT. The total fleet grand total of delivered vessels is 584 million GT, 43,878 vessels, 8.1 million TEUs and 847 million DWT. With regard to GT, national registries reach a 37.8% of world total and 57.2% of number of vessels. 61.8% of TEU carrying capacity is flagged under open registries. The largest Flag States in terms of GT are Panama (22%), Liberia (9%), Bahamas (6%) and Greece (4.7%). Japan is the second largest Flag State in terms of number of vessels, which reflects its large fleet of inter-island cargo and passenger transport and also fishing vessels. One third of the world’s tonnage is registered in Latin America and the
Determinants of Vessel Flag
177
Fig. 2. Top 30 Operator Countries, GT, January 2003. Source: Authors, based on data provided by LRFairplay. Notes: “National” and “Foreign” refers to the vessel’s flag. “Foreign” includes second registries.
Caribbean. As sample cases, this paper will look at eight major registries of this region in more detail (marked “∗ ∗” in Table A1). Of these Latin American and Caribbean registries, Bahamas has the largest vessels (average 26,082 GT per unit), whereas Honduras specializes in the smallest vessels (1,719 GT).
3. MODEL In the basic Model (Eq. (1)), the choice of a foreign flag for the ith vessel by its operator is posited to be a function of vessel age, size and carrying capacity; country of build; characteristics of the country where the operator is domiciled (i.e. the “operator country”); vessel type; and its classification society. Specifically: FOREIGNi = F(VAGEi , VSIZEi , VCAP−DWTi , VCAP−TEUi , BUILTOPERi , OPERDEVi , CARGOi , CONTi , CIACSi ) (1) . . . where FOREIGN is a binary variable equal to 1 for a foreign flag and 0 for a national flag for the ith vessel; VAGE is vessel age; VSIZE is vessel
178
JAN HOFFMANN ET AL.
size in GT; VCAP-DWT is the vessel’s carrying capacity in deadweight tons (DWT); VCAP-TEU is the vessel’s container carrying capacity in twenty-footequivalent units (TEUs); OPERDEV and BUILTOPER are binary variables representing a vessel operator country that is a developed country and the country where the vessel was built, respectively; CARGO and CONT are vessel-type binary variables that represent vessels that can transport cargo and containers, respectively; and CIACS is a binary variable denoting that the vessel is classed by a member of the International Association of Classification Societies (IACS). In a more detailed model (Eq. (2)), by substituting alternative measures for the vessel type, classification and operator country variables, CARGO, CONT, VCAP-TEU, CIACS and OPERDEV, the FOREIGN function can be rewritten as: FOREIGNi = G(VAGEi , VSIZEi , VCAP−DWTi , BUILTOPERi , OPCASUALi , OPOPENREi , OPIMONUMi , OPGDPCAPi , OPPOPULAi , OPLITERAi , OPLIFEEXi , VGCARGOi , VCONTi , VLBLKi , BDBLKi , VPASSi , VROROi , VREEFi , VOROILi , VORSHOREi , VFSTUGi , CABSi , CBUVi , CCCSi , CDNVi , CGELi , CKORi , CLLRi , CNIKi , CRINi , CRUSi )
(2)
The variables describing the operator’s country of domicile OPCASUAL, OPOPENRE, OPIMONUM, OPGDPCAP, OPPOPULA, OPLITERA and OPLIFEEX represent the operator country’s past national flag casualty rate (log); its characterization as open registry (binary); the number of ratified IMO conventions; the GDP per capita (log); the population (log); the literacy rate; and the life expectancy, respectively. The binary vessel-type variables, VGCARGO, VCONT, VLBLK, VDBLK, VPASS, VRORO, VREEF, VOROIL, VORSHORE and VFSTUG represent general-cargo; container; liquid-bulk; dry-bulk; passenger; roll-on/roll-off; reefer; ore/oil combined; off-shore; and fishing, tug or diverse miscellaneous service vessels, respectively. The binary vessel-classification variables, CABS, CBUV, CCCS, CDNV, CGEL, CKOR, CLLR, CNIK, CRIN and CRUS, represent the American Bureau of Shipping, the Bureau Veritas, the China Classification Society, the Det Norske Veritas, the Germanischer Lloyd, the Korean Register of Shipping, the Lloyds Registry, the Nippon Kaiji Kyokai, the Registro Italian Navale and the Russian Maritime Register of Shipping classification societies, respectively.
Determinants of Vessel Flag
179
The choice of a specific foreign flag for a vessel by its operator is also posited as a function of the same explanatory variables as found in Eqs (1) and (2). These specific foreign flag choices are analyzed for eight Latin American and Caribbean open registries. The binary foreign-flag variables, PANAMA, BAHAMAS, ST. VINCENT AND THE GRENADINES, BERMUDA, ANTIGUA AND BARBUDA, BELIZE, HONDURAS and BOLIVIA, represent the Panama, Bahamas, St. Vincent and the Grenadines, Bermuda, Antigua and Barbuda, Belize, Honduras and Bolivia open registries, respectively.
4. DATA Variables used in the estimations of Eqs (1) and (2) and their specific measurements appear in Table 1. Descriptive statistics (mean and standard deviation) of the data for our variables also appear in this Table. The mean statistics reveal that 45.7%, 12.9%, 3.1%, 2.4%, 0.3%, 1.9%, 1.2%, 0.7% and 0.1% of the vessels were foreign, Panama foreign, Bahamas foreign, St. Vincent and the Grenadines foreign, Bermuda foreign, Antigua and Barbuda foreign, Belize foreign, Honduras foreign and Bolivia foreign flagged. The average age and size of the vessels are 18.1 years and 13,428 GT; the average vessel carrying capacity of vessels in DWT and TEUs are 19,483 and 186, respectively. The observations for GT range between 300 and 261,453 and for DWT between 0 and 564,650. The mean statistics also reveal that 56.3% of the vessel operators were from developed countries, 33.0% of the vessels were built in a shipyard of the operator country, 92.9% of the vessels have cargo carrying capacity, and 20.7% can transport containers. Thus, 76.4% of the vessels were classed by an IACS member and 10.9%, 9.4%, 2.7%, 10.4%, 8.8%, 2.1%, 10.3%, 14.0%, 1.9% and 6.0% were classed by the American Bureau of Shipping, the Bureau Veritas, the China Classification Society, the Det Norske Veritas, the Germanischer Lloyd, the Korean Register of Shipping, the Lloyds Registry, the Nippon Kaiji Kyokai, the Registro Italian Navale and the Russian Maritime Register of Shipping, respectively. Regarding the information about the operators’ countries of domicile, the information in Table 1 is complemented by country level data (Table 2). Whereas the statistical measures in Table 1 are in fact weighted averages based on the number of vessels operated in each country, Table 2 provides simple averages and Std. Devs. based on the 114 available country values for each variable. The average number of ratified IMO conventions by country is 24.8 (Table 2), yet the number of IMO conventions that applies to the average vessel is 32
180
JAN HOFFMANN ET AL.
Table 1. Variable Definitions and Descriptive Statistics. Measurement Dependent variables FOREIGN PANAMA BAHAMAS ST. VINCENT AND THE GRENADINES BERMUDA ANTIGUA AND BARBUDA BELIZE HONDURAS BOLIVIA Independent variables Vessel characteristics VAGE VSIZE VCAP-DWT VCAP-TEU
Mean (Std. Dev.)
1 if vessel flag country is foreign, i.e. different from the operator’s country of domicile, 0 otherwise 1 if vessel flag foreign country is Panama, 0 otherwise 1 if vessel flag foreign country is Bahamas, 0 otherwise 1 if vessel flag foreign country is St. Vincent and the Grenadines, 0 otherwise 1 if vessel flag foreign country is Bermuda, 0 otherwise 1 if vessel flag foreign country is Antigua and Barbuda, 0 otherwise 1 if vessel flag foreign country is Belize, 0 otherwise 1 if vessel flag foreign country is Honduras, 0 otherwise 1 if vessel flag foreign country is Bolivia, 0 otherwise
0.457 (0.498)
vessel age, years vessel size, GT vessel carrying capacity in DWT vessel carrying capacity in TEU
18.1 (11.4) 13428 (23049) 19483 (40732) 186 (669)
Vessel operator country characteristics OPERDEV 1 if a vessel operator’s country is a developed country, 0 otherwise OPCASUAL Log of the operator country’s 1997–99 vessel causality rate OPOPENRE 1 if the country where the vessel’s operator is domiciled is on open registry, 0 otherwise OPIMONUM Number of conventions ratified by the country where the vessel’s operator is domiciled. OPGDPCAP Log of the GDP per capita (purchase power parity, ppp) of the country where the vessel’s operator is domiciled. OPPOPULA Log of the population size of the country where the vessel’s operator is domiciled. OPLITERA Literacy rate of the country where the vessel’s operator is domiciled. OPLIFEEX Life expectency of the country where the vessel’s operator is domiciled.
0.129 (0.335) 0.031 (0.173) 0.024 (0.153) 0.003 (0.051) 0.019 (0.138) 0.012 (0.108) 0.007 (0.084) 0.001 (0.033)
0.563 (0.496) 0.150 (0.619) 0.029 (0.167) 32.0 (8.18) 9.57 (0.813)
17.5 (1.76) 0.935 (0.102) 73.0 (8.65)
Determinants of Vessel Flag
181
Table 1. (Continued ) Measurement Vessel construction information BUILTOPER 1 if the vessel operator’s country is where the vessel was built, 0 otherwise Vessel type CARGO CONT VGCARGO VCONT VLBLK VDBLK VPASS VRORO VREEF VOROIL VOFSHORE VFSTUG Vessel classification CIACS CABS CBUV CCCS CDNV CGEL CKOR CLLR CNIK CRIN CRUS
1 if a vessel has cargo carrying capacity, 0 otherwise 1 if a vessel has container carrying capacity, 0 otherwise 1 if a general-cargo vessel, 0 otherwise 1 if a container vessel, 0 otherwise 1 if a liquid-bulk vessel, 0 otherwise 1 if a dry-bulk vessel, 0 otherwise 1 if a passenger vessel, 0 otherwise 1 if a roll-on/roll-off vessel, 0 otherwise 1 if a reefer vessel, 0 otherwise 1 if an ore/oil combined vessel, 0 otherwise 1 if an off-shore vessel, 0 otherwise 1 if a fishing, tug or miscellaneous vessel, 0 otherwise 1 if vessel is classed by a member of the IACS, 0 otherwise 1 if vessel is classed by the American Bureau of Shipping, 0 otherwise 1 if vessel is classed by the Bureau Veritas, 0 otherwise 1 if vessel is classed by the China Classification Society, 0 otherwise 1 if vessel is classed by the Det Norske Veritas, 0 otherwise 1 if vessel is classed by the Germanischer Lloyd, 0 otherwise 1 if vessel is classed by the Korean Register of Shipping, 0 otherwise 1 if vessel is classed by the Lloyds Registry, 0 otherwise 1 if vessel is classed by the Nippon Kaiji Kyokai, 0 otherwise 1 if vessel is classed by the Registro Italian Navale, 0 otherwise 1 if vessel is classed by the Russian Maritime Register of Shipping, 0 otherwsie
Mean (Std. Dev.)
0.330 (0.470)
0.929 (0.256) 0.207 (0.405) 0.167 (0.373) 0.069 (0.254) 0.193 (0.395) 0.197 (0.398) 0.068 (0.251) 0.046 (0.209) 0.041 (0.197) 0.004 (0.064) 0.069 (0.254) 0.146 (0.353)
0.764 (0.424) 0.109 (0.312) 0.094 (0.292) 0.027 (0.161) 0.104 (0.306) 0.088 (0.284) 0.021 (0.143) 0.103 (0.304) 0.140 (0.347) 0.019 (0.135) 0.060 (0.237)
182
JAN HOFFMANN ET AL.
Table 2. Descriptive Statistics for Vessel Operator Country Characteristics. Variable
Max
Casualty rate 1997–1999 Variable: OPCASUALa OPOPENRE OPIMONUM GDP percapita, USD ppp, 2000 Variable: OPGDPCAPa Population Variable: OPPOPULAa OPERLITERA OPERLIFE
Min
Mean
Std. Dev.
23.33
0.20
2.26
2.80
1 48 35894
0 0 498
0.22 24.80 10719
0.42 11.03 9432.69
1273111290
27649
49175678
156029150
1.00 80.80
0.31 36.40
0.85 70.00
0.17 9.17
Sources: OPCASUAL: Alderton and Winchester, 2002; OPENRE: Authors; OPIMONUM: www.imo.org; OPGDPCAP, OPPOPULA, OPERLITERA and OPERLIFE: www. mrdowling.com/800gdppercapita.htm. a In the regression, the natural log of the ∗ marked variable values is used.
(Table 1), indicating that countries with a larger operated fleet tend to ratify a higher number of IMO conventions. Equivalently, the average casualty rate by country (Table 2) is far above the weighted average by fleet size (Table 1), implying that larger registries have lower casualty rates. Twenty-two percent of countries can be considered open registries (Table 2), but only 2.9% of vessels are operated by operators that are domiciled in these countries (Table 1). Among the types of vessels, 16.7%, 6.9%, 19.3%, 19.7%, 6.8%, 4.6%, 4.1%, 0.4%, 6.9% and 14.6% were general-cargo, container, liquid-bulk, dry-bulk, passenger, roll-on/roll-off, reefer, ore/oil combined, off-shore and fishing, tug or diverse miscellaneous service vessels, respectively. Within these broad vessel types, the composition of specific sub-types is given in Table A2 in the Annex.
5. EMPIRICAL RESULTS 5.1. Foreign Flag Equations (1) and (2) are estimated via binominal probit analysis rather than ordinary least squares (OLS). Unlike OLS, probit analysis restricts the predictions of FOREIGN to lie in the interval between zero and one. The probability of observing FOREIGN = 1 in the probit model may be expressed as: Prob(FOREIGN = 1) = ( x)
(3)
Determinants of Vessel Flag
183
Table 3. Vessel Foreign Flag Choice: Eq. (1) Probit Estimates. Variable Vessel age, size and capacity VAGE VSIZE VCAP-DWT VCAP-TEU
Coefficient (t Statistic) −0.005 (−8.36) 0.001 × 10−2 (11.8) −0.008 × 10−4 (−1.38) −0.003 × 10−2 (−2.52)
Vessel operator country characteristics OPERDEV 0.730 (51.9) Vessel construction information BUILTOPER Vessel type CARGO CONT Vessel classification CIACS Constant
−0.786 (−51.0) 0.439 (15.5) 0.373 (19.3) 0.221 (13.7) −0.970 (−28.1)
Marginal Probability −0.002 0.005 × 10−3 −0.003 × 10−4 −0.001 × 10−2 0.289 −0.311 0.174 0.148 0.088 −0.384
# Observations: 43140 2 statistic: 8999
. . . where is the standard normal distribution function;  is a vector of parameters; and x is a vector of explanatory variables as found in Eqs (1) and (2). The marginal probabilities in the probit model for the x explanatory variables are: ∂E[FOREIGN]x = ( x) ∂x . . . where is the standard normal density function.
(4)
5.1.1. Eq. (1) Estimates Table 3 reports the probit estimation results (coefficients and marginal probabilities) for Eq. (1). The probit estimate fits the data well. Its chi-squared statistic is 8,999, well above the 21.7 critical value necessary for significance at the 0.01 level for 9 degrees of freedom. The estimation results suggest that the likelihood that an operator will choose a foreign flag for his vessel decreases with vessel age. This result is somewhat surprising in that open registries tend to be associated with substandard ships, which in turn are more likely to be older. Possible explanations include pro-active attempts by major registries to attract younger tonnage, giving them discounts and other benefits. Also, older vessels may have difficulties in complying with
184
JAN HOFFMANN ET AL.
international standards and thus do not trade internationally. Further, cabotage and inter-island services in many countries tend to be undertaken by older vessels that have to be nationally flagged as a result of cargo reservation regimes. In the United States (U.S.), the Jones (cabotage) Act requires domestic water traffic to be transported by U.S. flagged, built and crewed vessels, making the renewal of the fleet more expensive, thereby leading to an above-average aged nationally flagged fleet. The probit marginal probability estimate for VAGE indicates that an increase in vessel age by one year decreases the probability by 0.002 that it will be foreign flagged. The likelihood that a foreign flag will be chosen versus a national flag increases with the size of the vessel, but decreases with the carrying capacity of the vessel, all else held constant. Having adjusted for vessel carrying capacity, the positive relationship with respect to vessel size may simply reflect the fact that larger vessels are more likely to trade internationally than smaller ones. For fishing vessels, for example, the coastal fleet of smaller vessels will almost always use the national flag, whereas larger vessels that catch fish in international waters may find it more convenient to use a foreign flag. The same may apply to other non-cargo vessels such as tug boats, dredging vessels, or passenger ferries. An increase in vessel size by one GT increases the probability by 0.000005 that the vessel will be foreign flagged. The negative coefficient for VCAP-TEU suggests that the larger the carrying capacities for container vessels that trade internationally, the lower is the likelihood that they will be foreign flagged. One possible explanation is that crewing costs are a smaller proportion of vessel operating costs for larger than smaller container vessels. An increase in container vessel carrying capacity by one TEU decreases the probability of a vessel being foreign flagged by 0.00001. The insignificance of VCAP-DWT may be attributed to a multicollinearity problem involving VCAPTEU and VSIZE. The likelihood that the flag choice is foreign increases if the vessel operator’s country is a developed country, but decreases if the vessel was built in the vessel operator’s country. One possible explanation for the former is that wages are higher in developed countries, motivating vessel operators to reduce crewing costs by flagging out. Bergantino and Marlow (1999, p. 30) found that for United Kingdom operators, labour and crewing factors account for 39% of the “factors affecting the use of a foreign flag.” Also, environmental and safety controls tend to be less strict in developing countries, which would reduce the motivation for substandard ships to move to possibly a laxer open registry. Further, some of the largest nationally flagged fleets are from developing countries such as China, India or Indonesia, where markets are less liberalized and state-owned shipping companies still exist, making it simply impossible to use a foreign flag. When the vessel operator’s
Determinants of Vessel Flag
185
country is a developed country, the probability of the vessel being foreign flagged increases by 0.289. One explanation for the negative relationship between FOREIGN and BUILTOPER is that an operator who buys or charters a nationally built vessel is likely to have other national linkages, thereby increasing the likelihood of a national flag. Also, it may be that the national shipyard was chosen because of governmental financial incentives, which are linked to the obligation to use the national flag, at least during the initial years after construction. Further, government-operated vessels are more likely to be built in national shipyards, and thus are almost certain to use the national flag. Among categorical variables, BUILTOPER has the largest marginal probability effect on a vessel being foreign flagged. When the vessel operator’s country is where the vessel was built, the probability that the vessel will be foreign flagged decreases by 0.311. The positive coefficients for CARGO and CONT suggest that cargo vessels, and in particular those with container carrying capacity, are more likely to be foreign flagged. This follows from the clear positive correlation between a vessel trading internationally –typically cargo vessels – and being foreign flagged. Passenger vessels tend to have shorter journeys, thereby trading in just one country, which is usually the operator’s domicile, and flying its flag. Operating cargo and container vessels increases the probability that a vessel will be foreign flagged by 0.174 and 0.148, respectively. The positive coefficient for CIACS suggests that foreign flagged vessels are more likely to be classed by an IACS member than nationally flagged vessels. For a vessel to trade internationally, it is obliged to be “classed,” i.e. inspected and certified regularly by a classification society, 10 of which are a member of the IACS. Many non-IACS-member classification societies are more likely to work at the national level for nationally flagged vessels. The IACS is a self-regulated body whose members claim that their inspections tend to be of high quality, which would suggest that a vessel classed by one of the ten IACS members is more likely to comply with international safety and environmental standards. A vessel classed by an IACS member increases the probability of vessel foreign flagging by 0.088. Based on the available data for this paper, our “forecast” concerning the choice of a foreign flag was correct in 72% of the 43,140 observations when estimating Eq. (1) and in 76% of the 41,470 observations when estimating Eq. (2), below. 5.1.2. Eq. (2) Estimates for all Vessels Probit estimation results for Eq. (2) are found in Table 4. Unlike Table 3, the regressions include only significant explanatory variables. The probit estimate fits the data well. The 2 statistic of 14,383 is well above the 41.6 critical
186
JAN HOFFMANN ET AL.
Table 4. Vessel Foreign Flag Choice: Eq. (2) Probit Coefficient Estimates. Variable Vessel characteristics VSIZE VCAP-DWT
Coefficient (t statistic) 0.002 × 10−2 (15.0) −0.005 × 10−3 (−8.72)
Marginal Probability 0.006 × 10−3 −0.002 × 10−3
Vessel operator country characteristics OPCASUAL −0.132 (−10.1) OPOPENRE 0.163 (3.51) OPIMONUM 0.020 (20.5) OPGDPCAP 0.402 (33.3) OPPOPULA −0.029 (−5.76) OPLIFEEX 0.009 (10.8)
−0.052 0.065 0.008 0.160 −0.012 0.004
Vessel construction information BUILTOPER
−0.902 (−49.9)
−0.358
Vessel type VGCARGO VCONT VDBLK VPASS VREEF VFSTUG
0.212 (9.92) 0.273 (8.72) 0.299 (14.8) −0.732 (−22.6) 0.362 (9.89) −0.810 (−31.2)
0.084 0.108 0.119 −0.291 0.144 −0.322
Vessel classification CBUV CCCS CDNV CGEL CKOR CNIK CRIN CRUS
0.267 (10.8) −0.138 (−2.85) 0.207 (8.46) 0.319 (12.0) 2.34 (24.6) 0.525 (24.0) −0.294 (−5.56) −0.517 (−15.9)
0.106 −0.055 0.082 0.127 0.929 0.208 −0.117 −0.205
Constant
−4.68 (−28.8)
−1.86
# Observations
41470
2
14383
statistic
Note: Although the sum of the binary vessel type variables equals 1 for all observations, which would not allow to include a constant term, as in this regression the non-significant variables were excluded a constant could be included.
value necessary for significance at the 0.01 level for 23 degrees of freedom. The coefficient and marginal probability estimates for VSIZE, VCAP-DWT and BUILTOPER are similar to those found in Table 3. VAGE is not estimated to be significant any more.
Determinants of Vessel Flag
187
The coefficients of the vessel type variables are statistically significant, except those (not reported) for VLBLK, VRORO, VOROIL and VORSHORE. As expected, the coefficients for the cargo vessels are positive and the coefficient for passenger vessels (VPASS) and Tugs (VFSTUG) are negative. Among the vessel type variables, a reefer vessel has the largest positive marginal probability effect on a vessel being foreign flagged, followed by dry-bulk and container vessels, respectively. Specifically, container, reefer and dry-bulk vessels increase the probability that a vessel will be foreign flagged by 0.108, 0.144 and 0.119, respectively. The coefficients of the vessel classification variables are statistically significant, except those (not reported) for CABS and CLLR. Among the classification variables, the Korean Register of Shipping (CKOR) has the largest positive marginal effect on a vessel being foreign flagged, followed by Nippon Kaiji Kyokai (CNIK) and Germanischer Lloyd (CGEL). Specifically, vessels classed by Nippon Kaiji Kyokai, the Korean Register of Shipping and Germanischer Lloyd increase the probability that a vessel will be foreign flagged by 0.208, 0.929 and 0.127, respectively. All vessel operator country variables except the Literacy rate are significant. Population and Flag casualty rate have negative signs, while Open registry, Number of ratified IMO conventions, GDP per capita and Life expectancy have positive sign. The marginal probability of OPCASUAL, OPOPENRE, OPIMONUM, OPGDPCAP, OPPOPULA and OPLIFEEX are estimated as −0.052, 0.065, 0.008, 0.160, −0.012 and 0.004 respectively. A higher past casualty rate OPCASUAL may be an indicator of a lax national safety regime, which appears to encourage ship operators to choose the national flag. If port state control inspections or insurance premiums reflected past casualty rates, then one might have expected the contrary. In that case, a high past casualty rate should encourage operators to choose other, foreign flags in order not to be affected by the stigma of a bad national flag. In reality, the opposite is true; a strong national maritime safety regime appears to scare away national operators into foreign flags. Each increase of the past casualty rate of 1% decreases the probability of choosing a foreign flag by 0.00052. Accordingly, a doubling of the past casualty rate decreases this probability by 0.036. Note that taking logs of the original variable leads to a change in the interpretation of the estimated parameters. A constant change in the explanatory logged variable is equivalent to a constant percentage change in the original variable. It is then necessary to reduce the parameter values by two digits in relation to the percentage change of the independent variable. For example, if the estimated parameter “” is −0.052, a change of 1% of the dependent variable leads to change in the probability of using a foreign flag by −0.00052. In order to calculate specific cases, the following approach needs to be taken: Multiplying the explanatory variable by a factor Z
188
JAN HOFFMANN ET AL.
leads to a change of the probability to flag out by b (−LN (Z)). Again, by way of example, doubling the casualty rate leads to a reduction of the probability to flag out by (−0.052(–LN(2))) = 0.036044. Similar to the previous point, a high number of ratified IMO conventions OPIMONUM also appears to drive national operators towards the use of foreign flags. The estimated parameter can be interpreted as each additional ratified IMO convention increases the probability of choosing a foreign flag by 0.008, i.e. almost 1%. Having ratified 30 instead of 10 conventions, by way of example, thus increases the probability to flag out by 0.16. If the operator is domiciled in an open-registry country OPOPENRE, then the country per definition becomes his “national flag.” In fact, there exist, for example, Panamanian nationals who are domiciled in their country and own and operate vessels. In other cases, internationally active ship operators may move their “domicile” to a country which also hosts an open registry. Such operators, however, are now more likely to choose a different (sic) foreign flag than the “national” flag of their country of domicile. If an operator is domiciled in any of the countries identified as “open registries” in Table A1, he is more likely to choose a different flag than if he is domiciled in a national registry country, i.e. a country not denominated as “open registry.” Although this may come as a surprise if we expect that operators move to have their domicile in countries where they also flag their vessel, in reality it appears that open registry countries tend to attract operators who are not inclined to use any particular flag. If an operator is domiciled in an open registry country, the likelihood of using a foreign flag increases by 0.065. To maintain a registry – be it “national” or open – implies high fixed costs. There thus exist economies of scale, and it may be advantageous to be a “large” country to maintain a national fleet. Taking the population OPPOPULA as an indicator for “size,” the estimated parameters confirm that operators from large countries are more likely to maintain their national flag as compared to operators from smaller countries. A 1% increase in the operator country population decreases the probability for an operator to choose a foreign flag by 0.00012. Being operator in a country of 100 million inhabitants instead of one from a country of just one million leads to a decrease of the probability to flag out by 0.055, i.e. around 5%. GDP per capita OPGDPCAP and life-expectancy OPLIFEEX are closely correlated and are main indicators for the level of a country’s general development. Just as OPERDEV in the regressions on Eq. (1), they have a positive impact on the likelihood to flag out. A high level of development tends to coincide with high wage levels, strict security standards and corresponding labour regimes. All these factors appear to encourage operators to use foreign flags. An increase of the GDP per capita of 1% increases the probability to flag out by 0.0016. Doubling the GDP
Determinants of Vessel Flag
189
per capita leads to an increase of this probability by 0.11, i.e. more than one out of ten, which is a very strong impact given the large GDP differences that exist between countries. 5.1.3. Eq. (2) Estimates by Vessel Type Table 5 summarizes the results of regressions with Eq. (2), by vessel type, and Table 6 includes the marginal probability results. Only statistically significant variables are included. General cargo vessels VGCARGO: Estimated parameter signs coincide with those for the general estimation of Eq. (2) as presented in Table 4. The choice of a foreign flag for a general cargo vessel is particularly strongly influenced by its country of built. Container vessels VCONT: Parameter signs are the same as in the general case, except for OPOPENRE, where a negative sign is estimated. The DNV classifications society is particularly strong in foreign flagged container vessels. Liquid bulk vessels VLBLK: Parameter signs coincide with the general FOREIGN case, except for the vessel size component measured in GT. Here, larger liquid bulk vessel are less likely to be foreign flagged than bigger ones. Liquid bulk vessel operators appear to be particularly strongly reacting to a bad past safety record of its national registry, being encouraged to stay in the national registry if its casualty rate has been high, with the estimated parameter being 2.5 times higher in its value than in the general FOREIGN case. Older liquid bulk vessels are more likely to choose the national flag. Dry bulk vessels VDBLK: All parameter signs coincide with the general case. Vessel size is particularly relevant as larger dry bulk vessels are more likely to be foreign flagged than smaller ones. Passenger vessels VPASS: In the case of passenger vessels, several parameter signs are different from the general FOREIGN case. In particular, this vessel type is the only one where a larger number of ratified IMO conventions appears to decrease the probability to choose a foreign flag. Passenger vessels are more likely to be classed by the CCS than other vessel types. RoRo vessels VRORO: All parameter signs coincide with the general FOREIGN case, except for CRUS, which is positive for RoRo vessels. The GL and BUV classification societies are particularly strong in foreign flagged RoRo vessels. Reefer vessels VREEF: Estimated parameter signs for VSIZE and VCAP-DWT have opposite signs from the general FOREIGN case. The Japanese classifications society is particularly strong in foreign flagged reefer vessels. Miscellaneous other vessels VOROIL, VOFSHORE, and VFSTUG: All parameter signs are as expected, except for OPPOPULA, which has a positive sign for VOFSHORE, and CDNV, which has a negative sign for VOROIL.
190
Table 5. Vessel Foreign Flag Choices by Vessel Type: Eq. (2) Probit Coefficient Estimates. VARIABLE
VGCARGO
Vessel characteristics VAGE VSIZE VCAP-DWT
–
–
– 0.283 ×
VCONT
– 10−2
Vessel construction information BUILTOPER −1.05 (−24.4)
–
VLBLK
VDBLK
−0.010 −0.005 (−6.21) (−2.81) −3 −0.005 × 10 0.005 × 10−2 (9.62) (6.31) – −0.002 × 10−2 (−5.76)
−0.726 −1.03 (−11.0) (−23.4)
0.011 (5.16) 0.002 × 10−2 (9.63) –
−1.13 (−14.9)
−0.098 (−3.06) –
–
0.042 (16.5) 0.400 (14.4) −0.025 (−2.09) –
−0.010 (−2.34) 0.619 (10.0) 0.079 (3.81) −0.818 (−1.98) −0.019 (−4.96)
0.016 (7.14)
–
VRORO
VREEF
0.019 – (5.08) −2 0.003 × 10 −0.001 × 10−1 (8.49) (−3.67) −0.003 × 10−2 0.002 × 10−1 (−3.99) (7.55)
VOROIL
VOFSHORE
–
0.017 (6.46) 0.004 × 10−2 (6.40) −0.001 × 10−2 (−4.96)
0.004 × (2.94) –
10−2
VFSTUG 0.006 (3.46) 0.002 × 10−2 (3.97) –
−0.559 (−6.78)
−0.387 (−3.60)
–
−0.886 (−14.4)
−0.964 (−18.2)
−0.285 (−4.71) –
–
–
–
–
−0.234 (−4.77) –
−0.084 (−2.41) –
0.020 (3.52) 0.381 (4.95) –
–
0.114 (2.92) 2.02 (3.26) –
0.007 (1.80) 0.272 (5.56) 0.621 (3.98) 1.51 (4.38) –
–
2.42 (3.79) –
0.595 (8.33) −0.056 (−2.65) −1.58 (−2.90) 0.016 (2.77)
– –
0.263 (7.61) −0.109 (−7.78) – –
JAN HOFFMANN ET AL.
Vessel operator country characteristics OPCASUAL – −0.269 −0.328 (−4.48) (−11.2) OPOPENRE – −0.417 0.252 (−1.96) (2.20) OPIMONUM – 0.022 0.032 (4.71) (13.4) OPGDPCAP 0.193 0.134 0.266 (16.7) (1.85) (8.31) OPPOPULA −0.016 −0.053 −0.067 (−4.08) (−2.41) (−5.47) OPLITERA 0.172 1.34 0.780 (1.93) (2.05) (3.31) OPLIFEEX 0.006 0.037 0.006 (5.58) (10.7) (3.83)
−0.580 (−13.2)
VPASS
CCCS
–
CDNV
–
CGEL
0.274 (5.83) 2.18 (5.52) −0.245 (−4.00) 0.418 (6.83) –
CKOR CLLR CNIK CRIN CRUS Constant # Observations 2 statistic
−0.237 (−3.47) −5.23 (−16.3) 6995 2073
−0.354 (−4.02) –
1.64 (3.96) −0.340 (−3.89) 0.454 (4.69) −0.841 (−3.15) −0.980 (−5.25)
0.278 (4.51) 0.349 (5.23) −0.263 (−1.92) 0.444 (7.55) 0.354 (4.19) 2.07 (9.00) 0.111 (1.86) 0.495 (10.0) −0.238 (−2.28) −0.692 (−7.75)
−4.24 (−5.60)
−3.40 (−8.67)
−0.755 (−4.93) 0.675 (2.76) –
2738 579.1
8149 2500
−0.305 (−5.80) – −0.483 (−5.79) – – 1.46 (8.46) −0.192 (−3.64) 0.306 (6.98) −0.880 (−5.59) −0.633 (−9.41) −5.38 (−13.9) 8171 2318
−0.0.383 (−2.35) 0.477 (4.98) 0.459 (1.74) 0.395 (4.64) 0.259 (2.04) 2.60 (5.35) – – – – −5.84 (−8.98) 2833 826.1
−0.907 (−5.57) 0.783 (6.55) –
−0.418 (−1.80) –
–
–
–
0.195 (1.91) 0.820 (6.43) 2.03 (4.43) –
–
−1.92 (3.01) –
0.517 (4.93) −0.500 (−2.93) 0.412 (2.22) −7.18 (−10.9) 1917 572.6
–
–
0.363 (4.84) 0.638 (5.82) −0.992 (−3.35) 0.489 (5.94) 0.454 (3.33) –
−0.199 (−2.35) 0.327 (5.04) –
–
– –
0.954 (2.50) –
–
0.620 (5.64) –
–
0.182 (1.73) –
–
–
0.270 (1.70) –
−0.539 (−5.00)
–
−0.558 (−2.86)
−0.423 (−4.54)
−4.49 (−5.59)
−25.5 (−3.97)
1658 766.6
–
169 193.7
−5.87 (−11.8) 2945 697.5
Determinants of Vessel Flag
Vessel classification CABS −0.724 (−7.82) CBUV –
2.96 (14.3)
−1.48 (−3.49) 5895 1582
Note: t-statistics are in parentheses.
191
192
Table 6. Vessel Foreign Flag Choices by Vessel Type: Eq. (2) Probit Marginal Probability Estimates. VARIABLE
VGCARGO VCONT
Vessel characteristics VAGE – VSIZE – VCAP-DWT 0.001 × 10−2
– – –
VLBLK
VDBLK
VPASS
VRORO
VREEF
VOFSHORE
VFSTUG
−0.004 −0.002 0.003 0.007 – – 0.007 0.001 −0.002 × 10−3 0.002 × 10−2 0.006 × 10−3 0.001 × 10−2 −0.004 × 10−2 0.001 × 10−2 0.001 × 10−2 0.003 × 10−3 – −0.008 × 10−3 – −0.001 × 10−2 0.001 × 10−2 – −0.006 × 10−3 – −0.220
−0.286
−0.221
−0.147
Vessel operator country characteristics OPCASUAL – −0.090 −0.131 OPOPENRE – −0.140 0.100 OPIMONUM – 0.007 0.013 OPGDPCAP 0.193 0.045 0.106 OPPOPULA −0.016 −0.018 −0.027 OPLITERA 0.172 0.448 0.311 OPLIFEEX 0.006 0.012 0.003
−0.037 – 0.157 0.151 −0.009 – 0.006
– – −0.003 0.156 0.020 −0.206 −0.005
−0.113 – 0.008 0.150 – 0.957 –
– – – 0.225 −0.021 −0.597 0.006
Vessel classification CABS −0.289 CBUV – CCCS – CDNV – CGEL 0.109 CKOR 0.870 CLLR −0.098 CNIK 0.167 CRIN – CRUS −0.095 Constant −2.09
−0.115 – −0.183 – – 0.552 −0.073 0.116 −0.333 −0.240 −2.04
−0.097 0.120 0.116 0.100 0.065 0.655 – – – – −1.47
−0.358 0.309 – 0.077 0.324 0.802 – 0.204 −0.198 0.163 −2.84
−0.158 – – – – 0.362 – 0.235 – −0.204 −1.70
0.111 0.139 −0.105 0.177 0.141 0.824 0.044 0.197 −0.095 −0.276 −1.36
−0.352
−0.205
– – 0.033 0.582 – – –
−0.093 – 0.003 0.108 0.025 0.599 –
−0.018 – – 0.056 −0.023 – –
– – – 0.553 – – – – – – −7.34
0.145 0.254 −0.395 0.195 0.181 – 0.072 – – −0.222 −2.34
−0.042 0.070 – – – 0.630 – 0.057 – −0.090 −0.316
–
JAN HOFFMANN ET AL.
Vessel construction information BUILTOPER −0.421 −0.243 −0.411
−0.118 0.019 −0.252 0.226 – 0.550 −0.114 0.152 −0.281 −0.328 −1.42
VOROIL
Determinants of Vessel Flag
193
Although differences in the magnitude of estimated parameters persist, in general, the main conclusions concerning the direction of impact of different variables on the probability to flag out are not modified if we look at individual vessel types; the one exception being passenger ships, where several parameters are estimated with a different sign from the general FOREIGN case.
5.2. Latin American and Caribbean Foreign Flags In this section, we investigate determinants of the foreign flag choice for eight specific Latin American and Caribbean open registries. Probit estimation results for Eqs (1) and (2) for these registries are presented in Tables 7–10. 5.2.1. Eq. (1) Estimates In Table 7, probit coefficient estimation results for Eq. (1) are reported and probit marginal-probability estimates are reported in Table 8. For all eight probit estimates of Eq. (1) found in Table 7, the estimates fit the data well. The 2 statistics are well above the critical value necessary for significance at the 0.01 level for 9 degrees of freedom. The Panama registry, the world’s largest, has similar characteristics to that of the overall foreign flagged fleet (see Table 3), albeit with a special emphasis on Asian operators. In Table 7 the Panama probit estimated coefficients of the variables VAGE, VSIZE, VCAP-DWT, CARGO and CIACS have the same signs as those found in Table 3. Younger vessels and those with greater gross tonnage are more likely to be registered in Panama, but this likelihood decreases the greater the DWTcarrying capacity (adjusting for gross tonnage). Cargo as opposed to non-cargo and IACS as opposed to non-IACS classed vessels are more likely to be registered in Panama. Coefficient signs differ in Tables 3 and 7 for the container vessel variables, VCAP-TEU and CONT, and country where vessel was built (BUILTOPER). The likely reason lies in the fact that Panama specializes in flagging Asian operators, in particular vessels classed by the Japanese and Korean Asian societies, Nippon Kaiji Kyokai and Korean Register of Shipping (more information about individual classification societies will be incorporated in Table 10). Since Asian operators are particularly strong in global container shipping and since Panama has a relatively large share of larger container vessels, a priori signs for VCAP-TEU and CONT for the Panama registry are positive. In Table 7 VCAP-TEU has the expected positive sign, but CONT has a negative sign. The Asian countries, Korea, Japan, China and Hong Kong, are the world’s largest ship builders, which increases the likelihood that a ship constructed in one of these countries is also registered in Panama. In Table 7 BUILTOPER has the expected positive sign. The coefficients
194
Table 7. Vessel Open Registry Choices: Eq. (1) Probit Coefficient Estimates.* Variable
Panama
Vessel age, size and capacity VAGE −0.007 (−9.44) VSIZE 0.001 × 10−2 (12.6) VCAP-DWT −0.003 × 10−3 (−5.41) VCAP-TEU 0.003 × 10−2 (2.06)
Bahamas 0.005 × 10−1 (0.41) 0.002 × 10−2 (13.8) −0.006 × 10−3 (−9.49) −0.001 × 10−1 (−6.70)
Vessel operator country characteristics OPERDEV 0.012 0.559 (0.67) (18.4)
St.Vinc. and the Gren. 0.016 (11.9) −0.003 × 10−2 (−5.45) 0.001 × 10−2 (3.86) −0.004 × 10−1 (−4.48)
Bermuda
Antigua and Barbuda
Belize
Honduras
Bolivia
0.014 (8.10) −0.009 × 10−2 (−5.94) 0.003 × 10−2 (3.74) −0.001 × 10−1 (−0.52)
0.024 (12.4) −0.002 × 10−1 (−6.75) 0.007 × 10−2 (4.76) −0.008 × 10−1 (−1.00)
0.014 (3.72) −0.007 × 10−2 (−1.77) 0.003 × 10−2 (1.27) −0.002 (−1.09)
1.27 (19.4)
−0.196 (−4.65)
−0.244 (−4.41)
−0.297 (−2.56)
0.003 −0.016 (0.87) (−7.68) 0.001 × 10−2 −0.002 × 10−1 (6.55) (−12.9) −0.004 × 10−3 0.007 × 10−2 (−3.36) (7.27) −0.008 × 10−2 0.009 × 10−1 (−1.68) (12.41) 0.397 (4.81)
−0.598 (−16.8)
−0.717 (−15.4)
−0.875 (−5.21)
−0.622 (−13.1)
−0.350 (−6.47)
−0.266 (−3.87)
−0.347 (−2.21)
0.272 (7.35) −0.145 (5.90)
0.163 (2.31) 0.287 (8.18)
0.233 (3.37) 0.331 (7.63)
0.124 (0.59) 0.266 (2.84)
1.01 (3.34) 1.08 (23.3)
0.222 (2.64) −0.160 (−1.88)
−0.165 (−2.08) 0.062 (0.44)
−0.134 (−0.81) 0.266) (1.04)
Vessel classification CIACS 0.249 (11.8) Constant −1.65 (−36.9)
0.351 (8.09) −2.76 (−31.6)
−0.100 (−3.06) −2.30 (−27.4)
0.119 (1.05) −3.51 (−12.9)
0.296 (4.73) −3.98 (−12.6)
−0.112 (−2.74) −2.28 (−22.1)
−0.190 (−3.76) −2.30 (−21.4)
Vessel construction information BUILTOPER 0.201 (11.2) Vessel type CARGO CONT
# Observations 2 statistic ∗
43140 1532
t statistics are in parentheses.
43140 1291
43140 895.5
43140 188.8
43140 3156
43140 587.7
43140 597.6
−0.334 (−3.35) −2.72 (−12.7) 43140 94.1
JAN HOFFMANN ET AL.
0.131 (4.47)
Variable
Panama
Bahamas
Vessel age, size and capacity VAGE −0.001 0.002 × 10−2 VSIZE 0.002 × 10−3 0.008 × 10−4 VCAP-DWT −0.006 × 10−4 −0.003 × 10−4 VCAP-TEU 0.005 × 10−3 −0.007 × 10−3 Vessel operator country characteristics OPERDEV −0.002
St.Vinc. and the Gren.
Bermuda
Antigua and Barbuda
0.005 × 10−1 −0.001 × 10−3 0.004 × 10−4 −0.001 × 10−2
0.001 × 10−2 0.004 × 10−5 −0.001 × 10−5 −0.002 × 10−4
−0.001 × 10−2 −0.001 × 10−4 0.005 × 10−5 0.007 × 10−4
Belize
Honduras
Bolivia
0.001 × 10−1 0.001 × 10−2 0.005 × 10−3 −0.008 × 10−4 −0.001 × 10−4 −0.002 × 10−5 0.003 × 10−4 0.004 × 10−5 0.009 × 10−6 −0.001 × 10−3 −0.004 × 10−4 −0.005 × 10−4
0.026
0.004
0.001
0.009 × 10−1
−0.002
−0.001 × 10−1 −0.010 × 10−2
−0.028
−0.024
−0.002
−0.004 × 10−1
−0.003
−0.002 × 10−1 −0.001 × 10−1
0.053 −0.029
0.008 0.013
0.008 0.011
0.003 × 10−1 0.008 × 10−1
0.007 × 10−1 0.008 × 10−1
0.002 −0.001
−0.010 × 10−2 −0.004 × 10−2 0.004 × 10−2 0.009 × 10−2
Vessel classification CIACS 0.049 Constant −0.324
0.016 −0.128
−0.003 −0.078
0.003 × 10−1 −0.010
0.002 × 10−1 −0.003
−0.001 −0.019
−0.001 × 10−1 −0.001 × 10−1 −0.001 −0.001
Vessel construction information BUILTOPER 0.040 Vessel type CARGO CONT
Determinants of Vessel Flag
Table 8. Vessel Open Registry Choices: Eq. (1) Probit Marginal Probability Estimates.
195
196
Table 9. Vessel Open Registry Choices: Eq. (2) Probit Coefficient Estimates.* Variable
Panama
Bahamas
St.Vincent/ Grenadines
Bermuda
–
–
–
0.001 × 10−2 (9.97) −0.002 × 10−3 (−4.15)
0.001 × 10−2 (9.87) −0.005 × 10−3 (−6.75)
0.015 (10.1) −0.002 × 10−2 (−9.20) –
Vessel characteristics VAGE VSIZE VCAP-DWT
Vessel construction information BUILTOPER −0.152 (−6.41)
−0.526 (−13.7)
−0.015 (−6.37) −0.009 × 10−2 (−7.02) 0.004 × 10−2 (4.11)
Belize
Honduras
0.016 (9.17) −0.001 × 10−1 (−6.32) 0.004 × 10−2 (4.19)
0.028 (12.8) −0.002 × 10−1 (−4.69) 0.006 × 10−2 (2.32)
Bolivia
0.029 (7.44) – – –
−0.660 (−13.3)
−0.778 (−4.31)
−0.596 (−10.7)
−0.390 (−6.65)
−0.467 (−5.45)
−0.634 (−2.27)
−0.041 (−1.69) −0.473 (−5.19) 0.021 (9.54) –
–
0.348 (6.23) 1.06 (5.39) 0.043 (7.52) 0.382 (4.05) 0.346 (16.1) 2.47 (2.29) 0.031 (6.71)
−0.129 (−3.99) 0.412 (4.63) −0.010 (−3.90) 0.133 (4.78) –
–
−0.189 (−2.99) –
−0.094 (−9.60) −1.30 (−8.66) 0.006 (2.85)
– 0.029 (5.10) 0.320 (3.42) – – −0.026 (−8.47)
– −0.004 (−1.86)
0.415 (4.14) – 0.258 (5.62) – −0.901 (−2.94) −0.012 (−4.10)
– – −0.145 (−4.98) – –
JAN HOFFMANN ET AL.
Vessel operator country characteristics OPCASUAL −0.098 – (−6.00) OPOPENRE −0.208 0.518 (−2.76) (7.32) OPIMONUM −0.011 – (−8.38) OPGDPCAP 0.252 0.371 (13.4) (10.8) OPPOPULA 0.028 – (4.23) OPLITERA −0.374 1.58 (−2.51) (5.08) OPLIFEEX 0.013 −0.009 (11.2) (−6.07)
0.009 × 10−3 (8.62) 0.004 × 10−2 (4..11)
Antigua/ Barbuda
VCONT
–
VLBLK
0.203 (5.88) –
VDBLK
–
VPASS VRORO VREEF
−0.470 (−10.0) –
0.392 (9.48) VOROIL – VOFSHORE 0.216 (6.11) VFSTUG −0.717 (−17.3)
−0.387) (−6.37) −0.671 (−8.79) −0.665 (−11.1) −0.531 (−9.05) −0.379 (−5.47) −0.651 (−8.18) – – −0.459 (−6.87) −1.01 (−12.6)
0.424 (9.99) 0.184 (2.03) – 0.523 (11.9) – 0.418 (5.73) 0.363 (5.14) – 0.264 (4.23) –
−0.576 (−1.99) – −0.582 (−4.26) –
0.826 (14.9) 0.958 (12.4) – –
–
–
0.442 (3.53) 0.525 (3.86) – –
0.831 (7.74) –
–
–
– –
–
−0.582 (−4.23) –
0.672 (7.17) 0.564 (2.38) 0.502 (5.06) 0.521 (4.69) 0.336 (2.44) –
0.210 (2.58) – –
0.723 (4.96) – –
–
−0.225 (−3.61)
–
–
– −0.147 (−2.65) –
0.304 (2.33) –
Determinants of Vessel Flag
Vessel type VGCARGO
0.344 (2.66) – – –
– –
197
198
Table 9. (Continued ) Variable
Vessel classification CABS CBUV CCCS CDNV CGEL CKOR CLLR CNIK CRIN
Constant # Observations 2 statistic ∗t
Bahamas
St.Vincent/ Grenadines
Bermuda
0.478 (4.76) –
–
−0.077 (−2.25) 0.646 (1.83) –
0.289 (6.01) 0.388 (8.22) –
−0.199 (−5.34) −0.463 (−10.7) 0.616 (11.5) −0.123 (−3.43) 0.846 (31.2) −0.215 (−2.80) −0.671 (−9.71) −4.48 (−20.8)
0.576 (13.9) −0.188 (−2.97) −0.946 (−3.25) 0.364 (8.21) – –
−0.290 (−4.09) 0.115 (2.43) 0.559 (7.60) −0.246 (−4.05) −0.313 (−5.20) −0.458 (−3.06) −0.263 (−4.52) −0.233 (−3.75) –
−1.08 (−4.93) −6.05 (−19.6)
−0.189 (−2.97) −0.479 (−1.77)
41470 5613
41470 1943
41470 1307
statistics are in parentheses.
– – – – 0.353 (3.72) – –
Antigua/ Barbuda −0.840 (−2.73) 0.406 (5.21) – – 1.18 (20.7) – − −0.990 (−3.16) –
Belize
Honduras
Bolivia
−0.259 (−3.01 −0.189 (−2.62) –
−0.314 (−2.01) –
–
–
–
−0.240 (−2.67) −0.146 (−1.96) −0.246 (−1.72) −0.277 (−3.21) –
−0.392 (−2.51) −0.271 (−2.63) –
–
−0.201 (−1.83) –
−
–
–
–
–
– −1.48 (−3.04) 41470 117.1
−5.43 (−5.81)
−0.657 (−2.40) −18.4 (−20.2)
−2.81 (−11.0)
−0.718 (−3.57) −3.90 (−10.4)
41470 377.4
41470 4115
41470 632.8
41470 702.5
–
– –
–
JAN HOFFMANN ET AL.
CRUS
Panama
Variable
Panama
Vessel characteristics VAGE – VSIZE 0.002 × 10−3 VCAP-DWT −0.004 × 10−4 Vessel construction information BUILTOPER −0.023
Bahamas
St.Vincent/ Grenadines
– 0.004 × 10−4 −0.002 × 10−4
0.004 × 10−1 −0.004 × 10−4 –
0.017
Vessel operator country characteristics OPCASUAL −0.015 – OPOPENRE −0.032 0.016 OPIMONUM −0.002 – OPGDPCAP 0.038 0.012 OPPOPULA 0.004 – OPLITERA −0.057 0.050 OPLIFEEX 0.002 −0.003 × 10−1
Bermuda
Antigua/ Barbuda
Belize
Honduras
Bolivia
– 0.008 × 10−6 –
−0.008 × 10−3 −0.006 × 10−5 0.002 × 10−5
0.009 × 10−1 −0.007 × 10−4 0.003 × 10−4
0.006 × 10−3 −0.004 × 10−5 0.001 × 10−5
0.001 × 10−2 – –
−0.017
−0.006 × 10−1
−0.004 × 10−1
−0.003
−0.001 × 10−1
−0.004 × 10−1
−0.001 −0.012 0.005 × 10−1 – −0.002 −0.034 0.002 × 10−1
– – 0.002 × 10−2 0.003 × 10−1 – – −0.002 × 10−2
−0.002 × 10−1 0.007 × 10−1 0.003 × 10−2 0.002 × 10−1 0.002 × 10−1 0.002 0.002 × 10−2
−0.009 × 10−1 0.003 −0.007 × 10−2 0.010 × 10−1 – – −0.003 × 10−2
– 0.009 × 10−2 – 0.006 × 10−2 – −0.002 × 10−1 −0.003 × 10−3
−0.001 × 10−1 – – – −0.009 × 10−4 – –
Determinants of Vessel Flag
Table 10. Vessel Open Registry Choices: Eq. (2) Probit Marginal Probability Estimates.
199
200
Table 10. (Continued ) Variable
Panama
Bahamas
Vessel type VGCARGO VCONT VLBLK VDBLK VPASS VRORO VREEF VOROIL VOFSHORE VFSTUG
– 0.031 – – −0.072 – 0.060 – 0.033 −0.109
−0.012 −0.021 −0.021 −0.017 0.012 −0.021 – – −0.015 −0.032
0.011 0.005 – 0.014 – 0.011 0.010 – 0.007 –
−0.005 × 10−1 – −0.005 × 10−1 – – 0.004 × 10−1 0.004 × 10−1 – – –
0.005 × 10−1 0.006 × 10−1 – – – 0.005 × 10−1 – – – –
– – −0.001 – −0.004 – 0.002 – – −0.002
0.002 × 10−1 0.001 × 10−1 0.001 × 10−1 0.001 × 10−1 0.008 × 10−2 – 0.002 × 10−1 – – –
Vessel classification CABS CBUV CCCS CDNV CGEL CKOR CLLR CNIK CRIN CRUS
−0.012 0.010 – −0.030 −0.071 0.094 −0.019 0.129 −0.033 −0.102
0.009 0.012 – 0.018 −0.006 −0.030 0.012 – – −0.034
−0.008 0.003 0.015 −0.006 −0.008 −0.012 −0.007 −0.006 – −0.005
0.004 × 10−1 – – – – – 0.003 × 10−1 – – –
−0.005 × 10−1 0.003 × 10−1 – – 0.008 × 10−1 – – −0.006 × 10−1 – −0.005 × 10−1
−0.002 −0.001 – −0.002 −0.001 −0.002 −0.002 – – –
−0.007 × 10−2 – – −0.009 × 10−2 −0.006 × 10−2 – −0.005 × 10−2 – – −0.002 × 10−1
– – – – – – – – – –
0.113
−0.099
−0.013
−0.012
−0.020
−0.009 × 10−1
−0.009 × 10−1
Bermuda
−0.004
Antigua/ Barbuda
Belize
Honduras
Bolivia
0.002 × 10−1 – 0.002 × 10−1 – – – – – – –
JAN HOFFMANN ET AL.
Constant
St.Vincent/ Grenadines
Determinants of Vessel Flag
201
of OPERDEV are positive in both tables; the coefficient is highly significant in Table 3 but statistically insignificant in Table 7. The latter may reflect the fact that the Asian countries, Korea, Taiwan and China, do not appear as developed countries in OPERDEV. Among the categorical explanatory variables, a cargo vessel has the largest marginal probability effect on selection of the Panama open registry, followed by an IACS classed vessel and country where vessel was built. A cargo vessel increases the probability of choosing the Panama flag by 0.053, while an IACS classed vessel and country where vessel was built increases this probability by 0.049 and 0.040, respectively. Bahamas is the second largest Latin American and Caribbean register. The Bahamas flag tends to be used by cruise and other passenger ship operators and their vessels are often classed by the American Bureau of Shipping. The Bahamas probit coefficient estimates for Eq. (1) found in Table 7 have the same signs as those for the overall foreign flagged fleet found in Table 3, except for vessel age. The Bahamas coefficient for the latter is positive, but statistically insignificant. Among the categorical explanatory variables, country where vessel is built has the largest marginal probability effect on selection of the Bahamas open registry, followed by a developed country and an IACS classed vessel. If the vessel operator’s country is where the vessel was built, the probability of choosing the Bahamas flag decreases by 0.028, while if a vessel operator’s country is a developed country and if it is an IACS classed vessel, this probability increases by 0.026 and 0.016, respectively. The third largest Latin American and Caribbean register is St. Vincent and the Grenadines. This registry has a bad reputation and features high on Port State Control black lists (Winchester, 2003). A comparison of the St. Vincent and the Grenadines probit coefficient estimates in Table 7 with those for the overall foreign flagged fleet in Table 3 reveals that the former has vessel characteristics unlike that of the latter. Specifically, the St. Vincent and the Grenadines registry attracts older vessels, smaller gross-ton vessels, and non-IACS classed vessels. Among the categorical explanatory variables, BUILTOPER decreases the probability of a St. Vincent and the Grenadines flag by 0.024, whereas a container and cargo vessel increases this probability by 0.011 and 0.008, respectively. The Bermuda registry is closely linked to British operators. The Bermuda probit coefficient estimates for Eq. (1) have the same signs as for the Bahamas coefficient estimates. However, unlike the Bahamas results, the coefficients of CARGO and CIACS are statistically insignificant. If the operator country is also the country of build, this decreases the probability of a Bermuda flag by 0.002, whereas a developed operator country and a container vessel increases this probability by 0.001 and 0.0008, respectively. The Antigua and Barbuda registry specializes in German container and general cargo operators. Like the overall foreign flagged fleet, vessels under this flag tend
202
JAN HOFFMANN ET AL.
to be younger, but unlike the former, vessels tend to be smaller. Further, unlike the former, vessels tend to have larger carrying capacities (adjusting for gross tonnage). Among the categorical variables, OPERDEV has the largest marginal probability effect (0.0009) on Antigua and Barbuda flagging, likely reflecting the high incidence of German operators, followed by container (0.0008) and cargo (0.0007) vessels. The registries, Belize, Honduras and Bolivia, are relatively small open registries and appear on the black lists of many Port State Control organizations. The signs of their probit estimated coefficients are similar, except those for CARGO and CONT. The Belize registry attracts cargo vessels, whereas the Honduras and Bolivia registries attract non-cargo vessels. The Belize registry attracts non-container vessels, whereas the Honduras and Bolivia registries tend to attract container vessels. Unlike that for the overall foreign flagged fleet (see Table 3), vessels in these registries tend to be older, smaller in size, come from a developing rather than a developed country, and to be classed by non-IACS societies. Among categorical variables, country where vessel was built has the largest marginal probability effect on the Belize and Honduras registry, decreasing the probability Belize and Honduras flagging by 0.003 and 0.0002 respectively. A comparison of the probit marginal probability estimates (Table 8) for the eight Latin American and Caribbean open registries provide further insight into the dissimilarities among these registries. An increase in a vessel’s size by one gross ton increases the probability of the vessel being flagged in Panama, Bahamas and Bermuda by 0.000002, 0.0000008 and 0.00000004, respectively, but decreases the probability of the vessel being flagged in St. Vincent and the Grenadines, Antigua and Barbuda, Belize, Honduras and Bolivia by 0.000001, 0.0000001, 0.0000008, 0.0000001 and 0.00000002, respectively. A vessel classed by the IACS increases the probability of the vessel being flagged in Panama, Bahamas, Bermuda and Antigua and Barbuda by 0.049, 0.016, 0.0003 and 0.0002, respectively, but decreases the probability of the vessel being flagged in St. Vincent and the Grenadines, Belize, Honduras and Bolivia by 0.003, 0.001, 0.0001 and 0.0001, respectively. 5.2.2. Eq. (2) Estimates In Table 9 probit coefficient estimation results for Eq. (2) are reported and probit marginal-probability estimates are reported in Table 10. Unlike Table 7, the results include only statistically significant explanatory variables. For all eight probit estimates of Eq. (2), the estimates fit the data well. The 2 statistics are well above the critical value necessary for significance at the 0.01 level. The statistically significant coefficient estimates for VAGE, VSIZE, VCAP-DWT and BUILTOPER and their marginal probability estimates are in general similar to those found in Tables 7 and 8.
Determinants of Vessel Flag
203
5.2.2.1. Vessel types. For the Panama registry, the signs of the probit coefficient estimates of the vessel type variables in Table 9 coincide with those of the general FOREIGN case. As expected, the coefficients for the cargo vessels are positive and that for passenger vessels (VPASS) is negative. Among the vessel type variables found in Table 10, a reefer vessel has the highest positive probability of flying a foreign Panama flag, i.e. a reefer vessel increases the probability of a foreign Panama flag by 0.06. For the Bahamas registry, all vessel type variables except VREEF and VOROIL are statistically significant. Tugs and similar vessels are those least likely to fly the Bahamas flag. Among the vessel type variables for the St. Vincent and the Grenadines registry, VDBLK has the strongest positive impact on the likelihood to choose this flag. Vessel-type estimation results for Bermuda suggest that this registry attracts mostly reefer and roll-on/roll-off vessels and is unlikely to attract liquid-bulk vessels. Also – different from the general FOREIGN case – a general cargo vessel reduces the likelihood of being Bermuda foreign flagged. The Antigua and Barbuda foreign flag attracts mainly general-cargo, container and roll-on/roll-off vessels, with all other vessel type variables not estimated to be significant. A reefer vessel increases the likelihood of choosing the Belize register, and a passenger vessel makes this choice particularly unlikely. The Honduras open registry is more likely to be chosen if the vessel is a general cargo or a reefer vessel. The Bolivia open registry attracts relatively few different types of vessels, i.e. only the vessel types VGCARGO and VLBLK are statistically significant. 5.2.2.2. Classification societies. For the Panama registry, the probit coefficient estimates of the vessel classification variables are statistically significant, except those (not reported) for CCCS. Among the significant classification variables, only the coefficients of CBUV, CKOR and CNIK are positive, highlighting the fact that many Panama flagged vessels tend to be classed by the Asian classification societies of Korea and Japan. Vessels flying the Bahamas foreign flag are likely to be classed by the American Bureau of Shipping (CABS), Bureau Veritas (CBUV), Det Norske Veritas (CDNV) and Lloyds Register (CLLR). Among these societies CDNV has the highest estimated impact on the likelihood to choose the Bahamas foreign flag. All vessel classification variables for St. Vincent and the Grenadines are estimated as negative, except the coefficients for CBUV and CCCS. Hence, vessels classed by the French and Chinese classification societies are attracted to the St. Vincent and the Grenadines foreign flag, and vessels classed by other IACS members are less likely to choose the St. Vincent and the Grenadines flag.
204
JAN HOFFMANN ET AL.
For the Bermuda registry, only the classification societies ABS and Lloyds Register have significant marginal probability effects on this registry, both with a positive sign. The probit estimate for the Antigua and Barbuda flag reveals that only the classification societies Bureau Veritas and Germanischer Lloyd have positive effects on this registry. The latter is by far the strongest, reflecting the high proportion of German operators in this registry. For the Belize and Honduras open registries, none of the statistically significant vessel classification variables in the probit estimates are positive. This can be interpreted as negative indicator for the quality of these registries as far as their control and certification of vessel safety is concerned. No vessel classification variables were statistically significant in the Bolivia probit estimate. 5.2.2.3. Operator country characteristics. Vessel operators that choose the Panama flag tend to come from countries with similar characteristics as the general FOREIGN case. Exceptions are the variables OPOPENRE, where the sign is negative for Panama, OPIMONUM, where the sign is negative, and OPPOPULA, where the sign is positive, i.e. different from the general FOREIGN case. This indicates that vessel operators that choose the Panama flag are less likely to be domiciled in other open registry countries, but rather in relatively large countries without open registry. Also, Panama does not appear to attract vessels whose operators are trying to avoid a high number of ratified IMO conventions. Bahamas attracts vessels from countries with similar characteristics as the general FOREIGN case. The impact of a high GDP per capita is particularly strong. For St. Vincent and the Grenadines, all operator country variables except OPGDPCAP are significant, and all except OPOPENRE have the expected sign. For Bermuda, most country characteristics variables are not significant, except OPIMONUM and OPGDPCAP, which are estimated with the expected positive sign, and BUILTOPER and OPFILEEX, which are estimated with a negative sign. In the case of Antigua and Barbuda, again the strong participation of German operators appears to influence the estimated parameter values. Unlike in the general FOREIGN case, vessels in this registry do not come from countries with a bad national flag safety record and they are from relatively larger countries. They come from countries that have ratified a relatively high number of IMO conventions. Belize and Bolivia seem to attract vessels from countries with good national flag safety records, i.e. a relatively strong estimated negative parameter for OPCASUAL. Different from the general FOREIGN case, Belize and Honduras
Determinants of Vessel Flag
205
both appear to attract vessels from countries with a relatively low life expectancy. Could this be interpreted as implying that those who choose these flags attach a lower priority to safety of life at sea?
6. CONCLUSIONS 6.1. Determinants of a Vessel’s Flag The empirical results of our research suggest that older vessels and vessels that are not classed by an IACS member are more likely to be nationally flagged than foreign flagged. This may come as a surprise to some if the expectation was that foreign flagged ships would be “worse,” i.e. older and less stringently controlled, than nationally flagged ones. A determining factor for a vessel to choose a foreign flag appears to be the likelihood that it trades internationally, as do most cargo and larger vessels as compared to passenger or smaller units. Further, if a vessel is built in the operator country this increases the likelihood of remaining in the national flag registry. Another interesting conclusion is certainly the impact of socio economic variables. The operators from developed countries are more likely to choose a foreign flag than those from countries with a lower GDP per capita or with low indicators concerning human development such as the literacy rate or the life expectancy. Our empirical results suggest that higher wages and labour standards may scare operators away from national registries. If the operator’s country of domicile has ratified a high number of IMO conventions, and also if its national flag register has a positive past safety record, i.e. a low casualty rate, this increases the likelihood to use a foreign flag. This should of course not be interpreted in a way that ratification of IMO conventions should not be recommended – neither would we recommend to reduce the GDP per capita or the life expectancy in order to attract more nationally flagged vessels. However, international organizations and national registries should be aware of this situation where high national safety standards appear to encourage vessel operators to choose a foreign flag. There exist of course big differences between Flag States. We have shown how different Latin American and Caribbean registries have specialized in certain operator countries and vessel types and sizes. From the vessel types, their ages, and the chosen classification societies, one would expect that Panama and Antigua will be less likely to have their vessels detained, whereas St. Vincent and the Grenadines, Belize, Honduras and Bolivia should find their vessels with a higher probability of being inspected and detained by Port State Control authorities.
206
JAN HOFFMANN ET AL.
For Panama, the world’s largest registry by far, our data suggests that it has a fleet of relatively larger and younger vessels, more likely to be classed by an IACS member. This IACS member is most likely to come from Asia, and the same applies to the operator’s country of domicile and the country of build. Panama does not appear to be particularly attractive to operators from countries with a high number of ratified IMO conventions, indicating that Panama is not chosen as a safe haven to avoid having to comply with IMO’s rules and regulations. 6.2. Developed vs. Developing Countries – Who Benefits from the Open Registries? Operators from developed countries are most likely to choose a foreign flag. This allows them to remain competitive in a business environment where developing countries might otherwise have a competitive advantage due to lower wages and, perhaps, less stringent safety and environmental standards. The question then arises if the whole system of open registries is in general unfavourable for developing countries. Should international organizations such as UNCTAD continue to try to reduce the use of open registries, in the belief that the system is detrimental to developing countries’ interests? Against this background, there are three reasons why developing countries, including the eight countries which were particularly looked at in this paper, might actually benefit from the present system: (1) First, most open registries are based in developing countries (Table A1). For some of them, this is a relevant business, i.e. an export of a service. (2) As indicated also by our research, one of the main reasons why operators choose a foreign flag is related to crewing costs – and without open registries far fewer seafarers from the Philippines, Indonesia, Honduras etc. would find employment aboard. These countries would certainly be very much against a dismantling of the open registry system. (3) If the use of open registries really reduces costs, as we assume, this eventually leads to lower freight rates. Lower freight rates are particularly to the benefit of developing countries, who pay more on average for the transport of their imports and exports, and who depend more on maritime transport than do the developed countries.
6.3. Future Research Our research suggests that there exists a close relationship between the socioeconomic development of a country and the likelihood to attract nationally flagged
Determinants of Vessel Flag
207
vessels. The challenges and opportunities that arise for developing countries could be a topic for future policy-oriented research. Concerning maritime safety, more detailed research should attempt to shed light on the relationships between the determinants of flag choice and flag registries’ safety records. As has been shown, younger and IACS classed vessels are more likely to choose a foreign flag, which should be a positive indicator in relation to maritime safety. At the same time, however, a bad past safety record and a low number of ratified IMO conventions encourages vessel operators to choose the national register. Put differently, strict national controls appear to scare operators away into foreign flags. In this context, future research might look at the impact of specific IMO and other conventions to identify which conventions exactly are empirically most likely to attract or scare away vessels into foreign flag registration. Flag States could use this type of research to target specific markets to increase their fleet – or to defend their national flagged fleet. They might, for example, modify their pricing structure based on different variables that are estimated to have an impact on the likelihood of a vessel to choose a foreign flag. As a second step, Flag States could also take into account which variables are most likely to be linked to substandard vessels, and consequently apply statistically supported criteria when accepting or rejecting vessels to their registries.
NOTE 1. The authors gratefully acknowledge the supply of that information from Lloyds Register Fairplay. For further information visit http://www.ships-register.com/
REFERENCES Alderton, T., & Winchester, N. (2002). Flag states and safety: 1997–1999. Maritime Policy and Management, 29(2), 151–162. Bergantino, A., & Marlow, P. B. (1999). An econometric analysis of the decision to flag out. Cardiff: SIRC. Li, K. X., & Wonham, J. (1999). Who is safe and who is at risk: A study of 20-year-record on accident total loss in different flags. Maritime Policy and Management, 26(2), 137–144. OECD (2003a). Ownership and control of ships. Paris, March. OECD (2003b). Cost savings stemming from non-compliance with international environmental regulations in the maritime sector. Paris, February. Roberts, S., & Marlow, P. B. (2002). Casualties in dry bulk shipping (1963–1996). Marine Policy, 26, 437–450.
208
JAN HOFFMANN ET AL.
Talley, W. (1999). Determinants of ship accident seaworthiness. International Journal of Maritime Economics, 1(2), 1–14. Talley, W. (2002). Non-seaworthy risks of bulk ship accidents. International Journal of Transport Economics, XXIX(1), 3–15. Thanopoulou, H. A. (1995). The growth of fleets registered in the newly-emerging maritime countries and maritime crises. Maritime Policy and Management, 22(1), 51–62. Thanopoulou, H. A. (1998). What price the flag? The terms of competitiveness in shipping. Marine Policy, 22, 359–374. UNCTAD (1986). United Nations convention on the registration of ships. Geneva. UNCTAD (1994). The place of general average in marine insurance today. UNCTAD/SDD/Leg/1. Geneva, March. Winchester, N. (2003). Flag state audit 2003. Cardiff: SIRC.
APPENDIX A: OPEN AND NATIONAL REGISTRIES, JANUARY 2003
Registry (Ranked by GT)
GT
Units
TEU
DWT
Panama** Liberia* Bahamas** Greece Malta* Cyprus* Singapore* Norwegian International Register* United States China Hong Kong Marshall Islands* Japan Italy Russia United Kingdom Germany St. Vincent and the Grenadines**
128847021 52633013 34898752 27719307 26808191 23296834 21512511 19230527
5626 1645 1338 1051 1406 1351 1172 744
1480409 930017 349066 169562 212841 373280 350521 127979
193080967 79423906 48590858 46131822 43601920 36876291 33631765 29241576
17223973 15558825 14981444 13273846 13202619 9894662 8077939 7744380 7275160 6634756
1630 1645 569 370 2575 807 2140 929 584 1071
297566 187373 210680 145859 49845 127285 69910 200149 575675 78432
19443571 23230280 25187110 21918659 16235867 10674426 7978577 7371573 8664070 9572546
Determinants of Vessel Flag
Registry (Ranked by GT) Korea (South) India Isle of Man* Netherlands Turkey Malaysia Danish International Register* Bermuda** Philippines Antigua and Barbuda** Taiwan Norway Iran Brazil Kerguelen Islands* Sweden Indonesia Canada Denmark Cayman Islands* Kuwait Australia Cambodia* Thailand Saudi Arabia Canary Islands Finland Luxembourg Belize** France Egypt Netherlands Antilles* Vanuatu* Vietnam Ukraine Algeria
209
GT
Units
TEU
DWT
6390606 6285300 6155831 5760478 5487251 5483118 5307031
871 487 300 913 747 434 302
66747 21230 60621 237710 64180 74843 273298
9854163 10216716 9772486 6443630 8502229 7796908 6851899
4912086 4907163 4831335 4531387 4129964 4104917 3461569 3341217 3122840 2962851 2749531 2334417 2323992 2261556 2020625 1922554 1680996 1610243 1600745 1589632 1539974 1507086 1502177 1430899 1398140 1355979 1196432 1012633 944205
113 507 840 195 901 208 242 90 291 718 486 247 150 54 198 486 354 139 183 166 74 566 185 209 165 238 184 302 80
36074 24384 382680 111880 2835 34633 28126 55577 29889 33050 4828 61368 7292 15024 2970 16957 31223 24607 18964 18617 12702 10382 3100 16882 47747 5480 10887 9554 3580
7042220 7095152 6316247 7090064 4842823 7022548 5529012 5366928 1917543 4019601 3239376 2695496 3660906 3623585 2586501 2727326 2616066 2488681 1807459 1313656 2216857 2028624 1452865 2063956 1822668 1516700 1843977 987816 1053718
210
Registry (Ranked by GT) Madeira Bulgaria Gibraltar* Chile Croatia United Arab Emirates Mexico Venezuela Qatar Israel Poland Korea (North) Switzerland Honduras** Romania Nigeria Spain Brunei Darussalam Azerbaijan Syria Lithuania Myanmar* Barbados* Unknown Morocco Estonia Argentina Tonga Bangladesh Bahrain Lebanon* Libya Yemen Georgia Ireland Ecuador Pakistan
JAN HOFFMANN ET AL.
GT 907809 872407 857971 794177 784816 756495 753470 709570 656481 606510 591542 576428 570146 565427 538075 516487 478945 477229 452488 426312 399363 399105 394547 392071 371240 343025 324532 319749 283136 278120 271394 262676 260001 255112 246740 240833 225601
Units
TEU
DWT
134 81 98 111 91 179 117 85 56 24 85 97 27 329 120 66 233 17 110 151 76 47 54 89 81 75 84 130 50 32 70 25 10 82 75 45 20
9415 5639 25449 7216 10513 19974 283 735 14653 44595 541 3820 5798 2814 2425 752 2627 – 376 5472 8064 3988 4402 6743 7844 5653 1326 3908 10528 8472 4974 1477 – 2298 3050 – 7402
1343034 1290625 1234394 1062048 1175904 864185 1042418 1127547 986921 704614 806939 854755 1002749 768659 698109 881776 493744 442201 383646 638967 392774 599836 535320 508345 314367 218538 437498 433887 391772 385058 391743 359301 529475 361024 176760 404224 339249
Determinants of Vessel Flag
Registry (Ranked by GT) Iraq Belgium Portugal Jordan Wallis & Futuna Islands Tunisia New Zealand Faroes* South Africa Bolivia** Iceland Sao Tome et Principe* Comoros Latvia Peru Cuba Ethiopia Puerto Rico Cameroon Republic Sri Lanka Mauritius* Papua New Guinea Jamaica Uruguay Greenland Seychelles Tuvalu Maldives Angola Colombia Paraguay Equatorial Guinea* Austria Reunion Sudan Tanzania Ghana
211
GT 224116 200719 198703 192980 182423 179431 172071 158937 141281 138699 125755 124264 85660 85285 84995 84760 81782 77074 75227 72654 62780 55649 54114 52400 51699 50532 45979 45525 40856 35796 35364 33263 33182 31562 29854 29805 27351
Units
TEU
DWT
41 92 68 9 7 21 62 60 51 55 122 35 16 25 54 27 9 4 5 27 17 38 7 19 35 10 7 18 14 24 25 41 8 5 4 10 21
772 400 5212 1853 – 696 587 94 2423 862 724 660 408 681 316 392 2032 2868 – 1729 244 1579 188 438 – 364 705 1234 226 – 1107 67 2098 – 922 247 –
323083 256345 294510 370125 92719 160767 161472 174802 117560 187998 64412 175104 141551 60596 117220 110205 101375 53024 139593 97215 75189 69091 70926 32666 18470 63000 58709 62368 62330 50341 44132 28785 42223 45885 39084 32573 33943
212
JAN HOFFMANN ET AL.
Registry (Ranked by GT)
GT
Falkland Islands Turkmenistan French Polynesia Albania Eritrea Fiji* Belarus Namibia Madagascar Oman Trinidad & Tobago Kenya Samoa Mozambique Congo (Democratic Republic) Slovakia Cape Verde Islands Yugoslavia Micronesia Mauritania Senegal Gabon Hungary New Caledonia Guadeloupe Virgin Islands (British) Guyana Kiribati Costa Rica Laos Solomon Islands Guatemala El Salvador Macau Somalia Kazakhstan Dominican Republic
27319 26970 26840 23599 20010 20004 18871 17906 15401 14932 14594 10397 8951 8368 7469 5994 5895 5714 5362 5012 4976 4234 3784 3579 3370 3012 2946 2840 2509 2370 2303 2165 2109 2007 1916 1897 1880
Units
TEU
14 11 21 20 8 16 1 15 11 7 11 12 3 7 4 2 8 4 6 3 7 9 1 4 6 2 2 3 2 1 3 1 1 1 3 2 2
12 150 298 79 527 55 – – 24 74 32 – 650 77 – – 123 – – – – – 274 – – – 68 33 – – – – – – 260 820 –
DWT 18107 24669 15304 33744 25859 19101 14947 12648 19095 10055 9859 10844 8492 11295 10579 8900 8267 2132 4325 3363 1832 3041 5500 3840 214 5095 4507 2609 1108 3110 1004 1727 1650 2184 4539 1446 2581
Determinants of Vessel Flag
Registry (Ranked by GT)
213
GT
Units
TEU
DWT
American Samoa Suriname Niger Togo St Helena Island Cook Islands Haiti Ivory Coast Dominica Congo Slovenia Gambia Sierra Leone St Pierre and Miquelon Guinea Mayotte Total national registries Percent of world Total open registries* Percent of world*
1092 1 – 1200 981 1 – 1599 942 1 – 1002 836 1 – 1187 789 1 – 478 764 2 – – 613 1 – 793 489 1 – 1020 445 1 – 630 413 1 – 329 369 1 – 36 350 1 – 600 349 1 – 411 324 1 – 280 321 1 – – 300 1 – 30 220662866 25083 3087159 299985727 37.8% 57.2% 38.2% 35.4% 363217093 18795 4988121 547235423 62.2% 42.8% 61.8% 64.6%
Grand Total
583879959
43878
8075280
847221150
Source: Authors, based on data provided by LRFairplay. All vessels of 300 GT and above delivered prior to January 2003. ∗ Open registries. ∗∗ Latin American and Caribbean registries included in this paper’s research.
The Grand Total is slightly above the total included in the regressions as some vessels had to be excluded due to missing information such as unknown flag. The distinction between open and national registries is of course not always clear cut. Singapore and the Canary Islands, for example, may be border line cases, and we had to make a subjective choice when assigning them to one or the other group of registries. Vessels with an “unknown” flag were included in the “national registries” total. Comments and suggestions are welcome to
[email protected].
214
JAN HOFFMANN ET AL.
APPENDIX B: VESSEL SUBTYPES
% VGCARGO General Cargo Ship Multi-Purpose Ship Livestock Carrier Heavy Lift Ship Semi-Sub Heavy Lift Vessel General Cargo/Part Refrigerated Ship Log Tipping Vessel Deck Cargo Ship
64.97 31.83 1.25 1.17 0.41 0.33 0.03 0.01
VCONT Container Ship Container Ship/all Reefer
99.83 0.17
VLBLK Tanker Products Tanker Crude Oil Tanker Chemical Tanker LPG Carrier Chemical/Oil Tanker LNG Carrier Bunker Tanker Asphalt Tanker Replenishment Tanker Parcels Tanker Bitumen Tanker Wine Tanker Ethylene Tanker Storage Tanker Water Tanker Fruit Juice Tanker Sulphur Tanker
24.54 20.77 16.78 13.75 10.24 6.89 1.59 1.37 1.15 0.91 0.64 0.37 0.31 0.23 0.15 0.13 0.10 0.07
Determinants of Vessel Flag
215
% VDBLK Bulker Bulk Carrier Ore StrenGThened Bulk/Container Carrier Bulk Cement Carrier Bulk Wood Chip Carrier Bulker - Great Lakes Only Ore Carrier Bulk Vehicle Carrier Bulker Great Lakes only/Dumb
58.11 28.86 5.57 3.90 1.38 1.37 0.43 0.25 0.13
VPASS Passenger/Vehicle Ferry Passenger/Cargo Ship Passenger Vessel Multi-Hull Passenger Ferry Cruise Ship Passenger/Train/Vehicle Ferry Multi-Hull Passenger/Vehicle Ferry River Cruise Ship Passenger Hydrofoil Passenger Excursion Vessel Passenger Hovercraft Casino Ship Surface Effect Passenger Ferry Passenger/Vehicle Hovercraft Surface Effect Passenger/Vehicle Ferry
43.55 13.38 12.34 11.53 10.13 3.18 2.86 0.91 0.81 0.42 0.42 0.23 0.13 0.06 0.03
VFSTUG Fishing Vessel Tug Barge Hopper Suction Dredger Oil Barge Hopper Barge Salvage Tug Research Vessel Suction Dredger
37.51 15.92 5.13 4.17 4.10 2.20 2.16 1.98 1.93
216
JAN HOFFMANN ET AL.
% Buoy Tender Crane/Derrick Barge Cable Ship Training Vessel Oceanographic Vessel Dredger Fisheries Research Icebreaker Logistics Vessel Anti-Pollution Vessel Cutter Suction Dredger Survey Ship Bucket Dredger Fisheries Protection Vessel Pusher Tug Survey/Research Vessel Grab Dredger Live Fish Carrier Salvage Vessel Geophysical Research Vessel Pontoon Sail Training Vessel Sludge Carrier Museum Ship Container Barge Search And Rescue Vessel Maintenance/Utility Vessel Pleasure Craft Sand Loading Dredger Cable Repair Ship Icebreaker/Buoy Tender Oil Storage Barge Pilot Vessel Cement Storage Barge Accommodation Vessel Diamond Mining Vessel Weather Ship
1.41 1.34 1.28 1.22 1.18 1.09 1.04 0.98 0.93 0.88 0.84 0.79 0.75 0.73 0.71 0.59 0.50 0.49 0.49 0.46 0.46 0.46 0.44 0.35 0.34 0.32 0.28 0.26 0.26 0.24 0.24 0.24 0.24 0.21 0.19 0.19 0.18
Determinants of Vessel Flag
217
% Fisheries Training Restaurant Ship Ro-Ro Barge Floating Hotel Hospital Ship Spent Nuclear Fuel Carrier Waste Disposal Vessel(Liq) Icebreaker/Research Vessel Floating Dock Patrol Vessel Polar Research Vessel Whaling Vessel Cutter Dredger Fire Fighting Vessel Exhibition Vessel Floating Power Station Support Ship Drilling Barge Incinerator & Waste Disposal Vessel Naval Vessel Tank Cleaning Vessel Dipper Dredger Floating Car Park Radio Station Transshipment Vessel Asphalt Barge Dragger Dredger Floating Wave Powered Power Station Pile Driving Vessel Radioactive Waste Carrier Rocket Launch Vessel Sealing Vessel VRORO Ro-Ro Vehicle Carrier Pallet Vessel
0.16 0.16 0.16 0.15 0.15 0.15 0.15 0.12 0.10 0.10 0.10 0.10 0.07 0.07 0.06 0.06 0.06 0.04 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.01 0.01 0.01 0.01 0.01 0.01 0.01 48.67 28.86 6.40
218
JAN HOFFMANN ET AL.
% Ro-Ro/General Cargo Ro-Lo Ro-Ro/Heavylift Barge Carrier Ro-Ro/Cellular Multi Hull Ro-Ro Freight
6.15 5.90 1.55 1.20 1.20 0.05
VREEF Reefer Refrigerated Fish Carrier
56.47 43.53
VOROIL Ore/Bulk/Oil Carrier Ore/Oil Carrier Product/Ore/Bulk/Oil
51.96 35.20 12.85
VOFSHORE Anchor Handling/Tug/Supply Supply Vessel Self-Elevating Mobile Offshore Drilling Semi Submers Mobile Offshore Drilling Safety Standby Vessel Anchor Handling/Tug Seismic Survey Vessel FPSO Diving Support Vessel Crewboat Offshore Maintenance/Utility Vessel Offshore Cargo Barge Drillship Offshore Construction Vessel Offshore Support Vessel Floating Storage Offtake Survey Ship Rov Support Offshore Drilling Barge Pipe Laybarge Pipe Carrier/Platform Supply Oil Well Stimulation Vessel Offshore Accommodation Vessel
26.76 25.30 8.09 5.99 5.23 4.32 3.28 2.68 2.25 2.01 1.79 1.37 1.34 1.16 1.09 1.00 0.94 0.85 0.67 0.61 0.58 0.49
Determinants of Vessel Flag
219
% Aht/Salvage Floating Production Unit Rock Laying Ship Multi Function Service Vessel Pipelay Vessel Self-Elevating Production Unit Oil Well Production Test Vessel Trenching Vessel Semi-Sub Pipe Laybarge Source: Authors, based on data provided by LRFairplay.
0.46 0.46 0.43 0.24 0.21 0.18 0.09 0.09 0.06
7.
THE CONTAINER SHIPPING INDUSTRY AND THE IMPACT OF CHINA’S ACCESSION TO THE WTO
Kevin Cullinane 1. INTRODUCTION The objective of this paper is to present an overview of the current major trends in the world’s liner shipping and container port sectors and to evaluate the possible short to medium term outlook for both these industries. In so doing, the importance of China’s position in the world economy and, more specifically, the impact of its international trade cannot be ignored. In particular, the likely effect of China’s recent accession to the WTO is analysed as a specific influence on the future of the two industries.
2. LINER SHIPPING RATIONALISATION In the liner shipping sector, the most pronounced trends over the last few years have been greater rationalisation (at both corporate/strategic and operational levels) and the deployment of Super post-Panamax containerships. These two strategies are implicitly interdependent and, in consequence, their individual impact upon the container shipping market is very difficult to isolate. Instead, for ease of analysis, they must be considered as a single causal market development (Cullinane & Khanna, 1999). Shipping Economics Research in Transportation Economics, Volume 12, 221–245 Copyright © 2005 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0739-8859/doi:10.1016/S0739-8859(04)12007-6
221
222
KEVIN CULLINANE
Fig. 1. Control of Global Slot Capacity by Top 20 and Top 10 Carriers. Source: Based on data derived from Informa U.K. Ltd (2004).
The trend towards greater rationalisation is illustrated in Fig. 1 which shows that the percentage of TEU slots controlled by the top 20 operators has continued to increase dramatically over recent years. In fact, this figure could have been taken back to 1985 to reveal that at that time the top 20 carriers controlled just 36% of slot capacity. As can be seen in Fig. 1, however, by September of 2003 their share had increased to 65%. These top 20 operators are also responsible for approximately 62% of the slot capacity on order as that time. In addition, as at September of 2003, the top 20 operators control 84% of the cellular container fleet compared to just 70% in 1998.
2.1. Corporate or Strategic Rationalisation This has two major sources: M & A activity in recent years has been immense; examples include the mergers of P&O with Nedlloyd and Maersk with Sealand, the takeover of APL by NOL, the purchase of 70% of DSR-Senator by Hanjin, CMA’s acquisition of CGM, CP’s takeover of a number of small container lines etc. All of this has resulted in greater concentration in the industry.
The Container Shipping Industry
223
Greater understanding and co-operation among operators. In one manifestation or another, this has provided the basis of the economic structure of the container shipping industry since the 1960s (Deakin & Seward, 1973; Gilman, 1983; Jankowski, 1989a; Jansson & Schneerson, 1987; Neel & Gooding, 1997; Sjostrom, 1989a). It has also been the subject of great scrutiny by competition regulators (Gardner, 1997) and by academic investigations into restrictive practices, barriers to entry and fair competition (Brooks, 1993; Butz, 1993; Clyde & Reitzes, 1998; Davies, 1989; Fox, 1994; Franck & Bunel, 1991; Gilman, 1994; Jankowski, 1989b; Scherer, 2000; Sjostrom, 1989b, 1993; Yong, 1996). However, new world-wide alliances began to emerge at the end of 1995 and have evolved ever since. These strategic alliances represent a significant concentration of power and differ from their predecessors in that (theoretically) a long-term commitment exists, they provide global coverage and they extend to landside integration (Doi, Ohta & Itoh, 2000). Most crucially, alliances provide members with the opportunity to: (a) justify investments in new and larger ships; and (b) exert greater purchasing power in negotiations with ports and terminals. Increasing economies of scale and scope, together with greater rationalisation, has meant that barriers to market entry have risen so rapidly that small and mediumsized lines no longer have any real opportunity for moving up into the major league (Heaver et al., 2000). This is so because increasingly it is the attainment of a critical mass in operations that makes feasible the appropriate quality of service which guarantees market share. The increasing cargo concentration which naturally results from increasing ship size also implies that partners in an alliance will be much more interdependent as they seek to procure sufficiently high load factors to fully reap the benefits that come from economies of scale (Cullinane & Khanna, 2000). The continued predilection towards deploying large size tonnage has provided a catalyst for the proliferation of more agreements and the broadening of the power base of liner companies through alliances, mergers and acquisitions. An important corollary of increasing ship size, therefore, will be the further concentration of power into the hands of fewer main players so that only megacarriers will be in a position to effectively compete for mainstream liner business. Smaller operators certainly and perhaps even some of the big names will continue to be forced to retrench to niche markets. At the same time, however, rising entry barriers will still not lead to higher returns in the industry since the competition laws of the European Commission and the Federal Maritime Commission in the USA, together with the power exercised by the shippers’ councils, will not allow shipping lines to abuse any oligopoly power
224
KEVIN CULLINANE
they may retain in the future. Indeed, at the time of writing, the era of the liner shipping conference looks set to be coming to an end for voyages originating in, or destined for, the European Union (Anon., 2004; Haralambides et al., 2003; Porter, 2004). The rapid expansion of supply with only moderate annual increases in demand predicted will also mean that margins are likely to continue to be tight for some time. 2.2. Operational Rationalisation Intermodal and general developments in transport infrastructure and logistical systems, together with the growth in container vessel size, point to the obvious potential of operational rationalisation in terms of the number of port calls and the adoption of the load centre concept where, from a particular carrier’s perspective, certain ports become the foci of their operations within a region (Hayuth, 1988; Slack & Starr, 1994). As suggested by Robinson (1998) and Cullinane, Wang and Cullinane (2004), this also means that different ports tend to fulfil different types of role, depending upon the extent of their relative participation in international, regional or local liner shipping networks. The rationale behind the load centre concept is that not only does it allow economies of scale in the inland as well as the maritime transport system, but beyond that it also allows shipping lines to increase the utilisation of ships and reduce port time and charges. It is clear why the main proponents of the load centre concept are those carriers that deploy large containerships. Simultaneously, however, where the application of the concept leads to increased port volumes, then this will mean a better distribution of the high fixed costs of the port. The likely impact of the recent escalation in ship size on the pattern of port calls and the future application of the load centre concept is perhaps difficult to estimate. There is little doubt that the least cost solution is that large-sized containerships should be used on concentrated itineraries with the required feeder connection provided by rail, road or sea, depending on the circumstances. Historically, however, this seems to quite often contradict the practice of shipping lines. According to Gilman (1983), there are three reasons why this should be so: Basic demand characteristics in liner shipping suggest that cargo distributions are dispersed and not susceptible to aggregation in the same way as industrial processing. There are considerable secondary distribution costs. Feedership and inland transport costs are much higher per TEU-mile than the cost of mainline ships.
The Container Shipping Industry
225
The scale economies of liner shipping are not quite as powerful as expected and are not forsaken by multi-port calling, especially if the ports are geographically concentrated. It would seem that the point of balance between a transhipment solution and traditional multi-port calling has, in the past, been further to the advantage of the latter than had been expected. The recent phenomenal escalation in ship size, the fact that technologically there is little to impede further increases in the size of Super post-Panamax ships (Wijnolst, Scholtens & Waals, 1999) and the proposed and likely developments in mega ports and terminals all add up to the increased application of the load centring approach. Ever since 1980, transhipment traffic has been growing consistently and there is plenty of industrial optimism about the future of this niche business. Of course, we should also recognize that growing transhipment volumes also means that there is scope for reaping economies of ship size in the feeder business. In fact, there is still some merit in both approaches. The degree of load centring which provides a least cost solution is route specific and, as such, is difficult to generalise. It is determined by the trade-off between the costs associated with the level of cargo dispersion from the load centre and the extra cost of calling at an additional port. This, in turn, will depend on a complex interaction of factors including the size of the mainline ship, the amount of cargo at the diversion port, diversion distance, transhipment cost, port access time, handling cost, productivity of the port etc. (Adcock, 1995; Cullinane & Khanna, 1999; Gilman, 1983). Over recent years, at the same time as carriers have moved towards the load centre concept, container traffic has in fact become decentralised; indicating a trend towards a more evenly distributed cargo concentration between ports. This implies that the advantages of the load centre concept appear at the carrier level, but not necessarily at the port level. Thus, while shipping lines can gain by concentrating their operations in a limited number of ports and by reducing the number of calls which they make (particularly by big ships), this does not necessarily imply that all shipping lines will use the same port as their load centre (Slack, Comtois & McCalla, 2002). In summarizing the carriers’ position, the economics of large containerships are such that the adoption of the load centre concept by containership operators and the continued proliferation of hub ports as the foci for container operations is a trend which is unlikely to cease in the short to medium term. Indeed, given the highly positive correlation between the operation of Super post-Panamax vessels, company size and membership in a world-wide container shipping alliance, both industrial and hub concentration are likely to be self-perpetuating (Cullinane & Khanna, 2000).
226
KEVIN CULLINANE
3. PORTS AND TERMINALS There is an interdependent relationship between ship size and port design; improvements in cargo handling techniques in ports have made new types of ships and bigger ships more profitable while, on the other hand, changes in ship design have induced ports to develop. The latest generation of containerships place considerable demands on terminals and ports in the form of additional infrastructure, cranes, depth in port, productivity etc. This additional required capital expenditure has significant implications. As does the fact that high throughput and the adequate utilisation of resources are vital, not only for a container terminal to remain competitive in the long run, but also if it is to achieve a reasonable return on investment. Because ports will require sufficient infrastructure in terms of berths, depth of water and craneage, fewer ports will be in a position to compete for the larger ships that are entering service. In fact, the economics of containership operation are crucially dependent on port productivity and, in this respect, it is important to recognize that there is a proven relationship between the throughput of a terminal and its productivity (Cullinane, Song, Ji & Wang, 2004). To quantify this relationship, one study has shown that a terminal with a throughput of 11 million container moves per annum is expected to unload a vessel in 88% of the time taken by a port with only 5 million moves (Tabernacle, 1995). A high throughput of containers for a terminal means better productivity for ships which, in turn, will attract still more ships to call at the terminal. All this adds up to the fact that fewer ports will be in a position to compete for these larger and larger ships. Conversely, poor productivity at a port can prove detrimental to the attraction of traffic. Thus, high throughputs are self-reinforcing. With increasing ship size, it is inevitable that the bargaining power of the large ship operators will increase; a trend that has been very much enhanced by the formation of the large, worldwide alliances and the greater industrial concentration within the sector. Indeed, the major alliances wield enormous purchasing power over ports and terminals and, clearly, this ability will further erode the competitive status of small and medium size operators in liner shipping. Faced with fewer shipping lines (due to the combined effect of continued industrial concentration, market domination by a small number of large alliances and the deployment of larger vessels), mega-terminals become not merely a luxury but, rather, a necessity (Fossey, 1995). Investment in the ports sector reflects this tendency – especially in Western Europe, the western seaboard of the USA and in China (a country that, according to Frankel (1998) invested more in its ports in the 1990s than the rest of the world combined).
The Container Shipping Industry
227
All is not lost, however, for smaller and medium sized ports since opportunities exist for a second tier of ports and terminals that do not target ultra large containerships; an aspect implied by Robinson (1998). These ports can feed off their larger counterparts rather than competing with them, thus removing much of the need for investment in the most expensive infrastructure. To succeed, this set of ports will have to streamline their activities to target niche shipping markets, especially in feedering. Although the implications for ports are contentious in some cases (Baird, 1996, 1997; Notteboom, Coeck, Verbeke & Winkelmans, 1997), in the absence of financial incentives or any other form of market intervention, the economics of containership operation are such that operator and port specialisation are likely to become more common. Ports which are blessed with a favourable geographical location with respect to a trade route can probably expect to continue to enjoy direct calls by mainline vessels. They can also aggressively pursue transhipment business, a strategy which is facilitated by the fact that transhipment cargo and the hub factor have a generative effect on a port’s overall volume of business; a high volume of transhipment traffic attracts feeder services and a high frequency of feeders further reinforces calls made by mainline vessels (Port Development International, 1995). Ports which are geographically removed from main trade lanes will, on the other hand, become increasingly unable to attract larger size mainline vessels. With greater industrial concentration, service rationalisation and deployment of ever-larger containerships, there is a need for the hinterlands of load centre ports to expand. The larger the average size of ship calling at a specific load centre port (particularly if service frequency is maintained as ship size increases), the larger is the scale of required hinterland expansion. This will be facilitated by infrastructure investment and the establishment of appropriate logistical systems. Eventually, and especially in certain geographic areas, the potential hinterlands of different load centre ports will overlap and the “neighbouring” ports concerned will then be competing directly with each other for the same cargo base. This scenario will be exacerbated in situations where different carriers select different hub ports for their operations. A notable characteristic of investment in ports over the past few years has been the globalisation of the container handling sector. There are, of course, two prominent Asian players in the market. At the time of writing, the PSA Group (http://www.internationalpsa.com/) participates in 17 port projects in 11 countries: Singapore port; Antwerp and Zeebrugge ports in Belgium; Holland Terminals in Rotterdam;
228
KEVIN CULLINANE
Voltri Terminal Europa and Venice Container Terminal in Italy; Sines Container Terminal in Portugal; Tuticorin Container Terminal in India; Dalian, Fuzhou and Guangzhou Container Terminals in China; Incheon Container Terminal in Korea; Muara Container Terminal in Brunei; Eastern Sea Laem Chabang Terminal in Thailand; Hibiki Container Terminal in Japan.
At the time of writing, Hutchison Port Holdings (http://www.hph.com.hk/ business/ports/ports.htm) has an interest in 36 ports worldwide, as shown in Table 1. In the last five years or so, both these major independent port operators have dramatically expanded their sphere of influence. This strategy has been aided, of course, by the general policy over the past decade for states to privatise their ports sector (Baird, 1995; Cullinane & Song, 2002; Hoffman, 2001; Peters, 2001). This has occurred to such an extent that there now exists only limited scope for buying into newly privatised, but well-established container ports and terminals. It is true to say, however, that the perceived success of port privatisation policies does mean that they will continue to be implemented where they may be relevant and this, of course, provides more opportunities for PSA, HPH and other international container handling companies to expand their activities still further. The increased concentration within the liner shipping industry ostensibly means that this sector will exert greater power in the port services market. This has obvious implications for the competitive success of ports. Mirroring what has been happening in the carrier sector, however, there has also been a concentration of power within the container handling sector over recent years and this looks set to continue (De Souza, Beresford & Pettit, 2003; Notteboom, 2002). This increase in industrial concentration in the container handling sector, together with the emergent strategy of forming regional port alliances (Anon., 2002a, b, c; Song, 2003; Stares, 2002), both provide a method of combating the power exerted by carriers over ports. Those carriers that have sought to operate dedicated terminals have undermined the progress towards a global steady-state duopoly in the provision of container handling services. This vertical integration has a number of advantages from a carrier’s perspective but no company has sought to implement it more forcefully than the biggest carrier of them all, Maersk-Sealand. This company’s involvement in the container handling sector now extends to it taking the number 3 position in the rankings for container throughput handled in 2001 and the announcement in 2002 of not only a new brand name in APMoller Ports but also a new strategy in actively competing for third party cargoes with other port operators.
The Container Shipping Industry
229
Table 1. Hutchison Port Holdings Worldwide Port Interests (as at September 1st 2004). Hong Kong
Hongkong International Terminals, COSCO – HIT Terminals, Mid-Stream Holdings River Trade Terminal
Chinese Mainland
Jiangmen International Container Terminals Nanhai International Container Terminals Ningbo Beilun International Container Terminals Shanghai Container Terminals (Puxi) Shaghai Pudong International Container Terminals Shantou International Container Terminals Xiamen International Container Terminals Yantian International Container Terminals Zhuhai International Container Terminals (Gaolan) Zhuhai International Container Terminals (Jiuzhou)
Other Asia
Hutchison Korea Terminals Jakarta Container Port Karachi International Container Terminal Korea International Terminals Myanmar International Terminals Thilawa Thai Laemchabang Terminal KMT – Westport, Port Klang
The Americas
Buenos Aires Container Terminal Ensenada International Terminal Freeport Container Port Internacional de Contenedores Asociados de Veracruz Panama Ports Company – Balboa Panama Ports Company – Cristobal Port of Lazaro Cardenas Terminal International de Manzanillo
Europe
DeCeTe Duisberger Containerterminal Europe Container Terminals Harwich International Port Port of Felixstowe Thamesport Trimodal Container Terminal Belgium
Middle East & Africa
International Port Services, Saudi Arabia Tanzania International Container Terminal Services
230
KEVIN CULLINANE
4. CHINA’S WTO ACCESSION1 4.1. Background On December 11th 2001, following 15 years of negotiation, China officially became the 143rd member of the WTO. As China’s trade expands exponentially, no industry is more affected than the shipping sector. This is particularly the case as rising trade volumes are accompanied by easier access to China’s international and domestic logistics market and, in consequence, to a sharp increase in competition as overseas companies enter the market to carry China’s trade and to secure business at source. In 2002, China was ranked No. 4 in the top 35 most important maritime countries and territories, after Greece, Japan and Norway (United Nations Conference on Trade and Development, 2003). Its waterborne cargo volume in 2002 had reached 1,420 million tons and total demand had reached 2,751 billion ton/km; increases of 6.8% and 5.9% respectively over the equivalent figures for 2001. Out of this total, 300 million tons were carried by ocean shipping (2,173 billion ton/km). In total, the cargo volume and the total ton-kms carried by water transport made up 9.6% and 54.5% respectively of the overall national figures for freight transport across all modes (China Ministry of Communications, 2002). Figure 2 shows this increasing trend in China’s waterborne transportation and attests to the significance of the role that China plays in the world shipping industry. As of May 2004, there were 22 wholly foreign-owned shipping companies, 71 foreign-owned shipping company branches and more than 900 representative
Fig. 2. Waterborne Transportation in China 1952–2001 (million tons). Source: Based on data derived from the China Ministry of Communications (2003).
The Container Shipping Industry
231
offices of foreign shipping companies operating in China. These enterprises already compete directly in China’s cargo handling, shipping and inland container transportation, international freight forwarding and freight agency markets (China Ministry of Communications, 2003). There are high hopes that the industrial awakening of China will continue apace and overseas shipping companies have been waiting for China to enter the WTO for a long time. This is because there is a belief that China’s accession to the WTO will provide tremendous commercial opportunities for them. The question remains, however, what will be the real impact of China’s WTO accession on international trade, the international shipping industry and non-Chinese shipping operators (especially in the liner trades) and will the expected opportunities envisaged actually materialize in practice. Although its shipping sector was probably the most liberally regulated of all the service sectors even before WTO accession, WTO accession will impose a number of other commitments on China’s maritime transport services. Through its accession, China has committed to:2 Abide by the rules of international trade; Lower import tariffs; Open up its domestic markets; Enhance the transparency of its trade-related regulations. If adhered to, all this will obviously have an impact on how easy it is to do business within China and with Chinese firms. But what is in it for China? Why have they been so keen to become a member of the WTO? As with all these sorts of questions, the answer is multi-faceted. However, the single most important economic benefit to China comes from the lifting of export quotas that have limited its ability to export to the rest of the world’s markets. Whilst China had been granted “Most Favoured Nation” (MFN) status from the USA prior to entry into the WTO, this was on ad hoc basis and was subject to annual review. With WTO membership comes not only recognition as an MFN on a permanent basis, but also recognition from all the members of the WTO. This will particularly benefit China’s labour-intensive textile and clothing exports which will now fall within a commitment under the General Agreement on Tariffs and Trade (GATT) whereby all quota restrictions on imports of textiles and clothes within WTO member states will be eliminated by 2005. In consequence, the phrase “Made in China” is likely to become even more ubiquitous than it already is. The removal of import quotas on China’s products and the imposition of a more stable and transparent trading environment within China are also expected to attract Foreign Direct Investment (FDI) and the relocation of overseas manufacturing operations to China (Panayides, 2003). All this should add considerably to the
232
KEVIN CULLINANE
volume of trade in which China engages; a factor that will clearly have an impact upon both the international shipping and port sectors. Although benefiting from the removal of import quotas of other nations and despite its agreement to reduce a range of tariff and non-tariff trade barriers and other restrictive practices, it is most certainly not the case that China has “sold the family silver.” Although non-tariff barriers are to be eliminated, tariffs will merely be reduced (either immediately or over some extended period of time). Tariffs are not to be eliminated completely, therefore; they are to remain in place but at lower levels. This is important to reiterate since it is a common misconception. Similarly, there are various conditions and other protective measures that have been put in place to secure both the national interests of China and the commercial interests of its industry. A particular example of this is provided by China’s agricultural sector. Because it is inherently inefficient, in the absence of very severe tariffs and quotas on the importation of agricultural goods, China’s agricultural sector would be decimated by overseas competition in its domestic market. Clearly, this would be politically and socially unacceptable. However, even with WTO accession on the grounds negotiated and agreed by China, the country is still bracing itself for enormous unemployment in the agricultural sector. Opportunities for overseas companies arise as a result of a movement towards a system where there will be common treatment for both domestic and overseas companies. Some of the outcomes that are expected to emerge as a result of China’s accession to the WTO are as follows: A surge in China’s external trade and investment flows. An opening up of China’s mainland service sector, including trade and logistics, retail distribution, telecommunications, finance, professional services, tourism etc. Fair competition with equal treatment for domestic and overseas companies. The reduction of tariffs and a better trading environment will lower the import costs of equipment and raw materials. To this end, both domestic companies as well as overseas companies using China as a manufacturing base are set to benefit and become more competitive in export markets. Restrictions on domestic sales by foreign manufacturing companies (including Hong Kong companies) will be lifted so that there will ultimately be free access to China’s domestic market. 4.2. Conditions on China’s WTO membership The following are the main highlights of the membership agreement signed by China (Li, Cullinane & Cheng, 2003):3
The Container Shipping Industry
233
All non-tariff barriers will be eliminated. This will entail the elimination by 2005 of all import quotas (covering 383 import products at point of entry) as well as the need to obtain import licenses and to conform to tendering requirements (which were applicable to 107 product items pre-accession). The average import tariff on industrial products was approximately 14.8% prior to accession. China has agreed that this should be cut to an average figure of 8.9% with a range varying from 0 to 47%, with the highest rates applicable to products such as photographic film, cars and auto parts. For agricultural products, the average tariff pre-WTO membership was 18.9%. Given the sensitivity of this sector this will only be reduced to 15%. There will be greater information available, particularly on China’s subsidy programmes. China’s accession to the WTO agreement will have implications for four business activities that are likely to have a significant impact on international trade over the next few years. In analyzing the effect on each of these activities, it is important to recognize that China’s very bureaucratic license approval system gives no recognition to the existence of “logistics” as a holistic activity and that, therefore, different licenses have to be applied for to operate in either the trucking and/or warehousing sectors. This is despite a booming logistics market and the phenomenal popularity of ‘logistics’ as an academic area of study with China’s young people. 4.2.1. Trading Prior to signing the WTO agreement, China required overseas investors to include various provisions in their investment contracts for government approval. These related to the balancing of foreign exchange, local content and export share. Upon accession, these requirements have now been eliminated. The right to trade in China was, however, originally denied to overseas companies that have manufacturing operations in the country. However, 1 year after accession, full import/export rights (including the right to trade internally) was granted to joint ventures with a minority overseas share. After 2 years, this was to be extended to joint ventures with majority overseas shares and to all enterprises after 3 years following accession. 4.2.2. Freight Transport and Distribution With the exception of a few joint venture companies that had specific permission to operate under a form of pilot scheme, prior to WTO accession foreign companies were not allowed to distribute products produced by other enterprises in China or overseas and only Chinese nationals and Chinese-owned companies were permitted to actually conduct surface transportation within China’s borders. The
234
KEVIN CULLINANE
main exception to this was that foreign participation was allowed specifically for cross-border operations with Hong Kong and then only where the foreign participation took the form of a minority joint venture. With a few product exceptions, overseas companies were allowed to engage in the distribution of all imports and the domestically made products of overseas companies within one year of accession. Product exceptions to this provision include books, papers, magazines, pharmaceuticals, pesticides and certain sorts of film that were to be allowed 3 years from accession and chemical fertilizers, crude oil and oil products that will be allowed to be distributed inside China 5 years after accession. Distinct from this, upon accession a minority joint venture for any actual rail and/or road transport operation became viable. One year after accession, overseas companies were to be allowed to own majority stakes in road transport operations and 3 years after accession, these can then be wholly-owned. The equivalent dates for rail are 3 and 6 years respectively. Entry thresholds on capital investment and business track records will still be quite difficult for potential entrants to meet however. 4.2.3. Warehousing Prior to accession, foreign firms were permitted to own warehouses only in Free Trade Zones (FTZs), provided that the stored materials were limited to those required by their own production or service activities in China. Outside FTZs, no foreign firm was permitted to either own or manage warehouses. Upon accession, however, warehouses can be owned or managed by joint venture companies with a minority stake in overseas hands. This right was to be extended to majority joint ventures 1 year after accession and to wholly-owned companies 3 years following accession. 4.2.4. Freight Forwarding Prior to WTO accession, overseas companies could participate in freight forwarding operations only as minority joint ventures. Although some whollyowned companies did exist, these were merely exceptions to the general rule. Even then, the business of freight forwarding joint ventures was limited only to certain geographical areas with very few being allowed to handle domestic internal shipments. One year after accession, majority joint ventures will be allowed and wholly-owned freight forwarding operations will be allowed after 4 years. Also under the WTO agreement, after 1 year of operating in China, these overseas freight forwarding companies will then be allowed to open up branches, but they will be required to register additional capital of US$120,000 per branch.
The Container Shipping Industry
235
5. IMPLICATIONS FOR THE REGIONAL ECONOMY There is a general perception amongst economists within East Asia that China’s WTO membership spells extremely bad news for the developing nations within ASEAN (the Association of South-East Asian Nations). The rationale for this is based on the following beliefs: The export structures of ASEAN nations are similar to those of China and, therefore, there is little complementarity with China in the products exported. Obviously, this implies a probable deleterious impact on the exports of ASEAN countries to third party nations as their exports are substituted by those of China – so-called trade diversion. This is expected to be particularly the case for labour-intensive products where China has a significant advantage over ASEAN countries in terms of its unit labour costs. This tendency will be enhanced as a result of the greater trading confidence in dealing with China as a result of its membership of the WTO. Pre-WTO, China was a net importer of capital-intensive and certain agricultural products. However, the importation of these products is expected to increase with WTO membership. This is likely to lead to price rises for such products and, therefore, the ASEAN nations will be facing higher prices when importing them. Throughout the 90s, ASEAN’s share of FDI has been declining, largely as the result of the increasing attractiveness of China as a destination for foreign investors (Panayides, 2003). China’s accession to the WTO and its conformity with a worldwide set of standards on trade and investment can only boost the level of FDI which is moving into China. In particular, there is a strong sense that there will be a significant movement of manufacturing operations out of the ASEAN countries and into China as a result of the lifting of export quotas from the latter. It should be recognized that there is the possibility that the opportunities for ASEAN nations to export to China may actually counterbalance these perceived disadvantages. The sheer size of the potential market in China should not be forgotten. In this respect, empirical studies suggest that intra-Asian trade in commodities and products such as wood, rubber, tin, petroleum, palm oil, beverages, tobacco and vegetable oils would greatly expand thanks to China’s greater propensity to import these products from the ASEAN nations. Whether there is a net gain in value to the ASEAN nations depends, of course, upon not only the volume impact of China’s WTO entry, but also the impact on prices. At the same time, however, an expansion of China’s exports of labour-intensive products in the order of US$ 40 billion is expected to result from WTO membership.
236
KEVIN CULLINANE
6. IMPLICATIONS FOR LINER SHIPPING AND CONTAINER PORTS Simply as a result of the increase in trade to and from China that WTO accession will inevitably trigger, the prospects for all shipping sectors look bright.4 In liner shipping, there appears to be hope of expanded trade volumes over the EuropeFar East Trade, the Transpacific Trade and the Intra-Asian Trades. Of course, it is expanded volumes (rather than expanded values) that are good for shipping. It is up to the liner shipping sector to provide an appropriate response to the rising demand for container cargoes to be transported to and from China. Networks and schedules will need to be revised to cater for the greater demand that will come from the import of consumer goods and the export of manufactures in containers. At the moment, the imbalance in this trade is one of the problems that needs to be surmounted. However, expectations are such that this will be less of a problem after the policies associated with WTO accession yield their full effects. In addition, the huge container throughput that moves through China will mean that mainline calls and even load centre status can increasingly be enjoyed by China’s major ports. This is also something for the liner shipping sector to consider. Of course, liner shipping is set to prosper not only from China’s accession to the WTO but also from the natural move towards the enhanced containerization of China’s trade. At the moment, the container penetration rate is very low (at about 45%) and the use of non-standard containers is extremely common. These factors, together with the increasing propensity for the containerization of bulk cargoes (i.e. an expansion of what is containerizable), all suggest an explosion in container volumes with origin or destination China over the next 5 years or so. Obviously, however, it is generally the case that this is set to benefit the Asian based operators such as COSCO, OOCL and Evergreen more than those of Europe or the States, although this competition should also be subject to the level playing field that the WTO intends to establish for overseas vis a vis domestic companies. Having made this assertion, however, it is quite clear that Maersk-Sealand and P&O-Nedlloyd, in particular, have already gone to some efforts to access the burgeoning market, especially with their early FDI involvement in China’s inland logistics. The ports sector is, if anything, even more interesting than what may happen in liner shipping. In a bid for greater efficiency and international standing, there is general consensus that three China ports will emerge as the dominant hubs for different regions of China. Each of the three locations for hub ports corresponds to the geographical concentration of competitor ports: in the North around the bay of Bohai; in the Centre around the mouth of the Yangste River and; in the South at
The Container Shipping Industry
237
the mouth of the Pearl River. All three of these port concentrations have developed historically as the result of the location of industry in areas that are served by these port concentrations (Cullinane, Wang & Cullinane, 2004). With its recent massive year-on-year improvements in throughput and with the recent government decision to invest US$50 billion in a new offshore container terminal area within the port (involving the construction of an 8-mile bridge), it seems certain that Shanghai will emerge as the preferred hub port in the Centre region. This is despite the fact that Shanghai is infamous as a poorly located port that has to be constantly dredged and is difficult to navigate, while Ningbo, which is slightly to the south, is more favourably endowed with better natural berthing (Cullinane, 2003). Against a background of continuing economic growth and the likely impact of WTO accession, Shanghai’s port cargo forecasts (made in 2000) of 8m TEU throughput by 2010 has already been surpassed, and 13m TEU by 2020 looks positively ridiculous. The outcome of the other two hub contests is a little more difficult to call. Situated as it is next to the world’s largest manufacturing location in the Pearl River Delta, as long as Hong Kong maintains its efficiency advantage over the neighbouring Shenzhen terminals, then in the short term it is likely to maintain its competitive advantage and position as China’s main hub port. Under such a scenario, Shenzhen will continue to fulfill merely a niche role in handling less time-sensitive cargoes. However, the terminals in Shenzhen are much closer to the PRD cargo base and their use does not involve transiting the congested and overly bureaucratic border controls that lie between Hong Kong and its neighbouring areas in the mainland. In addition, as can be seen in Table 2, handling costs per container are much less expensive when transiting a Shenzhen port than when moving via Hong Kong.A recent study of freight route/mode choice Table 2. Comparison of Transportation Costs of FEU via Hong Kong and Yantian (from Donguan to Long Beach). Via Hong Kong
Via Yantian
Cost Differential
Trucking to new terminal Declaration fee Documentation fee Basic ocean freight charge Destination delivery BAF Origin receiving charge Terminal handling charge
360 50 14 1160 740 230 N/A 366
166 30 14 1210 740 230 269 N/A
194 20 0 −50 0 0 97
Total
2,920
2,659
261
238
KEVIN CULLINANE
in the PRD showed that the most important factors in selecting route and mode for containerized freight in this geographical context were frequency of sailings, customs clearance in ports and price in that order of priority (Cullinane & Kwan, 2000). It is this prioritization that keeps Hong Kong as the preferred hub port of choice. However, in the medium to long-term, productivity in mainland ports must continue to improve, especially since the Shenzhen terminals are managed as joint ventures with the involvement of all three of the big terminal operators in Hong Kong, HPH, MTL and CSX. Also, WTO accession will play its part in contributing to the easing of customs procedures, not only because of the need to comply with international standards on this, but also because of the reduction in work and associated bureaucracy that is concomitant with the removal of quotas and tariffs. In line with purchasing power parity theory (see Officer, 1982), as Shenzhen develops economically, it will generally possess less and less comparative advantage over Hong Kong. In particular, the price of land and labour will continue to move towards an equilibrium level that is comparable to the levels that exist in Hong Kong. The other hub port location which is difficult to determine is that in the north, where there is a strong contest for leadership between Dalian, Tianjin and Qingdao. Political favours in terms of government investment seem to be going to Tianjin, and its annual throughput has increased accordingly over recent years to virtually take the lead over its two local rivals. However, geographically it is not very well located at the very end of the Bay of Bohai. Indeed, ships need to sail past both Qingdao and Dalian to reach it. However, it is right on Beijing’s doorstep. As well as in ports, the government of China is investing heavily in air and, in particular in the development of air logistics hubs that integrate with port facilities. To this end, Tianjin may well be in a favourable situation as the chosen air freight hub for Beijing. Similarly, Shanghai has a brand new and expensive airport. If this helps explain which ports eventually do emerge as regional hubs, then this might mean that the case for Hong Kong is further supported. Whichever three ports emerge as the preferred regional hubs for the different parts of China, all of Hong Kong’s container terminal operators look set to benefit, but particularly HPH as the leading Hong Kong port operator in China, as elsewhere of course. They are already well diversified into China ports and their ownership opportunities look set to increase under the conditions by which China has acceded to the WTO. The PSA Group has also got some involvement in China but is perhaps more exposed to the risks associated with which ports emerge as regional hubs. However, it too has potential for further participation and greater diversification of its risks therefore. The China involvement of independent terminal operators and of shipping lines is not constrained to their own core businesses. In order to open up their hinterlands
The Container Shipping Industry
239
within China and to capture cargoes at source, many international operators are moving into the 3PL market within China. Maersk-Sealand and P&O Nedlloyd have established quite wide-ranging logistics hubs to secure cargoes inland and facilitate their movement to the ports. PSA and China Merchants Holdings have also formed a new logistics joint venture in China, while OOCL’s involvement even extends to a joint venture in rail. HPH has a joint venture in what is perceived to be a major rail hub for Shenzhen in Ping Yu in the South. These overseas international organizations invariably opt for joint ventures with large Chinese companies that already have international experience from their involvement in the international shipping scene. Typical partners will include Sinotrans, China Merchants and COSCO. Thus, those overseas operations with a track record in the international shipping or ports sector are finding themselves at an advantage when trying to enter China’s domestic logistics market (Panayides, 2003). WTO accession can only widen the opportunities in this field; a trend that is desirable from everybody’s perspective since the real bottleneck in trade flows into and out of China is the country’s archaic inland distribution system, particularly the transport infrastructure (Cullinane, Ji & Wang, 2002). Under the current system, these problems are quickly being addressed. Emergent trade patterns and manufacturing location decisions within China will have an effect upon logistics provider and port selection decisions. Infrastructure development trends, as well as government policies on industrial development, will also have an impact upon this. At the moment, China has a positive policy for the economic development of its currently impoverished western provinces. Roads, rail and inland waterways which access the region are being improved and built. Since costs are at a minimum in the western region, transport infrastructure development may induce overseas FDI into it. If this does prove the case and connectivity runs (as expected) east-west rather than north-south, then Shanghai may move into the position as dominant hub port for China, especially if such investment should stimulate the migration of manufacturing activities out of the PRD and into the Western Provinces. The final wildcard in the equation is the relationship of China to Taiwan. The impact of the restoration of full and complete direct shipping links will be monumental. It is virtually impossible to predict when, or even if, this will happen. However, should this ever occur in the future, the strategic geographic location of Taiwan’s container terminals on the mainline trades (particularly Kaohsiung) would mean a massive realignment of the cargo flows, with even Hong Kong’s official forecasts (GHK, 2000) suggesting that Taiwan will take all of the trade that moves through Hong Kong that is bound for or comes from North China (a small amount) and 65% of the South China trade (rather a large amount!).
240
KEVIN CULLINANE
7. CONCLUSIONS Irrespective of precisely how China’s trade should evolve over the next decade, its accession to the WTO is likely to go down in history as one of the most significant steps in sculpturing the global economic landscape of the 21st century. As an important specific influence on the future of the liner shipping and container port sectors, its impact can be married with that of the more generic influences that were analysed earlier in this paper. By so doing, some tentative predictions of the likely future of both industries can be made on the basis of extrapolating from the current state of these industries and the trends that have recently been exhibited within them. Containerships will continue to grow in size – the average continuously and the largest in phases. Evidence for this has recently been most forcibly provided by the marketing efforts of the Samsung shipyard to secure orders for ships of 12,000 TEU (O’Mahony, 2004). Container carriers will increasingly adopt the load centre concept, so that the continued proliferation of hub ports as the foci for container operations is a trend which is unlikely to cease in the short to medium term. In order to secure cargo volumes to justify the deployment of Super postPanamax tonnage and the application of the load centre concept, traditional shipping lines will increasingly diversify into shore-based logistics to compete head on with the more well-established “logistics” operators. This is, and will be, especially prevalent in Asia (and, in particular, in a post-WTO China), where the 3PL market is still in its infancy and where shipping is, in any case, the preferred mode of carriage. It can be observed, in fact, that in this fast developing market with the world’s biggest growth potential, international shipping lines are already competing very successfully in the land-based logistics sector against the more mainstream U.S. and European logistics companies. This diversification strategy of shipping lines will be supported by, and in some cases contribute to, the development of inland transport infrastructure to enlarge the hinterland of a particular load centre. This is already happening in China and is likely to continue. Because of the geographically fixed nature of ports and the floating allegiance of carriers, and with industrial and hub concentration increasing as well as the vertical integration of carriers into logistics and port operations, there can be no doubt that the carriers hold the upper hand in the container handling market. In other words, it is a buyer’s market. This is especially so on those occasions where the container shipping market itself is oversupplied and freight rates are low. Under such a relatively normal scenario, the pressure is on the port
The Container Shipping Industry
241
sector to compete, either through the reduction of charges or the enhancement of efficiency. This pressure is already manifest in the decisions of Maersk and Evergreen to move their transhipment business from Singapore across the border to the Port of Tanjung Pelepas in Malaysia. In addition, Giao Tauro in the Meditteranean has in the past been badly affected by carriers exerting the power they possess. With continuing improvements to inland logistical systems, hinterlands will continue to expand and overlap. The concept of a port as a regional monopoly with a highly defined hinterland and guaranteed traffic is moribund. To survive in tomorrow’s marketplace, ports will have to be highly competitive and innovative. The development by ports and terminals of EDI systems, the leasing of berths, volume discounts, productivity agreements, in-house feeder services, duty free parks etc. are all examples of sensitivity to their changing role. In particular, the ability to provide B2B e-commerce solutions will provide an important mechanism for service differentiation within both the container shipping and the port sectors. Independent port operators have already diversified their risks to a certain extent through their investments globally. In Asia, HPH in particular is now attempting to also vertically integrate into the logistics market, particularly in the burgeoning market for 3PLs in China. A question-mark exists, however, over how well dedicated container handling companies can compete with the more seasoned logistics campaigners that originated in either the shipping or land-based transport sectors. Container handling and the port and/or terminal ownership or operation which goes with it will probably emerge as an integral element within the integration strategies of the traditional shipping lines. Dedicated terminals will proliferate and carriers will increasingly expand their market share of the container handling sector. Thus, shipping lines themselves will in future provide an important source of competition for the dedicated port operator. This is not only through the carriers’ operation of dedicated terminals, but also through their entry into the common user market. This feature is especially important since PSA tends to specialise in the operation of transhipment hubs, while HPH tends to focus on locations where a hinterland cargo base is secured – thus, there is very little direct competition world-wide between these two main players. Development plans for mega-terminals constitute an important pre-cursor to the ultimate elimination of medium sized mainline container ports. Most fundamentally, the combination of a carrier’s ability to quickly swap allegiances, as well as the global diversification of the leading container handling companies, means that the role of a port as a planning instrument by which the economic prosperity of a nation or region can be ensured has been completely
242
KEVIN CULLINANE
undermined. Nowhere is this more evident than in Asia; the competition with Malaysian and Indonesian ports poses new challenges to Singapore and in Hong Kong, the competition from the port of Shenzhen, just across the border in mainland China, has prompted much debate and is having an influence over government policy. To conclude, it is likely that the pre-eminent position of Asia in international trade and as the focus of the world’s container shipping industry will be further reinforced over the next decade. This puts a great deal of pressure upon the industry’s many institutions to recognize this by the relocation or expansion of these influential maritime institutions into the Asian region.
NOTES 1. This section of the paper represents a summary of the main aspects of China’s WTO accession agreement that may be relevant to the future of the international liner shipping and port sectors. For a more comprehensive analysis, interested readers are referred to Li, Cullinane and Cheng (2003). 2. This part of the chapter is drawn from Cullinane (2002). 3. It should be recognised that the agreement did not contain anything on regulations relating to China’s patent, copyright and trademark laws. This was because these laws had already been revamped in preparation for accession so that they conformed already with the International Agreement on Trade-Related Intellectual Property Rights 4. In bulk shipping, the expected FDI in manufacturing in China looks likely to lead to greater imports, particularly of crude and oil products and possibly iron ore. On the other hand, China is already an important world source of Coal (an export sector that is likely to increase as quotas are removed) and is set to become a net exporter of steel very soon.
REFERENCES Adcock, G. (1995). Shipping lines see future; Its cloudy to others. World Wide Shipping (November), 39. Anon. (2002a). Promoting port understanding. Lloyd’s List (April 23). Anon. (2002b). Finnish ports count the gains of co-operation. Lloyd’s List (April 8). Anon. (2002c). Call for collaboration brings mixed response. Lloyd’s List (March 22). Anon. (2004). Loss of EU block exemption for liner conferences may drive mergers. Lloyd’s List (August 24). Baird, A. J. (1995). Privatization of trust ports in the United Kingdom: Review and analysis of the first sales. Transport Policy, 2(2), 135–143. Baird, A. J. (1996). Containerisation and the decline of the upstream urban port in Europe. Maritime Policy and Management, 23(2), 145–156.
The Container Shipping Industry
243
Baird, A. J. (1997). Rejoinder: Extending the lifecycle of container mainports in upstream urban locations. Maritime Policy and Management, 24(3), 299–301. Brooks, M. R. (1993). International competitiveness – Assessing and exploring competitive advantage by ocean container carriers. Logistics and Transportation Review, 29(3), 275–293. Butz, D. A. (1993). Ocean shipping economics: Free trade and antitrust implications. Contemporary Policy Issues, 11(3), 69–80. China Ministry of Communications (2002). China shipping development annual report. Beijing: Department of Water Transport. China Ministry of Communications (2003). China shipping development annual report. Beijing: Department of Water Transport. Clyde, P. S., & Reitzes, J. D. (1998). Market power and collusion in the ocean shipping industry: Is a bigger cartel a better cartel? Economic Inquiry, 36(2), 292–304. Cullinane, K. P. B. (2002). China’s accession to the WTO: The implications for the shipping and ports sector. Invited dinner speech to the ITMMA Forum, National Bank of Belgium, Antwerp, Belgium, April 22nd. Cullinane, K. P. B. (2003). Port competition in the Yangtze Delta: Plans and prospects for the future. Invited paper presented at the Seatrade London International Maritime Convention, ExCel, London, 16–18 September. Cullinane, K. P. B., Ji, P., & Wang, T. (2002). A multi-objective programming approach to the optimisation of China’s international container transport network. International Journal of Transport Economics, XXIX(2), 181–199. Cullinane, K. P. B., & Khanna, M. (1999). Economies of scale in large container ships. Journal of Transport Economics and Policy, 33(2), 185–208. Cullinane, K. P. B., & Khanna, M. (2000). Economies of scale in large containerships: Optimal size and geographical implications. Journal of Transport Geography, 8, 181–195. Cullinane, K. P. B., & Kwan, R. (2000). Freight mode choice in Guangdong Province. Occasional Paper, Department of Shipping & Transport Logistics, The Hong Kong Polytechnic University. Cullinane, K. P. B., & Song, D.-W. (2002). Port privatisation: Principles and practice. Transport Reviews, 22(1), 55–75. Cullinane, K. P. B., Song, D.-W., Ji, P., & Wang, T. (2004). An application of DEA windows analysis to container port production efficiency. Review of Network Economics, 3(2), 186–208. Cullinane, K. P. B., Wang, T., & Cullinane, S. L. (2004). Container terminal development in mainland China and its impact on the competitiveness of the port of Hong Kong. Transport Reviews, 24(1), 33–56. Davies, J. E. (1989). Impediments to contestability in liner markets. Logistics and Transportation Review, 25(4), 325–342. Deakin, B. M., Seward, T. (1973). Shipping conferences: A study of their origins, development and economic practices. Department of Applied Economics Occasional Paper No. 37, Cambridge University Press, Cambridge. De Souza, G. A., Jr., Beresford, A. K. C., & Pettit, S. J. (2003). Liner shipping companies and terminal operators: Internationalisation or globalisation. Maritime Economics and Logistics, 5(4), 393– 412. Doi, M., Ohta, H., & Itoh, H. (2000). A theoretical analysis of liner shipping conferences and strategic alliances. Review of Urban and Regional Development Studies, 12(3), 228–249. Fossey, J. (1995). Hub of the matter. Containerisation International (September), 81. Fox, N. R. (1994). An oligopoly model of ocean liner shipping. Review of Industrial Organization, 9(3), 343–355.
244
KEVIN CULLINANE
Franck, B., & Bunel, J.-C. (1991). Contestability, competition and regulation: The case of liner shipping. International Journal of Industrial Organization, 9(1), 141–159. Frankel, E. (1998). China’s maritime developments. Maritime Policy and Management, 25(3), 235–249. Gardner, B. (1997). Development in transport policy: EU competition policy and liner shipping conferences. Journal of Transport Economics and Policy, 31(3), 317–324. GHK (2000). Hong Kong port cargo forecasts 2000/01. Hong Kong: The Hong Kong Port and Maritime Board. Gilman, S. (1983). The competitive dynamics of liner shipping. Aldershot: Gower. Gilman, S. (1994). Contestability and public policy in liner and short sea shipping. In: H. J. Molenaar & E. Van de Voorde (Eds), Competition Policy in Liner Shipping (pp. 45–64). Antwerp: University of Antwerp. Haralambides, H. E., Veenstra, A. W., Fusillo, M., Sjostrom, W., & Hautau, U. (2003). Contract of services for the assistance in processing public submissions to be received in response to the consultation paper on the review of council regulation 4056/86. Final Report prepared for the European Commission Competition DG, http://europa.eu.int/comm/competition/antitrust/review/submissions/final report erasmus.pdf. Hayuth, Y. (1988). Rationalization and deconcentration of the U.S. container system. Professional Geographer, 40(3), 279–288. Heaver, T., Meersman, H., Moglia, F., & Van de Voorde, E. (2000). Do mergers and alliances influence European shipping and port competition? Maritime Policy and Management, 27(4), 363–373. Hoffman, J. (2001). Latin American ports: Results and determinants of private sector participation. International Journal of Maritime Economics, 3(2), 221–241. Informa U.K. Ltd (2004). Containerisation international yearbook. Informa U.K. Ltd. Jankowski, W. B. (1989a). The development of liner shipping conferences: A game theoretical explanation. International Journal of Transport Economics, 16(3), 313–328. Jankowski, W. B. (1989b). Competition, contestability, and the liner shipping industry. Journal of Transport Economics and Policy, 23(2), 199–203. Jansson, J. O., & Schneerson, D. (1987). Liner shipping economics. London: Chapman & Hall. Li, K. X., Cullinane, K. P. B., & Cheng, J. (2003). The application of WTO rules in China and the implications for foreign direct investment. Journal of World Investment, 4(2), 343–361. Neel, R. E., & Gooding, C. W. (1997). Service contracts, rate discounting, and the future of shipping conferences. International Advances in Economic Research, 3(4), 388–394. Notteboom, T. E. (2002). Consolidation and contestability in the European container handling industry. Maritime Policy and Management, 29(3), 257–269. Notteboom, T. E., Coeck, C., Verbeke, A., & Winkelmans, W. (1997). Comment: Containerisation and the competitive potential of upstream urban ports in Europe. Maritime Policy & Management, 24(3), 285–289. Officer, L. H. (1982). Purchasing power parity and exchange rates: Theory, evidence and relevance. Greenwich, CT: JAI Press. O’Mahony, H. (2004). Samsung plans to seek orders for 12,000 TEU super boxships. Lloyd’s List (August 19). Panayides, Ph. M. (2003). International market entry strategies in China: Lessons from ocean shipping and logistics multinationals. In: T.-W. Lee, M. Roe, R. Gray & M. Shen (Eds), Shipping in China (pp. 134–166). Aldershot: Ashgate. Peters, H. J. F. (2001). Developments in global seatrade and container shipping markets: Their effects on the port industry and private sector involvement. International Journal of Maritime Economics, 3(1), 3–26.
The Container Shipping Industry
245
Port Development International (1995). Northern exposure, transhipment supplement, July, T8. Porter, J. (2004). European lines offer to end collective pricing company. Lloyd’s List (September 2). Robinson, R. (1998). Asian hub/feeder nets: The dynamics of restructuring. Maritime Policy and Management, 25(1), 21–40. Scherer, F. M. (2000). On maritime shipping conferences. In: F. M. Scherer (Ed.), Competition Policy, Domestic and International: Economists of the Twentieth Century Series (pp. 161–177). Cheltenham: Elgar. Sjostrom, W. (1989a). On the origin of shipping conferences: Excess capacity and the opening of the Suez Canal. International Journal of Transport Economics, 16(3), 329–335. Sjostrom, W. (1989b). Collusion in ocean shipping: A test of monopoly and empty core models. Journal of Political Economy, 97(5), 1160–1179. Sjostrom, W. (1993). Antitrust immunity for shipping conferences: An empty core approach. Antitrust Bulletin, 38(2), 419–423. Slack, B., Comtois, C., & McCalla, R. (2002). Strategic alliances in the container shipping industry: A global perspective. Maritime Policy and Management, 29(1), 65–76. Slack, B., & Starr, J. T. (1994). Guest editorial. Maritime Policy and Management, 21(3), 185. Song, D.-W. (2003). Port co-opetition in concept and practice. Maritime Policy and Management, 30(1), 29–44. Stares, J. (2002). Roads crisis spurs Med shortsea drive. Lloyd’s List (February 13). Tabernacle, J. B. (1995). Changes in performance of quayside container cranes. Maritime Policy and Management, 22(2), 115–124. United Nations Conference on Trade and Development (2003). Review of maritime transport. Geneva: United Nations. Wijnolst, N., Scholtens, M., & Waals, F., 1999. Malacca-Max: The ultimate container carrier. Delft: Delft University Press. Yong, J. S. (1996). Excluding capacity-constrained entrants through exclusive dealing: Theory and an application to ocean shipping. Journal of Industrial Economics, 44(2), 115–129.
8.
LINER SHIPPING STRATEGY, NETWORK STRUCTURING AND COMPETITIVE ADVANTAGE: A CHAIN SYSTEMS PERSPECTIVE
Ross Robinson 1. INTRODUCTION The liner shipping market is an internationalised, service-provision market place; it is also exceptionally unstable. A rapidly globalising marketplace with production locations widely dispersed and footloose has underlined the need for fully connected, highly integrated distribution systems in which time competition has become as critical as price competition, if not more so, in many markets. The liner shipping response has been one of relatively rapid and continuing restructuring; large, powerful firms and extensive market concentration; extensive shipping networks of considerable complexity and specialisation offering increased access to main core markets and to niche markets within them; increased economies of scale and of network density which have spawned larger vessels, higher port productivities and the necessity for tighter information control and e-Business systems integration; and restructured landside logistics and supply chains modifying and altering network shape. But extended networks, greater market access and increased integration have not necessarily created competitive advantage for either very large or for small lines and in 2002, Panayides and Cullinane were moved to formally question what it is, in Shipping Economics Research in Transportation Economics, Volume 12, 247–289 © 2005 Published by Elsevier Ltd. ISSN: 0739-8859/doi:10.1016/S0739-8859(04)12008-8
247
248
ROSS ROBINSON
fact, that determines competitive advantage in liner shipping, what strategies lead to success and whether or not a micro-analytic, firm-specific approach rather than a wider market-based approach may offer greater insights (Panayides & Cullinane, 2002). Indeed, the issue was seen to be so central not only to shipping lines but also to continuing research and better understanding that the authors ranked it as a key contemporary issue in liner shipping economics. This degree of “puzzlement” is well founded and of interest in itself; but it is also seen to be symptomatic of a pressing need, as the authors themselves recognise, for a more adequate and insightful framework for explanation – for a new paradigm or a new “conceptual box,” in Kuhnian terms, that will offer better explanation (Kuhn, 1970). This paper shares this particular point of view. It argues that competitive advantage is a function of strategy; and that although shipping lines operate in networks they are not in fact in the business of network management or of “running ships.” Shipping networks change because shipping lines make decisions about their corporate strategies; for shipping networks are, in effect, artifacts of corporate strategy. Strategies decay; and they decay rapidly if the perceptions and the preferences of the shipper or the customer – “easily bored . . . and ever more promiscuous in their loyalties” (Hamel, 2000) – no longer value the firm’s output; if, in fact, the firm has “. . . fallen wildly out of step with marketplace realities” (Gerstner, 2002). Shipping lines are in the business of delivering value to buyers and sellers – and of capturing value to ensure they remain in business. Could it be that defining strategies on the basis of bigger and better networks as a basis for competitive advantage and for bigger market share might, in fact, be getting it wrong? Might it be that our conceptual focus on markets, on the marketplace, on what makes markets competitive and on how to define competitive advantage within the marketplace framework is no longer an adequate basis for explaining why some firms – and shipping lines – perform rather better than others? This paper explores these issues. Section 2 notes the elements of existing and important explanatory models of competitive advantage; but it also points to an alternative view that offers insights into the business success of third party service providers and particularly in this paper to that of shipping lines. Section 3 offers a framework for strategy definition for lines; and Section 4 details two case studies each of which sheds light on the underlying dynamics of network change at particular points in time as shipping lines and shippers seek to capture and deliver value. The first focuses on the Port Klang/Singapore hub feeder system, the second on the Hong Kong/Pearl River Delta and south China system.
Liner Shipping Strategy, Network Structuring and Competitive Advantage
249
2. DEFINING COMPETITIVE ADVANTAGE FOR SHIPPING LINES: AN ALTERNATE VIEW 2.1. Competitive Advantage and Competition in Markets: A Background Note The notion of competitive advantage, on the ways firms achieve advantage and on the bases of strategy definition for business success have certainly occupied the minds of management theorists and others for some considerable time. In this context we note, albeit briefly, two separate though essentially complementary analytical frameworks – that associated with the work of Michael Porter from 1979 and usually associated with the Harvard school (Porter, 1979, 1980, 1985, 1990, 1996) and that of Edith Penrose in the late 1950s (Penrose, 1959) and numerous others (Barney, 2001; Collis, 1994; Conner, 1991; Hunt, 2000; Mahoney, 1992; Peteraf, 1993; Priem & Butler, 2001; Wernerfelt, 1984, for example) that emerged as a resource-based view of the firm. Porter’s early paper (Porter, 1979) clearly reflected its industrial economics roots and it was an attempt to establish a systematic framework for explaining why some firms in an industry were more profitable than others. The framework, the now classic “five forces framework,” set out “the determinants of long term industry profitability” and how companies might influence them. Almost two decades later in a much less deterministic framework, the superior profitability of a firm was seen to reflect differences in the operational effectiveness of the firm and differences in its strategic positioning (Porter, 1996). Value to customers was delivered through the way the firm went about organising its activities – and, specifically, in activity systems focused on higher order strategic themes. There is no question that this 1996 paper offers a significant conceptualisation of competitive advantage, notwithstanding Cox’s view that it fails the test of predictive theorisation (Cox, 1997, p. 121) and that it “does not recognise the essence of entrepreneurial activities” (Cox, 1997, p. 119) – at least in Cox’s rather narrow view of what is and is not entrepreneurial behaviour. The notion that a company’s competitive or business success is closely related to its ability to exert control over and organise its particular set of resources – broadly defined – so as to achieve unique capabilities underlies the resource-based view of the firm. In 1991 Kathleen Conner noted “A resource-based approach to strategic management focuses on costly-to-copy attributes of the firm as sources of economic rents and, therefore, as the fundamental drivers of performance and competitive advantage . . . (and) . . . a firm’s ability to attain and keep profitable market positions depends on its ability to gain and defend advantageous positions
250
ROSS ROBINSON
in underlying resources important to production and distribution” (Conner, 1991). Somewhat later the view remained much the same and Rouse and Daellenbach noted the resource-based view “highlights how the development of unique and idiosyncratic organisational resources and capabilities can result in sustained superior performance . . . (and) . . . that sustained competitive advantage grows out of those valuable, rent-generating, firm-specific resources and capabilities that cannot easily be imitated or substituted” (Rouse & Daellenbach, 1999). The resource-based view of the firm has spawned a considerable literature and provides useful insights – though we note some discussion that argues the theoretical shortcomings of the approach (Cox, 1997, p. 233; Priem & Butler, 2001). Arguably, however, there is now considerable complementarity between the two schools of thought.
2.2. Competitive Advantage in Chains, Chain Systems and Supply Chains: A Radical Refocusing On what basis do shipping lines create and sustain competitive advantage? The concepts and frameworks developed by Porter and by the resource-based view of the firm shed much light on these issues. But are shipping lines not notably different across a number of dimensions from production firms? Are shipping lines not third party service providers intervening between buyers and sellers; are they not in the business of moving freight – and if there is no sale transaction between buyer and seller there will be no trade, nor will there be any transport service since the service is a derived demand? Further, if the line is to operate profitably and sustainably will it not need to deliver the value required of it by the buyer and/or seller, depending on the terms of trade, as well as delivering value to itself (or capturing value) to ensure its own viability? The fact is that the line is not only operating in a market but as a key player embedded in chains or supply chains – indeed, it carries out its entire day-to-day business operations within chains between seller and buyer and has no other reason for existence! Under these circumstances, how does a line achieve advantage? And superior profitability? And market leadership? Neither the Porter view nor the resourcebased view of the firm takes this chain or supply chain perspective; but Cox’s seminal work (Cox, 1997; Cox et al., 2002) argues that all firms exist not only in markets but also, and essentially, are linked into supply chains. If we are to understand the business success of firms – and of shipping lines as firms – we must “understand the structure of supply chains not the properties of markets” (Cox, 1997, p. 232) – and particularly, of course, how individual firms appropriate and accumulate value from their positioning within the supply chain (Cox, 1997,
Liner Shipping Strategy, Network Structuring and Competitive Advantage
251
p. 207). Again, as Cox notes, “markets record entrepreneurial action, not the basis for entrepreneurial success” (Cox, 1997, p. 232). For Cox, Porter’s focus on “markets rather than on supply chains” may have led us to “under-estimate the full range of contingent circumstances which . . . companies actually face externally” (Cox, 1997, p. 121). Similarly, he argues that the “resource-based school has stopped half-way, because it focuses primarily on the existing internal resources which the organisation possesses, and starts from a competitive market rather than a supply chain focus” (Cox, 1997, p. 233). This paper argues that a chain perspective is appropriate, even mandatory, if we are to understand the bases of competitive advantage for shipping lines and to define adequate and appropriate strategies for their business success. But it also takes the view that an appropriate conceptualisation is not self-evident given the particular functionality of third party service providers nor is it independent of its antecedent traditions of the Porter and resource-based view of competitive advantage. The paper outlines, in the following two sections, the essential elements of an analytical framework. 2.3. Competitive Advantage and Business Success in Liner Shipping: A Chain Systems Framework 2.3.1. The Elements of a Chain Systems Framework The framework discussed in the following paragraphs draws upon but extends the conceptualisation outlined as a “value driven chain systems approach” in two earlier papers which focused on the role and functionality of ports and port authorities though the underlying principles are consistent (Robinson, 2002, 2003). It is built around five key and separate but interdependent concepts and the particular implications that derive from them. They are: that shipping lines are third party service providers; that they are also networked third party service providers; that they compete in chain structures as well as in markets; that chains operate within and are structured by power and dominance relationships; and that business success, and strategies to achieve business success for shipping lines, are defined on the basis of “the acquisition and exploitation of supply chain and market power, and the pursuit of rents” so as to move from a position of weakness to a position of dominance or market leadership. The following sections expand on these points but must be regarded as indicative rather than substantive.
252
ROSS ROBINSON
2.3.2. Shipping Firms as Third Party Service Providers Shipping firms are in the business of moving freight. They contract with the buyer or seller (either directly or indirectly); they intervene, in effect, between the buyer and the seller as a third party and the price of their intervention is met, or redistributed, from the price paid by the customer. Freight moves only because in so doing it offers value and competitive advantage – to the shipper, to the buyer and to the service provider. Figure 1 suggests the key elements. Firm A in Market 1 sells to Firm B in Market 2; and product (freight) moves along logistics pathways, in this case including landside modes and through ports in shipping networks. Shipping lines will only derive competitive advantage by delivering the value that the customer will accept – not by operating on extensive networks, or by operating with larger and faster ships or by operating clever e-Business systems though these may be fundamental to the value proposition offered by the line and accepted by the customer. Competitive advantage and value are key concepts and Porter’s 1990 and 1996 papers (Porter, 1990, 1996), Phillips’ operationalisation of the concepts (Phillips, 1987) and Kaplan and Norton’s later contributions provide exceptional insights (Kaplan & Norton, 1996, 2001). For Porter, as we have noted, value and competitive advantage are delivered through the firm’s value chain (the activities the firm performs that provide customer value in his 1990 paper) or the fit of its “activity system” focused around higher order “values” (in his 1996 paper); in turn, these will be defined in terms of the unique and valuable position it chooses and will reflect its operational effectiveness as well as its strategic positioning. Phillips embeds these concepts within the notion of market-focused or value-delivery firms – firms whose strategic focus is superior value delivery to target customers at a cost allowing acceptable profit. Effectively, the firm will have a precise understanding of the value (its
Fig. 1. Freight Movement in Logistics Pathways.
Liner Shipping Strategy, Network Structuring and Competitive Advantage
253
value proposition) that it will deliver to the market – in fact, to a very precisely defined section of the marketplace, its target customers; how it will deliver that value; and how it will communicate that value to its customers (More simply, Phillips specifies the three steps as “Choosing the Value Proposition, Providing the value; and Communicating the value”). Phillips defines value simply as “benefits less price,” which is a useful operational definition (but which in fact requires careful specification). Kaplan and Norton offer further discussion and a quantifiable approach (Kaplan & Norton, 1996, Chap. 4). In any case, the notion and the explicit specification of value delivered to target customers or market segments is a critical one in the conceptualisation of shipping lines as third party service providers. 2.3.3. Shipping Firms as Networked Third Party Service Providers Shipping firms – like airlines – operate on non-tracked (though navigationally precise) networks. Expansion and contraction of networks under these circumstances ceteris paribus suggest the possibility of more rapid restructuring in both time and space. In effect, individual lines will likely operate on relatively restricted networks; and as joint ventures and/or alliances develop, these networks will become subnets of a larger net. Critically, and in the light of our earlier discussion, an individual line will only capture competitive advantage on the basis of the value it will deliver its customers – not on its level of integration into a network per se or simply because it may have access to a larger customer base. Shipping networks not only expand or contract in size but the development of new networks may reflect some segmentation in the market. In an earlier paper, and in the context of rapidly restructuring hub/feeder networks in southeast and east Asia, it was suggested that rapid trade growth and massive increases in container volumes were spawning new networks that reflected segmentation of the market on a cost, productivity and efficiency basis Conceptually and intuitively, at least, high order hubs were embedded in first-order, highly rationalised and even corridor networks supporting very large, fast and expensive vessels; and second and third and subsequent order nets handled trade volumes through smaller hubs embedded in networks supporting smaller and higher per unit cost vessels (Robinson, 1998). Recent changes suggest continuing complexity of network structure (Yap et al., 2003). In any case, shipping networks structure and restructure for what, prima facie, may appear to be exceptionally complex circumstances. But it is well to remember that shipping networks are artifacts of the corporate strategy of shipping lines; that lines themselves – or at least their management teams – decide, in the context of what value proposition they are offering to what customers at what level of profitability, whether or not the line will operate in any particular network. Lines
254
ROSS ROBINSON
that find themselves in networks in which advantage erodes rather than builds would wisely seek to redefine their strategic positioning. The conceptualisation of the shipping line as a networked third party provider explicitly recognises that although the line must deliver the value demanded by the customer, the network characteristics within which it operates, inter alia, will impose significant constraints on the value that the line is able to capture. There will, in fact, be constant tension between the value that can be delivered and the value that can be captured! 2.3.4. Shipping Lines Compete in Markets and in Chains Shipping lines compete in markets; markets differ in the degree to which they are competitive; and the strategy and strategic options defined by lines will be impacted upon by their relative position in the market and by the degree of competition that exists within it. Porter’s “five competitive forces” explanation remains a useful conceptual framework for assessing these relationships (Porter, 1990); and classical microeconomic theory has long offered powerful insights into monopolistic, perfectly competitive markets and oligopolistic markets, tightly structured or otherwise. Notions of contestability and contestable markets add to the richness of the broader conceptual framework. More recently, the emergence of competition policy as a formally legislated regulatory framework and legal implications for defining what is or is not a competitive marketplace have given rise to the concepts of “workable” and “effective” competition in competition law. In any case, how firms (or shipping lines) compete in marketplaces is well canvassed in Porter’s work, in the general framework of the structure, conduct and performance paradigm and in the resource-based view of the firm. But shipping lines (like other third party service providers) carry out their business not simply within marketplaces but also within corporately structured chains and supply chains reflecting particular logistics functions and spatial pathways. Figure 2 offers some insight into this conceptualisation though its focus is on port-oriented landside chain structures. Shipping lines intervene, as do numerous other third party service providers, between seller and buyer of product. Critically, the line (and the chain in which it is embedded) must deliver to the customer’s requirement; if it does not, and is itself not rewarded for so doing, that particular chain will not exist (Fig. 2 also suggests that the structure of landside chains may undergo significant rationalisation as powerful and/or opportunistic firms extend control up and down the chain). Note that the chain exists only on the basis that it not only delivers value for the customer but it also captures values for the players involved. Figure 3 illustrates the point. The chain delivers value and advantage to both customer and seller. Clearly, the buyer/seller transaction price will be acceptable to both and it will be
Liner Shipping Strategy, Network Structuring and Competitive Advantage
255
Fig. 2. Port-Oriented Value-Driven Chain Systems.
a proxy for “value” – though the customer is likely to identify value not simply in terms of price/cost but also in terms of reliability, convenience or some other value element. “Supply chain value,” in the diagram, is in effect the revenue redistribution from the price paid (including freight rates and other related movement costs) to the individual service providers in the chain.1 The relative share captured by particular service providers is, of course, of particular relevance to them and raises the issue of supply chain power in this conceptual framework.
Fig. 3. Delivering and Capturing Value in Value-Driven Chain Systems.
256
ROSS ROBINSON
The conceptualisation of competition in chains provides important insights and raises many issues – chains may be seen to compete with chains; interfirm arrangements may have significant impacts on chain efficiency; corporate business models may detract from chain efficiency, not enhance it; and chain structures may reveal more clearly the nature of margins and who captures them! 2.3.5. Power in Chains It is intuitively apparent that different firms exert differing levels of power in a supply chain and do so in different chains and at different times. But the work of Cox has provided insights into supply chain power and his thinking forms an important part of this conceptualisation (Cox, 1997; Cox et al., 2002). For Cox et al. firms are Janus-faced – they are both suppliers and buyers or customers and it is in this transactional relationship that they are able to exert varying levels of power. These dyadic relationships are shown in Fig. 3 by the curved arrows. Power is seen to be “. . . the ability of a firm (or an entrepreneur) to own and control critical assets in markets and supply chains that allow it to sustain its ability to appropriate and accumulate value for itself by constantly leveraging its customers, competitors and suppliers” (Cox et al., 2002). Critical assets and power resources are seen, therefore, as the basis for exerting dominance over players in the chain. Levels of dominance will vary and Cox et al. develop an eight-category framework for analysing dyadic relationships. For our purposes, we simply note that power may vary between players from a position of total dominance (dominant) in which the player has a critical asset (one that is scarce and of high utility) to one in which a player has no power (independent)! The more usual condition is one of “interdependence” – in which each partner has something to offer the other. For Cox et al., then, “. . . the ideal position for earning rents – or high levels of profit on a sustainable basis – is fairly simple to understand. When . . . a company is selling to customers the ideal must always be to have monopoly ownership of inimitable supply chain resources that are needed (not merely wanted) and highly valued by everyone. When . . . a company is buying from suppliers, the ideal must always be to be a monopsonist, who is able to source from suppliers located in highly contested markets in which there are low switching costs and low barriers to market entry” (Cox et al., 2002, p. 7). 2.3.6. Business Success and Strategy Definition In the Cox et al. conceptualisation, business success is concerned with the pursuit of rent – “earnings in excess of the firm’s costs of production that are not eroded in the long run by new market entrants . . . rents persist in long-run equilibrium while
Liner Shipping Strategy, Network Structuring and Competitive Advantage
257
profits tend towards zero.” The firm’s strategy, therefore, “. . . should focus on the acquisition and exploitation of supply chain and market power, and the pursuit of rents” (Cox et al., 2002, p. 6). This is a significant departure from a usual market power explanation for, as Cox notes, perfectly competitive markets degrade to zero profits given the impacts on competition from new entrants. Hamel makes a similar point, though in the context of strategy decay – “Do you remember Economics 101 and the idea of ‘perfect competition’ – when everyone in an industry followed an identical strategy and had similar resources? You probably also remember the textbook result: every company made just enough profit to survive and no more. It’s the business equivalent of a subsistence economy. That’s the inevitable result of convergent strategies” (Hamel, 2000). Rent seeking behaviour for shipping firms will focus on both market power and supply chain power – “The firm will use its market power over weaker and less effective competitors by closing the market to them. It will also use its supply chain power over dependent suppliers to extract cost and quality improvements, while using its power over dependent customers to increase, or at least maintain, its share of the total revenues earned in its market over the business cycle” (Cox et al., 2002, p. 6). We add, however, a caveat for it is all too easy to be seduced by the glib assertion that strategy is about power and the pursuit of rents if we, in so doing, fail to recognise that unless the chain is delivering the value which the customer demands there will be no chain and no rents! Chains exist to deliver value to the customer and capture value to ensure sustainability. Suppliers and buyers “close the market” when the value offered is the value accepted. In this conceptualisation, therefore, we recognise the explicit relationship between value delivery and value capture and its critical importance in defining strategies for business success.
3. STRATEGY DEFINITION FOR LINER SHIPPING FIRMS: A BUSINESS MODEL 3.1. The Model Elements: Overview How might these conceptual underpinnings be operationalised into a business model that provides guidance for strategy definition? Hamel’s recent work on strategy definition (and redefinition) – despite its sometimes unnerving focus on “radical innovation” and “revolution” – is especially useful and the following framework is loosely based on his thinking (Hamel, 2000).
258
ROSS ROBINSON
Fig. 4. The Elements of a Strategy Framework for Shipping Lines.
Figure 4 shows, in simple outline, the basic building blocks of the business model and the relationships between them. In effect, there are three orienting themes – the core strategy, the translation of the core strategy into the activities and organisation of the firm and the centrality of the value delivered to the customer. The firm’s charter and legal framework will give the fundamental strategy direction; and the level of business success is seen to be the outworking of the model. The core strategy articulated in the model is informed by two sets of crucial inputs – on the one hand, it must understand and interpret the firm’s charter; and on the other it will be based on a deep understanding of the freight systems and chains in which it is involved and on the specific value to be delivered and captured. Strategy definition will always be related to resource availability; but the core strategy will impact the resources required – as the core competencies, the core processes and the strategic assets. A precise understanding of strategy will underpin the line’s activity systems; and will be reflected in the organisational structure of the body; and it will be expressed in clear value propositions that will be the basis of market segmentation. Business success will reflect the capture and accumulation of value. We note, briefly, aspects of the individual elements of the model.
Liner Shipping Strategy, Network Structuring and Competitive Advantage
259
3.1.1. The Movement of Freight It is well to remember that it is the movement of freight that provides the essential rationale for the existence of any third party service provider given that the demand for transport and transport services is a derived demand. Clearly, it is a matter of importance that the types, volumes and patterns of freight movement are known with as much detail and accuracy as is appropriate; but it is of at least equal, if not greater, importance to understand the corporate structuring of the chain – the dynamics, the power relationships and the business and financial models of firms embedded in the movement patterns. 3.1.2. The Charter and Legal Framework of the Shipping Line Lines will differ in terms of the charter under which they operate, the legal requirements under flag of registry, ownership and corporate structures, for example, that will set general limits to how the firm operates and the expectations placed upon the management team. 3.1.3. Core Strategy The line’s core strategy is clearly related to its charter; but it will also be conditioned by the resources the line chooses to put in place, by the competitive frameworks of the freight industry – or more generally in Hamel’s terms, the domain or competitive domain – and by the line’s view of the value to be delivered and captured. Hamel’s discussion of core strategy in his “unpacked” business model is useful and applicable. In effect, core strategy is essentially about how the firm chooses to compete and attention must be paid to the overall objectives of strategy, a clear understanding of the business that the shipping line is in, and how it chooses to compete and differentiate itself from others. 3.1.4. Strategic Resources Every competitive advantage rests on unique, firm-specific resources that may be seen to include core competencies, strategic assets (and particularly “critical assets”) and core processes – the activities used in translating competencies, assets and other inputs into value for customers. 3.1.5. Freight System Structure The value driven chain paradigm conceptualises third party service providers as elements in chains – and import and export chains move cargo through particular logistics and corporate pathways from seller or shipper to buyer and end customer. Third party service providers exhibit market power, in the classical economic sense; but they are also endowed, in any particular chain, with some level of supply chain power that determines the firm’s ability to both deliver and capture value. Cox’s
260
ROSS ROBINSON
work on power in supply chains, on the criticality of assets and on the ability of firms to close the market is, in this context, of exceptional importance as we have noted. Note that Hamel’s inclusion in his “unpacked business model” of the notion of the “Value Network” (the network of firms that is linked to any firm, and which complement and amplify the firm’s own resources – and therefore allow it to capture added value) is not unrelated to the notions of chains in freight system structures; but it is different from it and should be seen as a point of departure in the two models. 3.1.6. The Line’s Value Framework At the core of the line’s strategy will be a clear understanding of the value that it will deliver to its target customers and the value it will capture in so doing. Phillips sees the choice of value proposition as a fundamental first step in framing strategy in market-focused firms; providing the value and communicating it are the two subsequent steps. Market segmentation will follow on the basis of the ability of the customer to share the line’s value proposition – not on location, or type or size of industry, or anything else; and value is interpreted as a benefit perceived by the customer less the price that has been paid. Kaplan and Norton’s generic model of the customer value proposition suggests that value, for the customer, relates to three components – one reflecting the attributes or characteristics of the service offered (or functionality), quality, price and time; the second relates to the image that the customer has of the firm and the value it is offering; and the third sees the importance of the customer’s relationship with the firm (Kaplan & Norton, 1996). For Cox, as discussed above, supply chain value is the value (revenue) captured by the firm from the role it plays in the supply chain. The revenue is redistributed from the price paid in the transaction between buyer and seller. The magnitude of value capture is a function of the power exerted by the firm in the chain; firms with scarce and valuable critical assets or firms which can close the market in one way or another exert considerable power, those without critical assets or power resources exert least. Clearly, and particularly for asset-based service providers where network efficiency is expensive and critical, strategies must find some balance between value delivered and value captured – and a balance between the benefits delivered to customers and cost to the provider reflecting varying levels of technical efficiency of the system and the benefits related to those efficiency levels. 3.1.7. Organisational Structure This element and the following one relate to the structure and organisation of the firm. Formal organisational structures differ; but too often historical
Liner Shipping Strategy, Network Structuring and Competitive Advantage
261
practice and precedent impose unclear reporting lines, duplication of activity and other efficiency-limiting conditions. There is a critical requirement that the organisational structure reflects clearly the value that the line is delivering and the value-capture mechanisms it embraces, the strategic intent of the line and the “activity systems,” in Porter’s terms, that are the basis of competitive differentiation of the line (Porter, 1996). 3.1.8. Activity Systems and Functionality We share Porter’s view that “the essence of strategy is in the activities” (with the proviso, of course, that for third party service providers including shipping lines these are critically defined within the framework of corporate structured logistics chains); that “competitive strategy is about being different.” It means deliberately choosing a different set of activities to deliver “a unique mix of value.” Porter’s reworking of his earlier view that firms deliver value through a “value chain” – a set of primary and secondary activities that mesh to provide value to a customer – to the delivery of value through “activity systems” provides an especially powerful conceptual and analytical tool. Such systems provide the basis for corporate positioning and differentiation and comprise sets of functions tightly integrated around “higher order strategic themes.” It is an analytical framework that can be operationalised to ensure that, in Hamel and Prahalad terms, the firm’s strategic intent is embedded into the functions and activities of the firm (Hamel & Prahalad, 1994) – or, in this case, the shipping line. Poor fit between strategy and activity systems delivers outcomes, in competitive markets, inferior to those delivered by market-focused firms with well-defined positioning strategies and activity systems that will deliver value associated with those strategies. 3.1.9. Customer and Market Segment Value Firms do business with those customers and those market segments that share their value proposition; but unless firms define their value proposition in such a way that it coincides with the customer’s value proposition there will, in fact, be no business done – at least in competitive markets! To some extent this element in Fig. 4 overlaps with the element “The line’s value framework”; but the focus is clearly on the nature of the value and the benefits that the customer wants from the line. In Hamel’s “unpacked business model” the notion of “customer interface” is not irrelevant and raises some issues about the way the firm reaches its customers, information about the customer, the relationship with the customer and the value/price alignment sought by the customer. In fact, of course, the coincidence of the firm’s value proposition with that of the customer must be a key element in business success.
262
ROSS ROBINSON
3.1.10. Business Success For Porter and Phillips business success is a product of the outworking of competitive strategy, of firms creating and sustaining competitive advantage. Hamel and Prahalad see the need to position with respect to the future and underline the need to recreate industries and to compete for opportunities. More recently Hamel has argued the case for radical innovation to overcome the problem of strategy decay; and Cox sees business success as the transition of the firm from weak power positions to positions of power and dominance in the supply chains in which they are positioned. More recently Gerstner, in describing the evolution of IBM from an internally-driven organisation to a market-focused, customer serviceoriented firm rightly underscores the critical importance of cultural change if firms are to successfully implement strategy.
3.2. The Point of Departure It is well to remember that “production” firms of one type or another – manufacturing firms, mining and agricultural firms for example – are fundamentally differentiated from service firms and, in this context, from third party service providers; and such differentiation provides a key to their success. In Gerstner’s terms “. . . in services you don’t make a product then sell it. You sell a capability. You sell knowledge. You create it at the same time as you deliver it” (Gerstner, 2002). This paper recognises that shipping lines operate in markets as traditionally understood; and key concepts from management and economic theory – competitive advantage, value, value propositions and value-delivery or market focused firms, strategy decay for example – are as relevant to shipping lines as they are to any other firm. Importantly, and the critical point of departure in this paper, is the recognition that, for third party service providers, day-to-day business is “fought out” in corporate structured chains in logistics pathways. The implication of this is that the notion of power and the work of Cox are of exceptional importance; and that the view of strategy as acquiring and exploiting chain and market power to achieve rents is an especially appropriate one.
4. BALANCING VALUE DELIVERED AND VALUE CAPTURED: THE COMPLEXITY OF CHOICE It is not unlikely that some shipping lines are in alliances and associated networks that lock them into strategic positions and chain systems that deliver them
Liner Shipping Strategy, Network Structuring and Competitive Advantage
263
break-even profits rather than monopoly-like rents; they will be price takers not price setters. Given the Cox et al. view that strategy is about “acquisition and exploitation of supply chain and market power, and the pursuit of rents” such a “lock-in” environment will be inherently unstable and the line might be expected to constantly review its strategic options. There is a great deal of complexity in the way in which shipping lines choose network ports and how shippers choose to use one line or another and one port or another to deliver and capture value and to achieve competitive advantage. A great deal of empirical research and research testing needs to be done to fully demonstrate the conceptualisation outlined in this paper; and in the two case studies which follow we do not pretend to do more than demonstrate the insights into value delivery, value capture and the bases of competitive advantage that emerge from a micro-analytical perspective on shipping network structuring in two particular, though rather different, shipping environments – one focusing on the SingaporeMalaysian port system in the mid-1990s, the other on the Hong Kong-south China system in the early 2000s.
4.1. Shipping Networks and Shipper Value: Container Operations in the Singapore-Malaysian Network2 In 1993 the Malaysian Government, anxious to minimise the loss of foreign exchange on freight services and at the same time to maximise the foreign exchange earnings of ports from shipping, set in place a policy that Port Klang be developed as the hub port for Malaysia – in effect, a policy to focus national cargoes through Port Klang rather than through the Port of Singapore. Clearly, such a policy would have implications for how shipping lines might route vessels and for how shippers might choose shipping services and the routeing of their cargo. For both sets of players, the move underlined issues of competitive advantage and of the likely impact on value capture and value delivery. This case study focuses on these issues at a particular point in time – the mid-1990s – and though the situation is somewhat different in the early 2000s it well-exemplifies aspects of shipper competitive advantage and its relationship to network structuring, particularly in relation to the movement of export containers. 4.1.1. Malaysian Container Cargoes through the Port of Singapore: The “Leakage” Problem In the mid-1990s Port Klang (210 nautical miles north of the Port of Singapore and about 26 hours steaming time at 15 knots) handled 1.4 million TEUs compared with Singapore’s 12 million TEUs. Of concern to the Malaysian Government was
264
ROSS ROBINSON
the leakage, in 1995, of about 300,000 export TEUs or about 56% of the premier port’s total export boxes to Singapore. This concern was further compounded by the fact that two other container handling ports in the west coast of peninsula Malaysia, Penang and Johor or Pasir Gudang, were also transshipping relatively significant volumes of containers through Singapore – about 136,000 TEUs and 75,000 TEUs respectively in 1995. In effect, of an estimated 909,000 export TEUs shipped from Malaysia’s west coast ports in 1995, 524,000 TEUs or 58% were transshipped through the Port of Singapore – an average monthly movement of about 44,000 TEUs (8000 TEUs from Johor, 24,000 TEUs from Port Klang and 12,000 TEUs from Penang). The Malaysian Government was also concerned about leakage of cargoes from East Malaysian ports (Kuching, Kota Kinabalu, Sandakan and Tauwau) amounting to 15,000 TEUs in 1995 or 28% of containers exported from the ports. The numbers were small, though of relative importance. The research revealed, at the outset, some considerable complexity in the feeder networks interlinking the ports, raising the issue of system-wide causality. Figure 5 suggests that in fact five “container movement corridors” operated with varying levels of independence – a cross-causeway link, the Port Klang/Singapore corridor,
Fig. 5. Container Movement Corridors.
Liner Shipping Strategy, Network Structuring and Competitive Advantage
265
the Penang-based flows, the East Malaysian corridors and the Johor/Singapore movement of containers. Regrettably, intractable data problems for the crosscauseway and for the East Malaysian movements focused attention on the movements of containers along the west coast of the peninsula, rather than on the whole system, to (and from) Singapore. 4.1.2. The Market Orientation of Export Containers Given the varying intensity of flows from the west coast locations the further issue that suggested itself was whether or not, again, there were systemic patterns in the export destinations of the container flows. Table 1 lists estimates of the proportion of export TEUs transshipped via Singapore in 1995 and though these should be regarded as indicative rather than definitive they reveal important clues to market orientation.
Table 1. Proportion of Export TEUs Transshipped via the Port of Singapore from Port Klang, Penang and Johor, 1995.a % Export TEUs via Singapore Port Thailand Indonesia Philippines Vietnam Hong Kong China Japan Korea Taiwan USA West coast East coast Gulf Canada U.K./Europe Middle East Australia/NZ Other a The
Klangb
Penangb
Johorc
98 32 100 100 31 – 25 20 17
47 19 42 – 32 – 39 – –
100 52 100 81 100 100 51 – –
99 100 100 100 93 66 22 69
100 100 – 100 100 100 47 97
100 100 100 – 95 100 100 56
figures have been calculated from a study for the Ministry of Transport by Coopers and Lybrand (1996) Study on Export Containers and Feeder Services – see Section 4. b For 10 months January–October 1995. c For the period July–October 1995.
266
ROSS ROBINSON
Note particularly that in all three ports Export containers to the U.S. and Canada and to U.K./Europe were transshipped via Singapore; but that In the case of Port Klang, for exports to East Asia (Hong Kong, Japan, Korea and Taiwan) direct services rather than transhipment services captured a significant proportion (about 70–80%) of the trade; For Penang direct services captured about 60–70% of movement to Japan and to Hong Kong respectively; and For Johor, direct services to Japan captured about half of the export movements; For all the ports, capture of ASEAN traffic was variable but Indonesia was served on direct services for about 50% of the traffic from Johor, for about 70% of the traffic from Port Klang and for about 80% of the traffic from Penang. Clearly some markets were well served by direct services from the three ports; and some markets were not served at all. This would not be surprising if volumes are low or availability is infrequent or the value of the cargo is low; but what if volumes for particular markets are relatively high, are of relatively high value and are non-seasonal – as is the case for the North American and U.K./Europe markets, particularly in Port Klang and to a lesser extent in Penang? These differences in the level of market capture by direct services rather than transhipment services prompt further inquiry and two key questions suggest themselves: Why is it that for some markets being served by Malaysian ports there is partial capture by direct services but not total market capture? And Why is it that some markets in Malaysian ports – and particularly though not only the U.K./Europe and the U.S./North American markets – are not being served effectively or at all by direct services but rely almost exclusively on transhipment services through the Port of Singapore? The following sections examine these issues in turn. 4.1.3. The Problem of Partial Capture of Markets on Direct Shipping Services Why did some Malaysian ports capture some but not all containers destined for specific markets on direct services? There is no simple answer, though prima facie there were two important factors: That there was insufficient shipping capacity and/or there was inadequacy of frequency of shipping services to the specific market in question; and/or That, for whatever reason, the transhipment port was seen to be more attractive by the shipper than the originating port.
Liner Shipping Strategy, Network Structuring and Competitive Advantage
267
In effect, there were both push factors (making the originating port less attractive) and pull factors (making the transhipment port more attractive) at work. 4.1.3.1. Limited shipping capacity and frequency. Shipping lines seek to make available just sufficient shipping capacity on any route or at any port to maximise profits or to satisfy corporate goals. Low absolute cargo volumes and/or seasonality or irregularity of cargo supply are likely to be sufficient reason for lines to maintain existing service levels and hold off an additional vessel call. In fact, of course, lines will add extra vessels or calls if revenues generated exceed costs incurred – and if an additional call is consistent with the line’s overall strategy – though in the short run lines may incur losses in order to establish market share. In highly concentrated shipping markets characterised by significant monopoly power it is arguable that the key player could manipulate the market in such a way as to maximise profits rather than provide reasonable shipping services. The contestability of the market will be an important factor. Did the volumes of containers leaking from Malaysian ports through the Port of Singapore as transhipment cargo represent excess demand or “overflow” cargo? Did cargo go to Singapore because there was insufficient space available on existing services? There was no evidence of exceptional pressure on existing shipping space. Certainly, for major destinations weekly services were available. Only in feeder services from Kuching to Port Klang did there appear to be pressure on capacity in 1996; and this arose late in 1996 when MISC diverted one of its two feeder vessels operating on the route in order to service new calls at Belawan. Rather, from Port Klang, Penang and Johor it is arguable that the supply of shipping space to those markets already being served by direct services (including but not only Japan, Korea, Taiwan and Hong Kong for example) was more-or-less equal to the existing cargo demand for that space. Why, then, if cargo was not overflow cargo (or cargo that was in excess of available shipping space) did it go to the Port of Singapore? 4.1.3.2. The port of Singapore as a more attractive option. A small proportion of cargo which was moved from major Malaysian ports by transhipment rather than on direct services was because of some special arrangements, usually though not necessarily contractual in nature – manufacturing firms with special arrangements with shipping lines operating out of Singapore; or manufacturing firms integrated into multinationals in which shipping decisions for whatever reason specified Singapore transhipment. Further, a proportion of container volumes for which higher frequency of service was, indeed, quite critical moved to Singapore. Justin-time cargo and high value electronics and other high value cargos must meet
268
ROSS ROBINSON
tight delivery or receival deadlines and/or must minimise high inventory costs. For these cargoes, high frequency connections out of Singapore may have met appropriate requirements. For some cargo, too, a wider spread of destination ports would have been a factor in service choice. There was no accurate measure of these volumes – no Port Authority or no shipping or shipper body had investigated this phenomenon and available data were seriously inadequate; but industry estimates suggested that no more than 20 or 30% of the “leaked” container volumes fell into these special arrangements and just-in-time/high value categories. By far the largest volume of “leaked” containers being transshipped via the Port of Singapore was not especially time-sensitive cargo; almost certainly, most containers were likely to contain cost-sensitive rather than time-sensitive cargo. Why, then, would shippers sustain the higher costs of transhipment via Singapore rather than use the direct services? 4.1.3.3. Transhipment: Who pays?. For the shipper we would expect that: If shipper inventory costs (or the costs incurred by the shipper in holding cargo in Penang, for example) exceeded the costs of feedering plus the costs of transhipment via the Port of Singapore then the shipper would use the feeder and transhipment service; If, however, feeder/transhipment costs exceeded inventory costs the shipper would hold the cargo and connect with a direct service from the port. Where significant penalties may accrue to the shipper if contractual arrangements are not met then the shipper may opt to pay the feeder/transhipment premium. In fact, however, for the Malaysian shipper whether shipping through Port Klang, Penang, Johor or East Malaysian ports: The feeder/transhipment cost was zero; and the cost of a direct service was under most circumstances the same as the cost of a transhipment service. The shipping line, not the shipper, met the costs of feeder and transhipment services. The line did not differentiate to the shipper the cost of shipping direct from Port Klang for example or via transhipment from the Port of Singapore. In economic terms the result was, of course, that the shipper was indifferent to whether containers were consigned on direct services from a Malaysian port or on transhipment services via the Port of Singapore; for the shipper the costs were the same. In effect, therefore, the shipper had a cost-neutral choice of shipping through the appropriate Malaysian port or shipping through the Port of Singapore.
Liner Shipping Strategy, Network Structuring and Competitive Advantage
269
4.1.3.4. The early payment option. Moreover, not only was there no additional cost to the shipper to use transhipment services there was in fact a considerable incentive for the Malaysian shipper to use the feeder/transhipment option if he could not benefit from an immediate connection with a direct service and would incur inventory costs awaiting the next scheduled departure of the direct service. This incentive arose from the fact that under normal commercial and operating conditions the shipper would initiate procedures for payment for cargo on presentation of the cargo Bill of Lading to the appropriate bank; and the Bill of Lading would be made available by the shipping line on confirmation of departure of the vessel. Given that feeder vessels departed with considerable frequency from Penang, Port Klang and Johor and with reasonable frequency from East Malaysian ports the shipper was able to obtain a Bill of Lading and initiate payment procedures with minimal delays. It is hardly surprising, under these conditions, that Malaysian shippers found significant competitive advantage in using the feeder/transhipment option via the Port of Singapore for not only was there no disincentive to transship cargo but there was, in effect, a significant revenue incentive to do so! The Malaysian shipper was very well served; his choice of ports was cost-neutral and was likely based entirely on connecting with an appropriate ship rather than on the costs of his actions. 4.1.4. The Problem of Capturing U.K./Europe and North American Markets on Direct Services The reason why Port Klang had not, in the mid- and late 1990s, captured containers to Europe and North American markets was directly attributable to the inherent and underlying pressures which were driving the globalisation of liner shipping and particularly, of course, international container shipping. It was not a simple function of the adequacy or efficiency of Port Klang per se – though this was not unimportant. Note two important points about international container shipping that at the time underlined Port Klang’s exclusion from shipping services to and from U.K. and Europe and to and from North America. First, the corporate policies of a number of the world’s major shipping lines had effectively locked these global container routes (Far East/Europe, TransPacific operations and Round-the-World service) into fleets of large vessels, including 5,000 and 6,000 TEU vessels. These vessels operated on tight, highly rationalised schedules that hubbed on ports that could concentrate very large volumes of cargo and that could handle container exchanges with high levels of efficiency and acceptable costs. The Port of Singapore, with extensive feeder shipping links concentrating large container volumes, was a high priority port of call and sustained high levels of mainline vessel calls. Port Klang, of the major east-west/west-east
270
ROSS ROBINSON
and Round-the-World services and with significantly lower volumes of containers was – under these circumstances – a lower priority call. In June 1996, for example, when the Grand Alliance implemented its new Europe/Far East schedule comprising four service strings, the Port of Singapore was included as a port of call in all eastbound and all westbound service strings. Port Klang, on the other hand, was included in one string (Loop A) as an eastbound call and in another string (Loop B) as a westbound call. Second, not only were these routes served by large vessels in highly rationalised services but they were also in the hands of a relatively limited number of very large operators the majority of whom were organised within new patterns of alliances. A small number of operators – such as Evergreen, with its Round-theWorld services, and Hyundai with transPacific services – existed outside these groupings but were very large and powerful operators. Clearly, high levels of market power and concentration characterised these routes; and arguably this factor had introduced some conformity of service patterns. This further underlined the priority role accorded to Singapore as a hub port. Under these circumstances, whether or not Port Klang could or would be included as a major port of call on these routes would depend on the corporate strategies of the major shipping companies involved on these routes; and that decision would be based, at least in part, on the relative levels of costs and revenues of diverting vessels from existing schedules and loops compared with the costs of feedering cargo to and from the Port of Singapore as a hub port. Clearly, shipping lines will have differing levels and structures of costs and different corporate philosophies that will determine their decision. 4.1.4.1. Diversion costs and feedering costs: The principles of trade-offs. Figures 6 and 7 indicate the elements of the shipping linkage patterns under diversion and non-diversion strategies on the Far East/Europe and Far East/North America routes respectively in the mid-1990s. The diagrams suggest that on the Far East/Europe route the container volumes from Penang, Port Klang and Johor were fed into the Port of Singapore to connect with eastbound and westbound operations (Fig. 6A). With diversion of mainline vessels to serve Port Klang (Fig. 6B) diversion costs would be incurred by diverting from a direct shipping route Singapore/Colombo or Singapore/Middle East. Feeder cost savings would be made by the elimination of Port Klang/Singapore traffic and a portion of the Penang/Singapore traffic. On the Far East/North America route the added diversion cost would be the cost associated with a two-leg extension to and from Singapore/Port Klang. Again, this cost would be offset against reductions in feedering costs Port Klang/Singapore and on a portion of the Penang/Singapore traffic (Fig. 7).
Liner Shipping Strategy, Network Structuring and Competitive Advantage
271
Fig. 6. Diversion Scenarios, Far East/Europe Services (6A – Top; 6B – Lower).
Note, however, that for each individual shipping line there would be an added cost or benefit associated with the dislocation of, and/or the adjustment to, the current shipping schedule. A final diversion decision would also take into account how well a diversion strategy would fit into the line’s overall corporate development strategies. Three sets of variables are critical in determining the costs of diversion and, though well recognised, bear repetition in this context. Port-related costs: Diversion costs are closely related to the costs incurred at Port Klang and, indeed, the relative levels of costs compared with those at the Port of Singapore. There are three variables or sets of variables: The volumes of containers available, and the value to the shipping line, of those containers. Lines are more likely to divert for high value rather than low value cargo though the overall vessel mix of high value to low value cargo will be important for the line; The operational efficiency of the port; and
272
ROSS ROBINSON
Fig. 7. Diversion Scenarios, Far East/North America Services (7A – Top; 7B – Lower).
The costs, both direct (tariffs) and indirect (delays, time) incurred at the port. These two variables (efficiency and costs) are obviously closely related. High cost container vessels sustain high daily operating costs and port inefficiency will add directly to these costs through delays or slow working and through disruption of schedules. Direct port charges usually represent a very small proportion of overall transport cost but inflexibility in pricing policy and poorly constructed tariff schedules may also be deterrents to lines. Ship-related costs: Vessel size and vessel characteristics will also be important elements in diversion costs. Very large, high cost vessels handling relatively low container exchanges will yield high per unit costs and find diversion unattractive; but smaller, lower cost vessels handling the same volumes of containers may be able to deliver an acceptable service to shippers. This is an important point in the question of diversion to Port Klang. Optimal vessel size for routes handling
Liner Shipping Strategy, Network Structuring and Competitive Advantage
273
lower container exchanges out of Port Klang is likely to be lower than the optimal vessel size on high volume routes focused on the Port of Singapore. Ideally, alliances and/or other lines would define service strings that would be based on smaller (3,000–4,000 TEUs) rather than large and very large vessels (5,000–6,000 TEUs). As the market develops and volumes increase vessel size would also increase. Costs related to the level of competition in the shipping market: Shipping lines or consortia, in tightly controlled markets with high barriers to entry and significant concentration of market power, are likely to be able to manipulate prices and the levels and structure of shipping services available. Whether or not the Port Klang/Singapore shipping range represented a contestable market is a question of interest. Certainly the alliances represented significant market power but the degree to which their presence represented a high barrier to entry and the degree to which they were able to hold vessels off a direct Port Klang call are open questions. The diversion issue was not a simple one for shipping lines operating on these trades; and although in a research framework it is tempting to want to define, precisely and for each line, it is neither possible nor, arguably, necessary. For Malaysian shippers, did it matter? 4.1.4.2. Options for Malaysian shippers: An example. There was in fact a range of options for Malaysian shippers to both Europe and North American destinations that may not have seriously disadvantaged the shipper even if containers were feedered from Port Klang, for example, or shipped on direct services from the port. Consider, by way of example, the following options for a Port Klang shipper exporting to a buyer in the U.K. (Southampton) choosing only Grand Alliance services (Table 2). The example was for a shipper having a dispatch date of around 23 September in 1995 and using vessels in the four Loop services of the Grand Alliance.3 Varying transit times resulted from using different combinations of direct and feeder services within the Loop structures; and varying extra costs resulted from combinations of costs for feedering to Singapore, transhipment at Singapore or transhipment at the Europe-end of the Loop services. Recall, however, that these extra costs would nave been borne not by the shipper but by the shipping line. The shipper need not consider only Grand Alliance services; and shipping lines may or may not accept high cost options under some circumstances. Nonetheless, the example underlines the availability of numerous options – at no cost to the Malaysian shipper!
274
ROSS ROBINSON
Table 2. Hypothetical Time and Extra Cost Options for Malaysian Cargo Shipped from Port Klang to Europe on Grand Alliance Services, 1995. Alliance Service
Transit Time (Days)
Extra Cost (US$)
Loop A 1 2
36 22
Loop B 3
21
Loop C 4 5
27 31
100
150 150
Loop D 6 7
22 22
100
150 150
Feeder Singapore
Transship Singapore
100
150
Transship Europe
Total extra cost
0 250 280
280
280
430 250 150 250
4.1.5. A Final Note This case study represents a static view of the dynamics of competitive advantage for shippers and of the capture and delivery of value by shipping lines in the mid1990s in the Malaysian peninsular ports and the Port of Singapore. It suggests that Malaysian shippers captured competitive advantage by shipping on directlink shipping networks in some markets and by transhipment, hub/feedering patterns for other markets. Clearly, shipping lines were able to deliver value to shippers and capture value internally by shipping in different network formats. On transhipment services the absorption of transhipment costs by the lines made the shipper indifferent to the network port of shipment. Indeed, there was some incentive for the shipper to use feeder services given that presentation of the Bill of Lading on consignment of containers to the feeder vessel initiated payment procedures! Malaysian shippers enjoyed competitive advantage under arrangements existing in the mid-1990s; whether or not they actually maximised competitive advantage is a somewhat different issue. Arguably, there may have been greater benefits to shippers if they could have used direct services from an adjacent Malaysian port and by so doing reduced the inventory costs that were associated with the time involved in using the transhipment option. Further, whether or not the market power of alliances maintained rates at levels higher than they might otherwise have been are moot points. Certainly the market
Liner Shipping Strategy, Network Structuring and Competitive Advantage
275
signals of rapid growth and expansion suggested an appropriately capacitated marketplace. In this context we have not focused on the port policy issue of “Malaysian cargo through Malaysian ports”; but as a postscript this case study suggests that Governments must take great care in offering policy prescriptions that could seriously destabilise the effective capture of competitive advantage by shippers and shipping lines.4
4.2. Shipper Options, Shipper Costs and Hub/Feeder Networks: Hong Kong and the Emerging Ports of South China and the Pearl River Delta5 Ports and port networks change; but some change more quickly than others and show remarkable ability to adapt, to innovate and to transform themselves. For more than three decades the port of Hong Kong has ranked as either the largest or second largest (with Singapore) container handling port in the world; but even more remarkably it has been unique in its adaptations to containerisation and in the development of coping mechanisms to handle explosive growth.6 In the late 1990s and in 2000 the port exemplified a port in transition and a port embedded in networks undergoing rapid transition. The intensification of new industrial development in southern China and in the Pearl River Delta adjacent to Hong Kong, the significant restructuring of the liner shipping industry into new alliances, joint ventures and partnerships and more recently still the rapid restructuring of logistics and supply chain systems to eliminate costs and capture control and efficiency have been conjoint and contemporaneous processes in port network restructuring. This case study notes, in its first section, the emergence of the new ports of Shenzhen (Chiwan, Shekou and Yantian) that along with the port of Hong Kong are restructuring container shipping networks in the region. It introduces global cost data to suggest the general importance of cost as a network-structuring variable. The second part of the case study focuses on these ports in their microeconomic setting in the Pearl River Delta and within the context of the “less formal” sector of the Hong Kong container market. It focuses on the complexity of chain patterns and the factors which impact shipper choice and which structure the movements of containers to and from Hong Kong locations and the Pearl River Delta. 4.2.1. Hong Kong and the Ports of Shenzhen in 2000: Networks in Transition In 1995 Yantian, Shekou and Chiwan (the largest ports in Shenzhen) handled less than 300,000 TEUs in international trade; by 1999 the ports were handling almost
276
ROSS ROBINSON
3 million TEUs.7 The port of Yantian had reached 1.6 million TEUs, Shekou 850,000 TEUs and the port of Chiwan handled about half a million containers. Much of the traffic (about 30%) was feedered to the port of Hong Kong but 1.85 million TEUs or 16% of the traffic contributed directly to “ocean traffic.” Importantly, also, significant imbalance existed between export/import boxes with more than three times as many loaded export boxes as import boxes; and 40% of the total traffic were empties (compared with 16% handled at Kwai Chung terminals). In 2000, the estimated throughput capacity of the three ports was 3.2 million TEUs; and projected capacities indicate throughputs of 6 million containers by 2005 and 8 million by 2010. Yantian, one of China’s nineteen “designated” international transhipment ports, will almost certainly emerge as a major network port, with an estimated 3.5 million TEUs capacity in 2005 and almost five times that capacity seen to be feasible in the full development of the port. Shekou and Chiwan have draft and landside limitations and are expected to operate as feeder ports to major hubs – including Hong Kong, Yantian and Kaohsiung for example. In 1999 the port of Hong Kong handled over 16 million TEUs – of which almost 10 million were directly related to export/import flows from Hong Kong and south China, 2.5 million TEUs regarded as transhipment traffic and 3.86 million TEUs generated as river traffic. It is this last component of traffic which is of particular interest in this context – not only because of its exceptional rate of growth (21% over the 1996–1999 period compared with a growth rate for direct traffic of 4.6% and for the whole port of 6.4% over the same period) but also because it has given rise to a somewhat unique “informal” sector not usually associated with the sophisticated terminals sector. In fact, in 1999 only about 20% of river vessels handled containers (about 800,000 TEUs) at Kwai Chung; the balance – around 3.1 million TEUs were handled at a variety of other locations including designated Public Cargo Working Areas (PWCAs), midstream buoys, private wharves and the purpose-built River Trade Terminal. Data available for loaded containers indicated that 26% were handled at Kwai Chung terminals, 46% at PWCAs, 21% at other berths and wharves and 8% at buoys and anchorages. Note also that a high proportion of empties moved in river traffic – 53% of inward traffic and 31% of outward boxes were empty in 1999. To what extent, and on what basis, have the Shenzhen ports become embedded in new networks? For some time “shuttle” services barged Pearl River traffic to Kwai Chung, either directly or via alternate handling locations as noted. By 1996 a small number of international shipping lines had introduced new service strings to include the Shenzhen ports with 8 sailings per week. By mid-2000 the number had increased to
Liner Shipping Strategy, Network Structuring and Competitive Advantage
277
42 per week for all three ports; and Yantian was supporting 23 calls per week. For Yantian, mainline services included Maersk SeaLand, the Grand Alliance carriers (MISC, P&O Nedlloyd, OOCL, NYK and Hapag-Lloyd), Hyundai and APL, China Ocean Shipping Group and Yang Ming; and a range of feeder services included the Grand Alliance carriers, Evergreen/Uniglory, Yang Ming, K Line, DSR Senator, Wan Hai and Lloyd Triestino. In mid-2000 Shekou sustained 11 sailings per week, attracting usually lower valued intra-Asian traffics and smaller vessels given the draft limitations of the approach channels; and Chiwan had 9 services per week. Clearly, despite limited sailings, shipping lines have been able to deliver value as well as capture it in new or alternately structured networks. Interestingly, and particularly in the Hong Kong/southern China context, carriers are concerned less about the size of vessel to be involved in servicing a new port than about whether or not the port and the line can sustain at least one fixed-day sailing per week. The criterion reflects the ability of new alliance structures to access cargo as well as to access a greater range of shipping options in service strings involving different port combinations. Under these conditions much larger vessels will operate in higher order networks with embedded high order hubs as Hong Kong, Singapore and Busan for example and alternate strings will mix and match ports and vessels. 4.2.1.1. Transport costs and cargo routeing: Some indicators. Door-to-door transport costs are not necessarily the definitive factors in cargo routeing decisions made by shippers or buyers or their agents; but they offer some insights and the following discussion is based on useful cost data compiled by the Hong Kong Port and Maritime Board.8 Figure 8 abstracts the Pearl River Delta ports into a set of points so as to underline their general relative location to each other and to the ports of Chiwan, Shekou, Yantian and Hong Kong, to recognise the east bank/west bank structure of the delta and to show the linkages by barge or truck that are typically recognised. Note that there are thee border crossing points linking the northern New Territories with Shenzhen (Man Kam To, Shau Tau Kok and Lok Ma Chau) though these are not shown on the diagram. Table 3 offers insights into the door-to-door transport costs of shipping containers from Dongguan, an east bank delta location, to Long Beach as a transPacific trade destination and to Hamburg as a European trade destination via either Hong Kong or Yantian. In both trade cases costs associated with trucking favour Yantian. From west bank locations (Table 4) barging from Zhongshan to Long Beach for example offers advantages to Hong Kong – though the situation is more balanced in the European trade. In a crudely cost-deterministic framework Yantian
278
ROSS ROBINSON
Fig. 8. Relative Locations of Pearl River Delta Ports and the Major Container Ports.
Table 3. Typical Door-To-Door Transport Costs from East Bank Delta Locations, Transpacific and European Trades, 2001. Truck
Barge
Hong Kong 20’
Yantian
40’
20’
Hong Kong
Yantian
40’
20’
40’
20’
40’
A. Dongguan to Long Beach (US$) 2725 3664 2540
3514
2544
3580
2732
3810
B. Dongguan to Hamburg (US$) 2131 3557 1756 2389 3815 2013
3324 3604
1950 1959
3473 3455
1948 1940
3617 3584
Table 4. Typical Door-To-Door Transport Costs from West Bank Delta Locations, Transpacific and European Trades, 2001. Truck
Barge
Hong Kong 20’
40’
Yantian 20’
Hong Kong
Yantian
40’
20’
40’
20’
40’
A. Zhongshan to Long Beach (US$) 2983 3922 2797
3797
2553
3562
2724
3777
B. Zhongshan to Hamburg (US$) 2389 3815 2013
3604
1959
3455
1940
3584
Liner Shipping Strategy, Network Structuring and Competitive Advantage
279
Table 5. Least Cost Options, Hong Kong and Yantian, September 2001. Least Cost Feeder Options US$ Hong Kong
West Bank Zhauqing – barge Jiangmen – barge Foshan – barge Zhongshan – barge Guangzhou Zhuhai – barge East Bank Dongguan – truck Huizhou – truck Shenzhen – truck Shantao – barge
Yantian
20’
40’
309 258 280 245 323 barge 245
399 382 370 329
355
20’
40’
291 truck 329
553
174 225 70 355
196 250 100 553
may be well supported by trucking operations, the port of Hong Kong by lower barging costs.9 Table 5 focuses on least cost options – either by barge or by truck – for movement to Hong Kong or to Yantian from west bank and from east bank delta locations. The patterns are distinctive, on this basis, with west bank locations gaining cost advantage from barging to Hong Kong, east bank gaining cost advantage to Yantian.10 Table 6 suggests the pattern of advantage between Hong Kong, Yantian and Shekou/Chiwan on least cost feeder terms. In this case, Shekou/Chiwan are seen to enjoy some advantage on trucking and barging costs from both west bank and east bank locations. 4.2.2. Chain Patterns and Shipper Choice: The Dynamics of Export Container Pathways from the Pearl River Delta Hong Kong’s adaptation to containerisation has been, since the late 1960s, not only a triumph of clever and adaptive planning but also of coping mechanisms, of inventiveness, of innovation in an exceptionally dynamic environment. River trade has long been a characteristic of the port – but the rapid industrialisation of the delta and adjacent areas in southern China has set in place a unique and more-or-less informal container handling “precinct” that is, one way or another, integrated with the exceptionally efficient Kwai Chung container terminals.
280
ROSS ROBINSON
Table 6. Least Cost Options, Hong Kong, Yantian and Shekou/Chiwan, September 2001. Least Cost Feeder Options US$ Hong Kong
West Bank Zhauqing – barge Jiangmen – barge Foshan – barge Zhongshan – barge Guangzhou Zhuhai – barge East Bank Dongguan – truck Huizhou – truck Shenzhen – truck Shantao – barge
20’
40’
309 258 280
399 382 370
223
290
355
553
Yantian 20’
Shekou/Chiwan 40’
255
250
355
553
20’
40’
206
206
206
206
145
160
60 355
60 553
Figure 9 indicates the elements involved and the more general pattern of linkages that exists to articulate the movement of containers. Note again the degree of concentration of river traffic – almost 4 million TEUs around 2000, with more than 3 million TEUs handled at PCWAs, the River Trade Terminal, at private wharves and midstream berths. Note also that the PCWAs handled 46% of loaded river trade containers, other berths and wharves handled 21 and 8% of the total was handled in midstream; and Kwai Chung handled about one in four loaded containers (26%). PCWAs have been a characteristic cargo handling element in the port of Hong Kong for some considerable time and emerged to handle an extraordinary mix of break-bulk river trade. Later they were seen to be important locations for handling factory-oriented and manufacturing product. Characterised by low levels of investment, limited backup land and parking facilities and often congested access routes, Government introduced administrative and management reforms after 1997. Restricted tendering introduced some efficiencies; and below-market prices for backup land and minimal levels of investment in basic technology have ensured low operating costs and tariffs. In 1999 there were eight PCWAs – five in Kowloon, one in the New Territories and two on Hong Kong Island. Importantly, three PCWAs – at New Yaumatei, Stonecutter Island and Rambler Channel – are within close proximity of the Kwai
Liner Shipping Strategy, Network Structuring and Competitive Advantage
281
Fig. 9. Elements in the Pearl River Delta Container Operations.
Chung terminals and have operated – and continue to operate – closely with the terminals (In 1999 the three PCWAs handled over one million TEUs and two million tones of break-bulk cargo). East of the Kowloon peninsula Cha Kwo Ling and Kwan Tong PCWAs offered cargo consolidation and were used by freight forwarders and small logistics operators. The two PCWAs on Hong Kong Island – Wan Chai and Western District – handled, together, about 35,000 TEUs in 1999 and about a million tonnes of break-bulk cargo. In 1993 Government moved to regulate river trade and in 1995 invited the private sector to tender for a new “River Trade Terminal,” the first purpose-designed container-handling terminal for river traffic. In 1996 the tender was let; and in due course the terminal was constructed adjacent to the mouth of the delta and close to Tuen Mun. In the event the terminal, by 2000, was unable to compete on costs with those PCWAs which were much closer to the Kwai Chung terminals, offered services that few shippers wanted and were prepared to pay for and could not compete with the rapid development of integrated logistics and supply chain services being offered by the large buyers, carriers and feeder operators able to bypass the terminal.
282
ROSS ROBINSON
Fig. 10. Shipper Chain Options for Pearl River Delta (River) Traffic.
Figures 10–12 provide some insights into both the operational and spatial complexity of the chain pathways operating in the movement of containers to and from the delta. They do not, of course, indicate the complexity of the power relations in the chains. Figure 10 suggests a number of alternative pathways which shippers use. Local manufacturers in Hong Kong and manufacturers or suppliers in the delta,
Fig. 11. Shipper Chain Options at the PCWAs.
Liner Shipping Strategy, Network Structuring and Competitive Advantage
283
Fig. 12. Chain Options for Cargo Moving to the River Trade Terminal.
particularly those on the east bank, have the simplest operational option of direct trucking to Kwai Chung (or to Chiwan, Shekou and Yantian from east bank locations) (Fig. 10A and B). Other options multiply handling and add costs; shippers in the delta may truck cargo to a feeder port, offload to a river vessel or barge or lighter which may then offload at Kwai Chung or at a midstream buoy, often for shipment to intra-Asian locations in vessels somewhat smaller than those using the main terminals. Alternately, the river vessel may unload to a PCWA, to the River Trade Terminal or to private berths for subsequent movement by barge or lighter or truck to Kwai Chung (Fig. 10C). There are, on closer scrutiny, further complexities and a number of other options are available from PCWAs (Fig. 11) and the River Trade Terminal (Fig. 12). Pearl River cargo unloaded at PCWAs may be trucked direct to a local customer; or trucked for storage to an ICD or to a CFS or warehouse for repacking or value adding, then trucked and/or barged to Kwai Chung or to midstream locations (Fig. 11). Cargo moving to the River Trade Terminal may be unloaded to berth and reloaded to alongside intra-Asian trade vessels; or to lighters to Kwai Chung or to midstream buoys; or to trucks for movement to private wharves for storage or repacking and then loaded to trucks for movement to Kwai Chung (Fig. 12)! Multiple handling adds costs; but it is clear that the “informal” container handling market that shipped about 4 million TEUs in 1999 (or 3 million TEUs excluding direct shipments from Kwai Chung) has established cost levels and pricing that are particularly acceptable to the marketplace. Obviously, low investment and low level technology solutions, lower wage rates and high volume throughputs, for example, are important; but even intuitively, as well as from observation and research, it is apparent that there are significant levels of integration, on one basis or another, within the chain systems. A great deal more
284
ROSS ROBINSON
research is needed to clarify these relationships but the more general principles are worth noting in the context of network structuring. The integration of chain activities is being driven by numerous players exerting widely different levels of power and by, or in response to, trading and business conditions impacted by cultural and political realities. Powerful customers, or buyers, may exert ultimate power over chain structures and dynamics and the routeing of cargo; and they may do this directly or indirectly via carriers that have restructured operations to meet the buyer’s specific needs (or value proposition). The large buyer, with significant bargaining power (attributable to high frequency of demand and high level of spend), may contract directly with lines for slots and freight rates; and may nominate freight forwarders to the shipper; or may choose to contract a feeder shipping company (or trucking company) to move cargo from production location to the port of loading. But with the emergence of the megacarrier – as a fully integrated end-to-end operator with effective control over the blue-water, port and end-dray operations to the final customer as well as the landside logistics linkages backwards to the supplier – the large buyer simply ‘buys’ the total package! The Pearl River delta and southern China container operations have been, and will continue to be, strongly impacted, operationally and in efficiency terms, by this development. The powerful buyers and the large and powerful shipping lines are also powerful integrators; by comparison, other players are not unimportant but for the most part exert less power – or that power is exerted over a cluster of logistics activities. Feeder carriers may have contractual relationships with the large carrier, with buyers, with sellers, with the Kwai Chung terminals and also with the individual elements in the “informal” sector of the port of Hong Kong – the River Trade Terminal, the PCWAs and with midstream operations. Major large, international freight forwarders have, not unusually, rationalised and refocused operations to ensure greater ability to integrate freight movements and are key players in chain restructuring. But there remain a large number of “traditional” freight forwarders who have varying levels of integrative power – some with major contractual arrangements with carriers and with ability to contract river vessels and trucks directly or through relationships with feeder operators, others who operate in a relatively ad hoc fashion. 4.2.2.1. Networks in transition: A final note. Since 1972 and the construction of its first purpose built container terminal the port of Hong Kong has been, and will continue to be, one of the world’s largest container shipping ports linked into high order shipping networks. Earlier research indicated some network restructuring with the emergence of south China ports including Yantian and to a lesser extent
Liner Shipping Strategy, Network Structuring and Competitive Advantage
285
Shekou and Chiwan. This case study focuses at an even finer spatial scale to suggest that Hong Kong, Yantian, Shekou and Chiwan are also embedded in complex hierarchical networks of somewhat lower order that include river ports in the Pearl River delta and loading points in the port of Hong Kong – PCWAs, the River Trade Terminal, midstream buoys and private wharves. The container handling “precinct” is unique; but the networks which it sustains are in transition, reflecting not only the changing value capture/value delivery relationships of shipping lines but also the changing patterns of competitive advantage sought by buyer firms. The commercial tensions between Hong Kong on the one hand and the south China ports on the other will persist. Banking, finance, customs, transport and other service advantages in Hong Kong will erode over time. The particular industry structure that has seen the production activities of Hong Kong-owned manufacturing firms (and probably many foreign owned firms) relocated to southern China but product consolidated and shipped from Hong Kong, rather than from southern China, will likely change over time. Certainly, too, changing contractual conditions of sale are encouraging shipment direct from south China and mainland ports. Freight has moved, traditionally, on an FOB (Hong Kong) basis in which the shipper or seller pays all the costs from factory to the ship’s rail at the port of shipment. More recently, shipments FOB (South China) encourage shipment from local ports rather than Hong Kong, though on these terms shipment through Hong Kong can still be made on a through Bill of Lading with transhipment in Hong Kong.11 For the seller, the ability to arrange landside transport allows the possibility of extra margin on production cost. For the buyer, it is clearly preferable to minimise the costs of transport to the port of shipment – and as we have noted, containers originating from the east bank of the delta may benefit the ports of Shenzhen. In any case shipping lines, and buyer firms, will likely see advantage in the continuing integration and rationalisation of export chains for high value goods and sophisticated markets; but the low cost “informal” sector is unlikely to be competed out of the marketplace for some considerable time yet!
5. SUMMARY AND CONCLUSIONS This paper has addressed the challenge, prompted by Panayides and Cullinane in 2002, of what it is that creates competitive advantage for liner shipping companies; and why it is that after a long and distinguished record of research in maritime economics, researchers have failed to clearly understand why shipping companies, despite aggressive expansion of networks and more ports of call, access to wider markets and often enhanced revenues, have often recorded low or negative profits
286
ROSS ROBINSON
and unimpressive business outcomes. In urging further research the authors, though without attempting to proscribe further efforts, suggested that a micro-analytic, firm-specific approach might in fact provide insights not available from the more usual market-based, industry wide view. This paper supports the call for a firm-specific approach; and it does so by offering a conceptual framework that not only draws heavily on the antecedent frameworks of Porter and the Harvard school as well as that of the resource-based view of the firm in explaining the bases of competitive advantage and business success but also adds a chain and supply chain view to the functionality and power of shipping lines as particular types of firms. It is easy to be beguiled by the corporate strategies of shipping lines that argue for bigger and better networks, for the inclusion of more ports of call, for larger numbers of customers and greater market share. But it is well to remember that networks are artifacts of corporate strategies; that networks not only expand but also often decay – and decay particularly rapidly if shippers no longer value the output of the line – or the services it has to offer. It is not the network size per se that is important but the value delivered by the line that happens to be operating in a network or networks. The critical issue is how shipping lines deliver value and capture value to ensure the line’s sustainability. This paper draws upon and extends a conceptualisation outlined in earlier papers as a “value-driven chain systems” approach to clarify the notions of competitive advantage and business success. It argues for a framework with five key concepts and a number of implications that follow. Particularly, it notes that shipping lines are third party service providers; that they are also networked service providers; that they compete in and carry out their every day operations in corporate chains and supply chains; that these chains are characterised by power structures that impact on the value delivered and captured by lines; and that business success is related to the firm’s ability to achieve supply chain and market power and to capture rents. The details are argued in the paper; but the concepts of value, value delivery to individual customers and to market segments and value capture by the line are key elements in the conceptualisation. Further, the particular value to be delivered is not self-evident but will be specific to individual customers or to market segments; and the degree to which value is captured by the line will reflect the strategy framework within which the line chooses to operate. The paper further clarifies the framework by offering a business model that makes explicit the relationships between core strategy, the line’s activity systems, the value to be delivered and the business success of the line. The paper is not offered, in any sense, as a final statement on the notion of value-driven chains and on defining strategies delivering competitive advantage
Liner Shipping Strategy, Network Structuring and Competitive Advantage
287
and business success for third party service providers generally, or in this context, for shipping lines particularly. The paradigm will require a great deal of empirical research and research testing. Certainly, the two case studies included in the paper provide detailed analyses of particular market and network configurations; but they claim to show, more particularly, that research at the level of the individual firm or the firms involved may be needed to reveal relationships between network structuring, the bases of competitive advantage for different players, the varying meanings of delivering and capturing value for different chain players and, particularly in the Hong Kong/South China study, the relative instability of the value delivery/value capture relationship for shippers and for shipping lines.
NOTES 1. Note that Cox (1997, pp. 207–208) refers to this redistribution process as the “value chain” – ‘the process by which money is exchanged through a supply chain.” Clearly, and as Cox notes, this definition differs from Porter’s notion of the value chain as a firm’s set of activities by which it creates value for itself and for its customers. My preference is to leave the term “value chain” to Porter and to recognise explicitly, as in the diagram, that value flows both ways; but here customer value may have a much richer meaning than is expressed in simple dollar terms. In fact, the serious flaw in Cox’s (1997) conceptualisation was the failure to recognise the critical importance of the “countervailing powers that customers . . . have over even monopoly suppliers” and that unless the customer values “what is being offered” the supplier has no leverage at all! In 2002 (Cox et al., 2002) attention has been paid to the issue but it is arguable that the critical importance of customer value in chain structuring remains understated. (For a statement of the differences between the 1997 and 2002 positions see the Preface to Cox et al., 2002 and particularly pp. xi–xvi.) 2. This case study draws on a number of research papers including Robinson (1998), Robinson, R. and Everett, S. E. (1997), Malaysian Cargo through Malaysian Ports: Defining Appropriate Policies, Unpublished Research Report prepared for the Maritime Institute of Malaysia (MIMA); and Robinson, R. (1998), Hub/feeder strategies for ports in complex shipping systems: Port Klang, Malaysia, Paper presented to the Pacific Rim Allied Economic Organisations Third Biennial Conference, Bangkok, June. 3. Loop A included Port Klang Eastbound and Loop B included the port westbound. The port of Singapore was included in all Loops. The port of Southampton was included in three of the four Loops – A, C and D. 4. For a recent statement see Mak, J. N. and Tai, B. K. M. (2001), Port development within the framework of Malaysia’s transport policy: some considerations, Maritime Policy and Management, Vol. 28, No. 2, pp. 199–206. 5. This case study draws from ongoing research into the development of the port of Hong Kong and, in particular, into the structure of export container chain systems. It draws also on unpublished preliminary research by Lawrence Wong in the context of a DBA program and on Lawrence’s encyclopedic understanding of the operations of the port. I acknowledge
288
ROSS ROBINSON
with gratitude his contribution. I acknowledge also the excellent data sources and reports available from the Hong Kong Port and Maritime Board. 6. See Robinson, R., and Chu, D. Y. K. (1978), Containerisation and the Port of Hong Kong in the 1970s, Australian Geographer, Vol. 14, pp. 98–111. 7. See Hong Kong Port and Maritime Board (2001), Hong Kong Port Cargo Forecasts 2000/2001, Final Report, Section 2.2.1. 8. The Board usefully defines door-to-door costs in the Hong Kong/south China context in considerable detail to include the following variables: Ocean freight rates including the Basic Freight Rate, Full Adjustment Factor, Bunker Adjustment Factor; Terminal Handling Charges including Original Receiving Charges as South China ports, Terminal Handling Charges in Hong Kong; Documentation fee; Declaration Fee (PRC); Import declaration fee (HK); Re-export declaration fee (HK); Trucking from factory to barge terminal or direct to Kwai Chung terminal; Destination Delivery Charge; Transport management fee; Barge freight rate. See Hong Kong Port and Maritime Board (2001), op. cit., Table 4.1. 9. Tables 3 and 4 have been abstracted from more detailed tables (Tables 4.4 and 4.5) in Hong Kong Port and Maritime Board (2001), op. cit. 10. Tables 5 and 6 have been abstracted from a series of figures (Fig. 4.1 to 4.4) in Hong Kong Port and Maritime Board (2001), op. cit. with cost values updated to September 2001. 11. For a useful discussion see Hong Kong Port and Maritime Board (2001), op. cit., Section 4.
REFERENCES Barney, J. B. (2001). Is the resource-based ‘view’ a useful perspective for strategic management research? Yes. Academy of Management Review, 26(1), 41–56. Collis, D. J. (1994). Research note: How valuable are organizational capabilities? Strategic Management Journal, 15, 143–152. Conner, K. R. (1991). A historical comparison of resource-based theory and five schools of thought within industrial organization economics: Do we have a new theory of the firm? Journal of Management, 17(1), 121–154. Cox, A. (1997). Business success. UK: Earlsgate Press. Cox, A., Ireland, P., Lonsdale, C., Sanderson, J., & Watson, G. (2002). Supply chains, markets and power. London: Routledge. Gerstner, L. V. (2002). Who says elephants can’t dance. Sydney: Harper Business. Hamel, G. (2000). Leading the revolution. Boston: Harvard Business School Press. Hunt, S. D. (2000). A general theory of competition. USA: Sage. Kaplan, R. S., & Norton, D. P. (1996). The balanced scorecard. Boston: Harvard Business School Press. Kaplan, R. S., & Norton, D. P. (2001). The strategy-focused organization. Boston: Harvard Business School Press. Kuhn, T. S. (1970). The structure of scientific revolutions. Chicago: University of Chicago Press. Mahoney, J. T. (1992). The resource-based view within the conversation of strategic management. Strategic Management Journal, 13, 363–380. Panayides, P. M., & Cullinane, K. P. B. (2002). Competitive advantage in liner shipping: A review and research agenda. International Journal of Maritime Economics, 4(3), 189–209. Penrose, E. (1959). The theory of the growth of the firm. Oxford: Oxford University Press.
Liner Shipping Strategy, Network Structuring and Competitive Advantage
289
Peteraf, M. A. (1993). The cornerstones of competitive advantage: A resource-based view. Strategic Management Journal, 14, 179–191. Phillips, L. (1987). Building market focused organizations. Stanford University (mimeo). Porter, M. E. (1979). How competitive forces shape strategy. Harvard Business Review (March–April). Reprinted in: M. E. Porter, On Competition (1998, pp. 21–38). Harvard Business Review Book Series, Boston. Porter, M. E. (1980). Competitive strategy: Techniques for analysing industries and competitors. New York: Free Press. Porter, M. E. (1985). Competitive advantage: Creating and sustaining superior performance. New York: Free Press. Porter, M. E. (1990). The competitive advantage of nations. London: Macmillan. Porter, M. E. (1996). What is strategy? Harvard Business Review (November–December), 61–78. Reprinted in: M. E. Porter, On Competition (1998, pp. 39–73). Harvard Business Review Book Series, Boston. Priem, R. L., & Butler, J. E. (2001). Is the resource-based ‘view’ a useful perspective for strategic management research. Academy of Management Review, 26(1), 22–40. Robinson, R. (1998). Asian hub/feeder nets: The dynamics of restructuring. Maritime Policy and Management, 25(1), 21–40. Robinson, R. (2002). Ports as elements in value-driven chain systems: The new paradigm. Maritime Policy and Management, 29(3), 241–255. Robinson, R. (2003). Port authorities: Defining functionality within a value-driven chain paradigm. Proceedings IAME Conference, Busan, September, pp. 654–674. Rouse, M. J., & Daellenbach, U. S. (1999). Rethinking research methods for the resource-based perspective: Isolating sources of sustainable competitive advantage. Strategic Management Journal, 20, 487–494. Wernerfelt, B. (1984). A resource-based view of the firm. Strategic Management Journal, 5(2), 171– 180. Yap, W. Y. et al. (2003). Developments in container port competition in East Asia. Proceedings of the IAME Conference, Busan, September, pp. 715–735.
AUTHOR INDEX Abbott, R., 149 Adcock, G., 225 Akatsuka, K., 6 Alderton, T., 174, 182 Alizadeh, A., 32, 36, 72 Anonymous, 151, 156, 161, 224, 228 Arditti, F., 117 Arkoulis, A. G., 6 Baird, A. J., 227, 228 Ball, R., 114 Banz, R., 113 Barney, J. B., 249 Basu, S., 113, 114 Beenstock, M., 3, 4, 19–21, 23, 26, 28, 60, 68, 71, 74, 75, 79 Beresford, A. K. C., 228 Berg-Andreassen, J. A., 35 Bergantino, A., 160, 161, 184 Bhandari, L., 114, 142 Bollerslev, T., 33 Boudoukh, J., 120 Brooks, C., 87 Brooks, M. R., 146, 152, 159, 223 Brownrigg, M., 160, 163 Bunel, J.-C., 223 Burmeister, E., 115 Butler, J. E., 249, 250 Butz, D. A., 223 Campbell, J. Y., 32 Chan, L., 114 Charemza, W., 24, 71 Chen, J., 116 Chen, N., 115, 116, 129 Chen, Y.-S., 6 Cheng, J., 232, 242 Cheng, P. C., 6 Clyde, P. S., 223
Coeck, C., 227 Collis, D. J., 249 Comtois, C., 225 Conner, K. R., 249, 250 Cox, A., 14, 249–251, 256, 257, 287 Cullinane, K. P. B., 6, 9, 12, 160, 161, 221, 223–226, 228, 232, 237–239, 242, 248 Cullinane, S. L., 224, 226, 237 Daellenbach, U. S., 250 Damodaran, A., 115 Darling, H. J., 145 Davies, J. E., 223 De Souza, G. A., Jr., 228 Deakin, B. M., 223 Dickey, D. A., 3, 84 Doi, M., 223 Drewry, 82 Engerman, M., 117 Engle, R. F., 3, 28, 29, 33 Eun, C., 120 Euromoney, 6 Eviews, 39 Fama, E., 114, 129 Ferson, W., 117, 120 Fossey, J., 226 Fox, N. R., 223 Franck, B., 223 Frankel, E., 226 French, K., 114, 129 Fuller, W. A., 3, 84 Gardner, B. M., 8, 158, 159, 163, 223 Gerstner, L. V., 248, 262 Gilman, S., 223–225 291
292 Glen, D., 25, 30, 35, 37, 72, 79 Gong, X., 6 Gooding, C. W., 223 Goss, R. O., 8, 158, 161, 163, 167 Grammenos, C. Th., 6 Granger, C. W., 3, 28, 29 Greene, W., 39, 43, 45, 51 Gronicki, M., 24, 71 Gyourko, J., 118 Hale, C., 25, 72 Hamao, Y., 114, 115 Hamel, G., 248, 249, 261 Haralambides, H. E., 6, 224 Harvey, C., 117, 120 Hawdon, D., 23, 67, 75 Hayuth, Y., 224 Heaver, T., 10, 223 Hoffman, J., 228 Holste, S., 9 Hunt, S. D., 249 Isimbabi, M., 117, 120 Itoh, H., 223 Jankowski, W. B., 223 Jansson, J. O., 223 Ji, P., 226, 239 Jin, D., 69 Johansen, S., 3, 29, 87 Jones, C., 110, 118, 119 Jordan, D., 116 Kahn, R., 118 Kane, E., 117 Kaplan, R. S., 252, 253, 260 Kavussanos, M., 6, 32–37, 46, 55, 58, 72, 80, 110, 111, 116, 118–125, 127–129, 131, 134, 137 Keim, D., 118 Khanna, M., 221, 223, 225 King, B., 112, 113, 117 Knudsen, K., 160, 163 Koopmans, T. C., 3, 4, 20, 67 Kuhn, T. S., 248 Kwan, R., 238
AUTHOR INDEX Lakonishok, J., 113, 114 Lambe, H. B., 146 Lanstein, R., 114 Lee, T.-W., 159 Leggate, H. K., 6 Li, K. X., 10, 12, 174, 232, 242 Lintner, J., 113 Mackinnon, J., 39 Mahoney, J. T., 249 Marcoulis, S., 4, 6, 110, 111, 116, 118–120, 124, 125, 127–129, 131, 134, 137 Marlow, P. B., 8, 146, 152, 158–161, 163, 167, 174, 175, 184 Martin, B., 35, 37, 72 Martinez, A., 115, 141 McCalla, R., 225 McConville, J., 156 McElroy, M., 115 Mirmiran, R., 33, 37 Morrell, P. S., 6 Neel, R. E., 223 Nerlove, M., 117 Neuberger, J., 117 Norman, V., 21 Norton, D. P., 252, 253, 260 Notteboom, T. E., 12, 227, 228 Officer, L. H., 238 Ohta, H., 223 O’Mahony, H., 240 Paine, F., 6 Panayides, Ph. M., 16, 231, 235, 248 Penrose, E., 14, 249 Peteraf, M. A., 249 Peters, H. J. F., 228 Pettit, S. J., 228 Phillips, L., 252 Poon, S., 115 Popper, K. R., 10 Porter, J., 224, 249, 252, 253 Porter, M. E., 14, 249, 261 Priem, R. L., 249, 250
293 Reid, K. R., 114 Reinganum, M., 113, 114 Reitzes, J. D., 223 Resnick, B., 120 Richardson, P., 8, 158 Roberts, S., 174 Robertshaw, M., 9, 160, 161 Robinson, R., 224, 251, 253, 287 Roe, M., 160 Roll, R., 113, 115, 116 Rosenberg, B., 114, 118 Rosenberg, J., 118 Ross, S., 115, 116 Rouse, M. J., 250 Roy, V., 118 Rubio, J., 115 Saunders, A., 117 Scherer, F. M., 223 Schneerson, D., 223 Scholtens, M., 225 Schubert, W. G., 157 Selkou, E., 160 Seward, T., 223 Seymour, J., 150 Shapiro, A., 113 Sharpe, W., 113, 116 Shiller, R. J., 32 Sjogren, H., 6 Sjostrom, W., 223 Slack, B., 224, 225 Sletmo, G. K., 9, 164 Slogett, J. E., 6 Song, D.-W., 226, 228 Sorensen, E., 116, 117 Stares, J., 228 Starr, J. T., 224 Stattman, D., 114 Stephenson Harwood, 6
Stokes, P., 17 Strandenes, S. R., 22, 71 Tabernacle, J. B., 226 Talley, W., 174 Taylor, S. J., 115 Thanopoulou, H. A., 174, 175 Tinbergen, J., 3, 4, 20, 67, 103 Tolofari, S. R., 159–161 Tsolakis, S. D., 36 Tvedt, J., 36 Unal, H., 117 Vanags, A., 25, 72 Veenstra, A., 3, 4, 25, 28, 30, 32, 36–38, 42, 44, 60, 72 Verbeke, A., 227 Vergottis, A., 4, 19–21, 26, 28, 60, 68, 71, 75, 79 Visvikis, I. D., 6 Volk, B., 70 Waals, F., 225 Wall, K., 115 Wang, S.-T., 6 Wang, T., 224, 226, 237, 239 Wasserfallen, W., 115 Wergeland, T., 21 Wernerfelt, B., 249 Wijnolst, N., 225 Winchester, N., 174, 182, 201 Winkelmans, W., 12, 227 Wonham, J., 10, 174 Wright, G., 33 Yap, W. Y., 253 Yong, J. S., 223 Yourougou, P., 117 Zannetos, Z., 24, 35
SUBJECT INDEX 128–130, 132, 133, 175, 221, 240, 3PL(s), 239–241 256 Academic(s), 3, 8, 110, 113, 116, 118, 120, Analys(es)is, 2–7, 9–11, 13–16, 19, 30, 33, 127–129, 157–159, 160, 163, 223, 41, 42, 58, 62, 69, 71–73, 77, 80, 84, 233 86, 87, 104, 110–112, 117–120, 122, Access, 7, 14, 153, 167, 225, 230, 232, 124, 125, 127, 130, 131, 145, 236, 239, 247, 253, 277, 280, 285 158–160, 165, 174, 182, 221, 242, 287 Accession, 13, 221, 231–240, 242 Analytic(al), 3, 5, 14, 15, 248, 249, 251, Accident(s), 17, 166, 174, 175, 207 261, 263, 286 Acquisition(s), 17, 78, 122, 152, 154, 166, Antigua and barbuda, 179, 201–205, 209 222, 223, 251, 257, 263 Antwerp, 227, 243, 244 Act, 135, 147, 149, 150, 152, 157, 162, APL, 222, 277 165, 170, 184 Application(s), 26, 28, 33, 34, 36, 48, 111, Activit(y)ies, 14, 27, 71, 107, 108, 117, 133, 155, 168, 224, 225, 240 120–123, 132, 144, 146, 147, 149, Arbitrage, 22, 26, 115, 116, 134, 135 154, 155, 162–164, 167, 168, 222, Arbitrage pricing theory (apt), 115, 116, 227, 228, 233, 234, 239, 249, 252, 134 258, 259, 261, 284–287 ARCH, 33–35, 37, 62, 72, 88, 104 Administration, 8, 9, 159, 174 ARIMA, 34, 72, 124, 128 Administrative, 160, 280 Arrangements, 21, 25, 149, 256, 267, 268, Aframax(es), 34, 79, 81, 97, 98, 100, 103 274, 284 Age, 11, 58, 71, 72, 175, 177, 179, 183, ASEAN, 235, 266 184, 201 Ashore, 8, 10, 153, 166 Agenc(y)ies, 166, 231 Asia(n), 2, 149, 164, 170,193, 201, 203, Agent(s), 26, 109, 277 206, 227, 235, 236, 240–242, 245, Aggregate, 3, 21, 34, 67, 79 253, 266, 277, 283, 287, 289 Aggregation, 224 Agreement(s), 134, 160, 232–234, 241, 242 Asia-Pacific, 149, 164 Asset(s), 4, 6, 8, 21, 25, 26, 32, 37, 65–72, Agricultur(e)al, 169, 232, 233, 235, 262 74, 80, 97–99, 104, 109, 110, 113, Air, 6, 119, 126, 133, 162, 238 115, 125, 127, 128, 130, 131, 133, Airline, 159, 160, 162 134, 136, 154, 157, 162, 167, 169, Akaike Information Criterion (AIC), 39 256, 258–260 Alliance(s), 223, 225, 226, 228, 253, 262, Assumption(s), 19, 25, 26, 30, 32, 33, 38, 270, 273–275, 277 53, 61, 71, 72, 79, 84, 116, 145 Allowance(s), 152, 155, 158 Augmented Dickey-Fuller (ADF), 34, 38, Alpha(s), 126–130, 137 39, 48, 51, 53, 60, 61, 84 American bureau of shipping (abs), 178, Australia, 157, 209 179, 201, 203, 204 Authorit(y)ies, 9, 205, 251, 268, 289 Analys(ed)ing, 1–3, 5, 10, 11, 13, 27, 29, Autocorrelat(ed)ion, 33, 55, 84, 88 67, 69, 80, 107, 110–112, 122, 295
296 Autoregressive (AR), 16, 28–30, 62, 66, 72, 84, 100, 103, 104 Autoregressive Conditional Heteroskedasticity (ARCH), 33–35, 37, 62, 72, 88, 104 Average(s), 23–25, 33, 34, 67, 70–74, 100, 104, 114, 118, 122, 123, 125, 126, 128–130, 132, 137, 174, 175, 177, 179, 182, 184, 206, 208, 227, 233, 240, 264 Bahamas, 176, 177, 179, 201–204, 208 Ballast, 80, 166 Baltic Exchange, 151 Bankers, 78, 124, 133 Banking, 16, 117, 134–136, 285 Bankruptcy, 17, 66, 83 Banks, 6, 83, 109, 117, 118, 134, 136 Bareboat, 25, 159 Barge(d), 276, 277, 279, 283, 288 Barging, 277, 279 Barra International, 117, 118, 129, 135, 136 Barriers, 159, 223, 232, 233, 256, 273 Bay of Bohai, 236, 238 Bear, 61, 125 Behaviour(s), 1, 5, 14, 19, 21, 23, 26, 27, 30, 32–35, 44, 46, 59, 60, 70, 72, 83, 84, 97, 103, 107, 125, 131, 135, 158, 249, 257 Beijing, 238, 243 Belgium, 157, 176, 211, 227, 243 Belize, 179, 202–205, 209 Benchmark(s)(ing), 77, 79, 81, 122 Benefits, 10, 145, 147, 156, 158, 160, 163, 164, 168, 169, 183, 223, 253, 260, 261, 274 Bergen, 22, 63, 105 Bermuda, 179, 201–204, 209 Berth(s)(ing), 226, 237, 241, 276, 280, 283 Beta(s), 125–130, 132, 137–140 Bias(es), 26, 113 Bill of lading, 149, 157, 269, 274, 285 Binary, 175, 177–179 Board, 145, 146, 169, 244, 277, 288 Bolivia, 179, 202–205, 211 Bonds, 109, 116
SUBJECT INDEX Book, 26, 30, 61, 77, 104, 110–112, 114, 116–119, 125, 127, 128, 130, 131, 134, 136, 141, 142, 289 Borrowing, 99, 109, 134 Boxes, 264, 276 British, 16, 156, 171, 201, 212 Broker(s)(age), 80, 122, 166 Brunei, 210, 228 Budget(s)(ing), 16, 133, 174 Building, 6, 16, 74, 88, 99, 108, 258 Bulk, 2, 3, 5, 6, 13, 15, 16, 19, 20–22, 26, 28–30, 32, 34–37, 46, 48, 61–63, 71, 72, 74, 79–81, 83, 85, 96–100, 103–105, 108, 132, 151, 174, 175, 178, 182, 187, 189, 203, 207, 208, 215, 218, 236, 242, 280, 281 Bulk carriers (bulkers), 74, 96, 99, 100 Bull, 125 Bunker(s), 20, 23, 24, 27, 34, 46, 48, 52, 53, 75 Buoys, 276, 283, 285 Bureau Veritas, 178, 179, 203, 204 Busan, 277, 289 Business(es), 14, 16, 17, 36, 63, 65, 82, 103, 105, 108, 116, 118, 121, 122, 132, 134–136, 146, 148, 149, 151, 155, 158, 170, 171, 173, 174, 206, 223, 225, 227, 228, 230, 231, 233, 234, 238, 241, 247–252, 254, 256–262, 284, 286–289 Buy, 66, 74, 134 Buyer(s), 93, 147, 240, 248, 250, 252, 254, 256, 257, 259, 260, 273, 277, 281, 284, 285 Cabotage, 151–153, 167, 184 Call(s), 10, 36, 224–227, 236, 237, 267, 269, 270, 273, 277, 285, 286 Canada, 8, 9, 117, 143–149, 151, 152, 157, 162–171, 209, 266 Canadian, 9, 10, 118, 136, 144–152, 157, 159, 162–164, 166–170 Capacit(y)ies, 5, 11, 21–23, 27, 66–70, 73, 76, 77, 79, 82, 83, 93, 97, 100, 104, 112, 131, 134, 151, 167, 176–179, 184, 185, 193, 202, 222, 245, 266, 267, 276
Subject Index Capesize, 37, 39, 44, 51, 53, 79, 80, 96, 97, 103, 212 Capital, 6, 16, 22, 25, 65, 69, 71, 73, 74, 75, 78, 98, 99, 103, 107, 109–111, 112, 116, 124, 131, 133, 134, 136, 146, 152, 159, 162, 163, 167, 226, 234, 235 Capital Asset Pricing Model (CAPM), 113, 122, 125, 126, 132, 133, 137 Capitalization, 110, 120, 122 Cargo(es), 2, 5, 19, 21–24, 26, 27, 33–35, 37–39, 41, 43, 44, 46, 48, 51–53, 55, 58, 60–62, 72, 74, 75, 80, 81, 100, 104, 105, 108, 109, 145, 151, 161, 165, 166, 168, 176, 178, 179, 182, 184, 185, 187, 189, 193, 201–203, 205, 214, 215, 218, 223–228, 230, 231, 236, 237, 239, 240, 241, 244, 259, 263, 264, 266–271, 275–277, 280, 281, 283, 284, 287, 288 Caribbean, 175, 177, 179, 193, 201, 202, 205, 213 Carriage, 2, 5, 240 Carrier(s), 2, 5, 6, 22, 23, 27, 71, 73, 74, 79–81, 83, 85, 96–100, 103, 104, 168, 174, 222–225, 227, 228, 240, 241, 277, 281, 284 Cash, 72, 82, 99, 115, 128, 133 Cashflow, 72, 99 Casualty, 175, 178, 182, 187–189, 205 Causal(ly)(ity), 3–5, 68, 70, 79, 221, 264 Certification, 151, 174, 204 Chain(s), 13–15, 250–252, 254–262, 275, 279, 282–287, 289 Charges, 224, 241, 272 Charter(s)(ing), 4, 5, 21–27, 30–35, 37–39, 41, 44–46, 48, 51–53, 60, 62, 63, 71, 108, 109, 134, 159, 161, 166, 185, 258, 259 Charterer(s), 21, 25, 60, 74, 75, 80, 134 Chemical(s), 81, 118, 234 Chicago, 136, 162, 288 China, 13, 15, 36, 97, 178, 179, 184, 193, 201, 208, 221, 226, 228, 230–244, 248, 263, 275–277, 279, 284, 285, 287, 288 Chinese, 6, 203, 231, 233, 239
297 Chi-squared, 183, 185, 193, 202 Chiwan, 275–277, 279, 283, 285 Clarksons, 3, 79, 81, 104 Class(ed), 174, 178, 179, 185, 187, 189, 193, 201–203, 205–207 Classification, 9, 81, 111, 116, 117, 119, 120, 121, 125, 127, 131, 174, 175, 177–179, 185, 187, 189, 193, 203–205 Classification society, 174, 178, 179, 189 Cluster, 10, 155, 284 CMA-CGM, 222 Coefficient(s), 22, 26, 27, 29, 32, 42, 43, 55, 59, 64, 67, 87, 99, 103, 125–127, 138–140, 183–187, 193, 201–203 Cointegrate(d), 29, 31, 34–36, 45, 46, 86, 87 Cointegrating, 29, 35, 45, 48, 51–53, 72, 87 Cointegration, 16, 19, 28, 29, 35, 37, 46, 51, 53, 62, 63, 72, 87, 104, 105 Commerce, 2, 135, 136, 149, 241 Commercial, 11, 145, 147, 154, 162, 175, 176, 231, 232, 269, 285 Commodities, 66, 99, 100, 235 Common Agricultural Policy (CAP), 33, 63 Community, 3, 12, 126, 156, 160 Compan(y)ies, 3, 6–8, 11, 13–15, 61, 83, 107–114, 116–122, 124–129, 132–134, 144, 146–149, 152, 154, 156, 158, 159, 161, 164, 168, 175, 184, 223, 225, 228, 230–234, 236, 239, 240, 241, 249, 251, 256, 257, 270, 284, 285 Companion form, 30–32, 64 Compete, 12, 74, 77, 119, 154, 163, 223, 226, 231, 240, 241, 251, 254, 256, 259, 262, 281, 286 Competition, 8, 16, 68, 77, 108, 147, 153, 154, 158, 160–162, 167, 170, 174, 223, 230, 232, 236, 241–245, 247, 249, 254, 256, 257, 273, 288, 289 Competitive(ness), 9, 13, 14, 17, 19, 22, 46, 66, 68, 70, 77, 79, 82, 146, 149, 151, 153, 154, 156, 162, 163, 165, 166–168, 170, 206, 208, 226, 228, 232, 237, 241, 243, 244, 247–254, 257, 259, 261–263, 269, 274, 275, 285–289
298 Competitive advantage, 13, 14, 17, 206, 237, 243, 247–253, 259, 262, 263, 269, 274, 275, 285–289 Competitor(s), 12, 14, 236, 256, 257 Complementarity, 235, 250 Compliance, 10, 161, 173 Computer, 39, 118 Concentrate(d), 161, 224, 225, 267, 269 Concentrating, 225, 269 Concentration(s), 13, 222, 223, 225–228, 236, 237, 240, 247, 270, 273, 280 Concept(s), 1, 4, 9, 10, 80, 109–111, 133, 151, 155–158, 162, 163, 165, 168, 224, 225, 240, 241, 250–252, 254, 262, 286 Conceptual(ization), 12, 14, 145, 248, 249, 251, 253–257, 259, 261, 263, 286, 287 Conditional mean(s), 33–35 Conduct, 2, 3, 13, 233, 254 Conference, 170, 224 Congested, 237, 280 Connectivity, 13, 239 Construction, 23, 67–69, 97, 167, 168, 185, 237, 284 Consumer(s), 61, 112, 116, 167, 236 Consumption, 112, 128, 134, 141 Container(s)(ized), 2, 11, 13, 15, 108, 132, 178, 179, 182, 184, 185, 187, 189, 193, 201–203, 214–216, 221–228, 231, 236–245, 253, 263–277, 279–285, 287, 289 Containerization, 236, 275, 279 Containership(s), 74, 221, 224–227, 240, 243 Contestability, 254, 267 Contestable, 254, 273 Contract(s)(ing)(ual), 2, 6, 13, 21–23, 68, 82, 97, 100, 233, 252, 253, 267, 268, 284, 285 Control, 12, 14, 16, 62, 124, 147, 149, 161, 165, 166, 168, 169, 173, 174, 176, 187, 201, 202, 204, 205, 207, 222, 247, 249, 254, 256, 275, 284 Convention(s), 11, 60, 148, 162, 165, 173, 175, 178, 179, 182, 187–189, 204–208, 243
SUBJECT INDEX Converg(es)(ent)(ence), 26, 30, 31, 144, 257 Corporate, 15, 109, 110, 116, 133, 134, 148, 152–154, 159, 161–163, 165, 173, 221, 222, 248, 253, 256, 259, 261, 262, 267, 269–271, 286 Corporation(s), 9, 116, 120, 146, 148, 149, 151, 162, 164 Correction(s), 16, 26, 28, 29, 46, 62, 66, 86, 87, 103 Correlated(d), 24, 65, 66, 71, 188 Correlation(s), 44, 84, 127, 135, 185, 225 Corridor(s), 253, 264, 265 COSCO, 236, 239 Cost(s), 5, 6, 8, 9, 12, 14, 16, 20, 22–27, 38, 46, 61, 66–71, 73, 75, 77, 78, 82, 88, 93, 96–100, 103, 104, 108, 109, 133, 134, 144–147, 152, 153, 155, 157–164, 166, 167, 170, 171, 174, 175, 184, 188, 206, 207, 224, 225, 232, 235, 237, 239, 252, 253, 255–257, 260, 267–275, 277, 279–281, 283, 285, 288 Countr(y)ies, 8, 9, 11, 16, 61, 68, 70, 77, 97, 115–117, 120, 122, 124, 125, 143–146, 149, 151, 152, 155–160, 162, 163, 167–169, 173–179, 182, 184, 185, 187–189, 193, 201, 202, 204–208, 226, 227, 230, 232, 233, 235, 239 Credit(s), 78, 131, 158 Crew(s)(ing), 25, 151, 152, 161, 162, 174, 184, 206 Cross - section(al), 114, 117, 135 Crude, 58, 61, 100, 234, 242 Cruise, 2, 108, 201 Cultural, 9, 262, 284 Culture, 13, 161 Currenc(y)ies, 2, 36, 75, 77, 79, 93, 97, 116 Customer(s), 1, 14, 15, 82, 109, 248, 249, 252–262, 283, 284, 286, 287 Customs, 150, 152, 159, 238, 285 Cycle(s), 4, 17, 20, 27, 36, 65, 67, 70, 103, 116, 124, 257 Cyclical(ity), 16, 62, 66, 67, 97, 105, 119, 124, 126, 134, 144
Subject Index Dalian, 228, 238 Danish, 155, 209 Data, 3–5, 11, 25–30, 33–39, 41–43, 46–48, 51–53, 55, 58, 60, 61, 72, 73, 77, 79–82, 84, 110, 111, 116, 119–123, 132, 135, 164, 175, 176, 179, 183, 185, 193, 202, 206, 213, 219, 265, 268, 275–277, 288 Data collection, 5, 16, 73, 80, 132 Database, 3, 11 Datastream International, 79, 104, 122 Deadweight (dwt), 37, 44, 47, 48, 67, 74, 79, 176, 178, 179, 184, 186, 189, 193, 202, 208–213 Decision(s), 7–12, 16, 80, 99, 100, 107–111, 119, 129, 135, 146, 149, 153, 157–161, 163, 165, 169, 173, 237, 239, 270, 271 Deep-sea, 144, 146–148, 163, 164, 168, 169, 171 Definition, 5, 75, 188, 248, 249, 253, 257, 258, 287 Degrees of freedom, 34, 36, 183, 186, 193 Delays, 269, 272 Deliver(y)(ies), 15, 23, 38, 66–68, 77, 82, 248, 250, 252–254, 257, 259–263, 268, 272, 274, 277, 285–287 Demand, 4, 19–24, 26, 27, 37, 46, 58, 61, 63, 66–71, 73–76, 78, 80, 83, 88, 97, 99, 100, 105, 109, 110, 112, 124, 134, 155, 224, 230, 236, 250, 259, 267, 284 Denmark, 156, 209 Density, 183, 247 Departure, 257, 260, 262, 269 Dependent, 5, 6, 11, 24, 34, 41, 46, 59, 67, 87, 158, 187, 226, 257 Depreciation, 23, 72, 77, 153, 159 Design, 82, 108, 226 Destination(s), 235, 236, 265, 267, 268, 273, 277 Det Norske Veritas, 178, 179, 203 Determinant, 76 Determinant(s), 7, 10, 22, 70, 74, 76, 111, 113, 114, 125, 127–130, 135, 161, 173, 175, 193, 205, 207, 208, 244, 249 Deterministic, 39, 84, 249, 277 Differentiation, 62, 79, 104, 241, 261, 262
299 Disaggregate, 5, 14, 37, 80 Disaggregation, 3, 79 Discount(s)(ed), 30, 32, 33, 42, 80, 82, 93, 99, 115, 133, 183, 241 Discount rate, 30, 32, 42 Distribution(s)(al), 16, 39, 55, 58, 81, 84, 85, 168, 183, 224, 232, 234, 239, 247, 250 Disturbances, 66, 127 Diversif(y)ication, 122, 131, 133, 238, 240, 241 Diversion, 225, 235, 270–273 Dividend(s), 62, 112, 115, 122, 123, 134, 146 Dock, 88, 99 Document(s)(ation), 108, 116, 145, 153 Dollar(s), 36–39, 41, 44, 46, 48, 60, 61, 74, 75, 77, 93, 97, 116, 117, 174, 235, 287 Domestic, 2, 8–11, 77, 144, 146, 151, 152, 157, 167–169, 184, 230–232, 234, 236, 239 Domicile, 175, 178, 179, 185, 188, 205, 206 Dominance, 143, 251, 256, 262 Dominant, 36, 155, 236, 239, 256 Drewry, 3, 80, 82, 104 Drilling, 119, 132, 217, 218 Driver(s), 26, 36, 68, 75, 249 Dry, 2, 3, 19–24, 26–28, 30, 32–37, 39, 41, 43, 44, 46, 48, 51, 52, 55, 58, 60–62, 72, 80, 97, 99, 100, 105, 151, 175, 178, 182, 187, 189, 207 Dry bulk, 2, 22, 26, 36, 37, 80, 97, 99, 100, 175, 189 Dsr senator, 222, 277 Dumm(y)(ies), 58, 59, 83, 96, 98, 99 Dynamic, 3, 29, 45, 71, 72, 145, 279 Dynamic(s), 3, 20, 21, 29, 44, 45, 62, 71, 72, 105, 145, 245, 248, 259, 274, 279, 284, 287, 289 Earnings/Price (E/P) ratio, 114, 134, 138, 140 Earning(s), 22, 23, 25, 37, 62, 71, 72, 74, 75, 100, 112, 114, 117, 123, 127, 130, 134, 136, 146, 256, 263
300 Econometric(s), 4, 16, 19, 20, 21, 23, 25–30, 33, 61–63, 77, 79, 103, 104, 134, 159, 207 Economic(s), 1–4, 6–9, 11–17, 21, 51, 60, 61–63, 66, 69, 70, 72, 77, 79, 83, 103–105, 107, 109, 111–113, 115, 116, 120, 122–124, 128–130, 134–136, 144, 145, 154–156, 163, 170, 171, 205, 206, 208, 223, 225–227, 231, 237, 239, 240, 241, 243–245, 248, 249, 257, 259, 262, 268, 285, 288 Economies, 67, 174, 188, 223–225, 247 Economies of scale, 67, 188, 223, 224, 247 Economist, 12, 151 Economist(s), 4, 12, 35, 37, 44, 46, 47, 60, 135, 151, 156, 160, 161, 169, 207, 235 Economy, 8, 10–12, 36, 78, 107, 110–112, 115, 124, 131, 221, 257 Effectiveness, 249, 252 Efficienc(y)ies, 13, 25, 26, 30, 36, 72, 79, 104, 236, 237, 241, 253, 256, 260, 261, 269, 271, 272, 275, 280, 284 Efficient, 22, 25, 30, 34, 72, 76, 108, 110, 124, 127, 128, 134, 153, 163, 166, 279 Elastic(ity)(ities), 20, 21, 24, 25, 28, 37, 69, 76, 83 Electricity, 119, 126, 130, 133 Electronics, 118, 267 Empirical, 3, 4, 7, 10, 11, 16, 22, 28, 33, 34, 37, 41, 51, 87, 107, 114, 125, 130, 134, 135, 161, 205, 235, 263, 287 Employed, 5, 21, 25, 27, 30, 34, 38, 39, 41, 42, 44, 51, 60, 80, 115, 116, 125, 126, 130–133, 153, 159 Employees, 70, 164 Employing, 9, 60, 72, 133, 138–140, 164 Employment, 8, 12, 22, 23, 25, 38, 147, 153, 154, 156, 164, 174, 206 Endogenous, 29, 67, 70 Enforcement, 12, 166 Enterprises, 99, 156, 231, 233 Entrepreneur(ial), 249, 251, 256 Environment(s)(al)(ally), 8, 10, 12, 14, 97, 127, 145, 147–149, 153, 155, 156, 159, 160, 162, 163, 165–169, 171, 173, 184, 185, 206, 231, 232, 263, 279
SUBJECT INDEX Equation(s), 4, 20, 23, 29–32, 38, 42, 43, 51–53, 58, 60, 67, 71, 72, 87, 88, 113, 115, 122, 127, 133, 161, 177–179, 183, 185, 188, 189, 193, 201, 202, 239 Equilibrium, 22, 67, 68, 73, 86, 87, 136, 238, 256 Equipment, 82, 97, 108, 118, 232 Equity, 9, 16, 114, 122, 123, 127, 128, 130, 131, 134, 163 Error(s), 16, 26, 28, 29, 33, 34, 39, 42, 46, 55, 59, 62, 66, 70, 82, 86–88, 103, 104, 113 Error Correction Model (ECM), 46, 87, 103 Estimate(d)(s), 4, 7, 11, 22, 23–29, 32–34, 38, 39, 42–44, 47, 53, 55, 58, 59, 60, 62, 66, 67, 72, 87, 100, 103, 104, 114, 118, 123–125, 127, 128, 131, 151, 156, 164, 182–189, 193, 201–204, 207, 224, 251, 264, 265, 268, 276 Estimating, 4, 41, 60, 87, 133, 185 Estimation(s), 5, 16, 19, 21, 22, 27–30, 34, 37, 38, 42, 53, 55, 62, 70, 72, 81, 83, 87, 105, 121, 124, 126, 130, 134, 179, 183, 185, 189, 193, 202, 203 Estimator(s), 16, 88 Europe(an), 8, 9, 17, 33, 81, 82, 104, 115, 136, 143, 149, 151–154, 156, 157, 160, 162, 164–167, 169, 170, 223, 224, 226, 236, 240, 242, 244, 245, 266, 269, 270, 273, 277 European Commission, 152–154, 160, 169, 170, 223, 244 European Community, 151, 153, 154, 155, 157, 162, 166, 170 European Union (EU), 33, 63, 82, 104, 154, 155, 157, 162, 167, 224, 242, 244 Evergreen, 236, 241, 270, 277 Eviews, 87, 104 Exchange(s), 5, 6, 15, 21, 25, 36, 66, 73, 74, 75, 93, 96, 97, 104, 107, 110, 115, 121, 122, 124, 125, 131, 132, 233, 263, 269, 272, 273 Exchange rate(s), 5, 6, 15, 36, 66, 73–75, 93, 96, 104 Exogenous, 4, 20, 23, 24, 26, 27, 29, 44, 46, 67, 69, 84
Subject Index Expectation(s), 14, 19, 25–27, 30, 31, 33, 35, 36, 55, 65–68, 71, 73–75, 77, 79, 82, 115, 116, 133, 155, 157, 205, 236, 259 Expertise, 6, 13, 143, 147, 153, 165, 166 Explanatory, 11, 38, 59, 98, 115, 116, 123, 126, 131, 179, 183, 185, 187, 201, 202, 248 Exploitation, 251, 257, 263 Export(s), 33, 97, 147, 206, 231–233, 235, 236, 242, 259, 263–266, 276, 285, 287, 288 Exposition, 4, 5, 27 Factor(s), 5, 7, 10, 11, 16, 23, 35, 44–46, 65, 66, 69, 70, 72, 80, 109–118, 125, 127–131, 133–136, 138–140, 153, 159, 161, 163, 169, 184, 187, 188, 205, 225, 227, 232, 236, 238, 266–268, 270, 275, 277 Fairplay, 121, 157, 170, 176, 207 Far East, 236, 269, 270 Fearnleys, 3, 79–81, 104 Federal, 146, 149 Fee(s), 25, 174, 288 Feeder(s)(ed)(ing), 2, 224, 225, 227, 241, 245, 248, 253, 264, 267–270, 273–277, 279, 281, 283, 284, 287, 289 Ferr(y)ies, 2, 108, 109, 132, 184, 215 Filter, 86, 88, 124, 128 Finance, 6, 7, 16, 17, 62, 75, 78, 104, 107, 109, 110, 113, 125, 134–136, 145, 148, 149, 169, 170, 232, 285 Financial, 3, 6, 16, 30, 66, 79, 83, 107, 112, 113, 120, 121, 134–136, 144, 147, 148, 185, 227, 259 Financing, 25, 82, 109, 134, 159, 166 Findings, 3, 5, 11, 25, 35, 99, 100, 111, 113, 125, 126, 128, 129, 145, 146, 159, 161 Finland, 156, 157, 209 Firm(s), 14, 113, 146, 158, 159, 231, 234, 247–250, 252–254, 256–262, 267, 285–289 First differences, 4, 28, 29, 38, 39, 41, 45, 60, 72, 85
301 Fiscal, 5, 8–10, 16, 143, 144, 146–166, 168–170 Fish(ing), 176, 178, 182, 184 Five forces, 14, 249, 254 Fixture(s), 25, 30, 32 Flag(s), 8–11, 17, 144–149, 151–155, 157–162, 166, 168–170, 173–179, 182–185, 187–189, 193, 201, 203–208, 213, 259 Flag(s) of Convenience (FoCs), 9, 144, 146, 153, 154, 157, 159–161, 166, 255 Flag states, 174, 207 Flagg(ed)(ing), 9–11, 144, 147, 151, 155, 159–161, 166, 175, 176, 179, 184, 185, 187, 189, 193, 201–203, 205–207 Fleet, 4, 8–11, 16, 20–22, 24, 27, 34, 46–48, 51–53, 67, 68, 71, 73, 80, 81, 96, 99, 104, 108, 143–145, 147, 154–156, 161, 174–176, 182, 184, 188, 193, 201, 202, 206–208, 222, 269 Flows, 37, 72, 115, 128, 232, 239, 265, 276, 287 Fluctuations, 6, 16, 62, 66, 73, 77, 80, 93, 96, 97, 100, 104, 105, 116, 131 Forecast(s), 36, 44, 46, 59, 60, 75, 100, 103, 157, 185, 237, 239 Forecasting, 4, 60, 66, 103 Foreign, 15, 119, 145–147, 151–153, 162, 166, 175–177, 179, 183–185, 187–189, 193, 201–203, 205–207, 230–235, 263, 285 Foreign Direct Investment (FDI), 231, 235, 236, 239, 242, 244 Forward, 16, 20, 44, 48, 68, 69, 72, 75, 97, 99, 152, 158 Framework(s), 14, 15, 36, 69, 73, 86, 153, 159, 167, 174, 248–251, 254–259, 261, 273, 277, 286, 287 France, 117, 157, 165, 170, 209 Freight, 2, 4, 6, 8, 12, 16, 19–21, 23, 24, 26–28, 30, 32, 34–37, 48, 61–63, 65, 66, 67, 70, 71, 76, 83, 93, 96–98, 100, 104, 105, 119, 161, 166, 167, 175, 206, 218, 230, 231, 233, 234, 237,
302 238, 240, 243, 250, 252, 255, 258–260, 263, 281, 284, 288 Freight forwarders, 281, 284 Freight forwarding, 166, 231, 234 Freight rate(s), 16, 20, 21, 23, 24, 26–28, 32, 34–37, 48, 62, 63, 65–67, 70, 76, 93, 96–98, 100, 104, 105, 167, 175, 206, 240, 255, 284, 288 French, 113, 114, 129, 135, 142, 157, 165, 203, 212 Frequency, 3, 5, 28, 35, 37, 227, 238, 266–269, 284 Fuel(s), 2, 4, 21, 23, 38, 112 Function(s), 2, 14, 21, 26, 27, 34, 46, 63, 69, 70, 73, 74, 84, 110, 111, 113, 124, 129, 177–179, 183, 248, 254, 260, 261, 269 Functionality, 251, 260, 286, 289 Funds, 7, 109, 110, 122, 124 Gain(s), 32, 127, 135, 145, 154, 159, 169, 174, 225, 235, 242, 249 Gas, 119, 126, 130 GDP, 11, 71, 175, 178, 187–189, 204, 205 Geared, 79, 128 Gearing, 7, 110 Generalised autocorrelated conditional heteroscedasticity (garch), 33–35, 37, 58 Geographic(al)(ally), 225, 227, 234, 236, 238–240 German(y), 146, 155, 201, 202, 204, 208 Germanischer Lloyd, 178, 179, 187, 204 Globalisation, 227, 269 Gold, 117, 118 Goods, 5, 14, 76, 112, 124, 232, 236, 285 Government(s), 10, 16, 78, 143–149, 155–160, 163, 168, 173, 185, 233, 237–239, 242, 263, 264, 275, 280, 281 Grain, 33, 63 Grand Alliance, 270, 273, 277 Greece, 154, 176, 208, 230 Greek, 99, 176 Gross (Registered) Tons (GT), 144, 175–179, 184, 189, 208–213 Growth, 6, 46, 55, 58, 65, 79, 112, 114, 116, 117, 119, 131, 141, 145, 155,
SUBJECT INDEX 156, 208, 224, 237, 240, 253, 275, 276 Guidelines, 153, 154, 160, 162 Handling, 13, 108, 109, 225–228, 231, 237, 240, 241, 264, 272, 275, 276, 279–281, 283, 285 Handy(max)(sized), 6, 37, 39, 43, 48, 52, 79, 81, 83, 85, 88, 93, 96–98, 103 Harmonize, 157, 159, 162 Heteroscedasticity, 19, 33, 55, 58, 72, 88 Hinterland(s), 227, 238, 240, 241 Hire, 24, 25, 27, 30, 38, 60, 75, 134 Honduras, 177, 179, 202–206, 210 Hong Kong, 15, 62, 152, 164, 193, 208, 232, 234, 237–239, 242–244, 248, 263, 266, 267, 275–277, 279–282, 284, 285, 287, 288 Hub, 225, 227, 236–240, 245, 248, 253, 263, 269, 270, 274, 287, 289 Hutchinson Port Holdings (HPH), 228, 238, 239, 241 Hypothesis, 22, 24–26, 33, 35, 36, 39, 41, 42, 52, 53, 60, 84, 87, 116, 134 Hypothesised, 4, 5, 7, 53 Hyundai, 270, 277 Import(s), 2, 97, 147, 206, 231–236, 242, 259, 276 Incentive(s), 158, 159, 185, 227, 269, 274 Income, 75, 112, 146, 148, 152, 153, 155, 159, 160, 164, 173–175 Incorporation, 148, 149, 168 Independent, 6, 11, 24, 34, 38, 87, 187, 228, 238, 251, 256 Index, 7, 24, 35, 37, 48, 61, 109, 111, 116, 118 India, 176, 184, 209, 228 Indices, 35, 61, 110, 118, 120, 122, 123, 134 Indonesia(n), 184, 206, 209, 242, 266 Industrial, 5, 7, 12–14, 34, 46, 52, 55, 66, 68, 77, 108, 110, 111, 115–120, 128–131, 133, 134, 162, 224–228, 231, 233, 239, 240, 249, 275, 288 Industr(y)ies, 1, 2, 4, 6–10, 12–17, 37, 60, 61, 65, 66, 68–70, 76, 77, 79, 98,
Subject Index 107–113, 116–135, 137–142, 144, 145, 147–149, 151, 155, 156–160, 162, 163, 168, 221–223, 228, 230–232, 237, 240, 242, 249, 257, 259, 260, 262, 268, 275, 285, 286 Inefficien(t)cy, 72, 136, 232, 272 Inelastic, 20, 21, 23 Inference(s), 5, 86–88, 121, 122, 130, 132 Inflation, 61, 62, 110, 111, 115–117, 128, 131, 134 Information, 5, 31, 38, 48, 60, 81, 82, 121, 122, 132, 149, 175, 176, 179, 193, 207, 213, 233, 247, 261 Infrastructure, 108, 109, 147, 157, 224, 226, 227, 239, 240 Inland, 2, 13, 224, 231, 236, 239–241 Innovation(s), 3, 35, 70, 128, 257, 262, 279 Input, 4, 23, 161 Input(s), 4, 23, 33, 69, 161, 163, 258, 259 Insignificant, 26, 70, 84, 97, 98, 99, 131, 138, 139, 140, 201 Inspections, 78, 161, 185, 187 Institution(s), 117, 166, 242 Instrument, 9, 25, 241 Insurance, 25, 159, 161, 166, 187 Integrated, 26, 28, 29, 31, 35, 61, 70, 71, 85, 131, 156, 247, 261, 267, 279, 281, 284 Integration, 27, 28, 35, 45, 46, 85, 131, 223, 228, 240, 241, 247, 253, 283–285 Interdependence, 4, 13, 62, 104, 256 Interdependent, 5, 221, 223, 226, 251 Interest rate(s), 5, 23, 25, 78, 116 Interests, 9, 134, 144–147, 160, 163, 164, 206, 232 International, 2, 6, 8–13, 16, 17, 62, 63, 70, 79, 104, 109, 110, 117, 118, 120, 122, 131, 132, 135, 136, 143–155, 157, 159, 160, 162, 164, 165, 167–170, 173, 174, 178, 184, 185, 205–209, 221, 224, 227, 230–233, 236, 238–240, 242–245, 269, 275, 276, 284, 288 International Association of Classification Societies (IACS), 178, 179, 185, 193, 201–203, 205–207
303 International Maritime Cluster (IMC), 10, 144, 150 International Maritime Organisation (IMO), 11, 173–175, 178, 179, 182, 187–189, 204–207 International Shipping Corporation (ISC), 9, 10, 147, 148, 150, 151, 157, 162–165 Inventory, 268, 269, 274 Invest(ed)(ing), 78, 100, 109, 131, 154, 163, 226, 237, 238 Investigate, 79, 98, 193 Investigation(s), 6, 7, 11, 66, 80, 97, 103, 110, 135, 166, 223 Investment, 16, 17, 65, 66, 68, 71, 74, 79, 80, 97, 99, 100, 108–112, 116–118, 120, 124, 126, 129–131, 133–136, 146, 158, 160, 170, 223, 226, 227, 231–235, 238, 239, 241, 244, 280, 283 Investor(s), 7, 65, 66, 69, 74, 75, 78, 107, 109–111, 115–119, 124–126, 131, 133, 156, 160, 163, 233, 235 Iran, 58, 176, 209 Ireland, 156, 210, 288 Ital(y)ian, 157, 178, 179, 208, 228 Japan(ese), 6, 36, 70, 74, 77, 79, 81–83, 114, 115, 117, 118, 134–136, 157, 159, 146, 176, 189, 193, 203, 208, 228, 230, 266, 267 Jobs, 69, 151, 152, 163, 164, 168 Johor, 264–270 Joint venture, 233, 234, 239 Joint venture(s), 233, 234, 238, 239, 253, 275 Jurisdiction(s), 152, 160 Just-in-time, 267, 268 Kaohsiung, 239, 276 Knowledge, 8, 19, 29, 37, 132, 143, 147, 163, 166, 262 Korea(n), 6, 36, 77, 79, 82, 97, 104, 159, 170, 178, 179, 187, 193, 201, 203, 209, 210, 228, 266, 267 Kuching, 264, 267 Kwai chung, 276, 279, 280, 281, 283, 284, 288
304 Labour, 2, 9, 11, 12, 24, 69, 70, 77, 112, 152, 155, 156, 158, 161, 167, 168, 173, 184, 188, 205, 231, 235, 238 Lag(s), 20, 23, 29, 39, 42, 67, 84, 85, 103 Lagged, 4, 25, 34, 38, 39, 41, 46, 59, 67, 84, 86, 87 Land, 223, 238, 240, 241, 280 Landside, 247, 252, 254, 276, 284, 285 Latin America(n), 175–177, 179, 193, 201, 202, 205, 213, 244 Law(s), 145, 146, 148, 167, 223, 242, 254 Lead, 11, 75, 84, 97, 118, 144, 160, 167, 223, 235, 238, 242, 248 Leadership, 238, 250, 251 Legal, 254, 258, 259 Legislation, 145, 148, 151, 156, 157, 162, 164, 165 Leisure, 108, 109, 118 Leverage, 14, 16, 114, 128, 130, 287 Liabilities, 153, 158, 162, 163 Liberalizing(ed)(ing), 152, 153, 184 Liberia(n), 159, 176, 208 Libor, 35, 74, 75, 79 Libya, 83, 210 Likelihood, 58, 174, 175, 183–185, 188, 193, 203, 205–207 Limitations, 33, 120, 166, 276, 277 Line(s), 16, 69, 72, 83, 97, 104, 114, 117, 126, 128–130, 151, 163, 213, 222–226, 238, 240–242, 245, 247, 248, 250–254, 258–263, 267–277, 284–287 Linear, 20, 24, 29, 34, 36, 38, 39, 45, 52, 67, 83, 84, 87, 113, 133 Liner, 2, 12–15, 17, 108, 147, 221, 223–226, 228, 231, 236, 240, 242–244, 247, 248, 251, 257, 269, 275, 285, 288 Linkage, 168, 173, 270 Linkage(s), 168, 173, 185, 270, 277, 280, 284 Liquid, 78, 175, 182, 189, 203 Liquidity, 30, 120 Listed, 6, 82, 107, 109, 110, 120–122, 124, 125, 127, 132 Literature, 3, 4, 7, 19, 33, 35–37, 103, 112–114, 125, 130, 157–160, 173, 250
SUBJECT INDEX Lloyds, 5, 17, 35, 44, 47, 81, 121, 170, 176, 178, 179, 203, 204, 207 Lloyds Register, 5, 81, 176, 203, 204, 207 Lng, 74, 214 Load centre, 224, 225, 227, 236, 240 Loading(s), 75, 96, 108, 284, 285 Location, 164, 227, 237, 238, 239, 260, 277, 284 Location(s), 164, 227, 236–239, 241, 247, 260, 265, 275–277, 279, 280, 283, 284 Logarithms(ic), 24, 27, 36, 42, 83, 85, 122, 178, 187 Logistical, 12, 109, 224, 227, 241 Logistics, 12, 13, 15, 16, 108, 230, 232, 233, 236, 238–241, 247, 252, 254, 259, 261, 262, 275, 281, 284 Loop(s), 270, 273, 287 Loss, 8, 17, 32, 154, 168, 207, 263 LRfairplay, 17, 175, 213, 219 Macroeconom(y)(ic), 7, 13, 16, 109–111, 113–118, 125, 128–131, 136, 139, 140 Maersk, 170, 222, 228, 236, 239, 241, 277 Mainland, 6, 232, 237, 238, 242, 285 Mainline, 15, 224, 225, 227, 236, 239, 241, 269, 270, 277 Maintenance, 145, 159, 166 Majority, 110, 117, 124, 125, 130, 148, 151, 161, 233, 234, 270 Malaysia, 15, 209, 241, 263, 264, 287 Malaysia(n), 15, 209, 241, 242, 263–269, 273–275, 287 Management, 9, 12, 16, 17, 143, 147–149, 154, 164–166, 168, 248, 249, 253, 259, 262, 280, 288, 289 Manager, 112, 129–131, 133 Manning, 155, 159 Manufacturing, 14, 152, 168, 231–233, 235, 237, 239, 242, 262, 267, 280, 285 Margin(s), 82, 224, 256, 285 Marginal, 10, 23, 25, 46, 183–187, 189, 193, 201, 202, 204 Marine, 1, 144, 147, 157, 166, 167, 169, 170, 171, 207, 208
Subject Index Maritime, 1, 8–13, 15–17, 61–63, 104, 105, 120, 121, 123, 132, 135, 143, 144, 147, 149–153, 155–157, 160, 162, 163, 165–171, 173, 178, 179, 187, 206–208, 223, 224, 230, 231, 242–245, 277, 285, 287–289 Market(s), 2–9, 13–17, 19–30, 33–38, 44, 46, 47, 55, 58, 61–63, 65–75, 77–80, 83, 87, 88, 97–99, 103–105, 107–118, 120, 122, 124–132, 134–140, 142, 145, 147, 153, 156, 161, 167, 168, 174, 184, 207, 221, 223, 226–228, 230–233, 235, 236, 239–241, 247–263, 265–267, 269, 270, 273–275, 280, 283, 285–287 Market capture, 68, 70, 79, 87, 239, 248, 253, 254, 257–261, 263, 266, 274, 275, 277, 285–287 Market entry, 66, 159, 223, 231, 233, 235, 241, 256, 273 Market place, 241, 247, 248, 253, 254, 275, 283, 285 Market share, 14, 47, 68, 223, 241, 248, 267, 286 Materials, 70, 93, 97, 108, 234 Matri(x)(ces), 29, 31, 32, 64 Maximise(d)(s), 21, 80, 88, 97, 99, 133, 263, 267, 274 Maximum, 34, 37, 51, 53 Measure(d)(s), 12, 22, 24, 27, 33, 34, 38, 39, 42, 44, 46, 48, 69, 77, 82, 97, 113, 116, 118, 123, 126, 131, 132, 134, 143, 144, 147, 153, 155, 156, 157, 163, 164, 167, 178, 179, 189, 232, 268 Measurement(s), 38, 179 Membership, 167, 225, 231–233, 235 Memorandum, 148, 165, 166 Mergers, 122, 222, 223, 242 Methodolog(y)(ies), 3, 4, 29, 32, 36, 37, 51, 72, 73, 84, 113, 127, 146, 159 Methods, 1, 5, 28, 103, 128, 134, 289 Microeconomic, 7, 110, 111, 113, 114, 115–118, 125, 127–130, 133, 135, 138, 254, 275 Midstream, 276, 280, 283–285 Minimise, 39, 263, 268, 285 Minority, 233, 234
305 MISC, 267, 277 Misspecification, 67, 70 Mobility, 6, 8, 165 Model(s), 3–7, 10, 11, 19–39, 41–47, 51, 53, 55, 58–64, 66–72, 74, 75, 77–79, 83–88, 96, 97, 100, 103–105, 109, 110, 114–118, 123, 124, 127–135, 144, 148, 157–159, 168, 178, 182, 183, 248, 256–261, 286 Modell(ed)(ing), 2–4, 16, 19–22, 24, 25, 27–30, 33–37, 44, 46, 48, 53, 60, 62, 63, 68, 72, 104, 105 Modes, 2, 109, 160, 167, 230, 237, 238, 240, 243, 252 Money, 2, 74, 100, 105, 109, 287 Monopol(y)ist(ic), 241, 245, 254, 256, 263, 267, 287 Morgan Stanley Capital, 120, 122, 131 Movement(s), 2, 3, 7, 68, 109, 112, 113, 126, 149, 167, 232, 235, 239, 255, 259, 263–266, 275, 279, 280, 282–284 MSCI, 120, 122, 123, 131, 132 Multidimensional, 110, 111, 127 Multifactor, 7, 114, 115, 116, 117, 118, 122, 123, 124, 129, 133, 134 Multinational(s), 162, 267 Multiplier, 112, 168 Nation(s), 8–12, 17, 65, 77, 143, 145, 146, 158, 162, 232, 235, 241 National(s), 2, 5, 8–11, 66, 68, 69, 144, 145, 146, 153, 155, 158–162, 164, 165, 167–169, 173–178, 184, 185, 187–189, 204, 205, 207, 213, 230, 232, 263 Nedlloyd, 222, 236, 239, 277 Netherlands, 16, 62, 105, 154, 155, 209 Network(ed)(s), 13, 15, 224, 236, 247, 248, 251–254, 260, 262–264, 274–277, 284–287 New territories, 277, 280 New Zealand, 157, 211 Newbuilding, 4–6, 22, 23, 27, 28, 66–78, 81, 82, 85, 88, 93, 96–100, 103, 104 Niche, 15, 223, 225, 227, 237, 247 Nippon Kaiji Kyokai, 178, 179, 187, 193 Non-stationary, 39, 45, 84
306 Normality, 55, 58 North America(n), 165, 266, 269, 270, 273 Norway, 63, 105, 149, 155, 157, 209, 230 Norwegian, 16, 21, 63, 74, 105, 155, 208 Null, 33, 39, 41, 51, 52, 53, 60, 61, 84 Objective(s), 10, 15, 103, 131, 147, 149, 151, 153, 154, 156, 158, 159, 162–166, 168, 221, 259 Observations, 27, 30, 37, 39, 47, 60, 80, 83, 100, 121, 179, 185 Ocean(s), 144, 147, 173, 230, 276 OECD, 82, 144, 160, 173, 174, 207 Offshore, 8, 109, 119, 132, 150, 164, 168, 218, 237 Oil, 7, 22, 24, 34, 37, 48, 53, 55, 58, 61, 62, 79, 80, 81, 83, 98, 100, 104, 105, 108, 110, 111, 116, 117, 128, 129, 130, 134, 141, 159, 178, 182, 234, 235, 242 Ols, 24, 28, 29, 33, 55, 67, 87, 88, 127, 131, 182 On board, 8, 9, 109, 153 OOCL, 236, 239, 277 Open registr(y)ies, 8, 12, 174, 176, 179, 182, 188, 202, 204, 206 Operating, 13, 21–23, 25, 27, 32, 33, 61, 71, 99, 108, 122, 124, 144, 146, 147, 149, 150, 160, 162, 164, 167, 169, 184, 231, 234, 250, 252, 267, 269, 272, 273, 280, 282, 286 Operational(ised), 25, 80, 221, 224, 249, 252, 253, 257, 261, 271, 282, 283 Operations, 61, 69, 108, 153, 164, 167, 223–225, 227, 231, 233–235, 239, 240, 250, 269, 270, 279, 284, 286, 287 Operator(s), 10–12, 16, 20, 29, 108, 152, 153, 158, 159, 161, 168, 173, 175–179, 182–185, 187–189, 193, 201, 202, 204–207, 222, 223, 225–228, 231, 236, 238, 240, 241, 243, 270, 281, 284 Optimal, 34, 39, 71, 131, 243, 272, 273 Optimis(e)(ing)(ation), 21, 108, 167 Optimum, 21, 157, 167 Options, 9, 81, 82, 149, 151, 152, 154, 160, 161, 163, 254, 263, 269, 273, 274, 277, 279, 283
SUBJECT INDEX Order, 1, 2, 12, 20, 23, 28–32, 38, 39, 41, 42, 45, 47, 58, 64, 67, 69, 70, 74, 76, 77, 80, 82–88, 97, 98, 100, 103, 107, 120–122, 131, 133, 147–149, 155, 157, 162, 167, 173, 174, 176, 187, 205, 222, 235, 238, 240, 244, 249, 252, 253, 261, 267, 277, 284, 285 Orderbook, 68, 70, 73, 74, 76, 80, 81, 93, 96, 98–100, 104 Ordering, 67, 74, 75, 97 Ore, 178, 182, 242 Organization(s)(al), 6, 145, 173, 202, 205, 206, 239, 250, 251, 258, 260–262, 288 Outcome(s), 164, 168, 232, 237, 261, 286 Output, 4, 12, 24, 76, 82, 87, 88, 99, 248, 286 Overseas, 8, 9, 230, 231, 232, 233, 234, 236, 239 Oversuppl(y)ied, 68, 100, 240 Owner(s), 21, 25, 60, 66, 75, 77, 97, 99, 100, 103, 108, 134, 143, 144, 146, 150, 151, 155, 158, 160–162, 169, 173, 174 Ownership, 7, 109, 120, 143, 147, 149, 155, 162, 166, 238, 241, 256, 259 P&O, 222, 236, 239, 277 Package, 87, 153, 155, 156, 159, 163, 169, 284 Panama(nian), 62, 63, 174, 176, 188, 179, 193, 201–206, 208 Panamax(es), 22, 37, 39, 44, 48, 79, 81, 88, 93, 96–100, 103, 104, 221, 225, 240 Paradigm, 248, 254, 259, 287, 289 Parameter(s), 20, 26, 28, 29, 42, 44, 69, 83, 87, 127, 183, 187–189, 193, 204 Partner(ship)(s), 144, 162, 168, 223, 239, 275 Passenger(s), 2, 108, 109, 176, 178, 182, 184, 185, 187, 189, 193, 201, 203, 205, 215 Pathways, 252, 254, 259, 262, 282 Payment(s), 5, 16, 21, 82, 154, 165, 170, 269, 274 PCWAs (Public Cargo Working Areas), 276, 280, 281, 283–285
Subject Index Pearl River Delta (PRD), 237–239, 248, 275–277, 279, 281–285 Penang, 264–270 Per capita, 175, 178, 187–189, 204, 205 Performance, 4, 6, 7, 36, 59, 60, 66, 100, 103, 107, 109, 110, 117, 118, 128, 133, 134, 174, 249, 250, 254 Petroleum, 100, 119, 126, 128, 133, 138, 235 Philippines, 206, 209 Piracy, 165, 166 Plates, 74, 77, 79 Polic(y)ies, 5, 8–10, 12, 16, 66, 68, 70, 77, 103, 104, 108, 143–147, 149, 151, 153, 155, 156, 158–169, 207, 228, 236, 239, 242, 254, 263, 269, 272, 275, 287 Political(ly), 2, 8, 9, 11, 13, 83, 96, 123, 163, 164, 232, 284 Population, 178, 188 Port(s), 1, 12, 13, 15, 17, 21, 75, 79, 108, 122, 151, 153, 161, 165, 166, 187, 201, 202, 205, 221, 223–228, 232, 236–245, 247, 248, 251, 252, 254, 263, 264, 266–277, 279, 280, 283–289 Portfolio(s), 71, 109, 110, 113, 116, 117, 120, 124–126, 130, 131, 133 Port Klang, 248, 263, 264, 266–273, 287 Portugal, 211, 228 Positioning, 14, 143, 249, 250, 252, 254, 261 Power, 11, 14, 15, 38, 61, 98, 115, 116, 126, 131, 223, 226, 228, 238, 241, 251, 255–257, 259, 260, 262, 263, 267, 270, 273, 274, 282, 284, 286 Practice(s), 146, 147, 223, 232 Practitioners, 110, 111, 113, 118, 120, 129, 133 Predict(ed)(ive), 26, 32, 154, 224, 249 Predictions, 10, 26, 44, 182, 240 Premi(um)(s)(a), 27, 30, 58, 66, 68, 69, 115, 134, 135, 161, 187, 268 Price(s), 3–7, 19–28, 32, 34, 35, 46, 48, 51–53, 55, 58, 61–63, 65–82, 84, 88, 93, 96–100, 103–105, 108, 110–114, 116, 117, 120, 122, 124, 127–130,
307 134–136, 141, 208, 235, 238, 247, 252–255, 260, 261, 263, 273, 280 Pricing, 62, 67, 68, 69, 70, 98, 104, 107, 110, 115, 116, 125, 133–136, 207, 245, 272, 283 Priority, 6, 173, 205, 238, 269, 270 Probabilit(y)ies, 11, 175, 182–189, 193, 201–205 Probit, 11, 182–185, 193, 201–204 Procedure, 36, 39, 51, 55, 112 Process, 26, 60, 109–112, 161, 166, 287 Product(s), 58, 77, 88, 99, 100, 108, 109, 112, 119, 231, 233–235, 242, 252, 254, 262, 280, 285 Production, 7, 27, 34, 46, 52, 55, 66, 74, 83, 110, 111, 115, 116, 128–131, 134, 234, 247, 250, 256, 262, 284, 285 Productivit(y)ies, 70, 225, 226, 238, 241, 246, 253 Profile(s), 116, 122, 124, 131, 133 Profit(s), 21–23, 25, 30, 32, 33, 71, 82, 107, 115, 153, 154, 158, 160, 163, 252, 256, 257, 263, 267, 285 Profitab(le)(ly)ility, 22, 27, 65, 66, 74, 77, 98, 145, 158, 226, 249, 250, 253 Programmes, 39, 233 Project(s), 75, 78, 133, 227 Protection(ist), 145, 151, 153, 156, 162, 165, 167 Provider(s), 239, 248, 250–255, 259–262, 286, 287 PSA, 227, 228, 238, 239, 241 Public Cargo Working Areas (PCWAs), 276, 280, 281, 283–285 Purchas(e)ing, 2, 61, 150, 222, 223, 226, 238 Qualitative, 112, 161 Qualitative, 16, 112, 169 Quality, 14, 79, 81, 153, 161, 165, 166, 174, 175, 185, 204, 223, 257, 260 Quantif(y)iable, 116, 226, 160, 175, 253 Quantit(y)ative(ly), 17, 24, 27, 63, 69, 70, 104, 105, 112 Questionnaire, 121, 132 Quota(s), 231–233, 235, 238, 242
308 Rail, 119, 126, 138, 224, 234, 239, 285 Rank(ed), 29, 208–213, 230, 248, 275 Rate(s), 4, 5, 7, 16, 20–28, 30–39, 41–46, 48, 51–53, 55, 58–62, 65–67, 70–79, 93, 96–100, 103–105, 113, 115–117, 123, 128, 131, 133, 152–154, 159, 167, 175, 178, 182, 187–189, 205, 206, 233, 236, 240, 255, 274, 276, 283, 284, 288 Ratification, 175, 205 Ratio(s), 16, 23, 32, 44, 80, 100, 110–112, 114, 117, 127, 128, 130, 134 Rational(ised)(isation), 7, 25, 26, 33, 36, 221–224, 227, 253, 254, 269, 270, 284, 285 Real estate, 118, 119, 126, 128, 130, 138 Redistribution, 255, 287 Reefer, 122, 178, 182, 187, 189, 203, 214, 218 Reference, 3, 4, 113, 154 Refining, 119, 126, 133, 138 Regime(s), 9, 83, 143, 144, 146, 148, 152, 154–159, 162–65, 167–169, 175, 184, 187, 188 Region(s)(al), 2, 69, 155, 156, 171, 177, 224, 228, 236–239, 241, 242, 275 Register(s), 8, 9, 17, 153, 155, 168, 173–175, 178, 179, 187, 201, 203, 205, 207, 234 Register(ed)(ing), 8, 9, 144–146, 148, 153, 156, 166, 167, 174, 176, 193, 208 Registration(s), 8, 9, 10, 16, 143, 146, 148, 161, 169, 173, 174, 207, 208 Registr(y)ies, 8–12, 143, 144, 149, 151, 157–162, 165, 169, 171, 173–179, 182–184, 187–189, 193, 201–213, 259 Regressed, 34, 38, 132 Regression(s), 28, 29, 33, 41, 84–88, 122, 126, 128, 131, 159, 185, 188, 189, 213 Regulated, 112, 185, 231 Regulation(s), 8, 10, 12, 17, 173, 206, 231, 242 Regulator(s)(y), 166, 167, 174, 223, 254 Relation(s), 3, 4, 23–29, 32–34, 37, 41, 43, 45, 47, 51, 55, 61, 82, 87, 127, 134, 152, 156, 163–168, 187, 207, 263, 282
SUBJECT INDEX Relationship(s), 4, 5, 7, 10, 11, 13, 14, 21, 22, 24–26, 29, 30, 33, 35, 37, 42, 44–46, 48, 53, 60–62, 67, 69, 70, 77, 79, 87, 114, 126–129, 131, 134, 138, 139, 140, 150, 184, 185, 206, 207, 226, 239, 251, 254, 256–259, 260, 261, 263, 284–287 Reliability, 82, 145, 255 Relocation, 231, 242 Rent(s), 249, 250, 251, 256, 257, 262, 263, 286 Report, 16, 23, 27, 43, 62, 63, 71, 81, 105, 144–149, 156, 244 Research(er)(s), 1–3, 5–8, 10, 12, 13, 15, 19, 21, 29, 30, 33, 35, 60, 63, 66, 68, 71, 75, 81, 103–105, 110, 111, 113, 115, 117–119, 125, 127, 130, 132, 133, 135, 153, 158–161, 169, 170, 205–207, 213, 215–217, 244, 248, 263, 264, 273, 283–289 Reserves, 152, 159 Residual(s), 28, 55, 58, 84, 113, 127 Resource(s), 14, 61, 78, 110, 118, 226, 249–251, 254, 256–260, 286, 288, 289 Restriction(s), 30–32, 35, 42–44, 64, 83, 127, 167, 231 Restrictive, 148, 223, 232 Restructure(d), 247, 253, 284 Restructuring, 83, 148, 245, 247, 253, 275, 284, 287, 289 Results, 4–7, 10, 11, 22–26, 32, 35, 36, 39, 41–44, 46, 48, 51, 53, 55, 58–61, 70, 72, 75, 80, 83–88, 93, 97, 103, 104, 117, 120, 124, 125, 127, 129, 130, 133, 134, 155, 183, 185, 189, 193, 201–203, 205, 223 Return(s), 7, 16, 33, 71, 74–76, 100, 107, 109–118, 120–123, 125–140, 151, 223, 226 Revenue(s), 2, 12, 23, 74, 97, 99, 100, 122, 149, 161, 255, 257, 260, 267, 269, 270, 285 Risk(s)(y)(iness)(ier), 6, 7, 12, 15–17, 22, 25, 27, 34, 35, 62, 65, 69, 71, 72, 78, 80, 104, 105, 110, 113, 115–118, 120, 122–128, 131–140, 165, 166, 174, 207, 238, 241
Subject Index River, 215, 275, 276, 279–281, 283–285 River Trade Terminal (RTT), 276, 280, 281, 283–285 Road(s), 224, 234, 239, 245 Roll-on/roll-off (ro-ro), 2, 178, 182, 189, 203 Root Mean Squared Error (RMSE), 59, 103 Roots, 26, 35, 249 Round-the-World (RTW), 269, 270 Route(s)(ing), 2, 35, 37, 38, 62, 75, 79, 104, 108, 109, 147, 225, 227, 237, 238, 263, 267, 269, 270, 272, 273, 277, 280, 284 R-squared (r 2 ), 47, 84, 87, 88 Rules, 30, 152, 155, 157, 206, 231 Safety, 8, 10, 11, 12, 17, 146, 152, 153, 156, 162, 165, 166, 168, 173, 174, 175, 184, 185, 187, 189, 204, 205, 206, 207 Sailings, 238, 276, 277 Sale(s), 2, 72, 80, 115, 123, 153, 161, 232, 250, 285 Salvage, 159 Sample, 11, 41, 43, 44, 47, 48, 58, 59, 60, 72, 116, 120–122, 124, 125, 126, 132, 161, 177 Samsung, 240, 244 Sanko, 83, 99 Savings, 9, 161, 174, 270 Scale, 76, 164, 165, 225, 227, 285 Schedule(s)(d), 2, 75, 236, 269–272 Schwartz criterion, 39, 41 Scrap(ped)(ping), 2, 22–24, 32, 66, 72, 159 Sea, 5, 10, 12, 13, 23, 80, 108, 119, 144–147, 150–152, 161, 164, 166, 168, 169, 205, 224 Seafarers, 143, 153, 155, 174, 206 Seafaring, 8, 10, 12, 156, 166 Sealand, 222, 228, 236, 239 Seasonal(ity), 19, 35, 58, 109, 266, 267 Secondhand, 2, 4, 5, 22, 23, 27, 35, 62, 63, 65, 66, 68–80, 83, 88, 97–100, 103–105 Sector(s), 2, 13–15, 22, 26, 27, 35, 37, 47, 53, 55, 59, 60, 62, 74, 83, 88, 99, 103, 105, 107–111, 117, 119–124, 127,
309 132, 133, 151, 154–156, 161, 166–168, 221, 226–228, 230–233, 236, 239–242, 275, 276, 281, 284, 285 Security, 6, 12, 109, 115, 128, 134, 156, 157, 165, 166, 173, 188 Seemingly Unrelated Regression (SUR), 42, 127, 129 Segmentation, 2, 131, 253, 258, 260 Segments, 4, 5, 27, 35, 39, 44, 53, 74, 79, 83, 96, 98, 99, 100, 103, 104, 108, 253, 261, 286 Sell(s)(ing), 66, 77, 252, 256, 262 Seller(s), 248, 250, 252, 254, 259, 260, 284, 285 Seminal, 3, 4, 20, 26, 29, 114, 158, 250 Sensitivit(y)ies, 116, 127, 129, 130, 133, 134, 233, 241 Series, 1, 16, 21, 26, 28–31, 33–39, 41–45, 48, 51, 53, 61, 72, 77, 81–87, 110, 115, 121, 124, 128, 143–145, 245, 288 Service, 1, 13, 21, 46, 108, 109, 119, 121, 122, 134, 162, 178, 182, 206, 219, 223, 226, 227, 231, 232, 234, 241, 244, 247, 248, 250–255, 259–263, 266–270, 272–274, 276, 277, 281, 285–287 Set(s), 8, 10, 11, 29–31, 33, 34, 38, 41, 43, 46, 51, 53, 58, 60, 87, 109, 115, 116, 120, 125, 128–131, 134, 145, 148, 151, 154, 156, 157, 160, 224, 227, 228, 232, 235, 236, 238, 242, 249, 258, 259, 261, 263, 271, 277, 279, 287 Shanghai, 237, 238, 239 Share(s), 5, 7, 47, 71, 109, 110, 114, 115, 120–122, 135, 145, 174, 193, 222, 233, 235, 241, 248, 255, 257, 260, 261, 267, 286 Shekou, 275–277, 279, 283, 285 Shenzhen, 237–239, 242, 275–277, 285 Ship(s), 2, 4–6, 8–13, 16, 17, 19, 21–28, 32–35, 37, 38, 44, 58, 60–62, 65–84, 87, 88, 96–100, 103–105, 108, 109, 134, 143, 144, 146–149, 151, 153–155, 158–162, 164–169, 173–176, 183, 184, 187, 188, 193, 201, 205, 207, 223–227, 238, 240, 248, 252, 269, 285
310 Shipbrok(ers)ing, 79, 81 Shipbuilder(s), 74, 82, 88, 99 Shipbuilding, 2, 5, 6, 22–24, 27, 62, 65–70, 73, 76, 77, 82, 83, 88, 93, 96, 97, 100, 104, 105, 144 Shipment(s), 2, 108, 234, 274, 283, 285 Shipowner(s), 8, 9, 21, 25, 30, 66, 74–76, 78, 80, 84, 98, 99, 108, 134, 146, 157, 158, 160, 161, 174, 207 Shipped, 264, 273, 283, 285 Shipper(s), 70, 108, 134, 145, 223, 248, 252, 259, 263, 266, 268, 269, 272–275, 277, 281–287 Shipping, 1–17, 19–21, 23–30, 35–37, 44, 46, 47, 60–63, 65–67, 69, 70, 72, 74, 79, 84, 86, 87, 99, 100, 104, 105, 107–111, 120–126, 128, 131–135, 143–149, 151–161, 162–171, 173, 174, 178, 179, 184, 187, 193, 201, 203, 207, 208, 221, 223–228, 230–232, 236, 238–245, 247–254, 257, 259, 261–263, 266–271, 273–277, 284–288 Shipyard(s), 5, 27, 66–69, 70, 73, 74, 76–79, 81–83, 88, 93, 97, 99, 100, 104, 179, 185, 240 Shore, 108, 151, 163, 164, 166, 168, 178, 182, 240 Sign(s), 15, 39, 55, 70, 74, 76, 78, 83, 99, 128, 130, 138–140, 187, 189, 193, 201–204 Significance, 5, 7, 35, 51, 53, 110, 116, 127, 183, 186, 193, 202, 230 Significant, 4–8, 10, 11, 15, 22, 34, 35, 43, 44, 55, 66, 67, 70, 76, 77, 79–82, 88, 93, 96–99, 103, 104, 110, 112, 115–117, 127, 129, 131, 132, 138, 144, 145, 155, 162, 185–187, 189, 201–204, 223, 226, 233, 235, 240, 249, 254, 256, 257, 264, 266–269, 273, 275, 276, 283, 284 Simulation, 21, 23, 33, 63, 105 Singapore, 15, 152, 208, 213, 227, 241, 242, 248, 263–271, 273–275, 277, 287 Size(s), 2, 11, 15, 20, 22, 24–26, 32–35, 37, 39, 41, 43, 44, 46–48, 51, 52, 53, 55,
SUBJECT INDEX 58, 60, 62, 67, 71, 72, 74, 76, 79–81, 83, 93, 96–100, 103, 105, 110, 112, 113, 117, 126–128, 132, 135, 137, 154, 161, 175, 177–179, 182, 184, 188, 189, 202, 205, 223–227, 235, 240, 253, 260, 272, 273, 277, 286 Skills, 8, 10, 153, 155, 156, 162 Slot(s), 222, 284 Social(ly), 1, 16, 153, 163, 170, 232 Societ(y)ies, 174, 175, 177, 178, 185 Space, 100, 164, 253, 267 Spain, 156, 165, 210 Spanish, 115, 135 Spatial, 254, 282, 285 Specialisation, 13, 108, 227, 247 Specification(s), 4, 5, 19, 27, 34, 35, 73, 80, 82, 84, 136, 253 Speculation, 66, 98, 104 Speculative, 26, 65, 68, 70, 72, 75 Spot, 4, 16, 21–27, 30–35, 37–39, 41–46, 48, 51–53, 55, 59–62, 75 Spread(s), 4, 25, 26, 30–33, 35, 38–44, 46, 55, 59, 60, 117, 268 Ssy, 79–81, 105 St. Vincent and the Grenadines, 179, 201–204 Standard deviation, 118, 179 Standard Industrial Classification (SIC), 111, 119, 120 Standards, 11, 152, 166, 173, 184, 185, 188, 205, 206, 235, 238 State(s), 11, 12, 25, 37, 46, 75, 110, 111, 132, 151, 153–157, 161–163, 165–170, 173–176, 184, 187, 201, 202, 205, 207, 208, 228, 231, 236, 240 Static, 44, 59, 60, 274 Stationarity, 28, 30, 34, 36–39, 41, 46, 51, 53, 60, 61, 84 Stationary, 28, 29, 31, 36, 38, 39, 41, 42, 44, 45, 46, 51, 53, 55, 61, 72, 83, 84, 85, 87 Statistic(s), 16, 38, 39, 52, 55, 81, 82, 84, 86, 87, 88, 123, 164, 175, 179, 183, 185, 193, 202 Statistical(ly), 3, 28–30, 38, 51, 55, 60, 67, 70, 84, 86, 87, 88, 93, 96–99, 113,
Subject Index 115, 125, 126, 137–140, 179, 187, 189, 201–204, 207 Steel, 24, 67, 70, 73, 77, 79, 112, 242 Stochastic, 35, 70, 72, 87 Stock(s), 7, 16, 17, 34, 71, 107, 109–122, 124–140 Stock exchange(s), 107, 110, 115, 121, 122, 124, 125, 132 Stock market, 7, 113, 117 Strateg(y)(ic)(ies), 13–15, 17, 37, 97, 117, 131, 133, 147, 153, 154, 156, 158, 160, 171, 221–223, 227, 228, 239, 240, 241, 243, 245, 247–249, 251, 252–254, 256–263, 267, 270, 271, 286–289 String(s), 270, 273, 276, 277 Structur(es)(ed)(al)(ing), 3, 4, 6, 13, 15, 19, 20, 24–26, 28, 30, 32, 35, 36, 41, 42, 45, 60, 61, 63, 79, 83, 108–112, 115, 128, 130, 133–135, 141, 145–147, 149, 156, 173, 207, 223, 235, 247, 250, 251, 253, 254, 256, 258–263, 270, 273, 275, 277, 284–287 Stud(y)ies, 2, 3, 6, 11, 13–15, 17, 19, 20, 22, 24, 26, 28–30, 32, 35, 37, 47, 55, 58, 60, 69, 72, 103, 110, 111, 113, 115–118, 120, 122, 124, 125, 127–129, 131, 132, 143–147, 149, 151, 152, 156, 157, 159, 161, 163, 166, 167, 207, 226, 233, 235, 237, 248, 263, 274, 275, 285, 287 Subsid(y)(ies), 23, 66, 68, 70, 76, 77, 78, 147, 155, 156, 160, 233 Substandard, 174, 183, 184, 207 Substitutes, 69, 73, 75, 76 Success, 9, 10, 14, 16, 65, 80, 99, 112, 155, 156, 161, 163, 164, 228, 248–251, 256–258, 261, 262, 286, 287 Suezmax(es), 34, 79, 81, 88, 93, 96–100, 103 Suppl(y)iers, 1, 12, 14, 15, 20–24, 26–28, 61, 66–71, 73, 79, 80, 83, 109, 110, 124, 207, 224, 247, 250, 251, 254–257, 259, 260, 262, 263, 267, 275, 281, 282, 286, 287
311 Supply chain(s), 12, 14, 15, 109, 247, 250, 251, 254–257, 259, 260, 262, 263, 275, 281, 286–288 Survey, 3, 4, 19, 20, 132, 152 Survive, 157, 241, 257 Sustainab(ility)(le)(ly), 250, 256, 257, 286, 289 Sweden, 157, 209 Switching, 22, 27, 161, 256 Switzerland, 117, 176, 210 System(s), 12–15, 20, 29–31, 33, 42, 82, 111, 112, 119, 120, 154, 155, 157, 158, 206, 224, 227, 232, 233, 239, 241, 247–251, 252, 258, 260–265, 275, 283, 286, 287, 289 Systematic, 113, 116, 118, 125, 126, 128, 132, 249 Taiwan, 201, 209, 239, 266, 267 Tanjung Pelepas, 15, 241 Tanker(s), 2, 3, 5, 16, 19, 21–28, 30, 32, 34, 35, 37–39, 41, 43, 44, 46–48, 51, 53, 55, 58–63, 67, 69–72, 74, 75, 79–81, 83, 88, 93, 96–100, 103–105, 108, 122, 132, 151, 161, 165, 166, 214 Tankship, 62, 67, 105 Tariff(s), 150, 152, 231–233, 238, 272, 280 Task force, 144, 146–149, 163, 164, 170, 171 Tax(ation)(able), 9, 10, 75, 143, 144, 146, 148, 149, 151–165, 167, 168, 169 Technical, 110, 260 Technique(s), 16, 19, 28, 33, 35–37, 42, 72, 226 Technolog(y)ical(ly), 66, 69, 70, 107, 108, 144, 163, 225, 280, 283 Term, 10, 13, 20–26, 28–30, 32–34, 39, 42, 45, 46, 51, 53, 66, 70–72, 84, 87, 105, 113, 115, 116, 128, 133, 134, 141, 151, 153, 156, 158, 161, 168, 221, 223, 225, 237, 238, 240, 249, 287 Terminal, 72, 226, 237, 238, 241, 281, 284, 288 Terminal(s), 72, 223, 225–228, 237–239, 241, 276, 279, 281, 283, 284 Terrorism, 165, 166
312 Test(s)(ed)(ing), 4, 5, 7, 16, 19, 25, 28, 30, 32–39, 41–44, 48, 51–53, 55, 58, 60–62, 72, 84–88, 114, 115, 126, 134–136, 148, 249, 263, 287 TEUs, 176, 178, 179, 184, 193, 208–213, 222, 224, 237, 240, 244, 263, 264, 265, 269, 273, 275, 276, 280, 281, 283 Textile, 118, 231 Thailand, 209, 228 Mean(s), 33, 35, 37, 58, 103, 147, 179 Theorem, 45, 67 Theoretical(ly), 26, 27, 32, 66, 103, 113, 134, 223, 243, 244, 250 Theor(y)(ies)(ists), 1, 4, 7, 22, 26, 27, 60, 61, 71, 85, 103, 104, 115, 116, 135, 136, 238, 249, 254, 262, 288 Third party, 228, 235, 248, 250–254, 259, 261, 262, 286, 287 Throughput(s), 226, 228, 236–238, 276, 283 Timecharter, 22, 25, 27, 30–32, 34, 37–39, 46, 48, 53, 60, 73–75, 85, 97–100, 103, 104 Timing, 65, 66, 69, 97, 100, 131 Ton(s), 37–39, 41, 44, 46, 48, 60, 178, 201, 202, 230 Tonnage, 9, 10, 27, 47, 66, 67, 74, 77, 78, 143, 144, 152, 154–157, 160, 162–165, 167–170, 176, 183, 193, 202, 223, 240 Tonne(s), 21, 27, 75, 176, 281 Tourism, 109, 232 Trade(s), 2, 7, 12, 13, 15, 24, 27, 33, 37, 58, 65, 68, 80, 81, 97, 108, 110, 120, 124, 145–147, 149–152, 154, 158, 160, 162, 163, 166–169, 175, 184, 185, 205, 221, 225, 227, 230–233, 235, 236, 239, 240, 242, 250, 253, 266, 273, 275, 277, 279–281, 283 Trading, 22, 27, 32, 35, 69, 72, 97, 99, 117, 144, 145, 147, 157, 160, 162, 163, 164, 167, 168, 185, 231, 232, 235, 284 Traffic, 149, 151, 184, 225–227, 241, 266, 270, 276, 277, 280, 281 Training, 143, 153, 155, 166 Transaction(s)(al), 25, 66, 250, 254, 256, 260
SUBJECT INDEX Transship, 269 Transshipment, 225, 227, 241, 266–269, 273, 274, 276, 285 Transshipp(ed)(ing), 264, 265, 266, 268 Transit(ing), 237, 273 Trans-pacific, 269, 270, 277 Transport, 1, 6, 8, 9, 12, 21, 61–63, 80, 104, 105, 107–109, 119, 121, 135, 138, 144, 145, 147–149, 152, 153, 155, 156, 162, 163, 168, 169–171, 176, 178, 179, 206, 208, 224, 230, 231, 233, 234, 239–245, 250, 259, 272, 277, 285, 287, 288 Transportation, 1, 2, 6, 7, 21, 22, 61, 62, 67, 100, 104, 105, 107–111, 119, 124–130, 133, 135, 137–142, 147, 149, 155, 157, 162, 167, 170, 171, 230, 231, 233, 243 Transport(ed)(ing), 149, 168, 184, 236 T-ratios, 28, 84 Treaty, 146, 154 Trend(s), 7–9, 12, 13, 28–30, 38, 39, 52, 53, 72, 83, 84, 87, 123, 143, 221, 222, 225, 226, 230, 239, 240 Truck(s)(ed)(ing), 119, 126, 133, 233, 277, 279, 283, 284 Tug(s), 108, 178, 182, 184, 187, 203 Uncompetitive, 152, 157 Underpric(ed)(ing), 117, 125, 126, 128–130 Unit(s), 4, 16, 26, 31, 35, 38, 39, 84, 86, 104, 177, 178, 205, 235, 253, 272 Unit root, 39, 86 United Kingdom (U.K.), 1, 16, 17, 58, 62, 104, 115, 117, 136, 146, 155, 156, 160, 161, 163, 168, 169, 170, 171, 175, 184, 208, 242, 244, 266, 269, 273, 288 United Nations Conference on Trade and Development (UNCTAD), 12, 148, 173, 174, 206, 208, 230, 245 United States (U.S.A.), 33, 34, 48, 58, 61, 63, 77, 82, 104, 114, 116–120, 123–126, 135, 137–139, 143, 146, 148, 152, 157, 162, 167, 170, 174,
Subject Index 184, 208, 223, 226, 231, 234, 235, 237, 240, 244, 266, 288, 289 Utilisation, 5, 20, 21, 135, 140, 169, 224, 226 Utilit(y)ies, 112, 158, 256 Valuation, 7, 109, 110, 115, 133 Value(s)(d), 4, 5, 14, 15, 20, 22–36, 38, 39, 41–44, 46, 48, 51, 52, 59, 62, 64, 66, 68, 71, 72, 74, 76, 81, 84, 85, 87, 98–100, 104, 110, 111, 113, 114, 117, 119, 120, 127, 128, 130, 132, 134, 136, 142, 159, 161, 166, 179, 183, 186, 187, 189, 193, 202, 204, 235, 236, 248–263, 266–268, 271, 274, 277, 283–289 Vancouver, 149, 150, 164, 170 VAR, 3, 4, 20, 29–33, 36–39, 41, 42, 44, 51, 60, 64, 72 Variability, 35, 44, 112, 131, 138 Variable(s), 4–6, 10, 11, 23, 24, 26, 28–32, 34, 35, 38–42, 44–46, 48, 51, 53, 55, 58–60, 66–68, 70, 72, 74–80, 83–87, 96–99, 103, 104, 112, 114–116, 122–124, 127, 128, 131, 135, 177–179, 183, 185, 187, 189, 193, 201–205, 207, 266, 271, 272, 275, 288 Variance(s), 32–35, 37, 44, 62, 83, 174 Variation, 44, 69, 117, 131, 135, 159 Vector(s), 4, 16, 28–31, 34, 35, 45, 46, 51–53, 62, 72, 183 Very Large Crude Carriers (VLCCs), 34, 37, 77, 79, 80, 88, 93, 96–100, 103 Vessel(s), 5, 9–11, 21, 22, 25–27, 32, 34, 37, 38, 48, 51, 52, 55, 60, 62, 66, 68–70, 72–75, 79–82, 84, 88, 93,
313 96–100, 104, 108, 134, 143, 144, 146, 149–153, 155, 158, 160, 161, 164, 165, 168, 169, 173–179, 182–185, 187–189, 193, 201–207, 213–219, 224–227, 247, 253, 263, 267, 269, 270–274, 276, 277, 283, 284 Volatility, 16, 32, 33, 34, 65, 66, 72, 75, 116 Voyage(s), 2, 21, 25, 30, 32, 37, 38, 46, 60, 75, 224 Wage(s), 11, 144, 153, 155, 157, 161–163, 175, 184, 188, 205, 206, 283 War, 58, 83, 109, 118, 144 Warehous(es)(ing), 108, 233, 283, 234 Water(s), 7, 8, 28, 30, 108–110, 119, 124–130, 135, 137–140, 165, 166, 184, 226, 230, 284 Waterways, 2, 239 Westbound, 270, 287 Wet, 2, 35, 63, 100 Wharves, 276, 280, 283, 285 White, 39, 165 Won, 74, 77 World Trade Organisation (WTO), 13, 221, 230–240, 242–244 Worldscale, 37, 39, 44, 48, 60, 75 Yangste river, 236, 243 Yantian, 275, 276, 277, 279, 283, 284, 285 Yard(s), 74, 77, 82, 109, 132 Yen, 36, 74, 79 Yield, 4, 10, 116, 122, 134, 236, 272 Yield(s), 4, 10, 20, 31, 64, 112, 116, 134, 136, 156, 236, 272