THE ROLE OF BANKS IN MONITORING FIRMS
Does the Anglo-American approach to the relationship between banks and firms hav...
28 downloads
547 Views
1MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
THE ROLE OF BANKS IN MONITORING FIRMS
Does the Anglo-American approach to the relationship between banks and firms have significant weaknesses compared with the German and Japanese approach? This book addresses issues in the current literature on corporate finance using historical evidence. In particular it looks at the role of universal banks in relaxing the credit constraints of firms, supervising managers and stabilising share prices. The key issue is whether Anglo-American assetbased financing is more effective than the main-bank approach used in Germany and Japan. Earlier studies have found that firms with a close relationship with a major bank have high market value compared to book value, although it is difficult to determine whether this is cause or effect. The case of the Crédit Mobilier—the first universal bank—is interesting because the bank failed. Had it been the case that links with the bank brought about high and stable share prices or relaxed credit constraints, the bank’s bankruptcy should have precipitated the loss of these benefits. In fact, the bankruptcy had almost no effect on the share prices or the investment behaviour of the relevant firms, casting doubts on the benefits of powerful banks. Elisabeth Paulet is Research Assistant at the European University Institute of Florence. Her main interest is in banking systems from the nineteenth century to the present day.
ROUTLEDGE EXPLORATIONS IN ECONOMIC HISTORY 1. ECONOMIC IDEAS AND GOVERNMENT POLICY Contributions to contemporary economic history Sir Alec Cairncross 2. THE ORGANIZATION OF LABOUR MARKETS Modernity, culture and governance in Germany, Sweden, Britian and Japan Bro Stråth 3. CURRENCY CONVERTIBILITY IN THE TWENTIETH CENTURY The gold standard and beyond Edited by Jorge Braga de Macedo, Barry Eichengreen and Jaime Reis 4. BRITAIN’S PLACE IN THE WORLD A historical enquiry into import controls 1945–1960 Alan S.Milward and George Brennan 5. FRANCE AND THE INTERNATIONAL ECONOMY From Vichy to the Treaty of Rome Frances M.B.Lynch 6. MONETARY STANDARDS AND EXCHANGE RATES Edited by M.C.Marcuzzo, L.Officer and A.Rosselli 7. PRODUCTION EFFICIENCY IN DOMESDAY ENGLAND, 1086 John McDonald 8. FREE TRADE AND ITS RECEPTION 1815–1960 Freedom and trade: Volume I Edited by Andrew Marrison 9. CONCEIVING COMPANIES Joint-stock politics in Victorian England Timothy L.Alborn 10. THE BRITISH INDUSTRIAL DECLINE RECONSIDERED Edited by Jean-Pierre Dormois and Michael Dintenfass 11. THE CONSERVATIVES AND INDUSTRIAL EFFICIENCY, 1951–1964 Thirteen wasted years? Nick Tiratsoo and Jim Tomlinson 12. PACIFIC CENTURIES Pacific and Pacific Rim economic history since the 16th century Edited by Dennis O.Flynn, Lionel Frost and A.J.H.Latham 13. THE POSTMODERN CHINESE ECONOMY Structural equilibrium and capitalist sterility Gang Deng 14. THE ROLE OF BANKS IN MONITORING FIRMS The case of the Crédit Mobilier Elisabeth Paulet
THE ROLE OF BANKS IN MONITORING FIRMS The case of the Crédit Mobilier
Elisabeth Paulet
London and New York
First published 1999 by Routledge 11 New Fetter Lane, London EC4P 4EE This edition published in the Taylor & Francis e-Library, 2002. Simultaneously published in the USA and Canada by Routledge 29 West 35th Street, New York, NY 10001 © 1999 Elisabeth Paulet All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data Paulet, Elisabeth, 1962– The role of banks in monitoring firms: the case of the Crédit Mobilier/Elsabeth Paulet. p. cm. Includes bibliographical references. 1. Corporations—Finance—Case studies. 2. Business enterprises— Finance—Case studies. 3. Banks and banking—Case studies. 4. Crédit Mobilier (France) —History. 5. Credit—Case studies. I. Title. HG4026.P329 1999 658.15´224–dc21 98–30729 CIP ISBN 0-415-19539-X (Print Edition) ISBN 0-203-21180-4 Master e-book ISBN ISBN 0-203-21192-8 (Glassbook Format)
It is necessarily part of the business of a banker to profess a conventional respectability which is more than human. Lifelong practices of this kind make them the most romantic and the least realistic of men. John Maynard Keynes
CONTENTS
1
List of figures List of tables Preface Acknowledgements
ix x xiii xiv
INTRODUCTION
1
AGENCY THEORY AND MONITORING A theoretical and empirical interpretation
7
1.1 1.2 1.3 1.4 2
Introduction 7 Presentation of the financial theories 8 The supervisory role of the bank in agency literature 15 General conclusion 24
THE CRÉDIT MOBILIER AND THE FRENCH STOCK EXCHANGE 1853–1914 An empirical perspective 2.1 Introduction 26 2.2 Financial capitalism and data on the Crédit Mobilier in France, 1853–1914 29 2.3 Data available over the period of the two Crédit Mobiliers: Companies involved and preliminary analysis of the data 40 2.4 Excess volatility tests 45 2.5 The effect of the bankruptcy of the Crédit Mobilier on the companies with which it was involved 53 2.6 Bankruptcy effect on the companies not affiliated to the Crédit Mobilier 65
vi
26
CONTENTS
2.7 Comparative analysis of the results obtained for affiliated and non-affiliated companies 71 2.8 Bankruptcy effect on the companies either affiliated or not affiliated to the Crédit Mobilier 73 2.9 General conclusion 74 3
CORPORATE INVESTMENT, CASH FLOW AND FINANCIAL CONSTRAINTS OF FIRMS The case of the Crédit Mobilier
76
3.1 Introduction 76 3.2 The value of association with the Crédit Mobilier on cash flow, investment and market value 78 3.3 Transformation of the raw data according to ‘traditional financial accounting’ 91 3.4 Extensions of the results: description and interpretation of a robustness test 98 3.5 The expression of all variables in real terms 104 3.6 General conclusion 108 4
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER Some interpretations
109
4.1 Introduction 109 4.2 The supervisory role of the Crédit Mobilier 111 4.3 Advantages and disadvantages of an association with a bank for the financing of productive projects 116 4.4 The Crédit Mobilier banks: a comparative analysis between France and Germany 124 4.5 The attempt by the Crédit Mobilier’s shareholders to prevent the bank’s bankruptcy 128 4.6 Conclusion 129 5
GENERAL CONCLUSIONS
132
APPENDIX A Data on the Crédit Mobilier
136
A.1 The Crédit Mobilier between 1852 and 1867 136 A.2 The Crédit Mobilier between 1903 and 1914 136
vii
CONTENTS
APPENDIX B Data relative to the general indices (GNP, share prices) for France between 1852 and 1914
140
B.1 General price indices 140 B.2 Share price series 141 APPENDIX C Data relative to the affiliated companies
142
C.1 The whole data 142 C.2 The Compagnie Immobilière 142 APPENDIX D Data relative to non-affiliated companies
146
APPENDIX E Monthly share prices for affiliated companies between 1866 and 1868
148
APPENDIX F Data for investment and cash flow tests
152
Notes Bibliography Index
161 166
174
viii
FIGURES
1.1 1.2 1.3 2.1
Moral hazard with hidden action Adverse selection models Signalling model Prices, dividends, and perfect-foresight fundamental sample 1870–1914
ix
10 10 11 46
TABLES
2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 2.11 2.12 2.13 2.14 2.15 2.16 2.17 2.18 2.19 2.20 2.21 2.22 2.23 3.1
Ownership of the shares of the Immeubles de la Rue de Rivoli Advances to companies, 1856–66 Fluctuations of share prices for affiliated companies between 1855 and 1856 Crédit Mobilier’s liquidation account in 1901 Market product prices by sector Volatility tests for the French, German and American stock markets between 1870 and 1914 Average asset prices for affiliated firms Relative impact of the bankruptcy of the Crédit Mobilier Relative impact of the bankruptcy of the Crédit Mobilier including a dummy variable d1867 Relative impact of the bankruptcy of the Crédit Mobilier including a dummy variable d1867/68 Regression for d1867 Regression for d1867/68 Results for 1866 Results for 1867 Results for 1868 Results for 1866–8 Relative impact of the bankruptcy of the Crédit Mobilier Relative impact of the bankruptcy of the Crédit Mobilier including a dummy variable d1867 Relative impact of the bankruptcy of the Crédit Mobilier including a dummy variable d1867/68 Results for d1867 Results for d1867/68 Cross-section results excluding the Compagnie Immobilière Cross-section results including the Compagnie Immobilière Companies included in the data set
x
32 32 34 36 41 49 54 55 56 57 58 59 61 62 63 64 67 68 69 70 70 73 73 80
TABLES
3.2 (a) (b) (c) 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10
3.11 3.12 3.13 3.14 3.15 3.16 3.17 3.18 3.19 3.20
3.21
3.22
3.23 3.24 3.25
Summary statistics comparing Crédit Mobilier and non-Crédit Mobilier firms Sample 1861–75 Sample 1861–7 Sample 1868–75 Investment regression, 1861–75 Investment regression, 1861–7 Investment regression, 1868–75 Investment regression equation for the whole sample Fixed effect model for 1861–7 Fixed effect model for 1868–75 Fixed effect model for the whole sample Summary statistics comparing Crédit Mobilier and non-Crédit Mobilier firms when introducing a depreciation rate of 5 per cent Investment regression for the whole sample 1861–75 Investment regression over 1861–7 Investment regression over 1868–75 Results for affiliated companies as presented in Becht and de Long’s paper Results for non-affiliated companies as presented in Becht and de Long’s paper Fixed effect model for the whole sample Fixed effect model for 1861–7 Fixed effect model for 1868–75 Fixed effect model for all firms Investment regression equation for affiliated companies and independent firms over the whole sample (a) Case when d = 3% (b) Case when d = 7% Investment regression equation for affiliated companies and independent firms before the bankruptcy (a) Case when d = 3% (b) Case when d = 7% Investment regression equation for affiliated companies and independent firms after the bankruptcy (a) Case when d = 3% (b) Case when d = 7% Cash flow parameters Sample 1861–75 robustness results for Crédit Mobilier firms without the Compagnie Immobilière Sample 1861–7 robustness results for Crédit Mobilier firms without the Compagnie Immobilière
xi
81 81 81 82 84 85 87 88 89 89 89
92 93 94 94 95 95 96 96 97 97 98 98 98 99 99 99 100 100 100 100 101 101
TABLES
3.26 Investment regression equation for all companies over the different samples (a) Investment regression equations when d = 3% (b) Investment regression equations when d = 7% 3.27 Cash flow parameters 3.28 Q parameters 3.29 T-statistics for sample 1861–7 as regards the cash flow parameters 3.30 Summary statistics comparing Crédit Mobilier and non-Crédit Mobilier firms when parameters are expressed in real terms 3.31 Investment regression equation for the whole sample 3.32 Investment regression equation for 1861–7 3.33 Investment regression equation for 1868–75 3.34 Investment regression equations for all firms 3.35 Regression for the whole sample without the Compagnie Immobilière 3.36 Comparison of cash flow and Q parameters 4.1 Financial participation of the Crédit Lyonnais in ‘big deals’ A1 Balance sheet of the Crédit Mobilier from 1853 to 1866 A2 Average profit of the Crédit Mobilier from 1855 to 1865 A3 Share prices of the Crédit Mobilier from 1853 to 1866 A4 Balance sheet of the Crédit Mobilier from 1903 to 1914 A5 Share price series A6 Data available for affiliated companies: share prices A7 Revenues of the Compagnie Immobilière A8 Expenses of the Compagnie Immobilière A9 Share price and dividend distribution, 1855–66 A10 Data for non-affiliated companies A11 Share prices for affiliated companies for 1866 A12 Share prices for affiliated companies for 1867 A13 Share prices for affiliated companies for 1868 A14 Data available for the railway companies A15 Data available for mining industries A16 Data available for the steel industries A17 The Compagnie Immobilière A18 New values of Q when introducing a depreciation rate of 5 per cent A19 Transformation of the data when changing origin A20 Data in real terms
xii
102 102 102 103 103 104
105 105 106 106 106 107 108 122 137 138 138 139 141 144 145 145 145 147 149 150 151 153 154 155 156 157 158
159
PREFACE
Economists seem to agree on the monitoring power of banks as regards their clients. Their arguments are based on the ability of these institutions to collect information on exposure to risk, and on their flexibility to diversify investment. In order to exert this monitoring role, a banker should always evaluate the trade-off between profitability and efficiency. This book examines how universal banks could help relax the credit constraints of firms, improve the supervision of managers and stabilise share prices. The case of the Crédit Mobilier, which went bankrupt in 1867, is a good illustration of the weakness of the monitoring role of banks. It exhibits the potential of enterprises to diversify their investments in order to avoid financial destabilisation. A historical approach adds evidence to this last assertion and casts doubt on the beneficial role of closed relationships with powerful banks. As far as bankruptcy is concerned, the attitude of the central bank is an interesting factor. At that time, the Banque de France and its governors were hostile to the founders of the Crédit Mobilier, the Péreires; consequently, its participation in rescuing the bank was almost non-existent. Beyond the historical context, this fact illustrates the importance of political pressures in financial decision-making. This debate remains very sensitive, particularly in view of the discussions concerning the future of the European central bank. As a whole, the questions mentioned in this essay are still valid today. This book was written in order to present some new ideas regarding the control and monitoring of financial institutions and, more specifically, universal banks. While the flexibility of other structures must be acknowledged, there are moves afoot to create a unified European banking system in which universal banks will have a privileged position on the financial markets. The justification for this concerns their perceived superiority as evaluators of information and risk. The conclusions of the empirical analysis discussed here have a direct bearing on the power of universal banks.
xiii
ACKNOWLEDGEMENTS
This book is based on my Ph.D. dissertation at the European University Institute in Florence, written between 1991 and 1995. I would like to express my gratitude to all the people who supported me at that period: many thanks to Robert Waldmann who has acted with great sensibility as a supervisor; the History Department at the Institute, and especially Peter Hertner, Albert Carrerras and Kirti Chaudhuri, who have always been interested in my research; William Kennedy, who kindly provided me with useful comments and encouragement; and last, but not least, Maurice Levy Leboyer who, through written exchanges, enabled me to form a new opinion on the personality and position of the Péreire brothers in nineteenth-century financial circles. My research led me to collect data in Paris. I am grateful to the personnel of the Paris Archives for their contribution to this work. Much of what follows is based on a simple idea: the Crédit Mobilier and its founders, the Péreires, failed because they encountered difficulties that were due more to external constraints than is usually acknowledged. In other words, environment matters. My own environment in the course of writing the present book has been as follows. For several years Francesc Relaño has endured my ramblings on the course of banking policy over the last century. My parents have provided support during the most austere period of the writing process. My last thought is to dedicate this book to my brother, who will never have the opportunity to read this piece. To Patrick.
xiv
INTRODUCTION
Bankers and investors influence industrial investment through the financial power they exert on the managers of firms. The purpose of this research is to study the interrelations that exist between these managers, their shareholders and the banks which finance investment projects when defining a debt-equity contract. The subject will be approached from two diverse perspectives: the first one is theoretical (Chapter 1), the second empirical (Chapters 2, 3 and 4). Theoretical arguments can be, and have been, advanced to describe how the conflict of interest between managers and owners of firms can be solved. The purpose of the first chapter is to survey the literature on principal-agent theory and the signalling approach which concentrates on these questions. In particular the power of banks will be debated. Two questions are considered: • •
Why might banks want and be able to obtain the power to influence and monitor managers? Why might shareholders benefit if the bank has such power?
Agency and signalling theory are concerned with the information and interest conflicts which can exist between the different economic actors associated with the firm: managers, shareholders and creditors. This theory holds that some agents (the managers) are better informed about the quality of the enterprise than the principals (shareholders, creditors). To resolve the conflict of interest, contracts (implicit or explicit) must be negotiated between the lenders, the creditors, the shareholders and the managers. The term ‘agency’ derives from the fact that corporate decisions are delegated to agents (e.g. managers) who act on behalf of other parties (external investors). The relation between investment and debt is explained by the choice between debt and equity for the financing of a project. The signalling approach makes the assumption that there is asymmetric information as between management and investors: the managers have more information than the shareholders on the value of their enterprise.
1
THE ROLE OF BANKS IN MONITORING FIRMS
The problem is the accurate transmission of such information, to give shareholders the incentive to finance productive projects. Both approaches feature the monitoring role exerted by lenders on borrowers: control remains one of the main factors of agency. The concept of monitoring will presently be examined, focusing on the role of banks in performing this control function. According to agency theory, the bank will have less information about the project than the managers; the role of collateral can be useful in dealing with the problem of adverse selection. Adverse selection arises when lenders cannot distinguish managers of firms who undertake low-risk projects from the ones who realise high-risk projects. Then a contract that is optimal when offered to less risky borrowers also attracts more risky borrowers and thus becomes unprofitable to the lenders. The use of collateral constitutes one way to control this risk. Several other tools will be discussed. Two examples are presented to illustrate the argument: Japan and Germany. The aim of this analysis will be to compare the conclusions reached for these two countries with the results obtained for the Crédit Mobilier. Historians have argued that the influence exerted on industry by banks such as the German Großbanken or the French Crédit Mobilier was substantial during the nineteenth century. The role played by the Großbanken—those banks which, like the Crédit Mobilier, took a great role in the industrialisation of Germany—in monitoring and supervising corporate management was an accepted part of German finance theory in the years before the First World War. The second step of my analysis is to gain, through empirical research, an insight into the role of banks in the financing of enterprises during the period 1850–1913. In particular, I study the Crédit Mobilier’s role in France during the pre-WWI period. This bank, founded by the Péreires in 1852, until it went bankrupt in 1867, contributed largely to French industrialisation. It failed a second time at the beginning of 1900 and merged in 1932 with the Banque de l’Union Parisienne. The Crédit Mobilier has been chosen for two reasons: •
•
Its organisation was similar to that of the German Großbanken. The comparison with these institutions can give some additional insight to the results mentioned in the literature. As the bank went bankrupt, it will be interesting to analyse the impact of this failure on the firms affiliated to it. If a real monitoring role were exerted by the Crédit Mobilier, affiliated firms should have faced difficulties when the bankruptcy occurred.
First the role played by the Crédit Mobilier on the stock market is determined. The level of volatility will be quantified in order to assess the
2
INTRODUCTION
French general economic context between 1852 and 1913. Share price series are then used to investigate the impact of the bankruptcy of the Crédit Mobilier on the stock market for affiliated and non-affiliated companies. Share prices did not fall when the bank failed. The conclusion is that the Crédit Mobilier provided liquidity to firms, but apparently did not seem to be supervising them, and its seal of approval was probably not worth as much as that of J.P.Morgan1 or the Großanken.2 Some remarks must be made concerning the data used to construct the test. During the period statistical sources such as share prices and balance sheets were not always kept according to strict rules. In addition the amount of statistical information available over the period was quite limited because managers were not obliged to publish their accounts. So the validity of our results has to be assessed in the context of a panel data of twenty five companies which provided information on their financial situation. As the Crédit Mobilier seems not to have succeeded in controlling firms, it would be interesting to measure the effect of an association with this establishment on investment policy. The basic theoretical idea is as follows. In their world of perfect information and no agency costs, firms can borrow and invest optimally; liquidity has no effect on the investment decision. In their classic paper, Franco Modigliani and Merton Miller (1958) showed under what conditions financial structure becomes irrelevant for real investment decisions. In a context of asymmetric information this proposition is no longer true. The next step in our research is to analyse the relationship between investment and cash flows between 1860 and 1880. The effect on investment policy of an association with this establishment will then be measured and compared to Großbanken policy during a similar period. This proposed analysis then tests the proposition that the Crédit Mobilier acted like a German-style universal bank. To that intent it examines the effects on the investment behaviour of French firms of affiliation with the Crédit Mobilier. Investment equations for a modified version of Tobin’s Q-model are estimated. Two samples are considered: the first consists of corporations that were affiliated to the Crédit Mobilier, the second includes companies which had to resort to other sources of finance. The focus of this construction is on contrasting the financing of the Crédit Mobilier’s affiliated enterprises with those which, for one reason or another, did not form part of the nexus. The chronological limits of this study are imposed by the data sources: after 1880 the Crédit Mobilier lost its powerful role in the banking sector and thus no records exist after this point. The sample includes fifteen companies: seven were affiliated with the Crédit Mobilier, eight with the Crédit Lyonnais. As each company, in any
3
THE ROLE OF BANKS IN MONITORING FIRMS
event, had a relationship with a bank, the interpretation of the results must be in some sense cautious. Using this evidence, the following two conclusions are reached: 1 2
companies that maintained close ties with the Crédit Mobilier were less liquidity constrained after the bank’s crisis in 1867 than before; unaffiliated companies experienced no more difficulty in financing projects than those which resorted to the Crédit Mobilier.
These results strongly contradict earlier findings. The interpretation will define the context in which they are valid. Some empirical extension of the results is provided. In particular, we intend to develop two different models: the first will try to re-evaluate the book value for all companies entering the sample by introducing a fixed depreciation rate of 5 per cent. This is because I suspect the account figures for book value changed only rarely from year to year. The second proposition is to divide all variables by the general level of prices. If the conclusions are not substantially different they support the arguments mentioned above. An interpretative chapter discusses whether the supervision of the Crédit Mobilier was really effective—if the firms with which it was involved had not anticipated the fall of the bank, and diversified their sources of financing corresondingly. Our empirical results prove that the bank did not really control firms. The regressions suggest that the affiliated firms were more liquidity constrained before the bankruptcy than after. This supports the argument that the firms did not really trust the Crédit Mobilier as a banking institution—they just used it to finance their investment projects. A large part of the economic literature considers that a bank-firm association is always profitable for both parties because: • •
the bank prevents the firm from undertaking projects that are too risky; the firm signals its reputation to the market by having a regular relation with a bank.
The Crédit Mobilier’s case gives no clear evidence to confirm this argument. The question is then: What is the most efficient way to control a firm: an interaction between firms and banks (which means substantially a bank-firm association), or a clear separation of banks and firms (as presented in our first chapter)? In other words, should banks supervise, or just choose whether or not to loan? In particular, we can ask if the principal (the Crédit Mobilier) was not an agent acting for firms which needed liquidity over the period. A comparison between France and Germany is then drawn. Other researchers 3 have shown the efficiency of the German Großbanken
4
INTRODUCTION
(which were explicitly modelled on the Crédit Mobilier). The diverse results obtained in the two cases enable me to say that in many countries the Crédit Mobilier was the inspiration for a new type of banking, but not the model. The last point will be a debate on the rationality or irrationality of the shareholders. In particular, some explanation for their passive attitude in the face of the bankruptcy will be advanced. The last chapter concludes.
5
1 AGENCY THEORY AND MONITORING A theoretical and empirical interpretation
For French business, whether medium or ‘big’, the obsession with independence vis à vis the lender (the bank) was constant. It was with reference to the level of their current bank accounts that the Schneiders of the nineteenth century, for example, calculated their degree of independence and their room for manoeuvre…. The domination of banks over business, if it existed was not achieved without resistance.1 Bouvier (1988:109)
1.1 INTRODUCTION In this chapter the basic concepts of agency and signalling theories applied to the financing of the firm are presented. In particular the chapter will study how a debt-equity contract is defined between the manager of a firm, the shareholders and the bank which provides liquidity. Agency theory, although not quite new, is a part of decision theory which is currently characterised by a dynamic development. The wellknown basic model of decision theory is related to a situation in which one person—the so-called decision-maker—has to make a decision as well as to bear its consequences. This assumption is often unrealistic. Therefore, in agency theory the decision-maker (the agent) and the ‘beneficiary’ of the decision (the principal) are all distinct. A typical example concerns the management of the firm: the agent, the manager of the firm, chooses an investment project. The principal—the shareholders of the firm and (or) the bank which finances the enterprise—partly support the risks involved in this investment. Hence, the common element is the presence of at least two individuals. The first (the agent) must choose an action from a number of alternative possibilities. The action affects the welfare of both the agent and the principal. The principal has the additional function of prescribing payoff rules; that is, before the agent chooses the action, the principal determines a rule that specifies the fee to be paid to the agent as a function of the
7
THE ROLE OF BANKS IN MONITORING FIRMS
principal’s observation. The problem acquires interest only when there is uncertainty at some point and, in particular, when the information available to the participants is unequal. As information is asymmetric between the economic agents, control is one of the main factors of agency. The second part of this chapter will be devoted to the definition of this control mechanism in order to determine which of the two principals, the shareholders of the firm or the bank with which the enterprise is involved, is able to exert effectively this monitoring role. 1.2 PRESENTATION OF THE FINANCIAL THEORIES (a) Basic concepts of agency and signalling theories The financial theory of agency focuses on the relationships between different groups of security holders (equity and bondholders) in the context of optimal financing of the firm. The standard economic theory of agency considers two individuals: a principal, who provides the capital, and an agent (the manager), who provides the effort. Both are assumed to be utility maximisers. Principals value end-of-period wealth which is derived from their share in the realised value of the firm. Agents value their end-of-period wealth stemming from the realised profit. Agency problems arise because, under the behavioural assumption of selfinterest, agents do not invest their best efforts unless such investment is consistent with maximising their own welfare. The agency model is basically a formulation of the principal’s problem of choosing the ‘best’ employment contract for the agent, where ‘best’ is defined in the context of Pareto optimality. A Pareto optimal contract is such that no other contract can improve the welfare of one party without reducing the welfare of the other. Observability of the agent’s effort is really the core of the incentive problem in this theory. While the effort level determines the level of output of the firm (i.e. the end-of-period value or cash flows), the output is also affected by random events that are beyond the control of the agent. An agency problem arises when the consequences of the agent’s efforts cannot be distinguished entirely from the consequences of other random events by observing output alone. In the tradition of the agency theory, the output (payoff) level is mutually observable by the principal and the agent, but the effort level is observable only by the agent. First best contracts induce an optimal risk-sharing between agent and principal. Further adverse selection is often introduced between the two parties. Consider an owner-manager who seeks to finance a project by
8
AGENCY THEORY AND MONITORING
selling securities, while the true nature of the return distribution of the project is unknown to the outside market. Management possesses valuable information about the project that is unavailable to the market. If this information were revealed fully to the market, the market would value the project at V . Otherwise, the market is unable to distinguish a this project from another less profitable project with a value V . This b asymmetry may be corrected, at a cost, through various ‘signalling mechanisms’. In the absence of an unambiguous signal, however, management will obtain less for securities sold than their ‘fair value’ reflected in the true nature of the project A. The difference between the ‘fair’ price and the actual price is the agency cost associated with informational asymmetry. It exists for the issuing of debt as well as new equity securities, provided that there is a differential probability of bankruptcy for the two projects. In their seminal paper, Jensen and Meckling (1976) proposed agency costs as a key tool in evaluating designs of a principal-agent relation. Agency costs were defined as the sum of: 1 the monitoring expenditure by the principal; 2 the bonding expenditure by the agent; and 3 the residual loss, i.e. the monetary equivalent of the reduction in welfare experienced by the principal due to the divergence between the agent’s decision and ‘those decisions which would maximise the welfare of the principal’ (ibid.: 308). The agent receives compensation for his work and supports part of the agency cost (bonding expenditure). Adverse selection can be described through three categories of model. The first is represented by a situation of moral hazard with hidden action. Under moral hazard, it is easier for the principal to observe the manager’s output than his effort. The principal offers a contract to pay the agent based on output which depends on the agent’s effort. We can represent this situation by a game tree. In the model, the principal (P) offers the agent (A) a contract, which he accepts or rejects. After the agent accepts, nature (N) adds noise to the task being performed. Figure 1.1 illustrates this situation. If the principal knows the agent’s ability but not his information level about the potential investment, the problem is moral hazard with hidden action. An example is given by F.Gjesdal (1982), whose paper considers the use of imperfect information for risk sharing and incentive purposes between bondholders and stockholders when perfect observation of actions and outcomes is impossible, making complete contracting infeasible. The incentive problem consists of two parts: the choice of information system and the design of a sharing rule based on the information in the agency problem.
9
THE ROLE OF BANKS IN MONITORING FIRMS
Figure 1.1 Moral hazard with hidden action
Figure 1.2 Adverse selection models
The second case concerns the largest part of the agency literature: adverse selection models. Adverse selection can be represented by Figure 1.2. Kose and Kalay (1982) illustrate adverse selection by analysing the conflicts between two groups (stockholders and bondholders) concerning the riskiness of company projects. They derive an optimal set of contractual arrangements which minimise the cost of this conflict. In that sense, these authors come back to the idea of Myers (1977), who has argued that stockholders who control the firm could attempt to transfer wealth from the bondholders, even by rejecting profitable investment projects. In other words they can choose policies which reduce the market value of the firm. Myers has indicated that self-imposed constraints on dividend pay-outs by stockholders are a possible solution to the problem of under-investment. The debt contract between a bank and the manager of a firm defines the adverse selection problem. According to the recent literature, such as Laffont and Freixas (1988), Gale and Hellwig (1985) and Diamond (1984), the assumption of asymmetric information leads the bank to prefer credit rationing, to an equilibrium between supply and demand at a specific interest rate. The agency relations between a bank and a firm prevent the lender from distinguishing sound creditor risks from bad ones. The requirement of a guarantee can solve such problems and enables the lender to separate good borrowers from the bad ones. In this context Deshons and Freixas (1987) establish two types of loan system: • •
the optimal separating contracts, where good credit risks are identified; the rationing credit solution, where neither guarantee nor interest rate make the selection role possible since the bank is unable to take the surplus of its loan customers.
10
AGENCY THEORY AND MONITORING
Figure 1.3 Signalling model
The latter situation, represented by Figure 1.3, is the signalling model. The agent sends a signal to the principal. For example, the managers of firms transmit information to the shareholders and the creditors of the firm they administer. This activity of signalling has two aspects: • •
the manager of a good firm sends information to the external investors to signal that the enterprise is a good firm; however, the manager of a bad firm might use the same signal to make investors think that the enterprise is a good one, but it is more costly for him to do so.2
A penalty system must be in place to prevent the manager of a bad firm from sending false signals. The problem of signalling in the determination of the debt level of a firm is expressed by the diversification of a security portfolio for a manager-shareholder. If the latter possesses a good productive investment, he will inform the market by the composition of his portfolio. If the project is really good, he will devote a significant part of his savings to it, to the detriment of other projects. As he is the best informed, his less diversified portfolio is a signal which can indicate the value of the project. Leland and Pyle (1977) conclude that the value of the enterprise is positively correlated to the proportion of capital held by the shareholder-manager. But the authors emphasise that this relation is a non-causal statistical relation between the value of the firm and its debt level. Every modification of the manager’s portfolio induces a change in the perception of the future liabilities by the market: it results in a new financing policy, and ultimately another value for the firm. In this context, the level of diversification in the assets held by a firm constitutes the most valuable information as regard the value of the firm. As observed by Jensen and Meckling (1976) or Grossman and Hart (1983), if a firm does not want to be indebted, its risk of bankruptcy is limited but the market will suppose that the maximum performance for the firm has not been attained. In the opposite case, a loan from a bank increases its risk of bankruptcy, which obliges the managers to be more efficient. This argument illuminates the divergence of objectives between the different partners: the
11
THE ROLE OF BANKS IN MONITORING FIRMS
shareholders, who seek dividends, the managers, who must signal the value of their organisation by diversifying their portfolio, and the banks, which grant credit and want to avoid losses if the firm goes bankrupt. To conclude these first two points, it seems that the application of agency and signalling theory to finance decisions is concerned with the influence of conflicts between shareholders and managers over the value of the firm, and the resolution of these conflicts by the use of debt. The latter involves an incentive scheme for performance. The debt increases the bankruptcy risk, which is a threat sufficient to encourage the manager to administer efficiently. Bankruptcy will mean for the manager not only loss of his work, but also being held responsible for the situation. In these models, debt is defined by an optimal contract for the firm, and the optimal structure of the enterprise is a combination of debt and stocks. So, the financial parameters justifying the limits of debt policy can be deduced by analysis of the shareholders’ and managers’ points of view (e.g. level of risk of the project, expected profit, etc.). However, these theories have a certain number of limitations. In particular, the same financial indicators have specific roles in each approach. The debt level constitutes a very manipulable variable for the administration of the firm. It signals the quality of an enterprise for Ross (1977). It is the origin of the agency cost for Galai-Masulis (1976) and Jensen and Meckling (1976). The results of these analyses do not seem to be very coherent: an increase in debt or in the part attributed to the manager and shareholders in the capital induces an increase in the market value of the firm, according to the signalling models. But the difficulties due to agency relations then become more and more pronounced. Myers (1977) has taken an interest in this problem. He explains that three financial parameters must be taken into consideration at the same time: the dividend distribution, the amount of investment and the level of debt. Agency costs appear there as one of the essential elements of a new kind of signalling model. For Myers, the existence of debt for an enterprise can generate a sub-optimal level of investment: the nature of equity and the financial structure can be modified such that the investors do not maximise the value of the firm but the value of the stocks, to the prejudice of debt.
(b) Common agency As indicated above, agency theory aims to explain the conflicts of interest and differing information that could exist between the different partners of the firm: managers, shareholders and lenders. Every partner determines the optimal level of debt according to his own interest. The manager wants to find the amount of
12
AGENCY THEORY AND MONITORING
funds he thinks sufficient. For the shareholders and the bank, investment is a way to realise profit. However, the latter actors bear different costs. The shareholders give the manager financial compensation for this work, while the banker pays interest costs for granting the credit (Laffont and Freixas 1988). In this section, possible areas of common interest between these three actors in the financing of the firm will be discussed. In particular the following question will be asked: are these different objectives sufficiently compatible such that the common determination of debt for the firm might be possible? Are there situations where the three partners are led to define jointly the very price to pay, so that the firm could finance its productive activities through an optimal level of debt? When several economic agents must take a decision in which they are all concerned (for example, the conclusion of a debt contract between bank and manager, preserving the solvency of the enterprise and the wealth of the shareholders) the economic argument is both normative (what is the optimal price of debt for a firm?) and strategic (which price results from the confrontation of these different interests?). In fact the shareholder is likely to be in favour of debt which gives him more profit and decreases his level of risk-sharing. However, if the manager invests in high-risk projects, the risk of bankruptcy increases. Lars Stole (1990) has considered the contracting problem facing multiple principals, each of whom wants to contract with the same agent. This has been termed the common agency. Common agency can either be delegated or intrinsic. Under delegated common agency, the choice of contractual relationship is delegated to the agent, who can choose whether to contract with both, one or none of the principals. Alternatively, when common agency is intrinsic, the agent’s choice is more limited: the agent can choose only between contracting with both or with neither. Consider a contracting environment with two principals and one agent. A first possibility is that each principal wants to share the profit of a given outcome which the agent produces. Each agent is represented by a type which describes the characteristic of the projects undertaken by the manager. A direct revelation mechanism incites the agent to announces his type to each principal separately. If the agent contracts only with one principal we return to the literature defining contracts between either shareholders and manager or between bank and manager. If the agent contracts with no one, as he has not enough money to realise the project, he cannot carry it out. The only interesting case is where all parties contract: •
Stage 1 The principals announce the contract which they wish to sign with the manager (agent). A contract specifies a project, a reward scheme for each configuration of the market (principal-agent) and the credit to be granted. We assume that offers between principals and agent are private information.
13
THE ROLE OF BANKS IN MONITORING FIRMS
•
•
Stage 2 The manager of the firm (agent) announces which of the offered contracts he is willing to accept. The announcement is public information (each agent must inform the principals whether he wishes to deal with that project exclusively or in conjunction with projects of other firms). Public observability of the announcement is not essential, but simplifies the description of an equilibrium. Stage 3 Each principal selects the manager which is willing to undertake the project and to accept the proposed contract. At the same time, principals announce the amount of capital they grant for its realisation. These choices are observable. Assumptions
First, all players are risk-neutral. While risk-neutrality on the part of the principal is not essential, a risk-neutral agent is indispensable in what follows. For example, as mentioned in stage 2, a manager can undertake several projects with differing levels of profitability, which enables him to spread risk. Second, principals are allowed to condition the terms of contracts upon the market. This feature of our model illustrates an important aspect of agency delegation. Since a principal’s decision to abandon a common agent (manager) constitutes an observable change in market environment, competitors may renegotiate their contracts in response. 3 Stage 1 of our model provides a simple way to represent this process of reactive renegotiation: the principal simply announces an amount of funds and an incentive scheme for every possible market configuration. Note, however, that, as long as a principal retains the same agent, other principals cannot respond directly to changes in its contracts since contract offers are private information. Consider now the model described above, but modify the description of the agency delegation process as follows. In stage 1, each principal makes one contract offer (credit incentive scheme) to the same agent (the manager of the firm). In stage 2 this single agent either accepts both offers or neither. If he rejects, principals earn zero profit and the game ends. In this simple model, common agency is intrinsic in the sense that the agent has no alternative to the pre-specified delegation. Bernheim and Whinston claim that: non co-operative behaviour induces an efficient (potentially second best) action choice if and only if collusion among principals would implement the first best level cost.4 In an earlier paper5 they discussed the possibility of collusion between shareholders and bank to reach an equilibrium. This leads us to consider the supervisory role of banks and shareholders in the agency theory.
14
AGENCY THEORY AND MONITORING
Agency as control is tied to compensation in at least three different ways. The first is a causal relation: compensation helps to define the agency relation, much more than the reverse. Compensation’s incentive role (giving the agent an incentive to act in the principal’s interest) is less important than its informative aspect: agents must know what their responsibility is (for example, as regards the solvency of their organisation) and how it changes each time the financial structure is modified. Performance-related compensation for managers may be primarily a vehicle for defining their duty. But if the bonus scheme or other form of incentive compensation is mechanical, conveying little about the kind of performance to be elicited, the scheme does not signal agency relations. This first connection between compensation and agency has to do with bringing agency into the formal organisation. The second connection brings agency into public markets, answering the question: in which direction should the cash flow component of compensation flow? For example, in a financial corporation this constitutes a way to achieve control on managers of firms, by giving them an incentive to maintain a reputation towards their lenders. As a third illustration of the relation between agency and compensation, consider the question of appointment of the different actors of the firm. The manager is the agent not only of the shareholders but also of the bank which finances its projects. Any study of corporate structure must therefore consider two very different interfaces: that between the shareholders and the manager, and that between the manager and the external investors. 1.3 THE SUPERVISORY ROLE OF THE BANK IN AGENCY LITERATURE The bilateral principal-agent question considers that one individual, called the agent (e.g. the manager of a firm) performs duties on behalf of another party, the principal (external investors: e.g. the bank and the shareholders). This theory explains 1 2
why debt was relied upon as a source of capital before debt financing offered any tax advantage relative to equity; why shares are issued when financing a project.6
In this section, I study the theoretical framework concerning the role of banks in monitoring the enterprises with which they are involved. More generally, how do financial institutions affect the allocation of funds for investment?
15
THE ROLE OF BANKS IN MONITORING FIRMS
Some authors who have considered this question have observed that financial systems differ significantly between countries and across periods, and the question arises how these differences between financial systems have affected the functioning of the different economies. In this section I am going to concentrate on two case studies, Japan and Germany, in order to contrast the results obtained for these two countries with those relating to the Crédit Mobilier in Chapter 2. As a first illustration of this concept, let us consider the paper of Mayer (1988). The author suggests there that there are systematic differences in performance between financial systems in which banks play a prominent role, and those financial systems in which banks are less prominent. He explains these differences in performance in terms of differences in the mechanisms that reduce or eliminate moral hazard between entrepreneurs and financiers. Specifically, he suggests that the more bank-oriented systems of Germany and Japan involve greater commitment on the part of the firm and the bank to a long-term relationship, which allows them to enjoy the benefits of long-term contracting. In this section, the following issue is considered: in what sense is ability to support long-term relationships the clue to assessing financial institutions? The focus of the discussion will be on financial institutions (more specifically, the role of banks in the provision of finance to firms), rather than financial instruments.
(a) Internal finance, external finance and bank finance: theoretical presumption in historical assessment Theoretical framework The manager-debtholder literature has two sub-literatures. First there is the issue of the manager-bondholder conflict. There are numerous theoretical discussions of the role of bond convenants in controlling the manager’s and (or) shareholders’ incentive to increase the risk of the firms. This literature has treated debt mostly as a homogeneous instrument creating incentives, in direct conflict with those associated with equity financing. Second, there is the borrower-lender literature which has developed from the simple examination of the incentive of the borrower to default on a bank loan, to the role of the bank as the delegated monitor of a portfolio of borrowers, i.e., a monitoring role for the bank’s depositors.7 These latter models have served to strengthen the argument that there is a unique information and monitoring role of bank loans and that the phenomenon of credit rationing is due to the nature of information asymmetry in the market. The argument discussed here is richer than the traditional literature briefly presented above. Essentially based on the model of Devinney and Milde (1990), it incorporates the role of ‘inside debt’. Inside debt, as
16
AGENCY THEORY AND MONITORING
defined by Fama (1985), is debt which explicitly incorporates monitoring by giving the debtholders access to otherwise proprietary information. The modelling framework used here is that of an asymmetric information game. Managers are viewed as choosing amongst alternative investment projects which require the same initial outlay but differ in terms of risk. Greater bank monitoring affects the project chosen by the manager, while the bank’s choice of loan contract terms affects management effort. Outside stockholders choose the managerial incentive contract and, hence, the level of risk-sharing between managers and stockholders. The major assumptions of the model presented by Devinney and Milde are the following: 1 The expected payoff of the project, µ is a function of management effort e with the properties µe > 0 and µee < 0. 2 Both debt L and equity S are used to finance the project. 3 All debt is inside debt (bank loan) as defined by Fama (1985). 4 The equity position consists of both ‘inside’ equity and ‘outside’ equity. Inside equity is owned by the management and is a fraction α of the total equity outstanding. The remaining fraction (1–α) is the equity position held by outside owners. 5 Outside shareholders determine the fraction of equity given to the managers, α, and the managers’ fixed income component H. Managers are prohibited from trading the equity of their corporation in the stock market. 6 Banks determine the level of monitoring activities m, the loan size L, and the (gross) loan rate R=(1+r) 7 Managers choose the project, i, and hence its riskiness, σ and the level of effort, e. 8 Banks can monitor the managers’ choice of project (risk) but monitoring is costly. Total monitoring costs are M=νm, where ν is the per unit monitoring cost. 9 Neither outside stockholders nor banks can monitor the level of managerial effort. 10 Stockholder and bank expectations are rational in the sense that they can calculate the management’s response to changes in their decision variables. 11 All decision-makers are risk-neutral. Under these assumptions, how can the bank-stockholders management equilibrium be determined? The manager’s responsibility is to accept or reject the projects and choose their effort. We consider that only one project can be undertaken. The chosen project is characterised by specific µ (its expected payoff) and its riskiness s is observable to the bank but not outside
17
THE ROLE OF BANKS IN MONITORING FIRMS
stockholders. By assumptions 6 and 8, managers cannot do anything but follow the bank’s instructions regarding s. The effort level choice e=e* is optimal if the marginal expected payoff to the manager is equal to his/her marginal opportunity cost. More specifically the relationship between e* and the decision variables determined by outside shareholders and banks are the following: • • •
an increase in the fraction of internal equity increases the manager’s effort; monitoring leads to higher effort levels by management; increasing the loan size and/or the loan rate will increase the effort until it threatens the solvency of the firm.
Banks can monitor the manager’s choice of s and/or participate in the decision-making process. By increasing monitoring and/or participation, banks can enforce their interest more and more, thus eliminating the tendency of management to undertake higher-risk projects. Then the bank’s decision-making process leads to three components characterising a loan contract: interest R*, loan volume L* and risk s*. The risk, s*, however, is implicitly determined by bank monitoring, m*. This contract specifies the second best solution and is consistent with the manager’s effort response function. The distinction between internal and external debt is similar to that between internal and external equity and can explain the monitoring role of the participants in the financing of investment projects. Theoretical presumption in historical assessment The economic theory stresses the notion of effort to explain the monitoring role of banks and stockholders towards the managers of firms. The main argument of the literature is based on the notion of ‘capital scarcity’. Capital is regarded as scarce because: •
•
in the late nineteenth century, a hundred years after the beginning of the industrial revolution, the adoption of up-to-date technology required large capital outlays; and the failure of some countries to participate in the earlier stages of industrialisation meant there was a lack of accumulated funds that could be used to finance these outlays.
Gerschenkron (1962) studied this notion of ‘capital scarcity’ and applied it in the context of German industrialisation from 1850 to 1914. He observes the prominent role of German banks in providing liquidity for heavy industry with large capital requirements.
18
AGENCY THEORY AND MONITORING
Gerschenkron’s argument is as follows: over the period, the firms’ retained earning were insufficient to finance desired investment. At this point, Gerschenkron raises two questions: 1 2
Why should an insufficiency of internal finance be seen as characteristic of a situation of ‘economic backwardness’? Why should the large-scale provision of external finance to industry require the intervention of banks or government?
The first question really concerns the role of internal finance in an industrially advanced country. It is clear that, in the second half of the nineteenth century, certain advanced technologies required large-scale investments. It is also clear that firms in the less developed countries had not had time to accumulate the funds required for such investment. However, it is less clear why firms in the industrially more advanced country should have retained earnings to such an extent that further industrialisation could be financed without the same level of recourse to outside finance. Gerschenkron’s interpretation is that the involvement of banks or government is needed to provide outside finance to industry on a large scale. External finance from the anonymous markets of our theoretical models is not seen as an alternative. Bank finance here is not necessarily just loan finance. During certain periods, especially prior to 1873, German companies obtained substantial amounts of equity finance from banks. The shares would be held by banks, or by clients of banks acting on the banks’ advice, and in many respects banks were as much involved in equity finance as in loan finance. While share markets were organised, they were certainly not anonymous and free for all. The question is to what extent bank involvement in equity and loan finance was actually necessary. According to Gerschenkron’s argument, bank finance could do something that anonymous organised financial markets could not. So we need to discuss what exactly the comparative advantage of bank finance is. This is the subject of the following section.
(b) The advantage of bank finance versus market finance The discussion of the advantage of bank finance over market finance implies that financial systems are institutions which reduce or eliminate problems of moral hazard or asymmetric information between firms and financiers. Financiers typically have less information about firms than entrepreneurs or managers. Moreover, investors are subject to various types of moral hazard: moral hazard concerning managerial effort, moral hazard concerning the
19
THE ROLE OF BANKS IN MONITORING FIRMS
riskiness of the firms strategies and moral hazard concerning reported ex post return realisation. These problems of moral hazard and asymmetric information cause difficulties for the provision of finance to industry.8 Intermediaries are taken to reduce these difficulties by engaging in monitoring and control activities. Diamond (1984) presents an explicit example in which intermediation successfully reduces the agency costs of outside finance under moral hazard. In his analysis, the feasibility of financial intermediation rests on two key propositions: 1
2
Monitoring and control of a firm involve natural scale economies: a single intermediary can monitor and control the firm as effectively as thousand of shareholders—but much more cheaply. If the intermediary has a well-diversified portfolio of firms that he finances, then relations between himself and his own financiers—the final investors—are not much affected by moral hazard and asymmetric information because his own return is approximately riskless, so for him, fixed interest debt finance is feasible and does not involve any moral hazard.
On the basis of these two propositions, Diamond shows that under certain circumstances incentive-efficient allocations cannot be made without intermediation. Diamond’s notion of financial intermediation as delegated monitoring (or delegated control) is closely related to Gerchenkron’s account of bank involvement in firms at the early stage of industrial development, although this is not explicit in Gerschenkron’s work. As emphasised by Mayer, bank initiative and bank participation in entrepreneurial planning may be a way to obtain enough information and control to reduce the moral and informational hazards of finance to a tolerable level. We may therefore look at the imperfect information approach to financial intermediation as the theoretical basis for Gerchenkron’s view that banks were needed to provide outside finance during the early stages of industrialisation in Germany, when capital was ‘scarce and diffuse’ and ‘the distrust of industrial activities…considerable’. In Diamond’s terminology, Gerschenkron’s assumption must have been that the sum of the monitoring costs of the banks and direct agency costs of bank deposits was less than the agency costs of direct finance, perhaps even that the agency costs of direct finance were so high that this was not a genuine alternative at all. According to this approach, supervision seems to be profitable for both the bank and the shareholders, which stand to benefit from higher profits by reducing their agency costs. It will be interesting to analyse which type of contract is induced to exert this control. In other words, are long-term commitments necessary to control firms?
20
AGENCY THEORY AND MONITORING
(c) Monitoring and the definition of contracts between banks and firms ‘Monitoring’ ought to be understood in a broad sense as any form of information-gathering about a firm, its investment prospects and its behaviour. This information is useful because it serves to sort out ‘bad’ projects and/or to punish ‘bad’ behaviour. The literature contains the following justifications for monitoring: • • •
Monitoring of the firm’s profit returns makes it possible to conclude contracts in which the financiers’ claims are related to the firm’s profits.9 Monitoring of the firm’s characteristics, i.e. a creditworthiness test, makes it possible to avoid bad loans.10 Monitoring of the firm’s behaviour during the loan application stage makes it possible to avoid loans to firms that are following too risky an investment strategy.11
In all these examples, the role of monitoring is straightforward. The main distinction is between monitoring that takes place before a contract is agreed, and monitoring of the execution of a contract. The former serves to reduce the proportion of bad loans, the latter serves to improve performance under the agreed contract. Monitoring that takes place before a contract is signed may give rise to an interesting problem of competition for contracts. In the following chapter, when describing the role of the Crédit Mobilier in French industrialisation, this competition will be referred to. In particular, the fight between the Péreires, founders of the bank, and the Rothschilds, as regards the railway companies, illustrate this point perfectly. More specifically, the competition between these two families discouraged monitoring: over 1861–7 the Rothschilds’ first objective was to prevent the Péreires from expanding their financial power. The Rothschilds’ reputation in the market won them preferential access to many companies, including the railways. This competition prevented the Péreires from exerting their control efficiently. Monitoring that takes places after a contract has been concluded stresses the length of the commitment between the firms and the bank that provides the funds. Mayer suggests that financial intermediation may lengthen the investment horizon of firms. He considers a model with two investment periods. There are two types of firms, ‘good’ and ‘bad’. Firms can choose between two investment strategies, short-term and long-term. Both strategies require identical amounts of funds in both investment periods. However, the long-term strategy has a relatively higher expected payoff in the second investment period. Model parameters are specified in such a way that ‘bad’ firms should not
21
THE ROLE OF BANKS IN MONITORING FIRMS
receive funds in either period. However, ex ante, there is no way for financiers to distinguish between ‘bad’ firms and ‘good’ firms. After the first investment period, a partial distinction may be possible because there are two sources of information: costless observation of first-period returns, and costly monitoring. In the absence of monitoring, banks interpret low first-period return realisations as an indication of poor quality, and discontinue financing. Anticipation of such behaviour induces firms to opt for the short-term investment strategy even though the long-term strategy may eventually be more profitable. This theoretical consideration can be illustrated by two examples: France and Germany. The policy of German banks as regards investment was essentially short-term and concerned enterprises where the expected profit was quite high (the mining industry). However, this does not mean that the German bank never undertook risky projects in the late nineteenth century. According to Tilly, 12 these banks used their information advantages and their freedom from liquidity worries to engage in investment strategies that involved high risks as well as high returns. The commitment of a German bank in a risky project induced a more careful control of the performance of the company. This idea will be developed further in the two following chapters when interpreting the empirical work proposed for the Crédit Mobilier. This last bank was quite different as regards the control it could exert on the companies with which it was involved. Despite the fact that it undertook risky projects, its supervisory role was very limited, and sometimes totally ineffective (as in the case of the Compagnie Immobilière). The real question is how monitoring fits into the overall relationship between the intermediary and the firm. Monitoring as a form of informationgathering about the firm is useful only because the information that is collected has consequences for behaviour and resource allocation within the relationship. The notion of financial intermediation as delegated monitoring must encompass such uses of information as well as the act of collecting the information. This consideration brings us back to the idea of incentive contracts presented in the preceding section. Let us recall the basic concept. From the theory of optimal incentive contracts, we know that even noisy monitoring will improve performance under an optimal incentive contract. 13 By combining this result with the procedure of Diamond that was outlined above, we obtain a new version of financial intermediation as delegated monitoring, one that emphasises the incentive effects rather than the sorting effects of monitoring. An interesting point will be to apply it to the financial systems of the late nineteenth century. Monitoring as an element in an incentive contract relationship is not at all that far from the accounts of Gerschenkron and Mayer.
22
AGENCY THEORY AND MONITORING
Consider Gerschenkron’s observation that banks played a key role in the formation of cartels in German industry. Any cartel is subject to the moral hazard problem that the individual firm has an incentive to under-cut and to serve more than its allotted share of the market. Banks that are involved with several firms in the same industry have an incentive to restrain such behaviour, in order to increase the aggregate gross return they can earn from the industry. In the following chapter we compare these considerations on the German case to that of the Crédit Mobilier. So far we have studied how the interdependence between information provision to firms and information collection about firms affects the functioning of financial intermediation as delegated monitoring. It will be interesting to debate how this delegation takes place effectively.
(d) Financial intermediation as a mechanism of commitment In this section the advantages of a close relationship between bank and firms will be described. Mayer (1988) considers that financial intermediation in the form of a close relationship between the bank and the firm constitutes a mechanism of commitment. Mayer suggests that Japanese banks are more willing to engage in corporate rescues than financiers elsewhere because the bank-firm relation in Japan involves a mutual long-term commitment. Some rather striking quantitative evidence on this point is provided by Hoshi et al.14 For a sample of Japanese firms they found that the cost of financial distress was significantly less for firms that had close relations to a ‘main bank’ than for firms that did not. To measure this relationship, they evaluated the degree of correlation between investment and cash flow for firms affiliated regularly with a bank, and for independent firms. Formally, they prove the following: that firms that have close banking ties appear to invest more and perform better than firms that do not have such ties. This proposition is crucial for the work presented here. In the next chapter, which is devoted to the empirical study of the Crédit Mobilier, the validity of this hypothesis will be investigated: as this bank went bankrupt in 1867, if the above affirmation were valid, the share prices of the companies with which the Crédit Mobilier was involved should have fallen relative to other firms; surprisingly, a formal test presented in the next chapter does not support this hypothesis. The question we should ask now is, to what extent should a system of corporate finance based on intermediation through a ‘main bank’ be regarded as internally stable? Mayer seems to believe that the superior performance of such a system is a guarantee of its persistence over time. In
23
THE ROLE OF BANKS IN MONITORING FIRMS
contrast, Gerschenkron (1962) regarded it as being transitory, with firms depending on outside finance through banks only until they have enough inside finance available. The empirical research reported below will test these assertions in order to clarify the financial role played by the Crédit Mobilier in French industrialisation. Similar patterns can also be observed in other countries. Thus Hoshi et al. report that in Japan, too, the larger, more profitable companies availed themselves of the newly developing organised markets to become more independent of their banks. In the long term, there is a certain tendency for firms to emancipate themselves from such a relationship, using markets, competition amongst banks and, above all, reliance on internal rather than external finance. The emancipation of firms from close banking relationships is certainly not costless. Recall the observation of Hoshi et al. that the costs of financial distress in Japan are significantly larger for firms without close banking relations than for firms with close banking relations. For firms that are not in financial distress, Hoshi et al. report a similar observation: hence the emancipation of a firm from a ‘main bank’ relationship is accompanied by a significant increase in the sensitivity of current investment to fluctuations in current earnings and liquidity. Bank loans are less used and/or available to smooth over fluctuations in earnings. From the perspective of Jensen and Meckling (1976) one might argue that internal finance has priority because the agency costs of internal finance are lower than the agency costs of external bank or market finance. This is also the explanation given by Myers and Majluf (1984) in the context of a model of asymmetric information and signalling. Internal finance is taken to have no agency costs (signalling costs, simple inefficiencies) because it represents the use of funds available to the firm itself. External finance has no agency costs because information asymmetries and externalities preclude the attainment of a first-best allocation in the arrangement between the firm and its outside financiers. From this perspective, the Hoshi et al. observation of investment sensitivity to current earnings should be seen as evidence of the agency costs of outside finance: investment projects that might be expected to be profitable under internal finance are deemed to be unprofitable under outside finance when the agency costs of outside finance are added to the mere opportunity costs of funds. The inefficiency in the allocation of funds across the firms that results when investment opportunities are less than perfectly correlated with earnings is nothing but an element in the overall agency cost of outside finance. 1.4 GENERAL CONCLUSION This first analysis describes the relationships that exist between the manager of a firm, its shareholders and the bank which finances the investment
24
AGENCY THEORY AND MONITORING
project, in the context of asymmetric information. As we have shown in the first section, each partner wants to define the debt contract according to his own interest. The manager is willing to find the amount of funds he thinks acceptable to undertake his investment project. For the shareholders and the bank, this commitment is a way to realise profit. However, the latter actors bear different costs. The shareholders give the manager a financial compensation for his work, while the banker supports the interest cost for granting the credit.15 As there is asymmetric information between the economic agents, we must determine which actor(s) is (are) able to control the manager and if necessary punish him when he does not truthfully reveal the necessary information about the productive project. There are two possible candidates: the shareholders of the firm, or the bank which finances the projects. Let us consider first the shareholders. In order to exert this control effectively, the shareholders have to agree unanimously, or at least on a majority basis. On the other side, the bank disposes of a number of tools to control its affiliated companies more efficiently. Our analysis shows the power of banks in screening the investment projects of firms with which they are involved. As pointed out in the survey of the literature, some authors carried out empirical studies of the efficiency of certain financial systems. More precisely, Gerschenkron for the nineteenth century and Hoshi et al. for the recent period have analysed the power of banks. They have stressed the importance of banks in controlling the investment policy of their affiliated companies. My aim is now to apply their methodology to the French case. As I have chosen the same period (1852– 1914), I intend to compare my results to those obtained for Germany and Japan. In particular, I shall investigate whether the power of banks in the French financial system was as strong as in Germany.
25
2 THE CRÉDIT MOBILIER AND THE FRENCH STOCK EXCHANGE 1853–1914 An empirical perspective The fate of the Crédit Mobilier is characteristic not only of the rivalry between financial groups, and of the dangers of speculation on the stock exchange: it highlights an ever present problem of industrial involvement.1 Bouvier (1988:109)
2.1 INTRODUCTION The aim of this work is to gain, through empirical research, an insight into the role of banks in the financing of enterprises during the period 1850– 1913 in France. Some authors (e.g. de Long (1989) and Ramirez (1992)) have already applied this analysis in the United States. The credit which a bank gives to a company enables it to exert an influence over the company. 2 J.Bradford de Long has illustrated the role of investment bankers in his article ‘Did J.P.Morgan’s men add value? A historical perspective on financial capitalism’. In the period of which he writes, financiers possessed strong voices in corporate management. The author illustrates his point of view by focusing on the J.P.Morgan Company. J.P.Morgan and his partners saw themselves, as did other participants in the pre-WWI securities industry, as fulfilling a crucial monitoring and signalling intermediary role between firms and investors, in a world where information about the underlying values of firms and the quality of management was scarce. De Long stresses that investment bankers played an important role, in that they took great care to make sure that the firms they watched had the right managers. This analysis fits our hypothesis that the bank can control the management of a firm and motivate it to reveal truthfully the information about the project it proposes to undertake. Another line of research has been made in the German context. The roles of four banks—the Deutsche, the Diskonte Gesellschaft, the Dresdner, and the Darmstädter—have been studied, producing the same
26
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
conclusion. Bradford de Long and Marco Becht (1992) have used price and dividends series to test the volatility of the German stock market. In particular, they examined whether the role played by the German Großbanken in the pre-WWI stock market might have been the cause of the low comparative volatility of the German stock indices before 1914. These ‘Great Banks’ were founded in deliberate imitation of the French Crédit Mobilier. There has been much discussion about the origin of the early German credit banks, and especially to what extent the Crédit Mobilier served as a model. The first significant joint-stock bank was created in 1848: the Schaaffhausen’schen Bankverein. During the crisis of the 1860s’ recession it faced a liquidity problem. Because of the bank’s importance and the strained atmosphere of the times, normal liquidation procedures did not occur. The Darmstädter was a direct result of the French Crédit Mobilier. It counted among its founders Abraham Oppenheim, who had been co-founder of the Crédit Mobilier. Much of the initial capital of the bank was supplied by the Péreire brothers. These banks had access to considerable capital with which to provide finance to corporate enterprises. They contributed to the intensity of the boom of the 1850s. Despite the crash in 1857 and the depression of the 1860s, most of the banks survived. The early German banks remained permanently tied in with the companies which they promoted, through representation on boards of directors and direct participation in ownership. How this cautious approach can be reconciled, at least in theory, with the role the banks wished to play in the development of industry, is clear from the Schaaffhausen’schen Bankverein for 1852 The management has proceeded from the principle that it is the function of a great banking institution not so much to call great new branches of industry into existence through large scale participation on its own account as to induce the capitalists of the country by the authority of its recommendations, based on thorough investigations, to apply their idle capital to undertakings which, properly planned, corresponding to real needs, and equipped with expert management, offer prospects of reasonable profits. 3 The case of the Crédit Mobilier will be analysed to see if, as in the American and German contexts, bankers possessed strong voices in corporate management. From 1853 to 1913 this bank took part in the financing of firms. In fact, two ‘Crédit Mobiliers’ existed. The first, founded by the Péreires, went bankrupt in 1867. After its bankruptcy it was reorganised, but it went bankrupt again in 1900. It reappeared in 1903 as the
27
THE ROLE OF BANKS IN MONITORING FIRMS
Crédit Mobilier Français. In 1932 it merged with the Banque de l’Union Parisienne. The most interesting banking model was, of course, the first Crédit Mobilier, because of its innovative characteristics and because of the power it had reached in financial markets and among banks. That was certainly one reason why the Péreires faced a certain amount of criticism. The first criticisms arose when they envisaged the possibility of founding a bank capable of issuing money. The Banque de France refused categorically to allow this. Their commitments in various risky sectors (for instance, the property business with the Compagnie Immobilière) contributed to their failure. However, they tried new banking techniques for financing industries. The advances they granted to the companies with which they were involved could be considered as innovative, matching the ideas the Péreires had on finance. To the Crédit Mobilier, the role of a merchant bank was to provide the necessary funds in order to contribute as fully as possible to the industrial development. This idea was new and sometimes unrealistic. The aim of this chapter is to verify whether, as de Long (1989) suggests, a close relationship with a bank is always beneficial for firms, increasing the market value of their shares. In that respect, the place of the Crédit Mobilier in the banking sector will be analysed. De Long maintains that an association with J.P.Morgan facilitated the assessment of company managers’ performance and the careful evaluation of projects. Having J.P. Morgan act as an intermediary partially resolved the principal-agent problem between shareholders and managers. A similar question will be asked about the Crédit Mobilier for the period 1853–1914; namely, ‘did the Crédit Mobilier increase the ratio of market value to book value?’ Attention will then turn to the private participation of the Péreire brothers in the financing of firms. In particular, we will investigate how the Crédit Mobilier may have increased the value of shares by providing liquidity. This question will be of interest at the point when the Crédit Mobilier went bankrupt. Bradford de Long maintains that investment banker representation on boards allowed bankers to monitor the performance of firms’ managers, quickly replace managers whose performance was unsatisfactory, and signal to ultimate investors that a company was well managed and fundamentally sound. If this argument is to be followed, the companies where the Crédit Mobilier was personally involved should have experienced difficulties in 1867, the year of its bankruptcy. In the second section of the chapter I will study the fluctuations of the share prices of the companies involved with the Crédit Mobilier between 1866 and 1868. According to de Long’s argument, these prices should
28
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
have fallen. A formal test will be presented to analyse the validity of this proposition for a sample including affiliated companies and nonaffiliated companies. The last section of this chapter will present the results obtained for a pooled panel including both affiliated and nonaffiliated firms. A comparison between France and Germany will be made. Differences and the analogies will be identified. As the German Großbanken were founded after the bankruptcy of the Crédit Mobilier, their success can be interpreted as a learning effect from the mistakes made by the Péreires. The Germans were more careful to screen projects according to their risk level, to avoid situations similar to that of the Compagnie Immobilière. 2.2 FINANCIAL CAPITALISM AND DATA ON THE CRÉDIT MOBILIER IN FRANCE, 1853–1914 (a) Some historical background about the Crédit Mobilier, the banking system and French industry The interpenetration of finance and industry was intensified by the countermovement of industrialists-turned-bankers. André Koechlin, a scion of the notable Alsacian family, head of Dollfus Mieg et Cie, and founder of the machine shops that produced the first French locomotives, became one of the most daring and successful promoters and investment bankers of the second Empire, both in France and abroad. Emile and Isaac Péreire went from railways to the Crédit Mobilier, and one of their principal associates, Baron Seillière, from a Vosges textile family, also helped to finance several large metallurgical establishments, including Le Creusot and Wendel in France and Krupp in Germany. In part, for a relatively small section of its total operations, the Crédit Mobilier functioned as an ordinary banking enterprise—that is, it accepted deposits, discounted commercial paper and made loans and advances on security. The Crédit Mobilier’s principal function consisted of lending to merchants and industrialists on ‘two-signature papers’ for a period of up to two years. (The Bank of France discounted only three-signature paper with maturities not longer than ninety days.) In an effort to obtain the sources for such loans, the bank proposed that it should be allowed to issue what would have been, in effect, interest-bearing bank notes (bearer bonds, interest bonds). The provision for interest (at 3.65 per cent per year—one centime per 100 francs per day) would have ensured circulation of bonds and if they were made receivable in payment of taxes, as the Péreires proposed, they would have acquired virtual legal tender status. (In some respects the proposal resembled a vast paper money scheme, in as much as the loans granted by the bank as well as their repayment might be in bonds.) To
29
THE ROLE OF BANKS IN MONITORING FIRMS
enhance confidence in the institution and in its attractiveness to potential investors and bondholders, the Péreires also proposed that the French government should guarantee it against losses up to twenty five million francs. In the Péreires’ own words, ‘The new establishment should be considered as a lending and borrowing office [“bureau de prêt et d’emprunts”] where industry will borrow from all capitalists on the most favourable terms through the intermediary of the richest bankers acting as guarantors, because the holders of the bonds, who are the true stockholders of the association, will find an easier and safer investment for their capital without the hazards of individual loans.’4 In the spring of 1853, soon after the Crédit Mobilier began to function, the Péreires and Fould requested an agreement for a ‘Caisse Générale des Sociétés de Crédit Mutuel’, intended as the nucleus of a network of mutual credit societies in various localities and different lines of industry. The Caisse Générale was intended to perform the same or similar functions for small firms and handicrafts that the Crédit Mobilier did for large-scale industry; thus the four-fold programme implicit in their abortive Saint Simonian scheme of 1830 would be complete: commercial credit (Comptoir d’Escompte), industrial credit (Crédit Mobilier), mortgage credit (Crédit Foncier), and mutual credit for small entrepreneurs. However, the Conseil d’Etat, reasserting its prerogatives in such matters, quietly shelved the project. Shortly afterwards, the Corps Legislatif passed a law to permit departments and communes to convert and consolidate their floating debts. The Crédit Mobilier requested permission to exchange its own bonds for the debts of these local government bodies, to have them stamped by an official of the government, made payable through the receivers-general of taxes, and accepted for purchase by the Caisse des Depots et de Consignations (the official government loan office). Such measures would, in effect, have advanced the Mobilier’s bonds to the status of fiduciary money. This was never realised in practice. The power of the Crédit Mobilier had an important effect on financial institutions at that time (P.Dupont Ferrier (1925)). Jean Bouvier, François Furet and Marcel Gillet (1965)5 drew up a balance sheet of the Crédit Mobilier from its foundation until its fall, with the average profit calculated over five years (see Appendix A), which supports this argument.
(b) The financing of enterprises during the period During the period the number of financial intermediaries providing shortterm credit6 varied. Probably there were around 400 institutions on average. It is difficult to gauge the extent of their activity, as only a limited number published their balance sheets. Only the report by the General Assembly together with the balance sheets of those companies with which the banks had relations provide evidence of the amount of credit devoted to the financing of production.
30
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
Statistics on the issue of shares and the interest rate on the financial market give an idea of the fluctuations on an aggregate level. 7 These statistics serve as a reference point for the analysis of the relations between the Crédit Mobilier and the companies with which it was in contact. Nevertheless, operating only with its own capital and deposits, the Crédit Mobilier performed prodigious feats of industrial promotion and financial manipulation. In its first major operation, it subscribed to 40 million francs in the Crédit Foncier’s bonds, with which it maintained close ties in the early years. In its first full year of operation the Crédit Mobilier financed, by making advances and underwriting bond issues, the Midi railway as well as the newly chartered Grand Central and the French Eastern, Mulhousen and Strasburg-Basle railways. The Crédit Mobilier also secured the merger of three short railways in the Loire coal basin with the Grand Central, reorganised the Loire coal industry and promoted the Darmstädter Bank. It played a leading role in the mergers which took place in 1855 in the French Western (railway), and promoted or financed the railways of the Pyrenees, Dauphine, the Ardennes and several shorter lines. It created, with generous government subsidies, the Compagnie Générale Transatlantique (French Line), the Compagnie Générale des Omnibus (Paris), and the Société de l’Hôtel et des Immeubles de la Rue de Rivoli (subsequently the Compagnie Immobilière). The Crédit Mobilier founded by the Péreires played a great role in French industrialisation. In every company financed by this bank, a member of the Crédit Mobilier’s board of directors was represented in the shareholders’ assembly. The annual report of April 1860 is very explicit on this subject: Our company has always considered as a principle of its high commercial morality never to open a subscription, nor recommend a firm, without first having a large proportion of interest in it and its administrators having become associated with it.8 A representative from the Crédit Mobilier was present on the board of directors of each of the companies asking for funds. Table 2.1 gives details of the ownership of the shares of the Immeubles de la Rue de Rivoli in 1854 (subsequently the Compagnie Immobilière) (Aycard, 1867). The par value of capital of the company was 24 million francs, divided in 240,000 shares. The three latter members of this list were subscribers of the Crédit Mobilier. Table 2.1 shows clearly that the affiliated companies belonged to the Crédit Mobilier, unlike the German case or J.P.Morgan company. Let us refer to the question J.Bradford de Long and Marco Becht (1992) have asked in their paper: ‘Did association with the Crédit Mobilier reduce
31
THE ROLE OF BANKS IN MONITORING FIRMS
Table 2.1 Ownership of the shares of the Immeubles de la Rue de Rivoli
Sources: Archives d’Entreprises Privées (General shareholders’ assembly and newspapers) of the Société Immobilière or Immeubles de la Rue de Rivoli, 1857–67
Table 2.2 Advances to companies, 1856–66
the variance of the ratio price to dividends?’ This question is an important one for the Crédit Mobilier. To be more precise in answering it, let us analyse how the Crédit Mobilier used to loan companies money. The Crédit Mobilier granted credit to the companies with which it was involved. Most of the time it gave advances to enable them to undertake their investment. Table 2.2 shows the aggregate advances made to companies between 1856 and 1866.9 Examples of advances being made without any special guarantee include the following: • • • •
In 1856 and 1857, 6 million fr, to the Compagnie Parisienne du Gaz. In 1857, 13 million fr. to the Compagnie Transatlantique. In 1855, 1856 and 1857, 29 million fr. to the Compagnie des Chemins de Fer du Midi. In 1865, out of 53,971,339 francs advanced, 52,080,709 francs went to the Compagnie Immobilière.
32
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
If we refer precisely to the balance sheet figures (see Appendix A), we observe that the advances are part of the investment account. These advances represent at least one third and sometimes half of the total amount of the account. This demonstrates the risky behaviour of the bank, since it did not ask for collateral. Collateral can be useful in dealing with the problem of adverse selection, as discussed in the preceding chapter. As mentioned earlier, the collateral in the debt contract has a dual role: in addition to providing financial cover in the event of bankruptcy of the firm, it acts as a signal transmitted between the parties.10 Any manager is ready to offer collateral if he thinks that the probability of bankruptcy is low. The Crédit Mobilier did not benefit from this information because it never asked for guarantees from its affiliated companies. Several hypotheses can be formulated to explain this practice. The first is that the Crédit Mobilier wanted to attract clients by offering them optimal conditions to finance their projects. Another argument refers to the situation of French industry over the period. Large numbers of companies were new. Many of the firms involved with the bank were in heavy industry (the railways, for example). They needed huge amounts of credit. The Crédit Mobilier had made the choice to invest in these sectors, thinking it could be profitable in the long term. In any case, if the French credit policy is compared to the German policy over the same period, we can see that German banks used similar practice as regards the financing of the productive projects of their companies. As the Crédit Mobilier often owned part of the companies’ equity with which it was involved, the bank was not only an intermediary but an active partner. The question is: did the Crédit Mobilier reliably screen and supervise companies? In other words, as the Crédit Mobilier took a great part in the enterprises where they granted credit, was it possible for the bank to correctly evaluate the risk of its investment? As the Crédit Mobilier gave credit without asking for any collateral, it increased the risk of bankruptcy (see the next section, concerning the fall of the Crédit Mobilier). It could be thought that, for affiliated companies, an association with the Crédit Mobilier did not always reduce the variance of share price towards dividends.11 The scale of the advances to companies emphasises the power of the Crédit Mobilier. This power was important among banks until 1865 (see Appendix A). After this date, the ‘Banque de Savoie affair’ provoked hostility among financial circles and at the Banque de France. Prominent investors such as the Rothschilds wanted to force the Crédit Mobilier out of the banking sector. However, it was the Crédit Mobilier’s position on the financial market and the Péreires’ investments in risky sectors (e.g. the Compagnie Immobilière) that eventually brought about the failure of the establishment. Before 1864 the
33
THE ROLE OF BANKS IN MONITORING FIRMS
Table 2.3 Fluctuations of share prices for affiliated companies between 1855 and 1856
Sources: Archives d’Entreprises Privées (General shareholders’ assembly and newspapers)
Crédit Mobilier had been a victim of strong speculative movements: for investors, the profit realised by the institution made investors believe in its stability. To illustrate this effect, Table 2.3 sets out the values of the shares of some companies involved with the Crédit Mobilier in 1856. These last remarks lead us to ask the same question as de Long and Becht: Did association with the Crédit Mobilier stabilise the share price compared to book value? The speculative movements indicated here cast doubt on this. In the next sections of this chapter, a formal test will be presented to analyse the impact of the Crédit Mobilier bankruptcy on the stability of assets of the companies with which it was involved. The aim is to see if the Crédit Mobilier ever managed to supervise firms properly and thus to stabilise share prices. Table 2.3 suggests that the bank never really achieved this.
(c) The fall of the Crédit Mobilier The fall of the Crédit Mobilier was essentially caused by the attempt to salvage the Société Immobilière. This company’s eagerness to participate fully in the extension of Paris led its managers to undertake huge investments with poor potential. From 1864–5 the Crédit Mobilier had to confront serious difficulties arising from the Banque de France’s decision to increase interest rates. The Crédit Mobilier decided to double its capital, to support advances for the Compagnie Immobilère. But in June 1866 its situation became worst and the fall of the share prices seemed to be the consequences of the depreciation of its portofolio. The Péreires redoubled their efforts to obtain cash. It was said at the time that a loan would be granted to the Crédit Mobilier by the city of Paris, but the deal was never realised. By the end of 1866, the Péreires were contemplating the possibility of a merger between the Mobilier, the Mobiliario Espanol and the Immobilière, with a loan from the Caisse des Consignations (an agency of the government), but they did not succeed. In addition to these difficulties, the Péreires faced a stockholders’ dispute
34
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
concerning the increase of the bank’s capital. The Crédit Mobilier’s true situation could no longer be concealed. In March 1867, for the first time in its existence, the prices of its shares fell below par. The Péreires were forced to resign from both the Crédit Mobilier and the Immobilière. Between 1867 and 1870, little information was available. The Crédit Mobilier went bankrupt, Emperor Napoleon III, who had supported the Péreires, fell and France had to face a war with Germany. In 1871, Baron Haussmann made a deal with the bank for the repayment of its debt. This involved formally dissolving the old company, and immediately forming a new one. The new bank, called the Société de Crédit Mobilier Français, was to take over the assets of the old Crédit Mobilier.12 The impact of this new bank on the financing of French industry was so small that it disappeared from the economic literature. The bankruptcy and loss of resources meant it played only a secondary role.
(d) The reconstitution of the Crédit Mobilier and its final bankruptcy After the fall of the Péreires in 1871, Baron Haussmann, an old member of the board of directors, became President of the new Crédit Mobilier. However, as mentioned previously, the bank had lost its earlier pre-eminent role in the financing of French industry. The first aim of the new bank was to solve the difficulties of the Compagnie Immobilière. The capital of the company was divided in two parts, between Paris and Marseille. Part of this credit came from the Crédit Foncier, which had contributed to the financing of the Compagnie Immobilière: 55 million francs came from the credit of the Paris division and 32 million from Marseille. Floating assets were liquidated. The Crédit Mobilier reduced its holdings of the Compagnie Immobilière’s debt to 51 million francs by selling part of its bonds and issued new assets for an amount of 76 million francs. Half of these assets were bonds with fixed revenue (5 per cent) and half were bonds with variable revenue to the amount of 5 per cent of the capital. The liquidity problems of the Compagnie Immobilière were solved by the partial liquidation of its debt. The Crédit Mobilier became a firm with a par value of only 40 millions francs, undertaking more projects abroad than in France. In 1875, Emile d’Erlanger, another member of the old board of directors (of the first Crédit Mobilier) became President. During this period the bank had relations with the Bank of Brussels and the Spanish government. No balance sheet was published in the reports to the shareholders’ assembly, despite the fact that they had been specifically requested by the shareholders. In August 1876, Baron Emile d’Erlanger submitted his resignation and Charles Wallut, the bank’s Vice-President, became President. The Crédit
35
THE ROLE OF BANKS IN MONITORING FIRMS
Mobilier continued to start new ventures with foreign companies, but little information was published relating to its activities within France. The bank participated in the Italian Ice House Society, as well as contributing to an Algerian government loan. In June 1881, Eugène Péreire, founder of the first Crédit Mobilier, was nominated Administrator of the bank. He proposed to add a bank service to the Compagnie Transatlantique. The Banque Transatlantique was founded with a capital of 50–60 million francs, of which one quarter was to be given immediately to the Company Transatlantique. Twenty thousand Comptoirs Maritimes shares were absorbed by the new bank. The Crédit Mobilier took 10,000 shares in this new society. In August 1881, M.Péreire was appointed to represent the Crédit Mobilier on the syndicate which was formed to issue the Banque Transatlantique shares. The Crédit Mobilier, with the Romanian government, undertook the constitution of a Romanian Crédit Mobilier. Economic relations were developed with the United States and Spain for railway construction. On 4 November 1882, M.Péreire raised the issue of the participation of the Crédit Mobilier in the profits of the Compagnie Transatlantique. He thought that the idea of a combination of capitalisation or repurchase of the right to the profit sharing should be examined. On 3 February 1883, Péreire submitted his resignation. This was followed, on 29 June that year, by the Crédit Mobilier asserting its rights to the benefits of the Compagnie Transatlantique. It was believed that the Crédit Mobilier should have 25 per cent of these profits and the shareholders 8 per cent. Subsequent court proceedings resulted in the Compagnie Immobilière being ordered to pay 367,000 francs to the Crédit Mobilier for 1881 and 1882. In 1886, the Crédit Mobilier was a company with 30 million francs of Table 2.4 Crédit Mobilier’s liquidation account in 1901
Sources: Crédit Mobilier: Procès Verbaux des Assemblées Générales; Procès Verbaux du Conseil d’Administration
36
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
capital. Its financial situation got worse and worse, until it became completely bankrupt in 1902. One consequence was that, in 1898, funds were cut to two main borrowers, causing the Puerto Rican railway to stop and the Rhone railway line to be liquidated. In 1901 the Crédit Mobilier’s liquidation account was as shown in Table 2.4.
(e) The new Crédit Mobilier After the first Crédit Mobilier’s bankruptcy and subsequent reorganisation, the Crédit Mobilier Français had to confront the scepticism of its clients. The strategy that was adopted was ‘to make the Crédit Mobilier Français become an intermediary between the mediumsized business and the public’ (Rapport sur l’exercice, 1903–4). Its capital of 7,525,000 francs was constituted of subscriptions in cash, all the capital of the Office des Rentiers, together with all the Crédit Mobilier capital, namely: • • • •
rights to the Compagnie Generale Transatlantique; the Société Publique Financière capital; the Société Civile Franco-Mexicaine capital; the Puerto Rico business.13
The Crédit Mobilier was anxious to prove that its portfolio was performing and it took part in business which was small and easy to realise. Between 1903 and 1904 the Crédit Mobilier began to develop: its capital reached 10 million francs and, the year after, 25 million francs. During this period the bank had business relations with the large financial house of Thalmann and Co. and had just concluded a South American business deal. The bank was then capable of entering, either alone or with the participation of other companies, more suitable financial operations (e.g. the Serbian national debt). The Crédit Mobilier maintained relationships with provincial banks to guarantee full realisation of share transactions. The report of 1911 states: We have sent more correspondents abroad and the portfolio fluctuation has increased. In this aim we have taken interest in the banks’ institutions…If this policy has your approval, we propose to improve our effort in this direction and to create a greater base of operations.14 From 2 million francs in 1909, the discount portfolio reached 20 million francs in 1910 and 39 million francs the year after. Financial advances to industry reached 12 million francs in 1910. At this point, the Crédit Mobilier succeeded in concluding a contract with the Banque de Paris and
37
THE ROLE OF BANKS IN MONITORING FIRMS
entered the bank consortium. By 1910, after an increase in capital to 60 million francs, the Crédit Mobilier had become a great financial establishment. In 1921 its capital was 100 million francs. The balance sheet, now officially audited, shows the first important variations: few elements of comparison exist between the preceding system and the new one. To illustrate this idea, let us analyse the balance sheet of the new Crédit Mobilier (see Appendix A). The credit side is constituted by monetary funds available in cash and funds resulting from liquidation on the stock exchange. Their use corresponds to financial operations such as bills in portfolio, reports and advances. The deficit account of investment banks consists, in the major part, of lending to companies in a direct relationship with the Crédit Mobilier, in order to guarantee their development. Credit for the assetsportfolio and financial participation during the period 1900–13 amounted to a maximum of 43.4% and a minimum of 10.3%, with the average value 24.6%. For the liabilities, with the exception of the reserves which have already been discussed, the most interesting aspect is the deficit current accounts. The Crédit Mobilier made a distinction between the debit accounts and the deposit account, but this distinction is not very useful from the eligibility point of view. The merchant banks had expanded considerably and as a consequence they experienced an increase in their balance sheets. The Crédit Mobilier’s balance sheet totalled 248 million francs in 1913 against 40 million francs in 1903. The average profit over the period 1904 to 1908 was about 5.3 per cent, and between 1909 to 1913 approximately 12.3 per cent. The distribution of the dividends was stable and from 1910–13 represented 7 per cent. This survey of the Crédit Mobilier demonstrates that the three different establishments of that name were not at all comparable. If we consider the first Crédit Mobilier and the second one, few points of comparison exist, except the fact the formal organisation of the firm was identical and a large number of the shareholders in the Péreires’ bank maintained their participation in the financing of firm’s projects, despite the bankruptcy of 1867. Concerning the relationship between the second Crédit Mobilier and the third the differences are much stronger as the beginning of the twentieth century heralded a new role for the merchant banks which was implied by the new role of the Banque de France. This latter exerted an increasing monitoring role over banks. However, in regard to the bankruptcy of the Crédit Mobilier, we can ask whether the central bank in France really fulfilled its interventionist role as the lender of last resort. Henry Thornton (1802) and Walter Bagehot (1873) developed the key elements of the classical doctrine of the lender of last resort in England.
38
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
They argued that the central bank should intervene for illiquid but solvent banks. Otherwise supporting failed companies would damage healthy competitors and weaken the discipline of a market economy.15 For Thornton, the Central bank’s responsibility in times of panic was to act as lender of last resort, providing liquidity to the market and discounting freely the paper of all solvent banks, but denying aid to insolvent banks no matter how large or important (Humphrey 1975, 1989). In other words, the lender of last resort must prevent illiquid but solvent banks from failing. 16 At the other extreme, Charles Goodhart (1985, 1987) advocates temporary central bank assistance to insolvent banks. He argues that the distinction between illiquidity and insolvency is a myth, since banks requiring lender of last resort support because of ‘illiquidity will in most cases already be under suspicion about…solvency.’ Furthermore ‘because of the difficulty of valuing [the distressed bank’s] assets, a Central Bank will usually have to take a decision on last resort support to meet an immediate liquidity problem when it knows that there is a doubt about solvency, but does not know just how bad the latter position actually is’ (1985:35). Solow (1982) is also sympathetic to assisting insolvent banks. He argues that any bank failure, especially a large one, reduces confidence in the whole system. To prevent a loss of confidence caused by a major bank failure from spreading to the rest of the banking system, the central bank should provide assistance to insolvent banks. However, such a policy creates a moral hazard, as banks respond with greater risk-taking and the public loses its incentive to monitor them. In the event, the Banque de France did not rescue the Crédit Mobilier. Its reasoning followed Thornton’s argument, according to which: It is by no means intended to imply that it would become the Bank of England to relieve every distress which the rashness of country banks may bring upon them: the bank, by doing this, might encourage their improvidence. There seems to be a medium at which a public bank should aim in granting aid to inferior establishments, and which it must often find it difficult to be observed. The relief should neither be so prompt and liberal as to exempt those who misconduct their business from all the natural consequences of their fault, nor as scanty and slow as deeply to involve the general interests. These interests, nevertheless, are sure to be pleaded by every distressed person whose affairs are large, however indifferent or even ruinous may be their state. (Thornton 1802:50)
39
THE ROLE OF BANKS IN MONITORING FIRMS
Hence the non-intervention of the Banque de France was profitable for the whole system, because an insolvent bank disappeared. From an historical and human point of view, a simple study of the main shareholders of the Banque de France sheds further light on this nonintervention. During the period, the Rothschild family, who disliked the Péreires for both personal and professional reasons (and may have been jealous of the power the Crédit Mobilier had obtained in the financial market) had contributed to the decision of the ‘Banque de France’ not to salvage the Crédit Mobilier. The only intervention that had occurred was to prevent a too sudden and huge fall of share prices on the stock market, which would have provoked a financial crisis for the whole economy. As regards this argument, it will be interesting to analyse the impact of the Crédit Mobilier’s bankruptcy on the share prices for both affiliated and nonaffiliated companies. 2.3 DATA AVAILABLE OVER THE PERIOD OF THE TWO CRÉDIT MOBILIERS: COMPANIES INVOLVED AND PRELIMINARY ANALYSIS OF THE DATA This section presents and discusses all the data sets that will be used in the empirical tests explained below. Three points are developed: the first aims to elicit the general context of the French economy and to analyse in more detail the situation of the financial markets; this will constitute the basis of our excess volatility test. The second and third parts describe the data available for companies that were affiliated with the Crédit Mobilier and independent ones; this will support our statistical research relative to the examination of the impact of the Crédit Mobilier’s bankruptcy for both sub-samples and for the whole sample.
(a) General context of the French economy during the period As Bradford de Long and Marco Becht (1992) did for the United States and Germany, I intend to analyse the excess volatility of the French stock market from 1852 until 1914. In this respect some general ideas about the French economy from 1853 to 1914 are given. 17 More specifically, general figures on share prices, market prices, interest rate and dividend are provided. The aggregate data are relatively simple. The money stock was stable. The huge saving rate compared to a reasonable demand maintained interest rates at a moderate level over the period: between 3 per cent and
40
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
4.5 per cent (Loutchitch 1930). The rate of discounting assets with fixed revenue varied between a minimum of 3.22 per cent in 1897 and 1898 and a maximum of 3.73 per cent in 1913. The official discount rate decreased to 2 per cent in 1896 and 1897. Its value was around 3 per cent from 1900 to 1910. This low return on credit gave the banks an incentive to invest abroad. Some explanations are required to explain how the indices are calculated. To construct the industrial, transport and trade series, fifty four separate assets have been taken into consideration for the years 1895 to 1913 and twenty four assets for the years 1870 to 1895. Amongst these were some issued by companies involved with the Crédit Mobilier. From 1895 to 1913 these were: • • • •
Electrometallurgie de Dives; Chemin de Fer PLM, Nord, Orleans, Midi; Compagnie Générale Transatlantique; Société Parisienne d’Eclairage.
From 1870 to 1895 they were: • • •
Chemin de Fer PLM, Nord; Compagnie Générale Transatlantique; Compagnie d’Eclairage par le Gaz.
From this series, the impact of the Crédit Mobilier’s bankruptcy on the level of production can be analysed. If the Crédit Mobilier had been a ‘Great Bank’, the impact of its bankruptcy on affiliated companies could have induced a relative effect for the sectors to which these firms belonged. As, for a large number of sectors (e.g. the railways), the Crédit Mobilier was an active partner in the financing of companies, it will be interesting to present some figures to support the argument. Table 2.5 compares market product prices by sector on an aggregate level. The sectors I have chosen are those in which the Crédit Mobilier Table 2.5 Market product prices by sector
Source: Lenoir (1919:87–8) basis average of the market prices for 1901–10
41
THE ROLE OF BANKS IN MONITORING FIRMS
was interested, such as railways and mining. A simple examination of these figures shows that the situation was very stable between 1867 and 1870. No big change can be observed when the Crédit Mobilier went bankrupt. The bank’s failure does not seem to have had a great impact on the companies. As regards the financial point of view, some general comments on these series can be made. From 1870 to 1872 wars (the Franco-Prussian, the Commune of Paris and a civil war) led to an increase in the interest rate. The rate of 3 per cent increased from 4.17 per cent in 1869 to 5.51 per cent in 1871 and 5.47 per cent in 1872. The yield on bonds went from 4.73 per cent to 5.72 per cent between 1869 and 1872. The dividend yield on stock passed from 5.07 per cent in 1870 to 7.28 per cent in 1872. Prices followed this trend between 1869 and 1872. The year 1873 was one of general crisis, particularly in Germany, Great Britain, and in the United States. France experienced economic difficulties until approximately 1879. Between 1882 and 1884, the interest rate increased, but a new crisis in 1890 led to a remarkable trend towards a reduced discount rate proposed by banks (see Appendix B: 2.18 per cent in 1878, 3 per cent in 1883, 2 per cent in 1896). During this period, the dividend yield decreased regularly, remaining one point below the interest rate on long-term bonds. A slow and almost regular increase of the interest rate characterised the period from 1897 to 1913. Shares did not experience any general marked movement. Goods price fluctuations were small,18 where the author studies the evolution of the general price level between 1809 and 1925 (see Appendix B).
(b) List of the companies affiliated to the Crédit Mobilier Here an analysis similar to that of Bradford de Long and Marco Becht (1992) will be employed. Companies affiliated and not affiliated to the Crédit Mobilier must be taken into consideration in order to answer the two following questions: • •
Did the Crédit Mobilier stabilise the ratio of market value to dividends of affiliated firms? Did the Crédit Mobilier increase the market value compared to book value of affiliated firms?
In addition, I want to study the power of the Crédit Mobilier among the banking sector, its policy towards investment and the importance of an affiliation with the bank as regards credit policy. P.Dupont Ferrier in his book Le Marché Financier sous le Second Empire provides a list of the societies founded by the Péreires between 1853 and 1867. These enterprises were:
42
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
• • • •
• •
The railway enterprises: Compagnie de l’Est, PLM, Compagnie du Midi, Compagnie d’Orleans and Compagnie de l’Ouest. Insurance companies: l’Union. La Compagnie Transatlantique. La Compagnie Immobilière de Paris. This company was very important for the Crédit Mobilier, Appendix C gives an indication of its financial situation in 1867, the year of the bankruptcy of the bank, and describes the market price of shares and dividends of the Compagnie Immobilière between 1855 and 1866. La Compagnie des Omnibus. La Société Parisienne d’Éclairage et de Chauffage par le Gaz.
The bonds issued by the firm were guaranteed by its capital. The received capital during the year could not be greater than the capital realised (Cameron 1991:142). Plenge (1903) gives the data on the Crédit Mobilier’s shares throughout its existence, their value in 1867 (the year when it went bankrupt) and the distribution of dividends (see Appendix A). The new Crédit Mobilier (1903–14) was initially involved with the Société métallurgique de Montbard-Aulnaye, in which the bank was largely represented, and with the Société d’Electro-métallurgie de Dives. The Crédit Mobilier also founded the Société Française des MachinesOutils. In varying degrees, the Crédit Mobilier participated in the Compagnie Parisienne de Distribution d’Électricité and in the reorganisation of the Compagnie Générale des Omnibus de Paris. It contributed to the development of the Société de Raffineries Say and to the expansion of the Société de l’Éclairage Électrique. The bank also concluded contracts abroad. In 1909–10 it participated in an Argentinian government loan whose interest rate was 4.5 per cent and in the Santa Fe business affair (interest rate 6 per cent). In summary, the firms in which the Péreires were interested were the railway companies, l’Union, the Compagnie Transatlantique and the Compagnie Immobilière de Paris. Two societies were involved with both the Péreires and with the Crédit Mobilier Français: the Société des Omnibus de Paris and the Compagnie d’Eclairage et de Chauffage par le Gaz. Firms involved with the new Crédit Mobilier included the Société des Machines Outils, the Société Montbard-Aulnoye, the Electro-metallurgie de Dives and the Raffineries Say. In contrast to the first Crédit Mobilier founded by the Péreires, the second one was not always represented on the boards of firms with which it was involved. In the Raffineries Say, however, the Crédit Mobilier was a financial partner of the company, while with the Société des Machines Outils, the Société Montbard-Aulnoye and the Electro-Metallurgie de Dives the Crédit Mobilier was strongly represented on the boards of
43
THE ROLE OF BANKS IN MONITORING FIRMS
directors, and sometimes also the principal shareholder, as with the Société des Machines Outils. However, the new Crédit Mobilier’s functioning was not very different from that of the Péreires, particularly when it was well represented on boards of directors. The close relations that the Crédit Mobilier maintained with the societies with which it was involved enabled it to be not only a financing establishment but also an active partner in the realisation of projects. As advances were granted to the societies where the Crédit Mobilier was officially represented, the risk of bankruptcy was increased. In this case we can ask the following questions: • • •
Did association with the Crédit Mobilier increase the value of shares compared to book value? Did association with the Crédit Mobilier reduce the variance of the price dividend ratios? Did association with the Crédit Mobilier always facilitate loans and advances to the companies? If so, in what measure?
More specifically, it will be interesting to study share fluctuations when the Crédit Mobilier went bankrupt. As in de Long (1989) I would like to extend this analysis to companies that were not affiliated to the Crédit Mobilier.
(c) List of companies not affiliated to the Crédit Mobilier This data set supports the hypothesis that companies not affiliated to a bank tend to be liquidity constrained (and face a larger gap between external and internal costs of funds) than those which are. Bertrand Gille (1968) provides us with a list of companies which were affiliated with the Crédit Lyonnais for the financing of their projects. These companies were • • • • • • • • • • • •
Alais; Schneider; Marine; L’home; Terrenoire; Denain; Fourchambault; Maubeuge; Givors; Firminy; Franche Comté; Chatillon.
44
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
Lenoir (1919) enables us to complete this list by considering the case of the Houillères du Nord and the textile sector, which was not affiliated to the Crédit Mobilier on the period. Appendix D presents the data available for these companies. The comparison of the results of the test for the two groups of companies will allow us to evaluate the effectiveness of an association with the Crédit Mobilier as opposed to an association with the Crédit Lyonnais (some of the firms, e.g. Marine, Fourchambourg, Firminy and Chatillon, were involved with the bank). Unfortunately, no data were available on firms which were affiliated with neither bank. From the historical data collected here, a test can be constructed to show if: • •
association with the Crédit Mobilier increased the value of shares compared to book value; association with the Crédit Mobilier always facilitated loans and advances to the companies. As advances (without collateral) were granted to the societies where the Crédit Mobilier was officially represented, the risk of bankruptcy increased. 2.4 EXCESS VOLATILITY TESTS
As in de Long and Becht’s paper, an analysis can be undertaken to test for excess volatility. More specifically, we can evaluate volatility of the prices relative to trend. Let us recall the idea of such of a test. Shiller’s19 first key insight was that the level of the stock market should be a forecast of the ex post perfect foresight fundamental. An investor who buys and holds, and pays less than the ex post fundamental, receives a supernormal return. Arbitrage, therefore, should push prices to an efficient forecast of the perfect foresight fundamental. If prices are too volatile to trend, investors can make better forecasts of ex post fundamentals by taking as their forecast some linear combination of the market price and the trend, and betting that returns will be low whenever the market price is above the trend. How can such a test be expressed more formally?
(a) Presentation of the test Following de Long and Becht’s argument, let us call p* the fundamental price, which can be expressed as: where p is the stock index divided by the (2.1)
45
THE ROLE OF BANKS IN MONITORING FIRMS
Figure 2.1 Prices, dividends, and perfect-foresight fundamental sample 1870–1914 Legend Lnrp* represents the natural log of the perfect foresight fundamental, Lnrp is the log of the real stock price and Lnrd for the real dividend
price level, d represents the dividends divided by the price level and r is chosen to represent either the an arbitrary real required rate of return or a risk premium paid on assets plus the risk-free rate minus the inflation rate. Here r is similar to the value taken by Becht and de Long (8 per cent) to facilitate the interpretation of the result. Other values for r between 5 per cent and 10 per cent would not have changed the shape of the graph in Figure 2.1. In the French case I have calculated the fundamental price using Dessirier’s series for the share prices and the dividend and Lenoir’s series for the general level of prices. All the series constructed are deflated by the GNP deflator from Lenoir. The same methodology used by Bradford de Long and Becht (1992) will be followed. As the general price series, the share price series and the dividends series do not refer to the basic year in the construction of the indices, normalisation of all these prices is necessary to avoid any bias problems in the test. Moreover, the dividend series presented in Dessirier’s paper was not used as such. We evaluated the ratio as dividend.100/share price. This series corresponds to a constant term times the ratio of dividend index to price index. After having normalised the dividend, a calculation of the volatility ratio, that is the price dividend ratio, is made in order to construct the test. Figure 2.1 provides individual surveys of the behaviour of prices, dividends, and perfect foresight fundamentals which are obtained by computing equation
46
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
(2.1). They plot for the period the logs levels of prices and the log ex post perfect-foresight fundamental. Shiller’s test says that: pt = E(p*t | It)
(2.2)
var pt < var p*t.
(2.3)
This formulation was the subject of much controversy because the definition of p* is said to be impossible. If so, investors will take t advantage of this to buy when the price is low, and sell when it is high. The basic idea of solving this problem is to consider a naive forecast noted p equal to a constant times dividends. The problem can be N reformulated and expressed by: var (ln pt – ln pN) < var (ln p*t – ln pN)
(2.4)
where pN=αd and α is a constant. Dessirier’s series and Levy-Leboyer’s series enable us to construct such a test. The question is then: are the conclusions obtained for Germany comparable with those we will obtain for France? To answer this question, let us consider the result represented by Figure 2.1 and that of de Long and Becht.20 Before discussing the similarities and the differences between these two graphs some explanations about the underlined data are worth making. The share prices series is taken from Dessirier (1928b) and presents some odd elements. The first remark is that the figures reported in the article are flat and the financial terms used to explain the data are confusing. For example, the values in column 2 of the share prices series retrace the dividend for assets with variable revenue. The French term used by Dessirier is ‘revenue des actifs à revenu variable’. For the description given in text it is clear that the two terms (‘revenu des actifs à revenu variable’ and ‘dividend’) reflect the same parameters. What is more ambiguous is on what basis the author built this series. Therefore the interpretation of the whole test is relative to the underlying data. The second argument is related to the figures themselves. When reading Dessirier’s data the symbol ‘is found in place of numbers. The problem is whether this a representation of unavailable data, or did the author choose not to report it as it was identical to the preceding value? Dessirier wrote four papers in three years and proceeded the same way each time. When a number was non available, he clearly stated this and when, for two subsequent years, numbers were identical, he always used the same symbol’. This clarification being made, arguments can be advanced to explain the existence or the absence of volatility in France over the period.
47
THE ROLE OF BANKS IN MONITORING FIRMS
As in the German case the fundamentals follow a trend. From 1890 to 1913 the German and the French results are quite comparable: dividends, prices and perfect foresight fundamentals follow a similar increase. Hence a test on excess volatility in this period would probably give similar results to those obtained by de Long and Becht. The second remark about the graph is that for France, between 1870 and 1890, the price/dividend ratio was increasing. How can we explain these differences? In 1882 France suffered a banking crash due to the failure of the Banque de l’Union Générale. All industrial activity decreased between 1881–4 and 1885–6, with some variations according to the sectors and the firms. There was then a slight increase, which was interrupted during 1887–9 due to the bankruptcies of the Compagnie du Canal de Panama and the Comptoir de l’Escompte. What is more interesting to analyse is the behaviour of the Banque de l’Union Parisienne just before the crash, that is at the end of the summer of 1881 and afterwards. Prior to this date, a revision of all their policy towards industry took place. They suspended part of their participation in the creation of enterprises. They decreased the supply of credit in the short and medium term. The crisis, which the bank anticipated, resulted in a decrease in their profit, which was at the origin of their defensive attitude. This situation became more pronounced during the second quarter of 1881. The price of money increased and most of the important investors and lenders followed a deflationary policy. According to Bertrand Gille in La Banque et le Crédit en France de 1815 à 1848, the disparity that was essentially responsible for the crises was an ‘excess of investment’21 inducing a relative gap in the ‘circulating capital’22 and hence the capital of firms. ‘The investment chains provoke the excess and the break in the equilibrium’.23 The crisis happens when ‘the investment ceases’ and this issue is unavoidable: The investments cease because the accumulation of capital is destroyed and there is no longer any liquidity; they cease because the scarcity of circulating money may increase its rate of interest; because certain projects are revealed to be bad or speculative.24 Certain bankers of the first half of the nineteenth century (and more later also such as Laffite, or, more importantly, the Péreires) thought and said that the progress in the banking system, in feeding payment and credit, enabled growth and avoided ‘commercial crises’. The facts contradicted their optimism. The promotion of banking and the modern form of credit did not prevent cycles. They furnished, on the contrary, a source of new turmoil: the rise of credit expansion and the confirmation of depression due to the increase of banking operations.
48
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
The Großbanken and the Crédit Mobilier were different. Several arguments can be adduced to explain these differences: •
•
•
The first one concerns the effective power of each bank. In particular the Crédit Mobilier and the Großbanken did not have similar powers with regard to screening companies’ projects. The German Great Banks were more cautious and evaluated the level of risk for any potential investment project. The second deals with the ability of the Crédit Mobilier to supervise the managers of the firms which whom it was involved. The differences described in the text between 1870–95 on the one side and 1895–1914 on the other side give a hint of the different policy which had been exerted by the German banks and the French Crédit Mobilier. The third refers to the economic situation in France during the period and, more specifically, the difficulties faced by the banking system because of its innovative character as evidenced by Saint Simon’s theory.
(b) Volatility ratios Presentation of the results Table 2.6 presents summary statistics on the volatility of the French stock market. For comparative purposes, it also reports similar statistics for the United States and Germany. If actual prices were rational estimates of fundamentals, they should exhibit less volatility relative to some ‘naive’ forecast than do the ex post perfect. As the values are Table 2.6 Volatility tests for the French, German and American stock markets between 1870 and 1914
49
THE ROLE OF BANKS IN MONITORING FIRMS
extremely flat before 1895 (some values are identical for subsequent years) and reported yearly after 1895, the ratios are evaluated for the whole sample and for 1895–1914. The first line of Table 2.6 shows the volatility of log prices (p) around the log of the perfect foresight fundamental (p*–p) calculated using an 8 per cent per year real discount. The second line shows the volatility of the log price-dividend ratio (p–d). The third line shows the volatility of the log ex post perfect foresight fundamental to dividend ratio (p*–d). If prices are more volatile relative to ‘naive’ forecasts than perfect foresight fundamentals are, relative to the naive forecast, investors could and should have constructed a better forecast: a weighted average of the market price and the naive forecast would have generated smaller forecast errors. Thus the values in the second line of the table, the volatility of the price-dividend ratio, should under the efficient market hypothesis be smaller than the values in the third line, the volatility of ex post perfect foresight fundamental. For France it is true after 1895 as for Germany. Between 1870 and 1914 the situation is more similar to the United States. Line four reports this volatility ratio. Line five of the table calculates another volatility ratio. The volatility of the perfect foresight fundamental about the actual price should be less than the volatility of the perfect foresight fundamental about the naive forecast. For both periods the price and the dividend are almost exactly equally good forecasts of the perfect foresight fundamentals. A final implication of the efficient markets hypothesis is that the two ratios of lines four and five—the sum reported in line six—should add up to one. If not (as is the case of France for the whole sample, and the United States) the log difference between price and the perfect foresight fundamental (p–p*) is correlated with the log price-dividend ratio (p–d). Profits could have been earned by trading on this correlation of the price-dividend ratio and value relative to price. Surprisingly, between 1895 and 1914 the situation of France is similar to that of Germany. The analysis of the results obtained for France leads us to the following conclusions: 1
Prices are volatile relative to dividend over the period 1870–1914. Prices are much more volatile than perfect foresight fundamentals relative to the naive forecast. Thus, tests based on market volatility ratios shows traces of excess volatility in the French stock market before the First World War. However, between 1895 and 1914 volatility tests present strong similarities between France and Germany 2 The situation described by the French stock market is much more similar to that of the United States for the whole sample than that of Germany. The symbol ‘ used in Dessirier’s data should lead to a bias which gives opposite results for the whole sample and the period 1895–1914.
50
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
Interpretation of the results How can we explain the similarity and the differences between France, Germany and the United States? We propose to give an interpretation of these results by explicating the monitoring role played by the banks and their relations with the stock market. The monitoring role of banks in France, Germany and the United States According to Becht and de Long, ‘the key position of the Großbanken in the German economy is linked to the absence of excess volatility. The Großbanken dominated pre-World War I financial markets in Germany.’25 The system involved well-developed analytical abilities, strong views regarding the proper levels of stock prices and an incentive to moderate fluctuations away from fundamentals. Vincent P.Carosso and Richard Sylla, when examining the role of J.P.Morgan said: Bankers generally undertook non financial responsibilities out of necessity—to protect their own interest and those of the investors to whom they sold the companies’ securities.26 Their role in the affairs of the corporations they sponsored sometimes required them to assume entrepreneurial responsibilities they generally would have preferred to avoid. The fact that they were capable and stood ready to undertake these tasks when the occasion demanded it gave the private bankers immense influence far beyond the sums of money they themselves commanded. In the United States these private bankers were largely responsible for facilitating the rise of giant enterprises, the consolidation of industries and the organisation of the earliest multinational companies, but did not monitor as thoroughly as the Großbanken. This strong monitoring role in the case of Germany explains the absence of volatility on the German stock market: the fluctuations of share price were reduced over the period. However, the stock market price seems to imply that this role was less efficient in the French case from 1870–95. The failure of the Crédit Mobilier and the bankruptcy of the Banque de l’Union Parisienne in 1880 can partly explain the high volatility ratios over the period. The banks and the Stock Market According to Becht and de Long, the Großbanken, as well as trying to create value for the investors to whom they sold newly issued securities,
51
THE ROLE OF BANKS IN MONITORING FIRMS
sought also to support the prices of the securities they issued. For Riesser, this support was not an attempt to produce an artificial market or to swindle the public. These authors wonder if the Großbanken had sufficient resources to keep asset prices close to their own internal estimates of fundamentals. As the Großbanken were by far the largest actors in the German economy, an affirmative answer can be given to this question. In the United States the situation was similar to that in Germany: the banks occupied a powerful position. To illustrate this argument, let us just quote Karl Erich Born (1983:93): What probably explains their success is that they had collaborated in the formation of big trusts at the end of 19th and the beginning of the 20th century thereby not only gaining opportunities for control and influence in industry and transport but also making considerable profit. Vincent P.Carosso and Richard Sylla (1991) added: Their chief contribution in the industrial development which was the ultimate basis of their strength and influence, was their access to the world’s principal source of capital. In France such a situation did not exist except for the ‘Haute Banque’ and more specifically the Rothschilds, whose aims was to fight the Péreires. The Rothschilds found the Péreires’ investment policy too risky to consider them as real bankers. On the other hand, they saw them as competitors they had to force out of the market. Evidence of this lack of supervision will be provided in the following section by analysing the impact of the Crédit Mobilier’s bankruptcy on the financial market.
(c) Conclusion It appears that the Crédit Mobilier benefited from a powerful position up until 1862–3. Even before the bankruptcy, however, it faced liquidity difficulties. Its place on the financial market was delicate, as most of the bankers of the period were hostile to the Péreires. The Péreires wanted to be innovative as regards credit policy: in particular they presented themselves as promotors of firms, whose principal aim was to provide liquidity to industry. But the control they exerted on firms was limited. When comparing the Crédit Mobilier with the German Großbanken, Engberg (1981:35) states that: The Crédit Mobilier never became a ‘universal’ bank as did the German banks soon after their organisation. The Crédit Mobilier
52
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
represented a bold and imaginative idea; it was designed to effect revolutionary changes on a grand scale of entire railway, industrial and credit systems. The German credit banks were called into existence to meet new and enormous demands for ‘finance’ in general. The Crédit Mobilier was created to implement change; the early German credit banks were created in response to changes of the society that were taking place. 2.5 THE EFFECT OF THE BANKRUPTCY OF THE CRÉDIT MOBILIER ON THE COMPANIES WITH WHICH IT WAS INVOLVED E.Nouette-Delorme (1865) pointed out the evolution of the prices of shares and bonds of these societies just before the bankruptcy of the bank. I have chosen to add to this analysis the evolution of the share prices in the year immediately following the fall of the Crédit Mobilier to answer the following question: what happened to the companies with which the Crédit Mobilier was involved between 1867 and 1869? Because of the bankruptcy of the Crédit Mobilier, the share prices of these companies should have fallen relatively to the share prices of other firms. Table 2.7, however, which reproduces the average asset prices for affiliated firms, shows the opposite. Table 2.7 suggests that the railways suffered least from the crisis of the Crédit Mobilier. The Rothschilds were a partner of one of the railway companies and came to the assistance of the railway companies during this difficult period. The importance of railways in the French economy, and the low level of competition that existed between the companies in the period, probably provided stability for these companies. Moreover, monthly share prices for affiliated companies between 1866 and 1868 show that the companies did not underperform in the market. In what follows, the impact of the bankruptcy will be analysed in three ways: 1 2 3
the effect of price level and dividend on share prices for the Crédit Mobilier’s firms; the correlation between the failure of the bank and the fluctuation of share prices for the affiliated companies; an evaluation of the association with the Crédit Mobilier between 1866 and 1868.
53
THE ROLE OF BANKS IN MONITORING FIRMS
Table 2.7 Average asset prices for affiliated firms
Sources: Archives d’Entreprises Privées (General shareholders’ assembly and newspapers); Société Immobilière or Immeubles de la Rue de Rivoli, 1857–67; Compagnie Générale Transatlantique, 1854–67; Compagnie Générale des Omnibus de Paris 1855–1914; Compagnie Parisienne d’Eclairage et de Chauffage par le Gaz, 1855– 1914. Notes a When the share price increased this is indicated by a minus sign (i.e. negative loss). * This must be considered as the loss per share before the reorganisation of the Crédit Mobilier and the issue of the new shares. After the fall of the Crédit Mobilier the Cie Immobilière went bankrupt: so the same comments as those made for the Crédit Mobilier apply in that context. ** After the bankruptcy of the Crédit Mobilier the Compagnie Transatlantique chose to reduce its partnership with this bank. For example, in 1864, the amount due to the Crédit Mobilier was 10,126,461.08 francs. The company was supposed to give 25 francs of dividends deductable from a reserve fund in case of insufficient profit over the year. In 1867 the dividend was only 15 francs per share and the Crédit Mobilier became a debtor of the company for an amount of 4,514,610.78 francs. In 1868, M.Dollfuss became President of the Compagnie Transatlantique; M.Péreire had to resign.
(a) The general influence of the Crédit Mobilier’s bankruptcy The aim of this section is to see if there was any marked effect on price level and dividends due to the bank’s failure. This effect will be examined by measuring changes in share prices for affiliated companies, correlated with price level and dividend. The ordinary least square regression is expressed by: ln(pt / pt–1) = constant + α ln(pyt / pyt–1) + β (divt / divt–1) 54
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
Table 2.8 Relative impact of the bankruptcy of the Crédit Mobilier
where py stands for the general level of prices for the whole economy, divt stands for general dividend series and pt represents the share prices for each of the affiliated companies. The sample considered is 1861–75. Table 2.8 does not reveal any significant figures regarding any influence of the failure that are distinct from the effects of price level or dividends. Dummy variables are introduced to concentrate the analysis on this specific aspect. The table provides us with very low coefficients for both the GNP deflator and dividends. If the figures produces by the second regression are much higher, they remain insignificant as most of them are close to 0. The regression equations which are going to be considered are: ln(pt / pt–1) = constant + α ln(pyt / pyt–1) + β ln(divt / divt–1) + γd18
67
where d1867=1 in 1867 and 0 otherwise. This test is carried out for each of the companies separately and over the same sample, that is 1861–75. The results are given in Table 2.9 (standard errors are in brackets). The statistical shown in Table 2.9 provide no more explanatory power than the preceding ones, which is quite surprising as the focus of the regression is the bankruptcy. The coefficients are not significant except for two companies: the Compagnie Immobilière and the Compagnie Transatlantique. Both these firms were confronted with financial difficulties, for different reasons. For the former, its simultaneous failure with that of the Crédit Mobilier (and causing it) explains the correlation between the different parameters. In the Compagnie Transatlantique’s case, the reason behind the failure is found in the struggle between the Compagnie, which wanted to sever all relations with the bank, and the Péreires, who wanted to maintain their position on the board of directors.
55
Table 2.9 Relative impact of the bankruptcy of the Crédit Mobilier including a dummy variable d1867
Table 2.10 Relative impact of the bankruptcy of the Crédit Mobilier including a dummy variable d1867/68
THE ROLE OF BANKS IN MONITORING FIRMS
To complete the argument the same regression can be carried out for d1867/ considering that d1867/68=1 in 1867 and 1868, 0 otherwise. The results are given in Table 2.10. Table 2.10 confirms the results obtained in the preceding test. Except for a relatively small number of firms such as the Compagnie Immobilière and the Compagnie Transatlantique, the bankruptcy did not lead to any great perturbation that could not be explained by the price level (e.g. the GNP deflator coefficient) or the financial market (e.g. the dividend coefficient). In what follows, these two parameters will be avoided: we concentrate on the correlation between the fluctuations of the share prices for affiliated companies and the bankruptcy, and we evaluate the validity of an association with the Crédit Mobilier. 68
(b) The impact on shares for affiliated companies To analyse the magnitude of the fluctuations of the share prices, we analyse two sub-samples: the first includes the railways companies, which suffered less from the failure (in total six companies); the second includes all other firms which were confronted with financial problems because of the policy of the Crédit Mobilier toward credit. Let us consider the six railway companies. An ordinary least square regression can be run by calculating ln(p /p ) for each of the companies t t–1 between 1862 and 1867. The equation of the regression will be then: ln (pt / pt–1) = c + βd1867 where d1867=1 in 1867 and 0 otherwise. The same regression can be done for d considering that d =1 in 1867/68 1867/68 1867 and 1868, 0 otherwise. The dependent variables are the share prices of the Compagnie de l’Est. Table 2.11 Regression for d1867
58
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
Table 2.12 Regression for d1867/68
The aim of the test is to analyse the impact of the bankruptcy of the Crédit Mobilier on the share fluctuations. Specifically we intend to test the null hypothesis: H0 β=0: ‘The bankruptcy of the Crédit Mobilier has no influence on the share prices’ against H1 β<0 The results of the test are given in Tables 2.11 and 2.12. These tables lead us to the following comments. • •
•
The coefficient β in both regressions is close to zero and positive. A negative sign was expected. Except in two cases in the first regression (North and West), and three cases in the second one (East, North and West), the significance level is greater than 0.60. All t-statistics are insignificant.
Some particular comments can been made as regards the companies themselves. For the Compagnie Immobilière and the Compagnie Transatlantique, the arguments are quite different as both firms tried to sever links with the Crédit Mobilier after its bankruptcy. In these two cases only, the bankruptcy of the Crédit Mobilier induced a decrease in share prices. If we examine the coefficient in the regression, we see that the Compagnie Immobilière has a negative coefficient because of its simultaneous bankruptcy with that of the Crédit Mobilier. For the Compagnie Transatlantique the coefficient is positive, but greater than that obtained for the railway companies, as this firm was liberated by the bankruptcy of the Crédit Mobilier. The T-statistic is larger compared to the railway companies but remains insignificant. The results for the Omnibus Compagnie and the
59
THE ROLE OF BANKS IN MONITORING FIRMS
two insurance companies are very similar to the those obtained for the railways. The only comment that can be made concerns the size of the regression coefficients. For the Omnibus Compagnie and for l’Union Life it is of the same order as the railways. But as they are all positive the conclusion is identical: H0 is not rejected by the data and there is no evidence that the bankruptcy of the Crédit Mobilier had any influence on the share prices.
(c) The evaluation of the association with the Crédit Mobilier between 1866 and 1868 The last point of our analysis will be to analyse the correlation between the fluctuations of the share prices for affiliated companies and the variations in asset prices for the Crédit Mobilier. With that aim in mind, monthly data from 1866–8 are used to run two regressions. The first is expressed as: Log(pt) = constant + β Log(pt CM). The second has as its objective the evaluation of the interaction of the change in asset price of the Crédit Mobilier on shares of affiliated companies. The regression is given by: Log(pt / pt–1) = constant + β Log(pt CM / pt–1 CM). All these tests are made separately for each year. A final regression will consider the sample as whole for 1866–8. The results are presented as a set of two tables representing the figures obtained for each of the two regressions each. Standard errors are in brackets. Some comments are worth making on these regressions. For the year 1866 the sign of the constant term for the Compagnie Immobilière changes between the two regressions: in the first regression it is positive and in the second one it is negative. What interpretation can we give to this phenomenon? In the upper tables the t-statistics are uninterpretable as, under the null hypothesis, the disturbances are log(p ) and are not independently t identically distributed or even probably stationary. This set of results as a whole leads to the following interpretation: except for the Compagnie Immobilière (where the β coefficient is strongly positive, as shown by the first table of each year) and the Compagnie Transatlantique, nothing serious happened when the Crédit Mobilier went bankrupt. What can be said when considering the sample as a whole? The regression figures are presented in Table 2.16 (standard errors are in brackets). Tables 2.13 to 2.16 reinforce the preceding arguments.
60
Table 2.13 Results for 1866
Table 2.14 Results for 1867
Table 2.15 Results for 1868
Table 2.16 Results for 1866–8
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
(d) Summary and general interpretation The preceding econometric results, as regards the impact of the bankruptcy for affiliated firms, show that the Crédit Mobilier provided liquidity to firms but did not seem to be supervising them (as witnessed by the bankruptcy of the Compagnie Immobilière). Since the Crédit Mobilier itself went bankrupt, its seal of approval was probably not worth as much as that of J.P.Morgan & Co or the Großbanken. According to Bradford de Long and Becht’s paper, ‘the Großbanken had the power to carry out whatever threats they might make to or conditions they might lay down for corporation managers. Their organisations controlled a large part of voting stock—and a still larger part of the stock voted’. In spite of the fact that the Crédit Mobilier often owned the biggest part of the shares of firms, in which it was involved, such threats were not applied. The authors add: ‘The key to the profitability of the Großbanken lay in [their] ability to securitize [their] credit commitments, and to place the bonds and shares of the corporations [they] sponsored at high prices…. The Großbanken thus sought to enhance their issue credit,…as their reputation as honest brokers selling sound securities, who would monitor corporations to protect both their own investment and the investments of those who had bought on the bank’s recommendation’. Thus, the formal tests prove that de Long’s proposition does not apply in the case of the Crédit Mobilier. According to this author ‘J.P.Morgan and Co’s approval of an issue had become…a large factor which inspires confidence in the investor and leads him to purchase’. Such was not the case with the Crédit Mobilier, whose active participation of this bank in a company did not ‘add value’ to the share prices. The inverse proposition seemed to be closer to the reality in this case. An analysis of the place of the Crédit Mobilier on the financial market enables us to give an interpretation of this result. From an economic point of view, this bank was at the height of its power on the Stock Exchange at the beginning of the 1860s. Around 1863, the innovative component of its investments made it lose part of its force in financial matters. In particular, the Crédit Mobilier was subject to speculative movements which contributed partly to its new fragility. However, it was its significant participation in the financing of firms such as the Compagnie Immobilière that was the origin of the fall of the Crédit Mobilier. 2.6 BANKRUPTCY EFFECT ON THE COMPANIES NOT AFFILIATED TO THE CRÉDIT MOBILIER To complete the analysis, it will be interesting to extend the study to the firms that were independent of the Crédit Mobilier. As in the preceding section,
65
THE ROLE OF BANKS IN MONITORING FIRMS
formal tests can be constructed to analyse the fluctuations of the share prices for the companies which were not affiliated to the Crédit Mobilier. Let us consider, then, the twelve steel companies and the Lenoir series to study the investment constraint, if any, of these companies. A methodology identical to that followed in the preceding section is used for the non-affiliated sample. In what follows, the impact of the bankruptcy will be analysed, taking into account the general effect of the bankruptcy on the financial market and the correlation between the failure of the bank and the fluctuation of share prices for affiliated companies.
(a) The general influence of the Crédit Mobilier’s bankruptcy for non-affiliated firms The aim of this section is to see if there was any marked influence on price level and on the financial market due to the failure of the bank. This effect will be examined in two stages: • •
first, in the absolute, by evaluating the effect of price level and dividend on share price for non-affiliated firms; second, by measuring the provoked modification in share prices for non-affiliated companies in correlation with the change in price level and dividend.
The ordinary least square regression is expressed by: ln(pt / pt–1) = constant + α ln(pyt / pyt–1) + β (divt / divt–1) where py stands for the general level of prices for the whole economy, pt represents the share prices for each of the affiliated companies and div is the general dividend series. The sample considered is 1861–75. The results are given by Table 2.17 (standard errors are in brackets). As previously, Table 2.17 does not produce any significant figures regarding any influence of the failure distinct from the effects of price level and dividends (the coefficient obtained for the logarithm of py and dividends being close to 0 or negative) dummy variables are introduced to concentrate the analysis on this specific aspect. The regression equations which are going to be considered are: ln(pt / pt–1) = constant + α ln py + β ln(div) + d18
67
ln(pt / pt–1) = constant + α ln(pyt / pyt–1) + β ln(divt / divt–1) + γd1867 where d1867=1 in 1867 and 0 otherwise.
66
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
Table 2.17 Relative impact of the bankruptcy of the Crédit Mobilier
This test is done for each company separately and over the same sample, that is 1861–75. The results are given by Table 2.18 (standard errors are in brackets). The statistical results given by Table 2.18 do not add any information to the preceding tables. To complete the argument the same regression can be done for d1867/68 considering that d1867/68=1 in 1867 and 1868, 0 otherwise. The results are given by Table 2.19. This comparison shows that there was no general increase in share prices in 1867 or 1868 in response to the Crédit Mobilier’s bankruptcy. For more than half of the sample the coefficients for d1867 and d1867/68 are significant and positive, while for the rest of the sample the same coefficients are slightly negative and thus not really significant. On the other hand the argument coefficients (ln(py) and ln(div) for the first regression, ln(pyt/pyt–1) and ln(divt/ div t–1 ) for the second one) are small and not significant. This shows (unsurprisingly, given the results for affiliated companies) that the Crédit Mobilier’s bankruptcy did not cause a general panic. In what follows, these two parameters will be avoided: we concentrate on the correlation between the fluctuations of the share prices for affiliated companies and the bankruptcy, and we evaluate the validity of an association with the Crédit Mobilier.
(b) Share price fluctuations while the Crédit Mobilier went bankrupt To analyse the magnitude of the share price fluctuations, an ordinary least square regression can be run by calculating ln(p t /p t–1 ) for 67
Table 2.18 Relative impact of the bankruptcy of the Crédit Mobilier including a dummy variable d1867
Table 2.19 Relative impact of the bankruptcy of the Crédit Mobilier including a dummy variable d1867/68
THE ROLE OF BANKS IN MONITORING FIRMS
each company between 1862 and 1867. The equation of the regression will be then: ln (p t /
p ) = c + βd t–1
1867
where d1867=1 in 1867 and 0 otherwise. The same regression can be done for d1867/1868 considering that d1867/68 =1 in 1867 and 1868 and 0 otherwise. The results are presented in Tables 2.20 and 2.21. These tables lead us to the following comments. Table 2.20 Results for d1867
Table 2.21 Results for d1867/68
70
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
The coefficient β is close to zero and positive except in the cases of Chatillon, Denain and Allais. These results must be put in perspective since these three companies had encountered difficulties during the recession of the steel industry at the beginning of the 1860s and around 1870. So the size and the sign for these three companies can be explained by factors other than a lack of association with the bank. In general, a negative sign would be expected if the failure of the Crédit Mobilier caused financial disruption. In particular, if the Crédit Mobilier were truly a great bank, as it is presented in the historical and economic literature, its bankruptcy should have had an influence on the other banks (like the Crédit Lyonnais, which was the financial institution of the companies quoted above) and subsequently on the companies affiliated to these banks. The results of the test give no support to this hypothesis. All t-statistics are low, and none of these coefficients are significantly different from zero. The conclusion is then that H is not rejected by the data and there is no 0 evidence that the bankruptcy of the Crédit Mobilier had any influence on the share prices for companies not involved with the bank.
(c) Conclusion The Crédit Mobilier did not cause general distress (unsurprisingly, given the results for affiliated firms). There is only weak evidence of a general boom which could mash the effect of the Crédit Mobilier’s bankruptcy. The results for affiliated and non-affiliated firms are very similar. This suggests that de Long’s proposition does not apply to the Crédit Mobilier. 2.7 COMPARATIVE ANALYSIS OF THE RESULTS OBTAINED FOR AFFILIATED AND NON-AFFILIATED COMPANIES In the preceding sections, we present regressions run for two samples: affiliated and non affiliated companies. The most interesting results are given by studying the fluctuations of share prices when the Crédit Mobilier went bankrupt. The aim of this section is to compare the results obtained for the two samples. Let us first recall the equations of regression: ln(pt / pt–1) = c + βd1867 or ln(pt / pt–1) = c + βd1867/68 where d 1867=1 in 1867 and 0 otherwise, d1867/68=1 in 1867 and 1868, 0 otherwise.
71
THE ROLE OF BANKS IN MONITORING FIRMS
The hypothesis tested is H0 β=0: ‘The bankruptcy of the Crédit Mobilier has no influence on the share prices’ against H1 β different from 0. Let us first compare the results obtained in the two cases for d1867 and d1867/68 as regards the coefficient β. Let us recall these values for d1867. For the railways companies we obtain 0.018<β<0.080 and all are positive. For all the affiliated companies except the railways the values are given by:
–0.044<β<0.322 and the magnitude of the coefficients is more important. For the non-affiliated companies (excluding Allais) we have:
–0.1125<β 0.097. For d1867/68 the railways coefficients ß are described by
0.014<β<0.092 and are all positive. For all the affiliated companies except the railways the values are given by:
–0.09<β<0.14 and the magnitude of the coefficients is more important. For the non-affiliated companies (excluding Allais) b is included in:
–0.1125<β<0.097. As regards the standard error, affiliated companies as a whole show for d1867 a value smaller than the ones obtained for d1867/68 and close to 0. The same consideration can be held for the non-affiliated companies. What kind of interpretation can we give to these results? The railways were best placed to resist bankruptcy for the following reasons: • • •
the Rothschilds were a financial partner; they benefited from government subsidies when facing difficulties; they were able to diversify their investment which enabled them to resist the crisis.
72
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
As a whole, the affiliated companies did not suffer from the bankruptcy of the Crédit Mobilier. The results obtained for the affiliated and non-affiliated companies are very similar (coefficients are slightly higher for d1867/68, but with a magnitude too small to be significant). In conclusion, we can affirm that the bankruptcy of the bank had no influence on either affiliated or non-affiliated companies. An association with the bank does not seem to have increased the values of shares. This expresses the lack of supervision of the Crédit Mobilier. 2.8 BANKRUPTCY EFFECT ON THE COMPANIES EITHER AFFILIATED OR NOT AFFILIATED TO THE CRÉDIT MOBILIER To extend the preceding results, a cross-section test is used, defined as: ln(pt / pt–1) = c + β(CM) where CM=1 if the companies were affiliated to the Crédit Mobilier, CM=0 otherwise. We run the regressions for the years 1866, 1867, 1868 and evaluate the ratios: ln p68 / p67 ln p68 / p66 The results, not significantly different for either ratio, are presented in Tables 2.22 and 2.23. Several comments can be made on Tables 2.22 and 2.23. Table 2.22 Cross-section results excluding the Compagnie Immobilière
Table 2.23 Cross-section results including the Compagnie Immobilière
73
THE ROLE OF BANKS IN MONITORING FIRMS
• • •
The figures obtained for the cross-section are not very different from those resulting from the preceding OLS. The coefficient β in both regressions is close to zero. The only negative sign appears when the Compagnie Immobilière is included. All t-statistics are insignificant.
Except when the Compagnie Immobilière is included, coefficients provide no evidence that the bankruptcy had any effect on the share prices of affiliated companies. Even when a negative sign is obtained, the magnitude of coefficients is too small to be significant. H0 is not rejected by the data. 2.9 GENERAL CONCLUSION The preceding tests prove the irrelevance of de Long’s proposition to the Crédit Mobilier case. According to de Long, ‘J.P.Morgan and Co’s approval of an issue had become…a large factor which inspires confidence in the investor and leads him to purchase’. The active participation of the Crédit Mobilier in a company did not ‘add value’ for the share prices. More specifically, our argument concerning the period 1861–70 proves that association with the Crédit Mobilier had no great influence on the share prices of the companies, as the failure of the Crédit Mobilier had little effect. The analysis of the situation of companies that were not affiliated to the Crédit Mobilier reinforces this hypothesis. Its bankruptcy did not cause a big enough financial perturbation to induce share price fluctuations on the important business concerns over the period. Between 1852 and 1870 the Crédit Mobilier does not appear to have been a bank capable of supervising properly the firms to which it granted credit. It essentially provided liquidity but without screening the firms’ managers. Moreover, the communication it entertained with other banks did not reflect the power of the Crédit Mobilier on the financial market. Numbers of banks over the period (like the Banque de France, the Société Générale and the Crédit Lyonnais) had serious doubts about the innovative ideas of the Crédit Mobilier concerning the financing of business projects. The results of the test as regards companies that were not affiliated to the Crédit Mobilier confirm this idea. All these findings lead us to question the concept of ‘control’: can a bank effectively control a firm through the credit it grants? No unequivocal answer can be given to this question. Edmond Baldy in his book Les Banques d’Affaires en France depuis 1900 bases his argument concerning effective control on the percentage of participation a bank has in a firm. Strictly speaking, the required participation for absolute control is 50 per cent plus
74
THE CRÉDIT MOBILIER AND THE STOCK EXCHANGE
one, but this is, of course, subject to the ownership of the assets between the shareholders. In particular, if one shareholder holds a huge amount of assets of the company, the bank, in order to exert a supervisory role, must acquire a majority, strictly larger than the external group. As regards the Crédit Mobilier, if the participation in the firms was always major, nevertheless it did not, typically, represent the largest proportion of the assets issued in the companies. Hence the monitoring role of the bank was diminished, according to this argument. However the strongest explanation of this lack of control remains the bank’s risk-prone investment policy (e.g. the Compagnie Immobilière). Beyond this economic interpretation, a historical explanation can be given. Around 1865 the Crédit Mobilier had lost some of its power. As early as 1863, the Péreires requested permission to double the capital of the Crédit Mobilier, but the government refused. With more and more of its assets tied up in advances to the Société Immobilière, in 1866 the Mobilier eventually persuaded the government to allow a doubling of its capital, in return for modifications in its statutes restricting still further its freedom of action. Naturally, the Péreires did not reveal to their stockholders the true situation of the bank, but spoke vaguely and grandly of the need for new resources to capitalise on the many opportunities for profitable operations which daily came its way. In fact, of course, the Mobilier undertook no new enterprises, but devoted all of the new capital to an attempt to salvage the Société Immobilière. Added to these difficulties, the Péreires faced a stockholders’ lawsuit against the increase of the bank’s capital. In 1867, the Crédit Mobilier went bankrupt. The first reactions of the governors of the Banque de France were outrage and indignation. But when the financial and political implications that the sudden failure of the Mobilier could provoke were pointed out, the governors agreed, in effect, to permit a more gradual liquidation of the enterprise. This constitutes one additional possible explanation of the lack of fluctuations for the share prices of the companies with which the Crédit Mobilier was involved.
75
3 CORPORATE INVESTMENT, CASH FLOW AND FINANCIAL CONSTRAINTS OF FIRMS The case of the Crédit Mobilier
French industry with its tradition of self finance had few direct links with the banking system…Whether or not this was so, and if so whether it reflected the strength or weakness of industrial enterprises, is still a subject for debate. Cameron and Bovykin (1991:15)
3.1 INTRODUCTION As shown in the preceding chapter, the Crédit Mobilier played a crucial role in the industrialisation of French industry over the period 1852–70. But contrary to the conclusions reached by other researchers studying other banks in other countries (e.g. de Long’s assertions about J.P. Morgan & Co), it does not appear to have been effective in facilitating the assessment of firms’ managerial performance and in the evaluation of projects. The Crédit Mobilier provided liquidity to firms but does not seem to have been supervising them. Since the Crédit Mobilier went bankrupt itself, the value of its seal of approval was correspondingly diminished. Moreover, the formal tests we have constructed for the year of bankruptcy of the Crédit Mobilier cast further doubt on the assertion that its seal of approval might have added value to French corporations. According to de Long, ‘J.P.Morgan and Co’s approval of an issue had become …a large factor which inspires confidence in the investor and leads him to purchase’. The evidence suggests that this was not true of the Crédit Mobilier. The active participation of this bank in a company did not ‘add value’ to the share prices. If anything, the inverse proposition seems to be closer to the reality. To analyse more deeply the role of the Crédit Mobilier in the financing of industrial companies, it would be interesting to see if an association with
76
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
the Crédit Mobilier reduced such companies’ costs of external borrowing and consequently made their investment less sensitive to cash flows. In a world of perfect information and no agency costs, firms can borrow and invest optimally; liquidity has no effect on the investment decision. In a classic paper, Franco Modigliani and Merton Miller (1958) demonstrated the conditions under which financial structure becomes irrelevant for the real investment decisions. Jorgenson (1992) has provided an extensive survey of the theory of investment behaviour, distinguishing between classical, neo-classical accelerator and Q theories; Modigliani-Miller (M-M) might apply to all of them. Today, however, there are many models of corporate finance that have moved away from the assumption of perfect information, stipulating some kind of information problem. The assertion that imperfect information invalidates M-M has many consequences:1 all these models predict that liquidity matters. Thus, our strategy is to see whether liquidity is a more important determinant of investment for independent, unaffiliated firms than for firms with close banking ties. Recent articles have explored the empirical relationship between corporate financial structure and investment.2 Fazzari, Hubbard and Petersen (1988) classify firms according to their a priori beliefs about whether a firm faces information problems in the capital market, and then test whether the information problems are presumed to be severe. Their basis of comparison is corporate dividend policy, and they argue that firms that retain more of their earnings are more likely to be liquidity constrained. Indeed, they find that investment is more sensitive to liquidity for firms that consistently retain a larger fraction of their earnings. Hoshi, Kashyap and Scharfstein (1989, 1991), on the other hand, show that investment expenditures of Japanese companies affiliated to a bank (e.g. a member of a zaibatsu) display almost no sensitivity to cash flows in contrast to the investment expenditures of firms which are not so affiliated. This finding supports the hypothesis that companies that are not affiliated to a bank tend to be more liquidity constrained (and face a wider gap between external and internal costs of funding) than those which are. This work follows their approach and investigates the assertion that the Crédit Mobilier played a role in the financing of French industry that was similar to that of J.P.Morgan in the United States3 and of the Großbanken in Germany. 4 In order to construct the empirical case study, two samples are considered: the first consists of corporations that were affiliated to the Crédit Mobilier, the second set of companies had to rely on other sources of finance. However, all firms were affiliated with a bank: independent companies used the Crédit Lyonnais to provide liquidity. This does not prevent us from assessing the impact of the bankruptcy on investment for
77
THE ROLE OF BANKS IN MONITORING FIRMS
affiliated and non-affiliated firms. The focus of this paper is thus on contrasting the financing of the Crédit Mobilier’s affiliated enterprises with those which, for one reason or another, did not form part of the nexus. In particular, the study of its investment policy before and after the bankruptcy will indicate the nature of the supervisory role of the Crédit Mobilier. To determine the appropriate procedure for deciding if firms were affiliated with the Crédit Mobilier, a short discourse into the general nature of the business relationship between banks and industry, as was suggested by contemporary sources like Jeidels (1905) and Riesser (1911), might be in order. 3.2 THE VALUE OF ASSOCIATION WITH THE CRÉDIT MOBILIER ON CASH FLOW, INVESTMENT AND MARKET VALUE There are several ways in which a bank and an industrial corporation can interact. A company might conduct its daily business through a current account with one or more banks. This was not the most common practice with the Crédit Mobilier. In the case of a private company, a bank might help to issue new equity or debt, for example bonds. A bank might take a direct stake in the company, or it appoint representatives to sit on the company’s board of the directors. This last arrangement was the most frequent for the Péreires’ establishment. In every company financed by the Crédit Mobilier, a member of its board of directors was represented in the shareholders’ assembly.5 The question of the advances to companies which were granted by the Péreires emphasises the direct participation of the bank. Lending, issuing shares, direct investment and supervisory board members helped the bank to monitor and in many cases influence decisionmaking in the industrial enterprise. Nevertheless, Jeidels (1905) carefully distinguishes between monitoring, in the sense of information-gathering, and direct influence on a corporation’s business decisions. Although collecting information seems to have been likely in the Crédit Mobilier’s case, any kind of direct influence over the decisions of management seems improbable (cf. the empirical results in Chapter 2). However, generally speaking, placing bank directors or members of the bank onto the board of corporations was common practice for the Péreire brothers and could have been the most effective means of influencing managerial decisions of affiliated corporations. In Chapter 2 we tried to analyse whether this supervision was effective in the Crédit Mobilier’s case. We intend to reinforce this argument by examining the investment policy of the bank. The aim of this section is then to analyse the relationship between investment and cash flows between 1860 and 1880. Two samples are
78
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
considered: the first consists of corporations that were affiliated to the Crédit Mobilier, the second includes companies which relied on other sources of finance. The focus of this construction is on contrasting the financing of the Crédit Mobilier’s affiliated enterprises with those which, for one reason or another, did not form part of the nexus. The chronological limits of this study are imposed by the data sources: after 1880 the Crédit Mobilier lost its powerful role in the banking sector and thus no records exist for this period. The sample is constituted of fifteen companies: seven were affiliated with the Crédit Mobilier, eight with the Crédit Lyonnais. As each company had a relationship with a bank the interpretation of the results must be in some sense cautious. Conscious of these limits, we discuss the validity of the actual research.
(a) Data selection The principal source for data underlying the statistical analysis comes from the General Assembly’s reports and some articles and books such as those by Jean Denuc (1934) and Bertrand Gille (1968). In his book, Bertrand Gille has reproduced a great number of company balance sheets, particularly steel companies; these figures make it possible to deduce cash flows, investment and book values. Jean Denuc provides values for other companies; the railways companies were reported separately. The different indices Denuc gives for all series enable us to deduce the book value and the cash flows for the different companies. The investment factor was given by an article issued by the Institut National de la Statistique which recapitulated the economic situation of France from 1850 until 1914. Here, we want to evaluate the correlation between cash flows CF, common equity Q and investment I. The regression specification was taken directly from Hoshi, Kashyap, and Scharfstein (1989, 1991) and was also used in Ramirez (1992). Table 3.1 lists the companies included in the sample taken into consideration and records their affiliation with the Crédit Mobilier. As can be seen, railways companies represent the major part of the Crédit Mobilier’s investments. We can note also that all companies were affiliated with either the Crédit Mobilier or the Crédit Lyonnais, both as powerful banks over the period. When the banks offered credit for business firms at this time, industrial enterprises were rarely involved because they had been warned by experience. Significant amounts of capital were provided by merchant banks especially to railways, shipping companies and urban transport projects, but industrial enterprises received little. As far as we are concerned with public utilities, the bank’s affiliation was omnipresent. This suggests that it may be difficult to deduce the effect of affiliation with a big bank by comparing
79
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.1 Companies included in the data set
Crédit Mobilier’s affiliated companies to non-affiliated companies. However, we should note that the degree of affiliation was totally different between the two banks. In the Crédit Mobilier’s case, the bank was substantially the principal shareholder of the client firms, whereas for the Crédit Lyonnais, affiliation only meant that the bank was a privileged financial partner of the firm involved. A useful approach is to contrast the effect of the Crédit Mobilier’s bankruptcy on those companies with which it was affiliated, and on Crédit Lyonnais affiliated companies. Hoshi, Kashyap and Scharfstein’s methodology used four parameters to construct a regression of the gross invesment I, the gross value of share K, the free cash flow CF, and Q measured as the market value of common equity divided by its book value. K constitutes a problem in the French context. It corresponds to the gross value of assets—in French, ‘le capital brut’. In the balance sheets of the companies taken into consideration in the sample, this factor seems to have be particularly stable over the period. The question arises: does this variable mean the same thing in the work presented by Hoshi et al. and in the French situation? It is hard to answer this question. Entrepreneurs over the period 1860–80 did not use sophisticated accounting and management techniques: what counted was the overall state of receipts and expenses seen in a concrete manner in the cash flow of the business. The next step of sophistication was the establishment of an annual budget, which was important because it allowed an assessment of net profit. But accounting rules were not imposed. Every entrepreneur had his own accounting conventions, to some degree, which makes it difficult to interpret
80
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
and compare these balance sheets. In order to avoid any ambiguity, this parameter will be called Cb, which should represent for the French case the gross value of shares. All variables, the market prices of common equity, the cash flows CF and the investment I, are normalised by the stock of depreciable assets at the beginning of the period. First, some summary statistics comparing Crédit Mobilier firms and non-Crédit Mobilier firms are presented. As the Crédit Mobilier went bankrupt in 1867, the mean is calculated in constant terms in regression, taking three samples: • • •
1861–67—the situation before the bankruptcy; 1868–75—the situation after the failure of the bank; 1861–75—for general information.
The Compagnie Immobilière is either excluded or included in the sample, as it went bankrupt with the Crédit Mobilier. Table 3.2(a) shows the differences in the means for Tobin’s Q, investment/ share value and cash flows/share value. Although some means are different, the standard errors would seem to suggest that no difference existed. In addition, the figures obtained for Q are quite low. An examination of the accounting rules over the period seems necessary. Financial reporting at the beginning of the nineteenth Table 3.2 Summary statistics comparing Crédit Mobilier and non-Crédit Mobilier firms*
81
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.2 Continued
Note: Averages are calculated for all firms and all years. Standard errors are in parentheses. Capital (Cb) and investment (I) are for depreciable assets. Sources and definitions of variables are described in the text. The sample for the Crédit Mobilier firms is constituted of seven firms, for non-affiliated firms eight.
century was characterised by the odd, scrappy profit calculation and balance sheet which contained assets valued in a wide variety of ways. In fact the main aim of the published balance sheet was to provide creditors and shareholders with a statement of company solvency and to reassure them that dividends had not been paid out of capital. There was no requirement for the publication of a profit statement, nor were there any minimum disclosure rules regarding the content of the balance sheet. During the nineteenth century the calculation of profit became the first stage in the preparation of the final accounts. The new approach to profit was the use of profits as the basis for dividend declarations. This adoption of the matching concept reduced the scope for management to manipulate profit; with valuations, there is a need not only to make estimates, but also to decide on the appropriate method of valuation from the various options available, including present value, selling price and replacement cost. This last parameter is of such importance that we should refer to James Tobin’s Q theory of investment.6 This can be summarised as follows. The rate of investment—the speed at which investors wish to increase the capital stock—should be related, if to anything, to Q the value of capital relative to its replacement cost. Over the period, replacement accounting delayed the recognition of capital consumption until expenditures were made for renewals. Furthermore, since periodical renewals would require substantial sums, there was a bias against renewals as costs because replacement reserves were not generally maintained. On the one hand, a firm might not have been able to charge substantial renewals to expenses without creating a deficit, while on the other hand, without sufficient internal funds available, it would be necessary to finance replacement expenditures with sources external to the firm. This discussion suggests that the application of replacement accounting tends to create serious liquidity problems and that renewals might have been deferred or treated as additions in many cases.
82
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
There was considerable scope for suspicion between bankers and industrialists over these years: bankers were cautious when dealing with industrialists because the former feared long-term commitments which were not thought profitable on a short term basis, while the industrialists considered the bankers as simply rapacious. Hence debt took the form of advances on capital, and was not reported as such in the balance sheets. On this basis it is hard, if not impossible, to approximate debt as such.
(b) The investment regression The selected regression specification takes the following form: It/Kt = α0 + α1 CFt/Cbt + α2 Q where: I Cb CF Q
= gross investment at time t, measured simply as the change in the gross value of depreciable assets (plant and equipment); = gross value of assets; = free cash flow, i.e. free surplus plus nominal allowances (mostly depreciation); = the market value of common equity stock divided by its book value.
Different sub-samples were chosen to run this regression: • • •
the whole sample; the period before the bankruptcy, 1861–7; the years following the bankruptcy.
The main results are presented in Tables 3.3, 3.4 and 3.5. The evidence presented in these three tables suggests that free cash flow was more important in both magnitude and statistical significance among independent firms. That means that free cash flow was an important determinant of investment expenditures among the independent mining and steel companies. This is also evidenced by the fact that the overall pattern of the regression for affiliated companies is significantly different from that of the nonaffiliated companies. That is, we cannot prove that the explanatory variables are explained by different investment expenditures among independent and affiliated companies. Included in these tables is an interaction dummy which takes the value 1 times cash flow when the corporation is affiliated to the Crédit Mobilier. This variable is introduced to estimate the difference and statistical importance of the cash flow coefficient between affiliated and non-affiliated companies.
83
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.3 Investment regression, 1861–75
Note: *CF.CM represents CF times Crédit Mobilier affiliation: affiliated firms 7, independent companies 8, all firms 15.
Also included in the tables is a regression for the whole sample, excluding the Compagnie Immobilière, and a regression for the sample taking into consideration this firm, which went bankrupt at the same time as the Crédit Mobilier. This last case produces different results depending on whether or not we consider the situation of the Compagnie Immobilière. This can be explained by the fact that the Crédit Mobilier invested a huge amount of capital in the firm which was irrelevant at the end and provoked its own bankruptcy. The different coefficients entering the investment equation were evaluated by identifying only whether or not a firm was affiliated to the Crédit Mobilier. The results obtained describe the general situation as regards the investment policy of the Crédit Mobilier. No change in investment is taken into consideration this time. The aim is to analyse the relationship between common equity and cash flow on investment for affiliated and non-affiliated companies. The question is then whether the structure of ownership and the behaviour of large shareholders differs between firms that were affiliated with a bank, and those that were independent of such groups. The empirical results compiled in order to answer this question are presented in Tables 3.3, 3.4 and 3.5. Table 3.3 enables us to identify the difference between affiliated and nonaffiliated firms and also to stress the importance of the Compagnie Immobilière in the investment policy of the Crédit Mobilier. If independent
84
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
firms were liquidity constrained by having to rely on cash flows to finance their projects, the coefficient CF/Cb must be much higher for these firms than for affiliated firms. Let us consider the interaction effect CF.CM. Everything else being constant, a unit increase in cash flows contributes about 0.02 units increase in investment expenditures among the Crédit Mobilier’s firms. Unlike the separate regressions, this suggests that affiliation with the Crédit Mobilier reduced liquidity. The results vary, as other coefficients must be the same for all firms as reported in columns four and five. The coefficient is then insignificant. An analysis of the impact of the bankruptcy of the Crédit Mobilier is now required. Two sub-samples of the initial sample (before and after the bankruptcy) are considered. A dummy variable reflects the affiliation of a company to the bank. For the sub-sample 1861–7, we suppress the Compagnie Immobilière: the absence of too many values prevents us from obtaining the relevant results. The results are presented in Table 3.4. Let us comment on the interaction coefficient, CF.CM. If we compare the result obtained for the whole sample and the one we have for the years preceding the bankruptcy, we note that the value is much larger for 1861– 7 than for the whole sample. That means that an association with the Crédit Mobilier was relevant during the period 1861–7. Remarkably the
Table 3.4 Investment regression, 1861–7
Note: *CF.CM represents CF times Crédit Mobilier affiliation: affiliated firms 7, independent companies 8, all firms 15.
85
THE ROLE OF BANKS IN MONITORING FIRMS
coefficient is significantly positive which suggests that association with the Crédit Mobilier reduced liquidity and thus increased the effect of the cash flow on investment. As regards the separate regressions, note that the cash flows are very similar for affiliated companies compared to the independent firms. What happened after the bankruptcy? Surprisingly, the coefficient on CF/ Cb for affiliated companies is tiny and insignificant. It is as if the failure of the Crédit Mobilier removed a liquidity constraint, the opposite of what theory would predict. The coefficient on the interaction CF.CM is negative, as if companies once affiliated to the new bank are less liquidity constrained than ever in the past. Note that for 1861–7 this coefficient was positive (Table 3.4 provides us with a coefficient equal to 0.101 for all companies including the Compagnie Immobilière and equal to 0.131 for the whole sample excluding the Compagnie Immobilière). The results, taken together, are the opposite of those predicted by the theory. It is as if affiliation with the Crédit Mobilier created liquidity problems. Conclusion The results obtained for France in the Crédit Mobilier case are very different from those presented for the United States and Japan by de Long and Ramirez, or for Germany by Marco Becht. Substantially, de Long (1989) asserts that the monitoring role of the Morgan Corporation was, in general, successfully achieved. The Pujo Committee of 1913 agreed that the House of Morgan had sufficient resources and influence in the financial world to render it the most powerful and active of the few investment houses. The econometric work of Ramirez (1992) suggests that association with J.P.Morgan relieved a credit constraint. The evaluation of Q and its fluctuations is not completely different from the results obtained by de Long and Ramirez. However, the participation of the bank in the firms’ projects was so non-uniform that a direct association with the Crédit Mobilier did not always contribute to an increase of Q for the affiliated firms (Table 3.5). It seems on the contrary that the bankruptcy of the Crédit Mobilier removed a liquidity constraint for affiliated firms. The comparison of the results obtained for 1861–7 on the one side and for 1868–75 on the other side indicates that, after the failure, the affiliated firms did not suffer from liquidity constraint as they had before. As a whole, our results seem to indicate that association with the Crédit Mobilier created liquidity problems. The next section intends thus to isolate the bankruptcy effect in order to evaluate more precisely the profitability of an association with the Crédit Mobilier (if any) on Q.
86
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
Table 3.5 Investment regression, 1868–75
(c) The fixed effect model The regression equations are evaluated by introducing year dummy and firm dummy variables. In other words, we determine investment for every year except for one, separately, and for every firm except one, separately. As in the preceding test we consider three samples: • • •
the whole sample 1860–75; the sub-sample 1860–7; the sub-sample 1868–75.
We run the regressions for affiliated companies, non-affiliated companies and all firms. The selected equation is modified as I/Cb = α0 + α1 Q + γ.CF/Cb where: I Cb CF Q
= gross investment at time t, measured simply as the change in the gross value of depreciable assets (plant and equipment); = gross value of assets; = free cash flow, i.e. free surplus plus nominal allowances (mostly depreciation); = common equity.
87
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.6 Investment regression equation for the whole sample*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors in brackets: affiliated firms with the Compagnie Immobilière 7, independent companies 8, all firms 15.
The regression equations are evaluated by introducing year dummy and firm dummy variables. In other words, we determine investment for every year except for one separately and for every firm except one separately. Let us consider first the Crédit Mobilier firms and non-affiliated companies. Tables 3.6, 3.7 and 3.8 present the main results. Some general remarks can be made regarding these tables: • • •
the coefficient of Q obtained for affiliated companies is quite high, whatever the sample considered; however, for these same companies, the coefficient on cash flows is low; generally speaking, the standard errors are quite large.
The evidence presented in these tables suggests that the coefficient on cash flow gains importance in magnitude among the non-Crédit Mobilier companies. Table 3.9 presents the results of the regression specification over the entire sample. The regression includes an interaction dummy which takes the value 1 times the cash flow when the corporation is part of the Crédit Mobilier sample. This variable is introduced to estimate the difference and statistical importance of the cash flow coefficient among affiliated and nonaffiliated companies. The findings are not very robust: except for the sample 1861–75, the interaction coefficient is negative and the standard errors are quite large. Crédit Mobilier firms relied less on cash flow to finance their investment expenditure over the first period than on other sources of credit. For the two latter samples, no evidence is shown by the table. Previous explanations of the importance of cash flow for non-affiliated companies from Japan and the United States are not really applicable to the
88
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
Table 3.7 Fixed effect model for 1861–7*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors in brackets: affiliated firms with the Compagnie Immobilière 7, independent companies 8, all firms 15. Table 3.8 Fixed effect model for 1868–75*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors in brackets: affiliated firms with the Compagnie Immobilière 7, independent companies 8, all firms 14. Table 3.9 Fixed effect model for the whole sample*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors in brackets. The number of companies is 15 for 1861–67 and 1861–75, and 14 for 1868–75.
89
THE ROLE OF BANKS IN MONITORING FIRMS
French case. The Crédit Mobilier did not act as a universal bank throughout the whole period. Because of asymmetric information, non-affiliated companies face much higher costs of external financing than did affiliated ones. However, this cost was partly compensated by the risky behaviour of the bank towards investment. How can we explain these differences between France and Germany? A series of eminent economic historians 7 had held that the major innovation in industrial financing in the nineteenth century was the Crédit Mobilier. Gerschenkron’s theory was elaborated into a model, a backwardness, in which banks, depending on how backward a country was, substituted liquidity for entrepreneurship. This had been the driving force in the industrial revolution in the United Kingdom, for example. But evolution proceeded rather differently in other countries. German banks took the example of the Crédit Mobilier and developed it to the utmost. Apart from Hamburg and Frankfurt, which clung longer to commercial banking, the German banks went in for close relations with industrial and mining firms. Theoretically, the Crédit Mobilier’s case illustrates perfectly the following dilemma. Adherents of rational expectations and efficient markets tend to believe that markets always work. Institutionalists, of course, take the opposite view, that institutions determine historical outcomes. Between the two positions, there is room for an eclectic view that markets mostly work but occasionally break down, and that sometimes institutions adapt to underlying changes in demand and supply conditions and sometimes they do not. The Crédit Mobilier was not capable of adapting to the change.
(d) Conclusion The main premise of this section is that, unlike the Morgan company, 8 the Crédit Mobilier was not fulfilling an effective monitoring role but only provided liquidity to firms whenever necessary. This bank does not seem to have alleviated the impact of the imperfections of the capital market. To strengthen the argument, we consider two extensions of this work in the following chapter. The first step will be to include in the book value a depreciation rate δ, equal to 5 per cent, in order to conform more to the Q theory. All details for the description of the parameters will be given in Chapter 4. Some additional value for δ will be considered to evaluate the robustness of the test. The next step will be to regress investment, cash flow and Q on the basis of their real value, that is by dividing the parameters by the general level of prices in France.
90
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
3.3 TRANSFORMATION OF THE RAW DATA ACCORDING TO ‘TRADITIONAL FINANCIAL ACCOUNTING’ As previously mentioned, the Q values are quite low because of the absence of replacement cost. The first transformation will be to introduce a depreciation rate in the evaluation of Cb and therefore Q in order to be consistent with James Tobin’s Q theory. This theory of investment can be summarised as follows. The rate of investment—the speed at which investors wish to increase the capital stock—should be related, if to anything, to Q, the value of capital relative to its replacement cost (Tobin 1971:330). Wilfried Feldenkirchen (1982:269) explains that over the period this replacement cost was not always reported. For German companies a depreciation rate was only introduced when the firms were making a profit. The same procedure occurred in France. The appendix describes the new evaluation of Q. The second proposition is to regress investment, cash flow and Q on the basis of their real value, that is by dividing the parameters by the general level of price in France.
(a) The introduction of a depreciation rate This section includes a depreciation rate in the evaluation of the parameter Cb in order to modify the factor Q in the investment regression. Why implement such a procedure? The purpose of the depreciation charge today is to spread the net cost of a fixed asset (original cost minus sales proceeds) over its estimated useful life. The aim is to ensure that the revenue arising during each accounting period bears a fair share of the total costs incurred. During the nineteenth century, depreciation involved no specific outlay of funds and could be ignored in the short term; even when recorded, many accountants apparently regarded depreciation as a segregation of profits rather than an expense. Depreciation reserves, a subject that occasioned considerable confusion in the nineteenth century, were viewed as a kind of surplus account. In early accounting texts, the term depreciation was often used to describe the difference between fixed asset valuations at two different dates. Depreciation, when charged, was viewed principally as a means of earmarking for retention resources which could be set aside to finance replacement, rather than as a bona fide cost of production. The data we have for Cb fits these parameters. Moreover, as early in the development of railways as 1852, it is evident that very few people had a good conception of the relationship between depreciation and net income. The methods used over the period for reflecting depreciation in the accounts seemed to receive little support in railway circles. An annual revaluation of properties was one of these; setting aside an annuity
91
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.10 Summary statistics comparing Crédit Mobilier and non-Crédit Mobilier firms when introducing a depreciation rate of 5 per cent*
Note: *Averages are calculated for all firms and all years. Standard errors are in parentheses. The sample for the Crédit Mobilier firms is constituted of 7 firms over 1861–7 and 1861–75, and 6 firms over 1868–75. The number of affiliated firms is 8.
which would accumulate to the desired sum by the time replacement became necessary was another. This explains the flat series for Cb. How can we transform it, and what is the definition of the new parameter? Let Cbk1861 represent the first element of the series which is defined as: Cbk1861 = I1861/δ where δ is constant and is chosen so that the series is slightly increasing. The whole series is given by: Cbkt+1 = Cbkt (1 – δ) + It . We call this new variable Cb1. Appendix F describes in detail the construction of the series and the new values obtained for Q called Q1. Table 3.10 presents some descriptive statistics to analyse the modifications produced by this alteration of the data. If we compare these numbers with those obtained in the preceding section, we note that they remain low. The problem then is to examine if this decrease is realistic or not. To complete the analysis presented here, can we explore the modification implied in the data set if we consider depreciation rates d both greater and smaller than 5 per cent (see Appendix F).
(b) The investment regressions The selected equation is modified as: I/Cb = a + a Q + ?.CF/Cb 1
0
1
1
1
where: I
=
gross investment at time t, measured simply as the change in the gross value of depreciable assets (plant and equipment);
92
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
Table 3.11 Investment regression for the whole sample 1861–75
Note: Standard errors are in brackets. *CF.CM represents CF times Crédit Mobilier affiliation: affiliated firms 7, independent companies 8, all firms 15.
Cb CF1 Q
= gross value of assets; = free cash flow, i.e. free surplus plus nominal allowances (mostly depreciation); = common equity including the depreciation rate given by Appendix F.
1
In the following section we present the results obtained and compare them to those of the preceding section. As before, we consider three samples: the whole sample 1860–75, the sample prior to the bankruptcy i.e. 1860–7, and the years after the bankruptcy. The results are given in Tables 3.11, 3.12 and 3.13. If we compare the results obtained here with those of the preceding section, we can make the following comments. Let us first consider the whole sample 1861–75. The coefficient for CF/Cb and Q are not very 1 1 different for affiliated and independent firms. The association with the Crédit Mobilier, even if it was effective (largest value for the dummy coefficient), was not important enough to cause a difference between the two sets of companies. As regards the sample 1861–7, a comparison with the results obtained by Becht and Ramirez (1992:28) will be interesting. Let us recall them briefly. Tables 3.14 and 3.15 only give us the coefficients obtained by running a regression similar to the one presented in this chapter. We deliberately choose not to report the standard errors, the aim being to analyse the analogies and the differences between the two countries.
93
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.12 Investment regression over 1861–7
Note: Standard errors are in brackets. *CF.CM represents CF times Crédit Mobilier affiliation: affiliated firms 7, independent companies 8, all firms 15. Table 3.13 Investment regression over 1868–75
Note: Standard errors are in brackets. *CF.CM represents CF times Crédit Mobilier affiliation: affiliated firms 6, independent firms 8.
If we compare these figures with the ones we obtain, we remark that the coefficients of the cash flows and the common equity are in exactly the reverse order for the Crédit Mobilier: •
CF/Cb is close to 1 for the affiliated companies and close to zero for 1 independent firms, which means that the affiliated companies were 94
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
Table 3.14 Results for affiliated companies as presented in Becht and de Long’s paper
Table 3.15 Results for non-affiliated companies as presented in Becht and
de Long’s Paper
•
liquidity constrained before the bankruptcy (Becht and Ramirez conclude from the test they build that an association with a universal bank limited the constraint towards liquidity); Q is very small (between 0.150 and 0.250), which does not represent a 1 significant parameter in the investment regression.
What kind of interpretation can we give to these results? The founders of the Crédit Mobilier cannot be consider as real ‘bankers’, but rather as promoters of enterprises. Their aim was the realisation of the highest profit possible and the construction of a solid reputation for the financiers who selected carefully the projects to which they granted credit. If we consider the situation after the bankruptcy (sample 1868–75) and we once again compare our results to those obtained by Ramirez and Becht, we note a strong analogy between the two cases. After the failure, the results are what one would expect if the Crédit Mobilier began to act as a big bank even though it had lost its powerful position among French financiers. The coefficients of the cash flows are negative for the affiliated companies, and Q conforms more closely to 1 the theory for the independent firms. We can conclude that the bankruptcy was the origin of a complete change in the investment policy: the Crédit Mobilier began to monitor effectively the firms with which it was involved. It will be interesting to see if these conclusions are confirmed by the fixed effect model.
(c) The fixed effect model The selected regression specification takes the following form: I/Cb1 = α0 + β1Q1 + ß2CF/Cb1 + γt + ai + εit 95
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.16 Fixed effect model for the whole sample*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors are in brackets. Affiliated firms with the Compagnie Immobilière 7, independent companies 8, all firms 15. Table 3.17 Fixed effect model for 1861–7*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors are in brackets. Affiliated firms with the Compagnie Immobilière 7, independent companies 8, all firms 15.
where α0=intercept, αi=(N–1) firms specific fixed effects, and γt=(T–1) year dummies and e is a residual term. The results from these estimations are given by Tables 3.16, 3.17, 3.18 and 3.19. If we analyse the coefficient we obtain for the cash flows, we notice that the Crédit Mobilier firms were more constrained before the bankruptcy (CF/ Cb =0.112 for Crédit Mobilier firm over 1861–7) than after (where CF/ 1 Cb =–0.103). This is the opposite of what is predicted by theory. The 1 collapse of the Crédit Mobilier should have created liquidity problems for affiliated firms. The most remarkable result of Tables 3.16 and 3.17 is the
96
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
Table 3.18 Fixed effect model for 1868–75*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include yearly dummies and firm dummies. Standard errors are in brackets. Affiliated firms 6, independent companies 8. Table 3.19 Fixed effect model for all firms*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors are in brackets. All firms with the Compagnie Immobilière 15.
difference that exists between a sample with the Compagnie Immobilière and one without it. Let us consider, for example, those samples, 1861–75 and 1861–7, as regards the coefficient on CF/Cb . It seems all affiliated companies 1 were liquidity constrained but that the Compagnie Immobilière was much more constrained than the others. The results presented for all companies corroborate this analysis, except that the difference between the set including the Compagnie Immobilière and the set excluding it is not so strongly marked. As a whole, the introduction of a depreciation rate in the data contributes to an increase in the value of CF dummy, which represents the effect of an association with the Crédit Mobilier on liquidity constraints. The pattern ‘affiliation causes liquidity constraints’ is strengthened.
97
THE ROLE OF BANKS IN MONITORING FIRMS
(d) Conclusion This econometric work provides some additional evidence suggesting that the Crédit Mobilier seems not to have functioned like a Großbank. The affiliated companies were more liquidity constrained before the bankruptcy than after. 3.4 EXTENSIONS OF THE RESULTS: DESCRIPTION AND INTERPRETATION OF A ROBUSTNESS TEST All the empirical work in this section is based on the assumption that the depreciation rate is equal to 5 per cent. This choice can be seen as arbitrary, but leads to a slightly increasing series for Q. To complete the analysis presented here we introduce the modification implied in the data set if we consider a depreciation rate d greater and smaller than 5 per cent. Appendix F presents the figures obtained for the coefficient Q when introducing a Table 3.20 Investment regression equation for affiliated companies and independent firms over the whole sample*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors are in brackets. Affiliated companies with the Compagnie Immobilière 7, independent firms 8.
98
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
Table 3.21 Investment regression equation for affiliated companies and independent firms before the bankruptcy*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors are in brackets. Affiliated companies with the Compagnie Immobilière 7, independent firms 8.
depreciation rate of 3 per cent and 7 per cent. In what follows, the new variables for Q and Cb are noted Q2 and Cb2 when δ=3 per cent, Q3 and Cb3 when δ=7 per cent. The aim of this section is then to measure the modification induced by a small difference in the depreciation when analysing the correlation between investment, cash flows and common equity. For that purpose we concentrate the econometrical analysis on the fixed effect model. The previous tests presented in section 3.2(a) are not going to be considered, in order to alleviate a fastidious examination of a large number of tables. The regression we run considers the set of affiliated companies on the one side and the set of independent companies on the other side over three samples: the whole sample 1861–75, the situation before the bankruptcy 1861–7, the event after the bankruptcy. We then evaluate the correlation between investment and cash flows for the whole set of companies over the same samples. The results presented in Tables 3.20, 3.21 and 3.22 successively report the figures obtained for the two values of depreciation rates in order to facilitate their interpretation.
99
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.22 Investment regression equation for affiliated companies and independent firms after the bankruptcy*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors are in brackets. Affiliated companies 6, independent firms 8. Table 3.23 Cash flow parameters
Note: *CF/Cb1 stands for cash flows including a depreciation rate of 5%; CF/Cb2 stands for cash flows including a depreciation rate of 3%; CF/Cb 3 stands for cash flows including a depreciation rate of 7%.
Some general comments can be made on this first set of results. In order to summarise the arguments we construct a small table (Table 3.23) reporting all the figures we obtain for the coefficient on cash flows.
100
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
As can be noted, the coefficients for the cash flows are very similar for affiliated companies when the Compagnie Immobilière is excluded. whatever level of deprecation rate is considered. The same argument can hold for the independent companies. As regards the coefficient on Q, the preceding tables exhibit very strong similarities for every set of companies (affiliated companies with or without the Compagnie Immobilière, as well as independent companies). The constant term always has the same sign and the same magnitude. The same conclusions follow for R2. Some additional comments can be made on the fluctuations of the t-statistics as regards the robustness of this test. In what follows we concentrate on the whole sample (1861–75) and on the years before the bankruptcy (1861–7) because the rest of our results exhibit no big change for the t-statistics. More specifically we consider the set of companies excluding the Compagnie Immobilière. The analysis is summarised by two tables (Tables 3.24 and 3.25) which reproduce the coefficient on CF/Cb in order to appreciate the degree of i liquidity constraints for affiliated companies over these two period. First, it is remarkable that the introduction or the exclusion of the Compagnie Immobilière should have an influence on the interpretation of the liquidity constraints that the affiliated firms had to suffer during these two periods. The Table 3.24 Sample 1861–75 robustness results for Crédit Mobilier firms without the Compagnie Immobilière
Table 3.25 Sample 1861–7 robustness results for Crédit Mobilier firms without the Compagnie Immobilière
101
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.26 Investment regression equation for all companies over the different samples
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include yearly dummies and firm dummies. Standard errors are in brackets. Affiliated firms with the Compagnie Immobilière 7, independent companies 8, all firms 15.
figures reported in these tables show that the coefficient on cash flows is more or less significant according to the level of depreciation rate included in the regression. As this situation does not represent the norm over the entire statistical work presented here, we can say the results are robust: the sign and the amplitude of the point estimates have not changed when the depreciation rate varies. Let us now consider the whole set of firms for each sample. The results are presented in Table 3.26. In order to compare the figures obtained for the three values of d we construct two tables (Tables 3.27 and 3.28) which summarise the results obtained for Q and CF/Cb . Some general comments i i will be made for the dummy coefficient and the R2 parameter. Let us first consider the coefficient obtained for CF/Cb . i The values reported in Table 3.27 exhibit strong similarities for the three values of d, except in one case, CF/Cb for all companies 1861–7 when the 3
102
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
Table 3.27 Cash flow parameters
Compagnie Immobilière is excluded. To complete the analysis we intend to Table 3.28 Q parameters
study the parameter Q and to examine if the same particularity happens over the same sample. As regards Table 3.28 it seems that the similarities for this coefficient are even stronger. If we now consider the dummy variable, we see that it always has the same sign and the same magnitude. The same line of argument can be made for the constant term. For all firms, R2 is included between: 0.647
≤ R2 ≤ 0.995
which means that the correlation between the variables is quite high. As before, we add some indication as regards the variation of the tstatistics, concentrating on the only sensible sample, 1861–7. Two coefficients are taken into consideration: the cash flow and the dummy variable, reported in Table 3.29. The most surprising element of this table is the small value for the tstatistics obtained for the dummy variable when the depreciation rate is equal to 5 per cent. If this number is compared to the whole set of figures reported in the table, it seems that this situation represents a marginal case and does not alter the validity of our results.
103
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.29 T-statistics for sample 1861–7 as regards the cash flow parameters
(a) Conclusion The robustness test constructed here reinforces the preceding conclusion, in the sense that no evidence is given that the Crédit Mobilier behaved over the period like a Großbank. As before it seems that affiliated companies were more liquidity constrained before the bankruptcy than after. This empirical work enables us, however, to consider that the chosen value for d does not have a greater influence over the results obtained in the test. In the statistical analysis presented below we then concentrate on a medium value for d equal to 5 per cent. 3.5 THE EXPRESSION OF ALL VARIABLES IN REAL TERMS As the figures presented above remain surprisingly low, an additional transformation is done in order to express all the factors in real terms. This means that we divide It by Pt which gives Bkt+1 in real terms. Q is then computed as: value of shares/(Pt.Bkt). The new series can be described as: Cbkt = Cbkt–1 (1–δ) Pt/Pt–1 + It.Pt/Pt–1. The appendix describes in detail the series retained for Pt and the method used to compute Q. Let us present some descriptive statistics to analyse the modifications produced by this alteration of the data. If we compare these figures with those obtained previously we note that they do not conform to what could be expected.9 Taking them as correct the point is then to evaluate the impact of this modification on the investment regression. The selected regression specification takes the following form: I/Cb2 = α0 + β1Q2 + β2CF/Cb2 + γt + αi + εit
104
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
Table 3.30 Summary statistics comparing Crédit Mobilier and non-Crédit Mobilier firms when parameters are expressed in real terms*
Note: *Averages are calculated for all firms and all year. Standard errors are in parentheses. Affiliated firms with the Compagnie Immobilière 7, independent firms 8. Table 3.31 Investment regression equation for the whole sample*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors are in brackets. Affiliated companies with the Compagnie Immobliere 7, independent firms 8.
where: α0=intercept, αi=(N–1) firms specific fixed effects and γt=(T–1) year dummies. The results from these estimations are given by Tables 3.31, 3.32, 3.33 and 3.34. Some comments can be made on these four tables. Let us first consider what happens over the whole sample and for 1861–7. The first observation is that the coefficient on Q is larger for affiliated companies and the second 2 that the results obtained when the Compagnie Immobilière is part of the sample are much different from the ones where it is excluded. As regards the independent companies, the coefficients are not well estimated. This difference appears more pronounced than in the preceding test. One reasonable explanation could be the lack of data: missing values are introduced for 1861–4.
105
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.32 Investment regression equation for 1861–7*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors are in brackets. Affiliated companies with the Compagnie Immobliere 7, independent firms 8. Table 3.33 Investment regression equation for 1868–75*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors are in brackets. Affiliated companies 6, independent firms 8. Table 3.34 Investment regression equations for all firms
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. Standard errors are in brackets. All firms with the Compagnie Immobilière 15.
106
CORPORATE INVESTMENT, CASH FLOW AND FINANCE
If we consider the period after the bankruptcy, the results do not provide us with new information regarding the correlation among the parameters. What seems to be more interesting is the figures relative to all companies. If the whole sample is compared with the years prior to the bankruptcy, it will be noted that the Q coefficient is much smaller when 2 the Compagnie Immobilière is excluded over 1861–7 than when it is part of the sample. It will be of interest to analyse the influence of the bankruptcy over 1861–75. For that purpose we introduce a new dummy variable we call CF.AFFI.67 which measures the influence of an affiliation with the Crédit Mobilier after the year of bankruptcy. With the interaction dummy which takes the value 1 times cash flow when the corporation is affiliated to the Crédit Mobilier, we multiply another dummy which takes the value 1 after the bankruptcy, 0 otherwise. We run the regression for the whole sample without the Compagnie Immobilière. The results are given in Table 3.35. Let us compare the figures obtained here with the ones we have in the preceding regressions (Table 3.6). When we exclude the Compagnie Immobilière from the sample and analyse the impact of the bankruptcy as regards the investment policy, we see that, even if there is an effect (see the last column of the table) the difference is either small (a decrease of 10.6 per cent in the coefficient on cash flow can be noted for a bank whose assets were largely liquidated) or non-existent (Q is increasing). 2 Let us consider the affiliated dummy itself. In the second regression the sign of this variable becomes negative, reflecting the failure of the Crédit Table 3.35 Regression for the whole sample without the Compagnie Immobilière*
Note: *The dependent variable is investment relative to capital stock at the beginning of the period. The regressions include year dummies and firm dummies. CF.AFFI.67 represents CF times Crédit Mobilier affiliation times dummy variable, 0 before 67, 1 afterwards. Standard errors are in brackets.
107
THE ROLE OF BANKS IN MONITORING FIRMS
Table 3.36 Comparison of cash flow and Q parameters
Mobilier. This again suggests that the failure of the Crédit Mobilier relaxed liquidity constraints for affiliated firms. 3.6 GENERAL CONCLUSION All these figures prove that the Crédit Mobilier’s firms were more liquidity constrained before the bankruptcy than after. The bank never managed to supervise firms. As long as stock values were appreciating, the Crédit Mobilier found it easy to participate in the stock market while diversifying its activities. The bank participated in railway buildings in Europe, international banking and French real estate. This multifaceted activity was in part due to the Péreires’ very broad view of the goals of the Crédit Mobilier, but it was also a response to the booming international capital market that occurred after mid-1850. The extraordinary degree of activity of the Crédit Mobilier meant that it operated with few reserves, especially given that it distributed a substantial part of its profit for nearly a decade. The absence of reserves made it impossible for the bank to survive the economic slowdown that followed 1865 and relieved credit constraints for affiliated firms. The next step of our argument will be to discuss the place of the Crédit Mobilier in the banking system, in order to explain the minor effect on companies when the Crédit Mobilier collapsed. We pay attention in our analysis to the differences between France, Germany and the United States over the same period.
108
4 THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER Some interpretations
The Crédit Mobilier has been, if I may be forgiven the term, a promoter of business, working on the premise that present interest almost invariably dominates future interest.1 Bonnet (1865:115) La revue des deux Mondes, November 1865
4.1 Introduction In this chapter we intend to use the different econometric results we have obtained to highlight the theoretical questions they suggest. In Chapter 2, we showed that over the period, prices were more volatile than perfect foresight fundamentals relative to naive forecasts would have suggested. Thus, tests based on market volatility ratios show traces of excess volatility in the French stock market in the period before the First World War. Concerning the fluctuations of the share prices of affiliated and nonaffiliated companies, the test constructed for the sample 1867–70 demonstrates that share prices for affiliated companies did not fall. My tentative conclusion concerning the period 1867–75 was that an association with the Crédit Mobilier did not have a great influence on the share prices of the companies with which the bank was involved. I then produced a series of statistical analyses concerning investment I, cash flows CF and common equity Q for affiliated and non-affiliated companies. As regards Q, we can say that before the bankruptcy, an association with the Crédit Mobilier was correlated with high value. But the figures we obtained in the Crédit Mobilier’s case prove that the participation did not add as much value as in the Morgan Company’s case. The standard interpretation suggests that the unaffiliated companies were not credit constrained compared to the affiliated ones, and that affiliated firms were more liquidity constrained before the bankruptcy than after. These results give rise to four different questions. The first concerns the supervisory role that the Crédit Mobilier had exerted over the firms with which it was involved. In regard to this, we study two main points:
109
THE ROLE OF BANKS IN MONITORING FIRMS
• •
the possibility of a financial crisis due to the Crédit Mobilier’s fall; the idea of a banking panic.
More specifically, if this supervisory role were effective and necessary, as the Crédit Mobilier went bankrupt, a financial crisis would have followed for the affiliated companies. Except for the bankruptcy of the Compagnie Immobilière, however, nothing spectacular happened. To support our argument we compare the example of the Ohio Life Insurance and Trust Co. in the United States to the statistical results concerning the fluctuations of the share prices of affiliated companies in the Crédit Mobilier’s case. We will then draw conclusions on the qualification of financial crisis in the Crédit Mobilier’s case. The second issue concerns the consequences of the bankruptcy for the banking sector: as the Crédit Mobilier played a great part in the industrialisation of France (providing a large amount of liquidity) we can ask why its failure did not provoke a banking panic. In regard to this, two definitions of banking panic are presented and applied to the Crédit Mobilier’s case to see if this crisis can be defined as a liquidity crisis. The empirical research concerning investment, common equity and cash flows will inform our argument. The question then asked is this: what is the most efficient way to control a firm—an interaction between firms and banks (which means substantially a bank-firm association) or a separation of banks and firms (as presented in Chapter 1)? In other words, should banks supervise, or just choose whether to loan or not? In the latter case, supervision is exerted by the credit policy of the bank toward enterprises. The power is derived from the firms’ need for funds to realise their productive investments. This will constitute the second point of our discussion. The third point of our argument will consist of a comparison between France and Germany. More specifically, some German Großbanken (in particular the Darmstädter) were modelled on the Crédit Mobilier. The research that Marco Becht and Carlos Ramirez (1993) have carried out on these banks demonstrates the efficiency of the system in contrast to the French experience. It will then be interesting to compare the two banking systems in order to explain the differences in the results obtained. The last point consists of a discussion about the role played by the shareholders (if any) when the Crédit Mobilier went bankrupt. In particular, we want to assess the rationality or irrationality of their behaviour when the firm was facing financial difficulties.
110
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
4.2 THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER If the Crédit Mobilier were truly a great bank, its bankruptcy would have provoked financial difficulties not only for the enterprises with which it was involved but also for the other banks on the market.
(a) Did the Crédit Mobilier’s bankruptcy create a financial crisis? According to Kindleberger (1984) and Miskin (1991), a financial crisis involves either sharp declines in asset prices, or failures of large financial and non-financial firms. To support this argument, Miskin in his article entitled ‘Asymmetric information and financial crises’, takes the example of the Ohio Life and Trust Company, a major financial institution with substantial investments in western land and railroads as well as in commodity futures, which failed on 24 August 1857. The most interesting part of this analysis is that the timing of events in the panics of 1857 seems to fit an asymmetric information interpretation of the financial crisis as in our preceding study of the relationship between bank, manager and shareholders. Rather than starting with the bank panic of October 1857, the disturbance to the financial markets seems to have arisen several months earlier with the rise in interest rates, the stock market decline, major failures of financial firms and the widening of the interest rate spread. The asymmetric information analysis provides an explanation of how the financial crisis could have led to a severe economic downturn. The rise in the interest rate and the stock market decline, along with the failure of Ohio Life and Trust Co. which increased uncertainty, would magnify the adverse selection and agency problems in the credit markets. Indeed, the stock market crash might have been linked to the general rise of interest rates, which would have lowered the present discounted value of future income streams. In this case, the panic of 1857 can be viewed as a financial crisis. The net result of the increase in the adverse selection and agency problems is that investment activity and aggregate economic activity would decline, causing low expectations of further economic contracting and business. This example is also interesting because the period during which it happened is not so far away from that of the Crédit Mobilier’s bankruptcy. If we compare the results given by the volatility ratios and by the test on the fluctuations of share prices for affiliated firms, we find that there is no evidence that the economic situation described above can apply to the French case. If the price-dividend ratio decreased during 1870–80 (just
111
THE ROLE OF BANKS IN MONITORING FIRMS
after the bankruptcy of the Crédit Mobilier), it is impossible to tell if the failure of the bank has anything to do with this movement. The FrancoPrussian War, the Commune of Paris and the Revolution of 1870 caused considerable economic disturbances over this period. However, the general crisis of 1866 that occurred in England, Italy and Spain was not a major event in France and thus had no influence on the Crédit Mobilier’s bankruptcy. Some authors (e.g. Friedman and Schwartz (1963)) have established the close connection of financial crisis (especially the bank failures) with changes in real output. Such a phenomenon did not happen in 1867. In that sense, we can affirm that the Crédit Mobilier’s bankruptcy did not create a financial crisis. The second strong argument concerns the relation between banks, managers and shareholders when financing a productive project. According to Stiglitz (1985), raising capital through banks results in more effective control over capital than raising it through an equity market.2 Because both adverse selection and moral hazard are inevitable, at least in realistic loan markets, credit will often be rationed, implying that firms will typically face borrowing constraints which really do limit their investment opportunities from time to time. Moreover, the tightness of these borrowing constraints for any particular firm will clearly depend on the amount of collateral that the owner-manager is able and willing to offer, and not just on the present discounted value of what he can afford to repay without surrendering collateral in the worst possible future his firm can face. Stiglitz points out the interdependence of the different partners in the firm:3 the shareholders and the bondholders all benefit from the control of the banks. In fact, the intention of banks to grant credit to a firm is an effective positive signal. The actions of firms which try to find funds on the capital markets because they cannot obtain them from their bank could be interpreted in an adverse way. According to this scenario, when the Crédit Mobilier went bankrupt, the share prices of the affiliated companies should have fallen: this would have proved that the Crédit Mobilier added some value to the shares for affiliated companies. But nothing like this happened to the share prices of these companies in the year of the bankruptcy. It could be argued that the investors had anticipated the fall of the bank and had already diversified their investments and discounted the effect it would have on affiliated firms. In his book La Naissance d’une Banque: le Crédit Lyonnais, Bouvier remarks that a great number of firms which were financed by the Crédit Mobilier before the bankruptcy turned to the Crédit Lyonnais for funds afterwards. In particular the Compagnie Transatlantique, which was one of the largest companies of the Crédit Mobilier, requested funds as early as 1865. Another explanation for the stability of the share prices is the
112
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
intervention of the Banque de France which prevented a crisis from developing; this encouraged other banks to lend funds to the companies affiliated with the Crédit Mobilier, except for the Compagnie Immobilière. This reinforces the idea that the failure of the Crédit Mobilier did not really induce a financial crisis. The third argument concerns the role of financial institutions in the crisis. Was the Crédit Mobilier’s bankruptcy in 1867 an isolated case, or was it a reflection of the fragility of capitalism over the period? The first remarkable element is that the bankruptcy of the Crédit Mobilier had nothing to do with the general crisis of 1866 in England and Italy. As Kindleberger pointed out in his book entitled Manias, Panics and Crashes, the crisis occurred in France in 1864 and was solved by an extension of the maturities of bills. Moreover, the plots of share prices of ten affiliated companies presented in the second chapter of that book prove that no particular event could be noted for this year. Thus the Crédit Mobilier crisis seems to have been an isolated event. Secondly, financial institutions stand between firms and households. To a large extent, the liabilities of firms are owned by financial intermediaries and the assets of households are largely liabilities of financial intermediaries. These intermediaries (here banks) are self-, or profit-, seeking institutions. Of these profit-seeking organisations, one set plays an exceptionally delicate role in capitalist economies. This set consists of the investment bankers who either as brokers—who bring buyers and sellers together—or dealers—who take financial liabilities into their own accounts—act as midwives to young companies and finance their continuing operations. Essentially, these operators have superior knowledge about those customers who need financing (they have a need for funds) and those customers who have a need for outlets in which money can be placed. They turn this private knowledge of the conditions under which funds are desired and the conditions under which funds are available to their own advantage, even as they perform the social function of selecting the investments that the economy makes. In the Crédit Mobilier’s case the problem was that the bank had strongly focused on short-term profit-seeking and neglected its selection role. Its desire to create a new style of credit banking prevented the Péreires from carefully studying the projects in which they participated. These financial intermediaries are of critical importance in determining the values attached to collections of capital assets held by firms. In a balance sheet, the difference between the sum of the values entered for capital and financial assets and the value of debts on the liability side is the book value of the owner’s interest in the firm. Dividing the book value of the owner’s equity by the number of outstanding shares yields the book value of a share. However, for the main companies in a large economy there is a large market for equity shares and this market value may be less than,
113
THE ROLE OF BANKS IN MONITORING FIRMS
equal to or greater than the book value. A major consideration in decisions to invest is that the market evaluation of the capital assets needs to exceed the supply price of the required investment goods. One consequence of the introduction into the markets of these layers of profit-seeking organisations, which determine the value of financial instruments, is that the value imputed to capital assets can and does vary independently of the cost of investment goods. Furthermore, the extent to which internal funds are available to finance investment depends upon the excess of anticipated cash flows over the amount needed to service liabilities that were issued to finance such acquisitions in the past. This scenario describes perfectly the situation of the Crédit Mobilier at its foundation. The power it had on the financial market influenced the value of assets and was at the origin of the speculative movements beginning in 1864. Thus capitalist techniques of valuing capital assets, the market determination of liability structures, and the possibility of sharp increases and decreases in the market price of capital assets and financial instruments lead to systematic increases and decreases in the price of assets relative to the price level of current output. Once current profits fall to the extent that cash flows to highly indebted operations are insufficient to meet commitments or liabilities, then the pressure of the need to validate debts and to meet withdrawals by depository institutions leads to a proliferation of attempts to make positions by selling out. The result can be a sharp fall in asset values. These considerations cannot be applied to the Crédit Mobilier’s case, as was shown in Chapter 2. The literature relative to investment constraint also suggests that a related transmission mechanism may operate through the banking sector: a reduction in bank liquidity makes it difficult for firms to obtain capital. As Bernanke (1983) has shown, the large fall in bank liquidity may help to explain the depth and the persistence of the Great Depression. This, however, does not apply to the Crédit Mobilier’s case. The fall of the bank did not provoke any financial crisis. This leads us to analyse the position of the Crédit Mobilier in the banking sector in France.
(b) Did the bankruptcy of the Crédit Mobilier provoke a banking panic? The term ‘banking panic’ is often used somewhat ambiguously and, in many cases, synonymously with events in which banks fail, such as recessions and stock market crashes. Historically, bank debt has consisted largely of liabilities which circulate as a medium of exchange, such as bank notes and demand deposits. The contract defining this debt allowed the debtholder the right to redeem the debt (in hard currency) on demand at par.
114
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
We define a banking panic as follows. A banking panic occurs when bank debtholders as a whole, or many banks in the banking system, suddenly demand that banks convert their debt claims into cash (at par) to such an extent that the banks suspend convertibility of their debt into cash. One theory of banking panics is based on identifying the conditions under which bank depositors would rationally change their beliefs about the riskiness of their bank. The core of the theory is that banking panics serve a positive function in monitoring banks in an environment where there is asymmetric information about bank performance. In an environment with asymmetric information, a panic can occur as follows. Bank depositors may receive information leading them to revise their assessment of the risk of banks, but they do not know which individual banks are most likely to be affected. Since depositors are unable to distinguish individual bank risks, they may withdraw a large volume of deposits from all banks in response to a signal. Banks then suspend convertibility and a period follows during which the banks themselves sort out which banks among them are insolvent. Indeed, it is possible to view panics as a means for depositors to force banks to resolve asymmetries of information through collective action (i.e. monitoring and closure) The efficiency of this mechanism derives from a supposed comparative advantage that banks possess. This last analysis explicitly utilises the notion of liquidity to define the banking panic. This parameter has been studied when analysing the relationship between investment and cash flow. In particular, we can ask the question: was there an increased sensitivity of investment to cash flow by the Crédit Mobilier’s affiliated companies in the period preceding and following the bankruptcy of the bank, related to the changes in the financing patterns that occurred at the same time? Answering this last point will enable us to give a positive answer to the question as to whether firms had anticipated the difficulties the bank was to face later. Is the following hypothesis as stated in Hoshi, Kashyap and Scharfstein (1991), true in the Crédit Mobilier’s case: ‘Bank relationships relax liquidity constraints’? The answer will be ambiguous, for different reasons invoked in the paper. As Q is greater for affiliated firms than for non-affiliated ones, the bankfirm relationship seems to be profitable. However, as nothing really important happened during the year of the bankruptcy, we can add the following question: ‘Do firms care who provides their financing?’ If the manager of a firm has to choose between banks and shareholders to finance a project, the answer will be positive, because the origin of the funds is different (the information cost in particular is different between bank and shareholders). If the manager has to choose between different banks the answer is a function of the debt policy and of the power of these banks to generate a good reputation for the firm. This second situation is illustrated
115
THE ROLE OF BANKS IN MONITORING FIRMS
by the Crédit Mobilier case. In particular, despite the facility to obtain liquidity from the Crédit Mobilier, some firms may have preferred more control to more flexibility. They may have chosen another bank to signal their honesty.
(c) Conclusion The results for the Crédit Mobilier’s case are different from the German Goßbanken or the J.P.Morgan Company. Becht and Ramirez conclude that ‘mining and steel companies affiliated with one of Imperial Germany’s “Great Banks” were not liquidity constrained’. Their paper provides evidence suggesting that this limit can be very high. The power of the Großbanken seems to have helped the development of the mining industry by alleviating the impact of capital market imperfections. The Crédit Mobilier’s case, however, proves the fragility of banks’ power; it shows that the supervisory role can be very limited and dispensable. The Crédit Mobilier was vulnerable to speculative movements. It is possible that the hostility of famous financiers like the Rothschilds contributed to its continuously decreasing power in financial circles. 4.3 ADVANTAGES AND DISADVANTAGES OF AN ASSOCIATION WITH A BANK FOR THE FINANCING OF PRODUCTIVE PROJECTS The bank-firm association can be debated especially as regards the results concerning the correlation of investment cash flows. More particularly, the fixed effect investment model proves that the interaction coefficient (which takes the value one times the cash flow when the corporation is part of the Crédit Mobilier sample) is very small and the standard errors are large except for the years 1861–7. The extensions of the basic model exhibit a liquidity constraint that is larger for affiliated companies than for nonaffiliated ones over 1861–7. The efficiency of a bank-firm association is, then, the object of discussion. We proceed here in two ways: 1 2
a general analysis of the bank-firm association; as in our sample the non-affiliated companies were in a relationship with the Crédit Lyonnais, a comparison of the behaviour of the two banks is drawn in order to explain the puzzling effects on investment policy of firms affiliated with the Crédit Mobilier’s case.
116
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
(a) The (dis)advantages of an association between a bank and firms: a theoretical analysis To begin our thesis, let us give some of the background to the banking sector in the nineteenth century. During the period of this study there were a number of sounder financial institutions than the Crédit Mobilier, which had an enormously swollen industrial portfolio greatly exceeding its capital, and which was dependent on favourable developments on the stock exchange to support the continuation of its activities. Universal banks such as J.P.Morgan and the Großbanken established the closest possible relations with their client industrial enterprises. Universal banks accompanied an industrial enterprise from the cradle to the grave, from establishment to liquidation through all the vicissitudes of their existence. Through the device of formally short-term, but in reality long-term current account credits and from the development of the institutions of the supervisory boards to the position of the most powerful organs within corporate organisations, the banks acquired a formidable degree of power over industrial enterprises, which extended far beyond the sphere of financial control into that of entrepreneurial and managerial decisions. This argument is strongly supported by evidence presented by de Long and Becht, Ramirez and Becht and Ramirez. Association with the Morgan company or with one of the Großbanken appears to have added value to enterprises. These were banks that took an active part in the decisions related to choosing a project, in order to defend their reputation in the financial markets. In the Crédit Mobilier’s case this argument is problematic, for one essential reason: as noted above, the bank extended its role to that of a shareholder (and often the principal shareholders of the firm). Therefore it suffered the same risks as an ordinary shareholder. The answer to the question: ‘Is it preferable for banks and firms to be associated or not?’ is neither always yes nor always no. For Aoki (1988) the bank can be a monitoring agent. Two essential reasons can be given to support this proposition: 1
2
The firm can raise investment funds by borrowing from the bank at a low cost, and does not have to rely on equity financing. Therefore, its strategic business decisions (corporate policy) can be made in the interest of individual stockholders. Individual stockholders of the firm are underprivileged in the sharing of economic returns to the firm, as they receive only a very small fraction of profit as dividend.
Hence the main bank is in a position to be briefed about the company’s general business and affairs in the capacity of a major stockholder and is
117
THE ROLE OF BANKS IN MONITORING FIRMS
able to scrutinise the company’s strategic plan in the capacity of a major lender. It often sends its representative to the company’s board of directors. As point out by Wrück (1990), highly concentrated bank borrowings are also associated with superior performance. For Japanese firms, close bank lending relationships and equity ownership go together. On average, the largest lender holds 23 per cent of the firm’s bank debt and 4 per cent of its equity. The implications of the results in Hoshi et al. for US firms are difficult to determine because the financial structures associated with superior performance by Japanese firms in financial distress are illegal in the US, where the Glass-Steagall Act prevents banks from holding large equity positions in firms, including the firms that borrow from them. Thus only the main bank-cum-major stockholder has sufficient ability to closely monitor its customer companies. This situation corresponds to the Crédit Mobilier’s Case. However, did it exercise its expertise and informational advantage only for its own gain? One possible scenario is as follows: in the normal course of affairs and business of the portfolio company, the individual stockholder, obviously disadvantaged in his/her capacity to gather information and monitor the management of the company, may feel more secure if the main bank closely and responsibly monitors management of the company to reduce the risk of bankruptcy. One may thus say that the individual stockholder is willing to delegate the monitoring function to the main bank-cum-major stockholder. The deviation from share price maximisation by the company at the expense of individual stockholders and in the interests of the bank may, in part, be regarded as an ‘agency fee’ paid by the individual stockholders to the bank for that service. If the above scenario is correct, then it is vital for the main bank to maintain its reputation as a competent and responsible monitor. This may explain an aspect of the rescue operation put forth by a main bank when it discovers serious financial difficulties in a company. The problem in the Crédit Mobilier’s case is that this argument was pushed to the extreme. When the Compagnie Immobilière faced difficulties, the bank tried everything to rescue the firm, eventually provoking the bankruptcy of the bank itself. This proves that the bank-firm association contains some risks and hence disadvantages. However, the Crédit Mobilier presents some peculiarites: as soon as the bank chose to rescue the Compagnie Immobilière whatever the risk, it stopped acting as a bank. This company had played a special role in the Crédit Mobilier’s story. This vulnerable young firm needed big loans. In anticipation of high returns, the Péreires chose to invest freely. This attitude was not totally unreasonable: other banks in the same period granted risky loans to support the financing of new companies. The problem as regards the Crédit Mobilier was that the profit took longer to be realised than had been foreseen.
118
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
The Crédit Mobilier had attempted a new style of banking, but never managed to apply it correctly: it still stuck to the old style, which consisted of holding 51 per cent of the assets of a company in order to participate actively on its financing. The companies affiliated to the Crédit Mobilier were joint stock companies managed like limited partnerships, which corresponded to the French tradition. Concerning the advantage for a firm of an association with a specific bank, Mayer (1988) suggests that there exist systematic differences in performance between the financial systems where banks play a primary role and those where the banks are not active in the same way. These differences must be explained by the mechanisms that each system provides to suppress the moral hazard in relations between entrepreneurs and financiers. More specifically, Mayer affirms that the systems proposed by the German or Japanese banks enable the firms and the banks to sustain long-term contracts which in turn enable them to extract the surplus associated with these contracts and to reduce moral hazard. The answer to the question of the (dis)advantage of bank-firm association will substantially depend on: •
•
•
the proportion of assets held by the bank: are the decisions taken by the shareholders’ assembly the expression of the majority or, as in the Péreires’ case, the will of the founders? The results of our tests prove that the second alternative is more realistic in the Crédit Mobilier’s case. In fact this argument leads us to analyse the position of shareholders when a bank holds a large amount of assets in a firm. Considering the Crédit Mobilier’s case, it seems that the role they played was limited. We will discuss this idea in the next section when comparing the French case to the German one; the expectations of financiers on the possible profitability of the firm (e.g. confidence on the potential revenue for the bank in the future); this reflects essentially the bankers’ point of view, whose aim is to make profit; the power the bank exerts on the financial market (e.g. capacity to solve the capital market imperfections). This last point corresponds essentially to the situation of the German Großbanken, in which case an association was profitable on average: for example adding value to the enterprise’s shares. Conclusion
It is relevant to point to the stress laid by Saint Simon and his followers upon industrialisation and the great task they had assigned to banks as an instrument for organisation and development of the economy. This, no doubt, appealed greatly to the creators of the Crédit Mobilier, who liked to
119
THE ROLE OF BANKS IN MONITORING FIRMS
think of their institution as a ‘bank to a higher power’ and of themselves as ‘missionaries’ rather than bankers. Despite the collapse in 1867, the Péreires profoundly influenced the history of banking in Europe from the second half of the last century onwards. As Gerschenkron (1962:13) noted, the number of banks in various countries created in the image of the Péreires’ bank was considerable. However, more important than their imitation was the creative adaptation of the basic idea of the Péreires and its incorporation in the new type of bank, the universal bank, which in Germany, along with most other countries on the Continent, became the dominant form of banking. The difference between banks of the Crédit Mobilier type and commercial banks in the most advanced industrial country of the time (England) was very great. The German banks, which may be taken as a paradigm of the type of the universal bank, successfully combined the basic idea of the Crédit Mobilier with the short-term activities of commercial banks. Over the period, the Crédit Mobilier had acted more as an investor than a banker which supervises and controls the firms with which it was involved. This justifies the idea that the Crédit Mobilier essentially provided liquidity to the enterprises and did not supervise them. A bank-enterprise association is then profitable for both parties, so long as both play an appropriate role. The bank must screen companies according to their expected returns; a firm can expect additional value from a bank in terms of the power it has to solve capital market imperfections.
(b) A comparative analysis of the Crédit Mobilier and the Crédit Lyonnais over the period When presenting the methodology of the investment regression we have stressed the importance for all companies of their sources of financing. On the one side, the affiliated companies were obviously involved with the Crédit Mobilier, on the other side the -affiliated ones were financially related to the Crédit Lyonnais. The statistical analysis proves that the nonaffiliated companies were not dramatically credit-constrained. As both samples were affiliated with a bank, it will be interesting to compare the finance policy in order to interpret the results of Chapter 2 and 3. With this aim, we ask four questions: • • • •
Were Were What Were
both of the banks merchant banks? the specific activities of the two banks similar? were the characteristics of the credit policies of the two banks? the dividend policies similar for the two establishments?
120
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
The definition of the two banks The argument developed here is based on a definition taken from the Encyclopaedia Universalis (1985). The way the Crédit Lyonnais was created was very different from the Crédit Mobilier. The Crédit Lyonnais was a limited liability company 4 founded for a period of thirty years from 1 July 1863 and whose main office was fixed at Lyon. It was founded first to receive only deposits and to use them to grant, in the very short term, credit to industrial establishments. Unlike the Crédit Mobilier, no long-term loans or equity holdings were contemplated. The objective of the Crédit Lyonnais is set out as follows by its founder, Henri Germain: The cash currently held in our bank was in great part hidden in the drawers, without any interest for those who kept it, without any utility for French society.5 The aim of the Crédit Lyonnais was quite different from the Crédit Mobilier. In 1854, Isaac Péreire had declared that the Crédit Mobilier was ‘a limited partnership’ for industries, 6 a lending bank and a bank of issue. The ambitions of this bank were wider than those of the Crédit Lyonnais. This diversification made the Crédit Mobilier powerful, but was also at the origin of its difficulties. Specification of its different activities will enable us to understand the functioning of the establishment. The specific activities of the two establishments As we have shown with the statistical tests presented in the two preceding chapters, the Crédit Mobilier’s main activity was to provide liquidity to firms in order to finance productive projects during the period of French industrialisation. For the Crédit Lyonnais, two types of activity could be distinguished: •
•
participation in the productive investment (like the Crédit Mobilier) which represents the ordinary transactions; in other words the ordinary affairs including overdrafts on current accounts; financial participation such as government loans, advances to foreign government and huge investments in large companies which necessitated a huge amount of liquidity. These were quite limited.
Table 4.1, after Bouvier (1968), illustrates this latter point. Bouvier asserts that during its early years, the Crédit Lyonnais was a kind of readjusted Crédit Mobilier, less ambitious, more prudent and more solid. If we study Table 4.1, we note that the ‘big deals’ only exceptionally
121
THE ROLE OF BANKS IN MONITORING FIRMS
Table 4.1 Financial participation of the Crédit Lyonnais in ‘big deals’
represented more than half of the total transactions of the bank. This enabled the bank to avoid liquidity problems during the French financial crises of 1870 and 1875. The Crédit Mobilier, on the other hand, was closer to what Sayous (1907) calls a financial firm. Talking of these firms he said: But they (the financial firms) generally play a more active role: they hold a block of shares, prepare the market or wait for any favourable circumstance and pass the assets to the public; they themselves transform the private enterprises into limited companies from which they later issue certificates. They go as far as to create industries or to organise business using an idea which was suggested to them. They are then the brokers, the traders of this modern merchandise called stocks: they facilitate and even organise the contact between savings and those who need it.7 This definition describes the Crédit Mobilier’s range of activities. Sayous adds that generally these financial institutions took a speculative attitude to the stock market. Large participations in firms and risk-taking were the order of the day. This can partly explain the empirical results obtained in Chapters 2 and 3: the affiliated companies were liquidity constrained before the bankruptcy because of a defensive attitude towards the Crédit Mobilier’s credit policy. Such an argument will not be valid in the Crédit Lyonnais’s case. This is the argument we intend to discuss now.
122
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
The credit policy of the Crédit Lyonnais This point is the area of most significant difference from the Crédit Mobilier. As we have already pointed out, the Crédit Mobilier granted credit to companies with which it was involved without asking for any collateral. This might seem to have been totally unreasonable when we consider the traditional economic view that the provision of collateral is a good signal for the bank as regards the quality of a project and the reliability of the enterprise which offers it. My hypothesis is that the Crédit Mobilier’s practice has to be considered in the context of a country which was at the beginning of its industrialisation and where the notion of risk was totally different. Investors and bankers had to accept risks in the short term in order to make profits in the long term. The policy of the Crédit Lyonnais was totally different to this. The most significant example is the establishment as early as 1870 (that is, only seven years after the foundation of the establishment) of a department in charge of the financial study of the different customers of the bank. Its duty was to establish secure information, brought regularly up to date, concerning the financial circumstances of the most important companies in France and abroad. However, it is perhaps significant that this department was established in 1870, after the bankruptcy and liquidation of the Crédit Mobilier. Probably the Crédit Lyonnais was responding to the experience of this failure by adopting a more prudent attitude. The policy towards the distribution of dividend The Crédit Mobilier was in the habit of distributing large dividends to its shareholders. I will support this claim in more detail below. The point here is to stress the difference between the Crédit Mobilier and the Crédit Lyonnais towards dividend policy. Henri Germain applied the theory of nondistribution of the exceptional benefits. In fact, he introduced to the bank a technique which was close to industrial self-financing, distinguishing between normal benefits arising from the short-term transactions (the classical banking services), non-distributable benefits (statutory reserves, increase of capital, etc.), and exceptional dividends. Of the total profit realised by the bank, only a small percentage was distributed to the shareholders: around 15 per cent. This illustrates the determination of the Crédit Lyonnais to maintain its solvency.
(c) Conclusion The points outlined above have a bearing on the validity of the econometric results set out earlier. All the statistical analysis has to be interpreted in the
123
THE ROLE OF BANKS IN MONITORING FIRMS
context of the Crédit Mobilier’s failure. The analysis relates to movements in investment, share prices and cash flows for affiliated and non-affiliated companies. Recent articles have explored the empirical relationship between corporate financial structure and investment.8 Fazzari, Hubbard and Petersen (1988) discussed the idea according to which the supply of investment finance is not perfectly elastic for firms that face asymmetric information problems in financial markets. They find that investment is more sensitive to cash flows for firms that consistently retain a larger fraction of their earnings. Hoshi, Kashyap and Scharfstein, on the other hand, show that investment expenditures of Japanese companies affiliated to a bank (i.e. a member of a keiretsu) display almost no sensitivity to cash flows, in contrast to the investment expenditures of firms which are not so affiliated. This finding supports the hypothesis that companies that are not affiliated to a bank tend to be more liquidity constrained (and face a larger gap between external and internal costs of funds) than those which are. The absence of any fluctuations for affiliated companies is a quite surprising fact that we have already discussed in the preceding section. As the non-affiliated Crédit Mobilier’s companies were involved with the Crédit Lyonnais, the preceding explanation of the attitude of the bank as regards credit and dividend helps to explain why the non-affiliated companies were not credit constrained. So, despite the small number of companies surveyed, we can advance the hypothesis that the analysis of a bank failure illustrates the capacity of firms to diversify their financing to avoid liquidity constraint. J.Bouvier provides evidence that supports this argument: after the bankruptcy the Crédit Lyonnais bought the insurance companies. The Rothschilds, who were investors in two railways companies (Nord and Quest) contacted the Société Générale to provide liquidity for further investment. 4.4 THE CRÉDIT MOBILIER BANKS: A COMPARATIVE ANALYSIS BETWEEN FRANCE AND GERMANY The next point we are going to consider is the divergence of the results obtained in the French and the German cases. As noted at the beginning of this chapter, the results obtained for Germany and France are very different. The aim here is to give some interpretative reasons for this difference. More specifically, we examine if the type of participation of the Großbanken in firms followed the same model as the Crédit Mobilier. The financial commitment of every ‘mixed bank’ consists in the promotion and financing of industrial firms (short- and long-term financing: the long-term corresponding to the renewal of short-term credit and advances). The German financiers were real entrepreneurs, facilitating access
124
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
of firms to the stock exchange. As Tilly (1986) mentions: ‘the German banks represented and reinforced not only the supply but also the demand for the investment funds’. The development of large firms and cartels was linked closely to the changing role of the banks in the German industrial system. The large corporate banks, which had appeared in the third quarter of the century, had first played a central role in the financing of industry in the boom of the 1850s and in the years around 1870. However, from the beginning of the 1890s, the co-operation between banks and industrial concerns increased steadily.9 Long-term credit became the main basis of banks’ relations with industry, to the extent that the expansions, mergers and conversions from private to joint-stock companies became more important and more frequent than the foundation of new companies.10 The issuing of shares and bonds, which was performed by the large banks for industrial concerns, was usually only the second step after co-operation had been established on a long-term basis. The banks sought to monopolise the financial arrangement of their industrial concerns and to service them in various way by developing a comprehensive policy towards them and being prepared to accept short-term financial losses in order to secure their longterm co-operation. How can we explain the difference between France and Germany? First, as mentioned in the above review of the relations between banks and firms, the main period of development occurred in Germany around 1870 and in 1890, that is after the Crédit Mobilier’s experience and after the crises that had occurred between 1850 and 1870. Secondly, the role of the central bank was stronger in Germany than in France over the period. In particular, the Imperial Bank imposed strict rules concerning the issue of bank notes and the emission of bonds on the financial markets. On the other hand, the opening of discount window as a role of a lender of last resort and a favourable policy as regards discount factors avoided liquidity problems for banks (Tilly 1986a, b). Thirdly, German banks had less liquidity than the French banks (Sayous 1899). In a normal context, such a situation represents an advantage because speculative movements have less power. But, as stressed by the author, this placed the German banks in a position of suspension of payment if any crisis were to occur. Fourthly, the participation of the German banks in the financing of enterprises was totally different from that of the Crédit Mobilier. As the German capital market was less developed than the French one, a credit had a stronger implication for firms than in a French context. In most cases, the Crédit Mobilier held the major part of the debts. In the German case, the situation was different. In particular, the supervisory board (Aufsichtrat) elected by the annual general meeting of the shareholders (or their representatives) was obligatory by law from 1870 and was strengthened by a law of 1884. Its functions were not limited to the appointment and
125
THE ROLE OF BANKS IN MONITORING FIRMS
supervision of the executive board (Vorstand), it could also make or influence the most important strategic decisions (especially investment decisions) at its quarterly or monthly meeting. However, the banks could threaten to withdraw credit, without being requested on the supervisory board. In addition, to hold a major part of the equity was a bad signal for a Großbanken. In such a case, the shareholders’ annual general meeting played a relatively peripheral role. To illustate the efforts of Großbanken to control their clients, let us quote Jeidels (1929:126) when he reported a letter which the Dresdner Bank wrote to the Executive of the North West German Cement Syndicate on 19 November 1900: According to the notice published by your Company in the Reichsanzeiger of the 18th, we have to reckon with the possibility that the decisions will be taken at the General Meeting held on the 30th which are likely to introduce changes in the scope of your business of a kind not agreeable for us. On this ground we are regretfully obliged herewith to withdraw the credit granted to you; accordingly we ask you to make no further drafts on us and at the same time politely request that you will repay the balance due to us, at the latest by the end of this month. Should, however, no decisions of this kind be taken at the General Meeting in question and safeguards be offered us in this respect for the future, we declare ourselves very ready to enter negotiations with you with a view to the granting of a new credit. If all these arguments seem to stress the prudent and rational attitudes as regards credit policy, one point should be mentioned which concerns the question of advances to companies. Sayous (1899) affirms that the economic development of Germany, in establishing a strict distinction between issuing banks and ordinary banks had, as a consequence, facilitated the credit and financial operations; the relationship between banks and industry became stronger and stronger as the activity increased. However, the politics as regards advances was not different at all. Simple loans based on certificates and assets did not play a crucial role under the classical form in the relations between trade and industry; more important were the credits granted on current accounts. If the name differed in Germany, there is no doubt that the results were identical. Both are advances granted to industries without any special guarantee. The success of the Großbanken may simply have been due to the fact that they developed around 1870, twenty years after the mistake of the first Crédit Mobilier.
126
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
(a) Conclusion As regards the principal-agent theory, the German banks really acted as principals which supervised the firms with which they were involved. In the Crédit Mobilier’s case it seems that the firms had used the bank to finance their productive projects and managed to avoid any kind of control. The bank, which should have acted as a principal, had become the agent of the societies to which it provided liquidity. An analysis according to the degree of affiliation with the Crédit Mobilier reinforces this argument. In the sample used for affiliated companies to construct the tests concerning the share prices fluctuations and the investment, two subsamples must be considered: •
•
companies where the Crédit Mobilier was an official shareholder (and sometimes the main shareholder) and where the Péreires were the directors of the companies; companies where the Crédit Mobilier was only a financial partner.
The eformer situation, where the bank held the main part of the assets in company, would have been impossible in the German case. For the Großbanken this would have signified difficulties for the companies in negotiating their titles on the equity market. Such a ‘signal’ would have induced non-participation for a German bank (Whale 1930). This distinction is very useful with regard to the power relationships between the firms and the Crédit Mobilier. For the first subsample, power was derived from the leading position the Crédit Mobilier held by virtue of the number of shares belonging to the Péreires. However, one surprising point is that the other shareholders were strong enough to force the Péreires to resign after 1867. How can this conundrum be explained? For this subsample (excluding the Compagnie Immobilière), the Crédit Mobilier served the firms by providing sufficient liquidity for the firms to grow during the period of French industrialisation. In terms of the principal-agent theory, it seemed that the bank—in appearance the principal—in reality was an agent which served the whole firm (that is, the other block of shareholders). This perhaps helps to explain the powerful situation the firms were in when the bank failed. For the second subsample, the roles were well defined: the bank (the principal) chose to grant credit to the firms because it expected thereby to earn some profit. However, something is odd in this case: the firms did not suffer from the bankruptcy. If an association with the Crédit Mobilier had been a signal of reliability, the shares of firms should have declined in value when the bank went bankrupty and thus lost its reputation for sound financial judgement. This suggests that the Crédit Mobilier never had such a reputation.
127
THE ROLE OF BANKS IN MONITORING FIRMS
The Crédit Mobilier specialised in company promotion and provided general financial services for the enterprises it patronised. In this regard, credit was granted on the basis of future profit, which prevented the bank from evaluating properly the risk of the project presented by the firms. The exceptionally high growth experienced during the French industrialisation probably constituted one of the reasons why the Crédit Mobilier lost control over the firms’ managers. Up to that point, the firms had enjoyed an ‘exploitative’ attitude toward the bank. 4.5 THE ATTEMPT BY THE CRÉDIT MOBILIER’S SHAREHOLDERS TO PREVENT THE BANK’S BANKRUPTCY
The shareholders are generally sheep; sometimes tigers.11 Henri Germain
The purpose of this section is to see which of these two categories of shareholders best described those of the Crédit Mobilier. The behaviour of the shareholders seems to have been passive in the Crédit Mobilier’s case. We can discuss the power of the shareholders and debate if their behaviour was rational or irrational when the bank went bankrupt. We may wonder why the Crédit Mobilier’s shareholders reacted so imperceptibly when the bank began to face its first liquidity problems in 1864. Several arguments can be made to explain this attitude: •
•
•
the power of the Péreires was so great in the shareholder’s annual general meeting that it prevented anyone from influencing their investment policy; most of the shareholders of the Crédit Mobilier were big bankers; the difficulties that began in 1864 were a sign of future troubles for the bank. They deliberately chose not to react in order to divide the spoils between themselves after the bankruptcy. This constituted, in fact, a way to control the Péreires, whose attitude towards risk was judged to be reckless; during the period when share values increased, instead of diversifying their portfolio, the Péreires had distributed their capital gain to the shareholders. Some 68.2 million francs of dividends were thus wasted in the four years between 1855–6 and 1862–3. In 1864, despite a new increase in the dividend, the Crédit Mobilier faced difficulties and managed to hold firm by the deposits of current accounts. As long as they received dividends, the shareholders adopted a passive attitude
128
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
towards the policy of the bank. We can discuss whether this high dividend policy did not correspond to a partial liquidation of the establishment. Perhaps, since shareholders received a high enough dividend for four years, they considered that, even if the worst happened, they would still gain. To refer again to Henri Germain’s definition, the shareholders were consciously ‘sheep’. This argument is reinforced by the discourse of K. Paris, who states: The shareholders under the magic of language, of promises and revenues, were not worried about the sudden shocks and relevant signs that reached the asset prices issued by the firm. What was the importance of an unexpected decrease? When the Crédit Mobilier fell down, it was like a balloon to rebound and go up higher! And the frenzy of speculation was maintained as tumultuous, as intense, as disorganised as before.12 This quotation points out the enthusiasm of the shareholders for a firm which had given them huge dividends and managed in the past to endure all speculative movements. According to the traditional theory of the firm, this attitude was irrational. The firm must, generally speaking, be considered as a profitmaximising entity, that is: • •
profit maximisation is in the best interest of shareholders of the firm; shareholders can and do create incentive schemes for managers that force managers to do what is best for the shareholders. So we conclude that the managers maximise profits in order to give the shareholder the highest dividend possible and to maintain the solvency of the firm.
The case of the Crédit Mobilier contradicts these two propositions. The Péreires did not act in the shareholders’ interests, but in their own interests. As they held the major part of the capital of the bank, the power of the other shareholders was very limited. Hence the control policy inside the bank was not effective. This constitutes a good example of the difficulties in supervising companies when financiers are at the same time decision-makers in the productive process. 4.6 CONCLUSION This interpretative chapter essentially has discussed relations between banks and industry. The subject of mutual relations between banks and industrial
129
THE ROLE OF BANKS IN MONITORING FIRMS
enterprises in either furthering or hindering industrialisation, in the process of concentration, in business cycles and in economic growth generally has been the cause of speculation and debate since the turn of the century. Renewed interest among economic historians was engendered by Alexander Gerschenkron’s essay on relative economic backwardness in historical perspective in the early 1860s. Recently, the rediscovery of Rudolf Hilferding’s Finanzkapital has once again brought vividly to our attention the issue of the interactions of banking and industry. In my comments the (dis)advantages of close relationship between industrial companies and banks are presented. In particular the bankruptcy of the Crédit Mobilier raises two issues: • •
the supervisory role of a bank over the companies with which it is involved; the association between bank and firms.
To supervise firms, the bank could have: • •
•
made the borrower’s actions part of the agreement and allowed repayment to depend on the outcome of the borrower’s project; written a simple agreement (payment of such amount to be made on such a date), then made the loan if it believed that the borrower was likely to repay; used a collateral.
In the Crédit Mobilier’s case nothing like this was done. The only task the Péreires accomplished was to provide liquidity and to try to realise an immediate profit, without really taking care of the solvency of their own establishment. Another issue concerns whether the bank-firm relationship consists of a domination by the banker, or a more flexible contract. As regards the Crédit Mobilier, the policy of placing a member of the bank on the board of directors in each company demonstrates the intention of the banker to place himself in a leadership position. But the results that followed do not bear out this assumption. Close relationships between banks and industry developed everywhere basically on the same lines during the nineteenth century. However, the evolution of these relationships proceeded rather differently in different countries. Geographically, my illustrative evidence is limited to Germany and the United States. The whole set of results exhibited by the Crédit Mobilier’s case contrasts with those obtained for the German Großbanken or the J.P.Morgan company. Considering then the existing theory concerning the investment policy of firms regularly associated with a bank, it is interesting to establish their predictive power. Our results have two specifics:
130
THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER
• •
they concern a bank, the Crédit Mobilier, which went bankrupt; the set of non-affiliated companies was closely linked with the Crédit Lyonnais, another bank.
They prove anyway that, contrary to the suggestion of Bernanke and Gertler (1989), capital market imperfections do not always contribute to excessive output fluctuations. For the pre-WWI period both assumptions, the scarcity of capital (illustrated in France by an underdeveloped capital market substituted by powerful financiers) and the decisive role of banks in furthering industrialisation were tested. The main findings with regard to the Crédit Mobilier serve to intensify the debate on the assumption of the banks’ supremacy over industry.
131
5 GENERAL CONCLUSIONS
This book has considered the supervisory role between 1852 and 1880 of the Crédit Mobilier, a bank which went bankrupt in 1867 and was reorganised subsequently. Economic historians have often studied the prominent role played by the Crédit Mobilier in industrial financing in the nineteenth century. Gerschenkron elaborated a model of backwardness, in which first banks, then government, depending on how backward a country was, substituted themselves for the entrepreneurship that had been the driving force in the industrialisation of the United Kingdom, called ‘the first industrial nation’ by Peter Mathias. Cameron explained in detail how the Crédit Mobilier and its imitators were responsible for the economic development not only of France but of Continental Europe as a whole. In order to evaluate the impact of monitoring on this bank, two approaches have been adopted. First, we have studied the impact of the Crédit Mobilier’s bankruptcy on share prices for affiliated and non-affiliated companies to ascertain if—as in de Long’s study of J.P.Morgan —an association with the Crédit Mobilier added value to the firms with which the bank was involved. The sample we have constructed rejects such a hypothesis. No significant effect of the bank’s bankruptcy could be detected either on the economy as a whole or on the particular sample of twenty five firms, half of which were involved with the Péreire brothers, the others being totally independent of this bank. Our conclusion is therefore that the Crédit Mobilier does not seem to have been supervising them properly (hence the bankruptcy of the Compagnie Immobilière). The Crédit Mobilier’s ‘seal of approval’ was probably not worth as much as that of J.P.Morgan & Co or the Großbanken. De Long in his paper mentioned that ‘J.P.Morgan and Co’s approval of an issue had become…a large factor which inspires confidence in the investor and leads him to purchase.’ In our case, the active participation of the bank in a company did not ‘add value’ to the share prices. Secondly, we evaluated the monitoring role of the bank as regards its investment policy. This study of corporate finance in France concentrates on
132
GENERAL CONCLUSIONS
three samples: companies affiliated with the Crédit Mobilier, non-affiliated companies and both. We also ran the regressions for three periods: the entire period 1860–75 and two sub-samples 1860–7 and 1868–75 to compare investment policy before and after the bankruptcy. The econometric work presented here is based on the theory of investment behaviour, and more specifically, the Q models (models of investment demands). In particular, we investigate the assertion that the role that the Crédit Mobilier played in the financing of French industry was not dissimilar to that of the House of Morgan in the United States1 or the postwar keiretsu and the pre-war zaibatsu in Japan.2 In other words, the Crédit Mobilier should have supervised firms and so relaxed the credit constraint for affiliated companies. The investment equation we use fits the Q models with possible liquidity effects. If the Crédit Mobilier had adopted the same attitude as the German Großbanken, the statistical results should have shown a lower correlation between cash flow and investment for affiliated companies. The data set has not provided any evidence that the Crédit Mobilier helped the development of French industry by alleviating the impact of capital imperfections. No significant differences as regards Q and CF/K coefficients can be identified between affiliated and non-affiliated companies: neither category was liquidity constrained. Moreover, the affiliated companies were more liquidity constrained before the bankruptcy than after. Hence, the results do not fit in the Hoshi, Kashyap and Scharfstein approach. Thus, despite the small sample, we reveal a situation where a bank was clearly not capable of controlling firms. This constitutes probably the most interesting point of the research. The Crédit Mobilier is a good illustration of the difficulties a bank can encounter when supervising a firm. Even if the financial literature insists on the fact that a bank can also put pressure on firms to enforce them to behave safely, these techniques are not efficient when the institution stops behaving like a real bank. Historians have been indulgent to the Péreires. They were not real bankers, but promoters of enterprises: they aimed to expand the industrial sector and to gain prestige and power through their ever-increasing role in financing firms. The Péreires’ attitude towards risk was inadequate. Even if the expected profit in the short term was sometimes high, they should have been able to evaluate the risk properly in order to assure continuity in their policy. If we admit the idea that the Péreire brothers were something other than bankers, we can question whether their establishment was really a merchant bank and discuss if, as mentioned in most of the literature, the Crédit Mobilier really was ‘a great bank’ which significantly contributed to French industrialisation. As regards the econometric work done in this research and the existing literature, it seems that the definition of the Crédit Mobilier has been
133
THE ROLE OF BANKS IN MONITORING FIRMS
unclear. When the bank was founded, Courtois fils, a specialist on financial questions and author of a Traité élémentaire des operations de bourse et de change3 wrote: The credit establishments which contribute to the conversion of fixed capital in stocks, are generally said to be Crédit Mobilier.4 This definition does not make clear the role and the limitation of such establishments. The merchant banks are distinct from other banks in the sense that they are, at least in France, the only institutions to offer the companies some help in their creation and organisation. The French merchant bank must evaluate the worth of the potential project that the firm wishes to undertake. It founds and reorganises the company around this idea. Its aim is not to keep control permanently: once the enterprise is well established in its industrial sector, the merchant bank will publicise its assets, mostly by its introduction on the stock market. The merchant bank does not expect any further profit from its participations: it will maintain a very small holding in the company and a representation in the shareholders’ assembly. This was not how the Crédit Mobilier saw its role: it was not a merchant bank but an investment bank: participation for the Péreire brothers signified a source of profit to be exploited in both the short and long term. The most serious restrictions to our conclusions are: • •
the short period over which the results are valid; the sample is constituted of firms which were affiliated either to the Crédit Mobilier or the Crédit Lyonnais. However this disadvantage (both groups of companies had relations with a powerful bank that helped French industrialisation) allowed us to compare the two institutions as regards their investment policy.
The Péreire brothers, with their initiative and their ambitions were themselves the cause of the liquidity difficulties. The search for immediate profit led them to make huge investments (e.g. the Compagnie Immobilière). Henri Germain of the Crédit Lyonnais was more careful: The Crédit Lyonnais was a long-term work and twenty years of trial and error were needed by its founder to make its foundations solid. More lucky than the Péreires, it had the chance and the ability to last. 5 Secondly, as Levy-Leboyer (1976) mentioned, the Crédit Mobilier’s failure has to be understood within the context of stabilisation of the market, with the Crédit Foncier between 1867 and 1869 willing to consolidate its liabilities especially towards the City of Paris and to reduce its short-term
134
GENERAL CONCLUSIONS
debt. This establishment had hastened the crisis by imposing a veto, on 3 August 1867, on new advances to the ‘Compagnie Immobilière’. The balance sheet of this society had, up to then, been maintained in equilibrium by an overvaluation of its stocks. Moreover the Banque de France chose not to fulfil its ‘lender of last resort’ role properly. Thus it seemed that the Crédit Mobilier was the victim of these establishments: both institutions considered the financial situation of the bank to be so catastrophic that they adopted a laissez-faire policy. In some sense, the Péreires were naive to believe that they could bring about radical change to French banking policy. Even if they intended to manage their firms as ‘limited liabilities’, they pursued the old techniques (limited partnership companies) which consisted in holding the majority of assets in a firm. They succeeded in most cases (e.g. the railways). However, one firm, the Compagnie Immobilière, provoked their downfall. This proves that being a banker requires not only stability and ingenuity, but also the ability to adapt to the rules of financial society.
135
APPENDIX A Data on the Crédit Mobilier
A.1 THE CRÉDIT MOBILIER BETWEEN 1852 AND 1867 The following data are presented in order to understand better the place of the bank in the financial market of the time. The balance sheet of the Crédit Mobilier is presented in order to discuss the policy regarding investment and dividends in the operations of the Crédit Mobilier (cf. Table A1). The investment account and the cash/balance account show us the power the bank had over this first period: these two accounts reflect the risky attitude of the founders. This is reinforced by the dividend distribution (described by the two last columns and the data below). The dividends declared were considerably higher than those distributed by other merchant banks during the same period. Further information on the bank is also disclose by a careful examination of the share prices fluctuations over the period. Let us first consider the prices reported in Table A3. These figures reflect the apparent power of the institution from 1852 to 1867. Some comments can be made on these prices. In 1856 and 1857 the fluctuations of the share prices are large; at that time the bank issued 240,000 new shares at a price of 500 francs. In 1862 the share price went from 705 francs in January to 1,285 francs in October. The general increase in liquidity in France was probably one reason why lots of speculators bought shares in financial markets in this period. In 1864 the increase of liquidity of the Banque de France caused the same phenomenon. A.2 THE CRÉDIT MOBILIER BETWEEN 1903 AND 1914 Table A4 reflects the improvement concerning the presentation of the private accounts. The Crédit Mobilier had become, over the period, a normal merchant bank, less powerful than the one of the Péreires.
136
Source: Dupont Ferrier (1925:234).
Table A1 Balance sheet of the Crédit Mobilier from 1853 to 1866 (in millions of francs)
THE ROLE OF BANKS IN MONITORING FIRMS
Table A2 Average profit of the Crédit Mobilier from 1855 to 1865
Source: Jean Bouvier et al (1965:449).
Table A3 Share prices of the Crédit Mobilier from 1853 to 1866
Source: (1903:111).
138
Source: Baldy (1922:234).
Table A4 Balance sheet of the Crédit Mobilier from 1903 to 1914 (in millions of francs)
APPENDIX B Data relative to the general indices (GNP, share prices) for France between 1852 and 1914
The aim is to present the data necessary to construct the excess volatility test. B.1 GENERAL PRICE INDICES The sources of data on prices are Chadeau (1988:216) and Levy Leboyer (1970:79–119). The general price index A is composed of the two following series: B agricultural price index and C industrial index. The repartition between the two sectors is implied by the weight attached to agriculture and industry in the GNP. For a detailed presentation see Levy Leboyer’s article pp. 115–17. B.2 SHARE PRICE SERIES Table A5 Share price series
140
APPENDIX B
Source: Dessirier (1928b:167). Notes: (a)According to the average annual market price of the rent 3% (average between the maximum and the minimum price until 1918). (b)Average rate: 1870–82 rent 3%, bonds of the railways 3% and industrial bonds 3%, 1883–92 the three latter and ground rent 3%, 1893–1913 in addition industrial bonds 4%. (c)According to the average of the discounted rate taken from the monthly balance sheets. (d)Average rate of eight groups of assets with fixed revenue: 3.84 (for the five quoted in note (b)and in addition ground rent 4% in 1913, railways bonds 4% and industrial bonds 5%).
141
APPENDIX C Data relative to the affiliated companies
C.1 THE WHOLE DATA The aim of this appendix is to present the data underlying the analysis presented in Chapter 2. The share prices between 1852 and 1867 presented here constitute an insight of the whole data set used to run the regression and will be provided in order to give an idea of the impact of the Crédit Mobilier’s bankruptcy on the share prices of companies. As regards Table A6 no fluctuations in the share prices can be noted for the set of affliated companies. C.2 THE COMPAGNIE IMMOBILIÈRE The Compagnie Immobilière was constituted from three societies. The first, founded in December 1854, was the Société de l’Hotel et des Immeubles de la rue de Rivoli. Its capital of 24 million francs was divided in 240,000 shares of 100 francs. E. and I.Péreire possessed at its foundation 42,220 shares; the Crédit Mobilier as a whole more than 182,195. E. Péreire was its President. In 1858, after a modification of its statute, the company was renamed the Compagnie Immobilière de Paris. In 1859, it merged with the Société des ports de Marseille whose capital of 1,500,000 francs was divided into 30,000 shares of 500 francs. In June 1863, the society reached an agreement with the Société de la rue Impériale de Marseille. A distribution of 125 francs of dividends was made to the shareholders of the Crédit Mobilier. The bankruptcy of the first Crédit Mobilier was largely caused by that of the Compagnie Immobilière. It is interesting to examine the situation of this company from 1870 up to 1874. As we have already pointed out, the Crédit Mobilier had granted this company many advances and loans, using the Crédit Foncier as intermediary. In 1870, the net profits that the shareholders could expect after having paid all the interests on loans would be:
142
APPENDIX C
in in in in
1871, 1872, 1873, 1874,
985,031.73 francs; 2,419,811.03 francs; 3,605,013.19 francs; 4,142,288.97 francs.
The reorganisation of the Crédit Mobilier after the fall of the Péreires obliged the society to liquidate part of the claims it had with its partners, that is the Crédit Foncier. The revenues of the company during the period were as shown in Table A7. The expenses were as shown in Table A8. In March 1881, the Crédit Mobilier reached an agreement with the liquidators of the Compagnie Immobilière to rebuy some property not yet liquidated. From this period on, all the claims with the enterprise were resolved. Table A9 shows the share price and the dividend distribution between 1855 and 1866.
143
Note: na stands for not available.
Source: Annales de la Statistique de Paris: Numéro Récapitulatif (1954) pp. 180 et seq. for the railways; private account 65AQ Archives Nationales de Paris for the insurance company l’Union, the Compagnie Transatlantique, the Compagnie Immobilière, the Compagnie Parisienne d’Eclairage.
Table A6 Data available for affiliated companies: share prices (francs)
APPENDIX C
Table A7 Revenues of the Compagnie Immobilière (francs)
Source: Private accounts on the company 65 AQ Archives Nationales de Paris. Table A8 Expenses of the Compagnie Immobilière (francs)
Source: Private accounts on the company 65 AQ Archives Nationales de Paris. Table A9 Share price and dividend distribution, 1855–66
Source: Private accounts on the company 65 AQ Archives Nationales de Paris. Note: *For 1864 no data are available.
145
APPENDIX D Data relative to non-affiliated companies
146
Source: Chadeau (1988:120), Gille (1968:233–79).
Table A10 Data for non-affiliated companies
APPENDIX E Monthly share prices for affiliated companies between 1866 and 1868
The most remarkable observation concerning these figures is the absence of fluctuations for the affiliated companies except for the Compagnie Immobilière and the Compagnie Transatlantique. The data show share prices remaining stable despite the failure of the bank.
148
Source: Le temps 1866. Note: The table reports monthly share prices evaluated the first day of each month.
Table A11 Share prices for affiliated companies for 1866 (francs)
Note: The table reports monthly share prices evaluated the first day of each month.
Source: Le temps 1867.
Table A12 Share prices for affiliated companies for 1867 (francs)
Note: The table reports monthly share prices evaluated the first day of each month.
Source: Le temps 1868.
Table A13 Share prices for affiliated companies for 1868 (francs)
APPENDIX F Data for investment and cash flow tests
In this appendix, the raw data and its transformation in order to construct the test will be described. To run the regressions, three parameters must be available for all companies in the sample: investment, cash flow, and Q. To estimate this last parameter the book values taken from the balance sheets have been used. Before reporting any data, some explanations and clarifications might be made. The investment variable is clearly specified in the balance sheets and exhibits few changes as regards the actual accounting . If we now consider the existing literature on Tobin’s Q, it appears that most authors take into consideration Q which is the ratio of the valuation of corporate physical capital in the stock markets to its estimated cost of reproduction. In the statistical work presented here we do not use this definition of Q but rather common equity Q. Because of insufficient data on the market value of debt, we could only approximate ‘Tobin’s Q’ with what we should be calling strictly speaking ‘common equity Q’. As regards the definition of the variable K it corresponds to the ‘capital brut’ (gross capital) which over the period did not include a systematic depreciation rate. In order to avoid any confusion it will be called Cb. Table A14, A15 and A16 describe the values of shares that we call p, investment, gross values of assets Cb and cash flows respectively for the railway companies, the mining industry and the steel firms. The sample cover the years 1860–75. As can be seen, some values are missing for certain years. For the unavailable data, missing values have been substituted. The situation of the Compagnie Immobilière must be considered separately. As this company went bankrupt at the same time as the Crédit Mobilier, little information was available over the period (from 1861 until 1869). However, the introduction of this element in the sample is relevant as it produces many elements of explanation and interpretation on the bank’s policy toward investment (the analysis of the results proves that the Crédit Mobilier did not behave like a Großbanken before the bankruptcy, but changed policy after 1868). The sources of our information in this special case are the balance sheets collected in the Archives Nationales de Paris and
152
Legend: p stands for value of share, Cb represents the gross value of assets, cf are the cash flows, i is the gross investment.
Sources: Balance sheets collected in the Archives Nationales de Paris
Table A14 Data available for the railway companies
Legend: p stands for value of share, Cb represents the gross value of assets, cf are the cash flows, i is the gross investment.
Sources: Balance sheets collected in the Archives Nationales de Paris.
Table A15 Data available for mining industries
Legend: p stands for value of share, Cb represents the gross value of assets, cf are the cash flows, i is the gross investment.
Sources: Balance sheets collected in the Archives Nationales de Paris.
Table A16 Data available for the steel industries
THE ROLE OF BANKS IN MONITORING FIRMS
also some financial newspapers over the period. The data are reported in Table A17. All figures represent real values in millions of francs. Some general comments can be made on Table A17. In comparison to the first three, we remark that the fluctuations of the parameters are larger and seem to be difficult to justify. They reflect in some sense the ‘irrationality’ of the Péreires as regards their investment policy. Up to the end (1868 represents a significant date), the bank continued to believe in the profitability of the company and provided a large amount of liquidity (e.g. the magnitude of investment i). The test presented for this company alone will support this argument. The last variable, Q, is simply the ratio of the market value to book value. This procedure is virtually identical to the one de Long used in his work. As the raw data do not conform to traditional accounting rules, some transformation has been made. First a depreciation rate δ is introduced for the book value. The first value noted Cbk1861 is chosen as I1861/δ where δ=5 per cent. Some words of explanation are necessary to justify the choice of departure for the calculation. If we refer to Jean Bouvier et al. (1965:165– 91) we can note that there was a lack of information for the preceding period. According to the data available for this work we then choose to begin in 1860. The series is then evaluated by: Cbkt+1 = Cbkt(1 – δ) + It where d is a constant and chosen such that the series is slightly increasing. This transformation of the data leads to an increase of Cb. Table A17 The Compagnie Immobilière
Sources: Balance sheets collected in the Archives Nationales de Paris. Legend: p stands for value of share, Cb represents the gross value of assets, cf are the cash flows and i is the gross investment.
156
Table A18 New values of Q when introducing a depreciation rate of 5 per cent
THE ROLE OF BANKS IN MONITORING FIRMS
Table A18 exhibits the new value of Q. Some remarks can be made on these figures. If the series exhibits numbers that are increasing slightly over the period, the figures obtained for Q are quite different from those obtained previously. In order to analyse this difference, we propose to see what would have been the data had another origin been chosen. Let us consider that the first value is Cbk1862=I1862/δ. The following value are then obtained by
The whole series is then given by: Cbkt+1 = Cbkt(1 – δ) + It. As the set of data as a whole has the same characteristic (slightly increasing) over the period we concentrate our argument on the first values for a few companies (three railway companies, for example). The figures are reported in Table A19. It appears that the first values have an influence on the data: if we compare the third and the fourth columns, we see that the figures for Q are increasing. The same phenomenon is to be noted for the other cases. However the biggest change comes from the introduction of a depreciation rate. The second step of our argument will be to choose other values for d. In order to see if our choice of d is valuable or not, we evaluate two other series by considering small variations for d. The idea is then to discuss the modification of data when choosing a depreciation rate smaller then 5 per cent and to see the fluctuations of the same parameter when d is larger. The figures obtained for Q remain low, but fit the theory (i.e. they are smaller and drecreasing when d decreases. The last step is to divide all variables by the general price index. Levy Leboyer (1970:79–119) provides us with a series. As all the figures are expressed in basis 1880, we transpose them in real terms by applying the formula:
Table A19 Transformation of the data when changing origin
158
Table A20 Data in real terms
THE ROLE OF BANKS IN MONITORING FIRMS
As before we evaluate Cbk by Cbk1861=I1861/δ. The whole series is given by: Cbkt+1 = Cbkt(1 – δ). Pt+1/Pt + It. Pt+1/Pt. The new values of Q are given in Table A20.
160
NOTES
INTRODUCTION 1 Cf. de Long (1989). 2 Cf. de Long and Becht (1992). 3 Cf. Becht and Ramirez (1993).
1 AGENCY THEORY AND MONITORING 1
Translation from the original: Chez l’industriel français qu’il soit ‘moyen’ ou ‘grand’ la hantise de l’indépendance vis à vis du créancier (la banque) fut constante. C’est au niveau plus ou moins élévé de ses comptes courents vis à vis des banques que les Schneider du XIXème siècle par exemple mesurent leur degré d’indépendance et leur capacité de mouvement…. La domination des banques sur les entreprises, si elle existait ne s’est pas faite sans résistance.
2 3 4 5 6 7 8 9 10 11 12
See in this respect Ross (1977) and (1978). A similar issue arises in the literature on insurance provision. Rothschild and Stiglitz (1976) and Miyazaki (1977) have proposed equilibrium concepts in which firms are allowed to react to changes in contracts offered by other firms. Bernheim and Whinston (1986:923). Bernheim and Whinston (1985). Cf. Jensen and Meckling (1976). Cf. Diamond (1984). Cf. Jensen and Meckling (1976). Cf. Diamond (1984) and Gale and Hellwig (1985). Cf. Broccklcr (1990). Cf. Diamond (1989), who points out that the requirement of a guarantee can address this problem and enables the lender to distinguish sound creditors from bad ones. This author has devoted a large part of his research to the study of German banks. More specifically he discusses the characteristics of these banks as regards investment policy and therefore justifies the control exerted by the banking sector on industry. As an illustration, let us quote two basic sources: Tilly (1986a) and Tilly (1986b).
161
THE ROLE OF BANKS IN MONITORING FIRMS
13 See Grossmann and Hart (1983). 14 Cf. Hoshi, Kashyap and Scharfstein (1991) and (1990). 15 See in this respect Laffont and Freixas (1987).
2 THE CRÉDIT MOBILIER AND THE FRENCH STOCK EXCHANGE 1853–1914: AN EMPIRICAL PERSPECTIVE 1 Translated from the original: Le sort du Crédit Mobilier n’est pas seulement caractéristique des rivalités entre groupes financiers et des dangers de la spéculation boursière: il rend sensible le problèms permanent de l’engagement industriel. 2 Cf. Baldy (1922). The author provides several examples, among which are the Crédit Lyonnais and the Banque de l’Union Parisienne. Some explanations are necessary on what could be understood by ‘control’. The literature on asymmetric information and incomplete contract presented in the first section indicates that debt does not provide control. Instead it provides access to privileged information, which the bank is then able to use in the process of monitoring and enforcing the contract. Control, however, stays with the managers of firms. As the Crédit Mobilier provided both debt and equity to its customers, it can be said that this bank clearly exerted control over the firms with which it was affiliated. 3 Huth (1918:11). 4 The quotation is taken from a memoir addressed to a group of bankers whose support the Péreires hoped to obtain. 5 The whole of Bouvier, Furet and Gillet’s Le mouvement du Profit en France au 19ème Siècle provides general information about the fluctuations of share prices over the period. It gives an indication of the increasing importance of the financial market over the period. 6 By ‘short-term’ we mean three to six months. 7 See Bouvier, Furet and Gillet (1965). 8 Translation from the original: ‘Notre société a toujours considéré comme un principe de haute moralité commerciale de ne jamais ouvrir une souscription, de ne jamais recommander une entreprise sans s’y interesser elle même dans une forte proportion, et ses administrateurs se sont fait un devoir de s’y associer.’ 9 In the balance sheets where these advances were disclosed, the Administrators of the Crédit Mobilier never mentioned that these advances to companies were not based on any collateral. ‘Shares and bonds on collateral’, to cover additional loans, had to be taken into consideration independently of the advances to companies. 10 Cf. Leland and Pyle (1977). 11 See de Long and Becht (1992). 12 Statutes of the Société de Crédit Mobilier, 8 November 1871. The new bank was capitalised at 80 million francs with 160,000 shares of 500 francs each; holders of the old stock obtained two new shares for five old and could subscribe at par for 48,000 additional shares. The founders of the new bank (including only two members from the original bank, Charles Mallet and Frederic Grieninger) took the remainder. 13 Numbers of events happened over the period, including the Spanish-American war and the US annexation of Puerto Rico. In 1900 the Crédit Mobilier had
162
NOTES
invested funds for the construction of around 70 miles of railways in Puerto Rico. 14 Translation from the original: ‘nous nous sommes appliqués à accroître le nombre de nos correspondants à l’étranger et le mouvement de notre portefeuille de banque. C’est notamment dans ce but que nous avons pris des intérêts dans les institutions de Banques…Si cette politique a votre approbation, nous nous proposons d’étendre encore nos efforts dans cette voie et de nous créer de plus en plus d’appui.’ 15 Bageholt has been criticised for not stating clearly when the central bank should intervene, for not giving specific guidelines to distinguish between sound and unsound banks, and for not realising that provision of the lender of last resort facility to individual banks would encourage them to take greater risks than otherwise. See in that respect Thornton (1802). 16 In part, Humphrey’s summary of the classical position is as follows: The lender of last resort’s responsibility is to the entire financial system and not to specific institutions…The lender of last resort exists not to prevent the occurrence but rather to neutralise the impact of financial shocks…[Its] duty is a twofold one consisting first, of lending without stint during actual panics and second, of acknowledging beforehand its duty to lend freely in all future panics…[It] should be willing to advance indiscriminately to any and all sound borrowing on all sound assets no matter what the type…In no case should the central bank accommodate unsound borrowers. The lender’s duty lay in preventing panics from spreading to the sound institutions, and not in rescuing unsound ones…. All accommodations would occur at a penalty rate, i.e., the central bank should rely on price rather than non-price mechanisms to ration use of its last resort lending facility…The overriding objective of the lender of last resort was to prevent panic-induced declines in the money stock. (Humphrey 1975:9) 17 18 19 20 21 22 23 24
Cf. Dessirier (1928a), (1928b) and Lenoir (1919). Cf. Levy-Leboyer (1970). Cf. Shiller (1981) and (1989). See de Long and Becht (1992:12). Gille (1959:373). Ibid. (p. 374). Ibid. Translation from the original: ‘Les investissements cessent parce que l’accumulation du capital est détruite et qu’il n’y a plus de disponibilités; ils cessent parce que la rareté de l’argent circulant en fait remonter le taux; parce que certaines affaires se sont révélées mauvaises ou spéculatives.’ 25 Becht and de Long (1992:22). The Großbanken are still present in the German economy today, although not nearly as important as they were before the First World War. Rudolf Hildering could say that all that was needed in order to attain socialism was the nationalisation of Germany’s six largest banks. 26 Cameron (1991:65).
163
THE ROLE OF BANKS IN MONITORING FIRMS
3 CORPORATE INVESTMENT, CASH FLOW AND FINANCIAL CONSTRAINTS OF FIRMS: THE CASE OF THE CRÉDIT MOBILIER 1 2
3 4 5
See Chapter 1 with reference to credit rationing literature, and Stiglitz (1981). The debate as regards the advantage of a strong relationship between banks and industry is one of the oldest in financial economics. The originality of Hoshi, Kashyap and Scharfstein (1989, 1990, 1991) was to use an econometric methodology in order to support the argument. Hoshi, Kashyap and Scharfstein (1989) is a basic reading which constitutes the foundation of the statistical work presented in this paper. De Long (1989) and Ramirez (1992). Hoshi, Kashyap and Scharfstein (1989, 1991). Annual report: April 1860. Our company has always considered as a principle of its high commercial morality never to open a subscription, nor recommend a firm, without first having a large proportion of interest in it and its administrators having become associated with it.
6 7 8 9
Cf. Tobin (1971). E.g. Gerschendron (1962). Cf. Ramirez (1992). As we have already pointed out throughout this chapter:. Economic logic indicates that a normal equilibrium value for Q is 1 for reproducible assets which are in fact being reproduced and less than 1 for others. Values of Q above 1 should stimulate investment, in excess of requirements for replacement and normal growth, and values of Q below 1 discourage investment. (Tobin and Brainard 1977:238) 4 THE SUPERVISORY ROLE OF THE CRÉDIT MOBILIER: SOME INTERPRETATIONS 1 The original: Le Crédit Mobilier a été, qu’on me pardonne l’expression, un lanceur d’affaires. Pour lui l’intéret du présent domine presque toujours celui de l’avenir. 2 These are not the only alleged control mechanisms in a capitalist economy. For instance, there are evolutionary arguments which suggest that firms which manage their resources efficiently will survive, while those that do not, will not. For some criticisms of this argument, see, for instance, Diamond and Stiglitz (1974), Stiglitz (1982). 3 In this article Stiglitz suggests that the firm should be considered as a problem of multiple agents and multiple principals. 4 In French, ‘Société anonyme à responsabilité limitée’. 5 Translation from the original:
164
NOTES
L’argent aujourd’hui déposé notre caisse était en grande partie enfoui dans les tiroirs, sans interêt pour ceux qui le gardaient, sans utilité pour la société. 6 7
In French, ‘Société Commanditaire pour l’industrie’. Translation from the original: Mais elles (les sociétés financières) jouent généralement un rôle plus actif: elles ‘prennent ferme’ un certain ‘paquet’ de rentes ou d’obliga-tions, ‘préparent’ le marché ou attendent quelque circonstance favor-able et font ‘infiltrer’ les titres dans le public; elles ‘transforment’ elles-mêmes les enterprises privées en sociétés anonymes dont elles émettent ensuite les actions. Elles vont jusqu’à ‘créer’ des industries ou ‘organi-ser’ un commerce, utilisant une idée qui leur fut insufflée. Ce sont done les courtiers, les commissionnaires, les négociants de cette marchandise moderne que l’on appelle valeurs mobilières: elles facilitent et même organisent le contact entre l’épargne et ceux qui en ont besoin.
8 9 10
11 12
See Hoshi Kashyap and Scharstein (1990, 1991). See Becht and Ramirez (1993). Cf. Weber (1903:337): of eighteen industrial corporations in Rheinland and Westphalia founded in the years 1896–1900 and quoted on the Berlin stock exchange in 1901, sixteen emerged from the conversion of previously existing private companies; only two were new foundations. Translation from the original: Les actionnaires sont en général des moutons, parfois des tigres. Translation from the original: Les actionnaires pris à la magie du langage, des promesse et des revenus ne s’inquiètaient même plus des brusques soubresauts et des signes révèlateurs qui atteignaient le cours des titres emis par la Société. Que leur importail une chute inatlendue? Quand le Crédit Mobilier tombait, c’était comme le ballon pour rebondir et s’élever plus haut! Et la frénésie de la speculation se maintenaient aussi tumultueuse, aussi intense, aussi désordonnée. ‘Le Crédit Mobilier et ses actionnaires’ (1867:50).
5 GENERAL CONCLUSIONS 1 2 3 4 5
See de Long (1989) and Ramirez (1992). See Hoshi, Kashyap and Scharfstein (1990, 1991). Courtois (1889:13). ‘Les établissements de crédit qui aident à la conversion des capitaux stables en capitaux mobiliers, sont dit, d’une manière générale, Crédit Mobilier.’ ‘Le Crédit Lyonnais fut une oeuvre de longue haleine et il fallut à son fondateur vingt ans de tâtonnement pour l’asseoir sur des bases solides. Plus heureux que les Péreires, il eut la chance et l’habileté de durer.’
Quotation extracted from Bouvier (1968:150).
165
BIBLIOGRAPHY
Aghion, P. and P.Bolton, (1986) ‘An “Incomplete Contract” Approach of Bankruptcy’, Mimeo, MIT Harvard University, March, 484. Akerlof, G. (1970) ‘The market for lemons: qualitative uncertainty and the market mechanism’, Quarterly Journal of Economics, 84: 488–500. Aoki, M. (1984) The cooperative game theory of the firm, Oxford: Clarendon. —— (1988) Information, Incentive and Bargaining in the Japanese Economy, Cambridge University Press. Aoki, K. and J.Kyle (1992) Corporate Investment in Japan: a Foreign Investor’s Guide to Making Strategic Alliances, London: Woodhead Faulkner. Autin, J. (1884) Les Frères Péreire: le Bonheur d’Entreprendre, Paris. Aycard, M. (1867) Histoire du Crédit Mobilier 1852–1867, Paris. Bagehot, W. (1873) Lombard Street: A description of the Money Market, London: H.S.King. Baldy, E. (1922) Les Banques d’Affaire en France depuis 1900, Paris. Barham, V. and O.Lefevre (1990) ‘Bank loans securitization with asymmetric information: a principal-agent solution’, Working Paper Core Louvain La Neuve, November, 9072. Baron, D. and B.Holmstrom, (1980) ‘The investment banking contract for new issues under asymmetric information delegation and the incentives problem’, Journal of Finance, December, 4: 1115–1138. Batchelor, R.A. (1986) ‘The avoidance of catastrophe: two nineteenth century crises’, in F.Capie and E.Wood Geoffrey (eds) Financial crises and the world banking system, New York: Saint Martin’s Press. Becht, M. and C.Ramirez (1993) ‘Finance capitalism, Großbanken and the German mining and steel industry 1906–1912’, Working Paper EUI, Italy, 93/26. Becker, E.B. and C.A.Olson, (1989) ‘Unionization and shareholders’ interests’, Industrial and Labor Relations Review, 42, 2: 246–261. Benston, G.J. et al. (1986) Perspectives on Safe and Sound Banking: Past, Present and Future, Cambridge: MIT Press. Bernanke, B. (1983) ‘Nonmonetary effects of financial crisis in the propagation of the great depression’, American Economic Review, LXXIII, 2: 257–276.
166
BIBLIOGRAPHY
Bernanke, B. and M.Gertler (1989) ‘Agency costs, net worth and business fluctuations’. American Economic Review, LXXIX, 1: 14–31. Bernheim, D. and M.Whinston (1985) ‘Common marketing agency as a device for facilitating collusion’, Rand Journal of Economics, 16, 2: 269–281. —— (1986) ‘Common agency’, Econometrica, 54, 4: 923–943. Besanko, D. and A.V.Thakor (1987) ‘Collateral and rationing: sorting equilibria and competitive markets’, International Economic Review 28, 3: 671–689. Bigo, H. (1947) Les Banques Françaises au cours du XIXème Siècle, Paris. Blee, E. (1882) ‘La Vérité sur la Crise Financière, Paris. Bonin, H. (1988a) Histoire Economique de la France depuis 1880, Masson. —— (1988b) ‘La Grande Depression Française de la Fin du Dix Neuvième Siècle: Reflexion sur sa datation et sur sa fonction’, Histoire, Economie et Société 6, 4: 509–533. Bonnet, V. (1865) La Revue des deux Mondes, Paris. Born, K.E. (1983) International Banking in the 19th and 20th century, Leamington Spa: Berg Publishers. Bouvier, J. (1960) ‘Les Péreire et l’affaire de la Banque de Savoie 1860–64’, Cahiers d’Histoire 3: 383–410. —— (1968) Naissance d’une banque: le Credit Lyonnais, Paris: Flammarion. —— (1972) ‘Systèmes bancaires et entreprises industrielles dans la croissance Européenne du XIXe siècle’, Annales ESC 1: 46–70. —— (1988) Histoire économique et histoire sociale, Geneva: Librairie Droz. Bouvier J., Furet F. and Gillet M. (1965) Le Mouvement du Profit en France au 19ème Siècle, Industrie et Artisanat, Mouton and Co. Brief, R. (1981) ‘Nineteenth century accounting error’, in J.R.Edwards (ed.) Company Legislation and Changing Patterns Disclosure in British Company Accounts 1850–1914, London. Brown, R. (1905) A History of Accounting and Accountants, Edinburgh: T.C. & E.C.Jack. Calomiris C.W. and Gorton G. (1991) ‘The origins of banking panics: models, facts and bank regulation’, in G.Hubbard (ed.) Financial Markets and Financial Crises, University of Chicago Press / National Bureau of Economic Research. Cameron, R.E. (1953) ‘The Crédit Mobilier and the economic development of Europe’, Journal of Political Economy, December, 4: 461–488. —— (1967) Banking in the Early Stage of Industrialization: A Study in Comparative Economic History, New York: Oxford University Press. Cameron, R.E. and V.I.Bovykin (1991) International Banking 1870–1914, New York: Oxford University Press. Caron, F. (1973) Histoire de l’Exploitation d’un Grand Réseau de Chemin de Fer: La Compagnie du Chemin de Fer du Nord 1846–1937, Paris: Mouton. Carosso, V.P. and Sylla, R. (1991) ‘US banks in international finance’, in R.Cameron and V.I.Bovykin (eds) International Banking, 1870–1914, New York and Oxford: Oxford University Press, 48–71. Cassis, Y. (1992) Finance et Financiers in European History 1880–1960, Cambridge University Press. Chadeau, E. (1988) ‘L’Economie Nationale aux XIXème et XXème siècles’, Annuaire Statistique de l’Economie Française aux XIXème et XXème siècles, Presses de l’Ecole Normale Supérieure, (1, 3).
167
THE ROLE OF BANKS IN MONITORING FIRMS
Chastenet, E. (1861) Le Crédit Mobilier et la Bourse, Dentu. Conant C.A. (1969) A History of Modern Banks of Issue, (Reprints of Economic Classics), New York: Auguste M.Kelley. Corrigan, E.G. (1987) Financial Market Structure: A Longer View, New York: Federal Reserve Bank of New York. Courtois, P. (1889) Traité élémentaire des Opérations de Bourse et de Change, 10th edn, Paris. Darrough, M.N. and N.M.Stoughton (1986) ‘Moral Hazard and Adverse Selection: the Question of Financial Structure’, Journal of Finance,, XLI, 2: 501–513. de Long B. (1989) ‘Did J.P.Morgan’s men add value? A Historical Perspective on Financial Capitalism’, Working Paper, Harvard University and NBER, 3426. de Long B. and M.Becht (1992) ‘“Excess volatility” and the German stock market 1876–1990’, Working Paper, Harvard University and NBER, 82. de Long B., A.Schleifer, L.H.Summers and R.Waldmann (1990) ‘Noise trader risk in financial markets’, Journal of Political Economy, 98, 4: 703–729. Denuc, J. (1934) ‘Dividendes, valeur boursière et taux de capitalisation’, Bulletin de la Statistique, XXIII, IV: 691–767. Deshons, M. and X.Freixas (1987) ‘Le Role de la Garantie dans le Contrat de Prêt Bancaire’, Revue Finance, 2: 7–32. Dessirier, J. (1928a) ‘Les Indices des Valeurs Mobilières en France et en Divers Pays’, Revue d’Economie Politique, 42: 202–231. —— (1928b) ‘La Prévision Statistique des Mouvements des Valeurs de Bourse’, Journal de la Société Statistique de Paris, 153–194. Devinney, T.M. and H.Milde (1990) ‘Agency Contracting and Inside Debt’, Schweizerische Zeitschrift für Volkswirtschaft und Statistik 1 2: 97–111. Diamond, D. (1984) ‘Financial Intermediation and Delegate Monitoring’, Review of Economic Studies 51: 393–414. Diamond, P.A. and J.E.Stiglitz (1974) ‘Increases in Risk and in Risk Aversion’, Journal of Economic Theory, 8(3), July, 337–360. Dupont Ferrier, P. (1925) Le Marché Financier sous le Second Empire, Paris. Edwards, J.R. (1981) Company Legislation and Changing Patterns of Disclosure in British Company Accounts 1850–1914, London. —— (1937) A survey of French Contributions to the Study of Cost Accounting During the 19th Century, London. Engberg, H.L. (1981) Mixed Banking and Economic Growth in Germany 1850– 1931, New York: ARNO Press. Fama, E.F. (1985) ‘What’s different about banks?’, Journal of Monetary Economics, 15, 29–39. Fazzari S., R.G.Hubbard a nd B.Petersen (1988) ‘Financing constraints and corporate investment’, Brookings Papers on Economic Activity, n.1. Feldenkirchen, W. (1982) Die Eisen- und Stahlindustrie des Ruhrgebiets 1879–1914: Wachstum, Finanzierung und Struktur ihrer Großunternehmen, Steiner Verlag GmbH. —— (1991) ‘Banking and economic growth: banks and industry in the nineteenth century and their changing relationship during industrialisation’, in W.R.Lee (ed.) German Industry and German Industrialisation, London: Routledge. Fohlen, C. (1981) Entrepreneurship and Management in France in the Nineteenth Century, Paris.
168
BIBLIOGRAPHY
Freixas, X. (1986) ‘Contrats de dettes en information imparfaite’, Working Paper, GREMAQ, Toulouse, Juillet. Friedman, M. and Schwartz, A. (1963) Monetary History of the United States 1867– 1960, Princeton University Press. Gale, D. and M.Hellwig (1985) ‘Incentive compatible debt contracts: the one period problem’, Review of Economic Studies, LII: 647–663. Gerschenkron, A. (1962) Economic Backwardness in Historical Perspective, Cambridge, Harvard University Press. Gille, B. (1959) La Banque et le Crédit en France de 1815 à 1848, PUF. —— (1968) La Sidérurgie Française au XIXème Siècle, Geneva: Librairie Droz. —— (1970) La Banque en France au XIXème siècle: Recherche Historique, Geneva. Gjcsdal, F. (1982) ‘Information and incentives: the agency information problem’, Review of Economic Studies, 49, 3: 373–90. Goodhart, C.A.E. (1985) The Evolution of Central Banks, Cambridge, MIT Press. —— (1987) ‘Why do banks need a central bank?’, Oxford Economic Papers (March) 39, 1: 75–89. Grossman, S.J. and O.D.Hart (1983) ‘An analysis of principal-agent problem’, Econometrica, 51, 1: 7–46. Guesnerie, R. and J.J.Laffont (1984) ‘A complete solution to a class of principalagent problems with an application to the control of a self-managed firm’, Journal of Public Economics, 25, 2: 329–369. Hammond, J.P. (1993) ‘Financial distortions to the incentives of managers, owners and workers’, Lecture prepared for the Conference on the Structure of Capitalism and the Firm in Contemporary Society, Milan, 18–20 March. Holmstrom, B. (1979) ‘Moral hazard and observability’, Bell Journal of Economics, 10: 74–91. Holzer, P.H. and W.Rogers (1990) ‘The origins and the developments of French costing systems (as reflected in the literature)’, The Accounting Historians Journal, 17, 2: 57–71. Hoshi T., A.Kashyap and D.Scharfstein (1989) ‘The role of banks in reducing the costs of financial distress in Japan’, Massachussetts Institute of Technology, 3435. —— (1990) ‘Bank monitoring and investment: evidence from the changing structure of Japanese corporate banking relationships’, in R.Glenn Hubbard (ed.) Asymmetric Information, Corporate Finance, and Investment, University of Chicago Press / National Bureau of Economic Research. —— (1991) ‘Corporate structure, liquidity, and investment: evidence from Japanese industrial groups’, Quarterly Journal of Economics, 1: 33–60. Hughes P.J. (1988) ‘Risk sharing and valuation under moral hazard’, in Gerald A. Feltham, Amin H.Amershi and William T.Zemba (eds) Economic of Information and Contracts, Kluwer Academic Publishers. Humphrey, T. (1975) ‘The classical concept of the lender of last resort’, Federal Reserve Bank of Richmond Economic Review 61, January: 2–9. —— (1989) ‘Lender of last resort: the concept in history’, Federal Reserve Bank of Richmond Economic Review, 75, March/April: 8–16. Huth, W. (1918) Die Entwicklung der Deutschen und Französischen Großbanken im Zusammenhange mit der Entwicklung der Nationalwirtschaft, Berlin: Franz Siemenroth.
169
THE ROLE OF BANKS IN MONITORING FIRMS
Innes, R.D. (1990) ‘Limited liability and Incentive contracting with ex ante action choices’, Journal of Economic Theory, 52: 45–67. Jacquillat, B. and M.Levasseur (1984) ‘Signaux, mandats et gestions financière: une synthèse de la littérature’, Finance, 5: 31–63. Jeidels, D. (1905) ‘Das Verhältnis der deutschen Großbanken zur Industrie mit besondere Berücksichtung der Eisenindustrie’, in Staats- und Sozialwissenschaftliche Forschungen, Band XXIV, Heft II: , 1–271, Leipzig: Duncker & Humblot. Jensen, M.C. and W.C.Meckling (1976) ‘The theory of the firm: managerial behavior, agency costs and ownership structure’, Journal of Financial Economics, 3: 305– 360. Johnston H.T. and R.S.Kaplan (1987) Relevance Lost: the Rise and Fall of Management Accounting, Boston, Massachusetts: Harvard Business School Press. Jorgenson, D. (1971) ‘Econometric studies of behavior: a review’, Journal of Economic Literature, IX: 1111–47. Kaufmann P. (1914) La Banque en France, Paris. Keely, M.C. (1990) ‘Deposit insurance, risk about market power in banking’, American Economic Review, 4: 1183–1200. Kindleberger, C.P. (1984) ‘Banking and industry between the two wars: an international comparison’, Journal of European Economic History, 39, 2: 7–26. Kocka, J. (1978) ‘Entrepreneurs and Managers in German Industrialization’, in Mathias P. and M.Postan (eds) The Cambridge Economic History of Europe, Vol. VIIThe Industrial Economics: Capital Labour and Enterprise Part I, Cambridge: Cambridge University Press. Kose, J. and A.Kalay (1982) ‘Costly contracting and optimal payout constraints’, Journal of Finance, XXXVII, 2: 373–390. Krause, D. (1991) ‘Zur Geschichle der Mitteldeutschen Creditbank’, Bankhislorisches Archiv, 17: 17–41. Kreps, D. and R.Wilson (1982) ‘Reputation and imperfect information’, Journal of Economic Theory, 27: 253–279. Kreps, D., P.Milgrom, J.Roberts and R.Wilson (1982) ‘Rational cooperation in the finitely-repeated prisoner’s dilemna’, Journal of Economic Theory, 27: 245– 252. Laffont, J.J. (1987) ‘Le risque moral dans la relation de mandat’, Revue Economique, 41, 1: 5–23. Laffont, J.J. and X.Freixas (1988) ‘Contrats Bancaires Optimaux’, Toulouse: GREMAQ. Laffont, J.J. and E.Maskin (1986) ‘The theory of incentives: an overview’, in W.Hildenbrand (ed.) Advances in Economic Theory, Cambridge University Press. Lambert, R. (1983) ‘Long term contracts and moral hazard’, Bell Journal of Economics, 14, 2: 441–452. Landes, D.S. (1956) ‘Vieille banque et banque nouvelle: la révolution financière du XIXème siècle’, Revue d’Histoire Moderne et Contemporaine, 3: 204–220. Leland, H.E. and D.H.Pyle (1977) ‘Informational asymmetries, financial structure and financial intermediation’. Journal of Finance, 32, 2: 381–387. Lenoir, M. (1919) ‘Le mouvement des cours des valeurs mobilières françaises depuis 1856’, Bulletin de la Statistique Générale de la France, Octobre, 65–92. —— (1922) ‘Les valeurs mobilières et la crise de 1920 (indices économiques compares en France et divers pays)’, Bulletin de la Statistique Générale de la
170
BIBLIOGRAPHY
France, Janvier, 171–221. Levy Leboyer, M. (1970) ‘L’héritage de Simiand: prix, profits et termes de l’echange en France au XIXème siècle’, Revue Historique, 493: 79–119. —— (1976a) ‘Le crédit et la monnaie: l’apprentissage du marché’, in Fernand Braudel and Ernest Labrousse Histoire Economique de la France Vol. III: ‘L’Avénement de l’Ere Industrielle’, Paris: Presse Universitaire de France. —— (1976b) ‘La spécialisation des établissements bancaires’, pp. 431–467 in Fernand Braudel and Ernest Labrousse Histoire Economique de la France Vol. III: ‘L’Avenèment de l’Ere Industrielle’, Paris: Presse Universitaire de France. Lindenberg, E. and S.Ross (1981) ‘Tobin’s q ratio and industrial organisation’, Journal of Business, January, 54: 1–32. Littleton, A.C. (1960) ‘Accounting Evolution to 1900, New York: Russell. Luce, R.D. and H.Raiffa (1957) Games and Decisions: Introduction and Critical Survey, New York: Wiley. Mackie-Mason, K. (1989) ‘Do firms care who provides their financing?’ Working Paper NBER, 3039. Mandelker, G. and A.Raviv (1977) ‘Investment banking: an economics of optimal underwriting contracts’, Journal of Finance, June, 2: 61–75. Mankiw, G. and M.Shapiro (1985) ‘Do we reject too often? Small sample properties of tests of rational expectations models’, Economics Letters, 20: 139–145. Mankiw G., Romer D. and M.Shapiro (1985) ‘An unbiased reexamination of stock market volatility’, Journal of Finance, , XL, 3: 677–689. —— (1991) ‘Stock market forecastability and volatility: a statistical appraisal’, Review of Economic Studies, 58: 455–477. Marriner, S. et al. (1978) Business and Businessmen: Studies in Business, Economic and Accounting History, Liverpool University Press. Mayer, C. (1988) ‘New issues in corporate finance’, European Economic Review,, 32: 1167–1188. Milgrom, P. and J.Roberts (1982) ‘Limit pricing and entry under incomplete information’, Econometrica, 50: 443–460. Miskin, F. (1991) ‘Asymmetric information and financial crises: a historical perspective’, in G.Hubbard (ed.) Financial Markets and Financial Crises, University of Chicago Press / National Bureau of Economic Research. Moussa, P. (1980) ‘A quoi sert une banque d’affaires?’, Banque, 394 Avril, 429– 433. Murphy K.M., A.Schleifer and R.W.Vishny (1989) ‘Industrialization and the big push’, Journal of Political Economy, 97, 5: 1003–26. Myers, S.C. (1977) ‘Determinants of corporate borrowing’, Journal of Financial Economics, 49: 993–1007. Myers, S. and N.Majlhuf (1984) ‘Corporate financing and investment decisions when firms have information that investors do not have’, Journal of Financial Economics, 77: 187–221. Nelson, J. (1990) ‘Principals and agents: does the assignment of information matter?’, Economic Letters, 23: 223–227. Neuberger, D. and M.Neumann (1991) ‘Banking and antitrust: limiting industrial ownership by bank?’, Journal of Institutional and Theoretical Economics, 147: 188–199. Nouette-Delorme, E. (1865a) Les sociétés de speculation: le Crédit Mobilier et ses valeurs, Paris: Castel.
171
THE ROLE OF BANKS IN MONITORING FIRMS
—— (1865b) Le Crédit Mobilier et ses Valeurs: Histoire de 12 ans, Paris: Castel. Page, F.H. Jr. (1989) ‘Incentive compatible strategies for the general Stackelberg games with incomplete information’, International Journal of Game Theory, 3: 409–421. Paris, K. (1867) Le Crédit Mobilier et ses Actions, Paris. Péreire, E. and I. (1879) La question des Chemins de Fer, Paris: Imprimerie de C. Motteroz. —— (1912–1920) Oeuvres, Paris: Alcan. Plenge, J. (1903) Gründung und Geschichte des Crédit Mobilier, Tubingen: These. Plessis, A. (1985) La Politique de la Banque de France de 1851 à 1870, Paris: Droz. —— (1993) ‘Bankers in the French society 1860s–1960s’, in François Crouzet (ed.) The Economic Development of France since 1870, Elgar Reference Collection Vol. II. Prowse, S.D. (1992) ‘The structure of corporate ownership in Japan’, Journal of Finance XLVII, 3: 1121–1140. Pujo Committee, US Congress, House, Committee on Banking and Industry (1913a) Report of the Committee Appointed Pursuant to House Resolutions 429 and 504 to Investigate the Concentration of Control of Money and Credit, Washington, D.C.: G.P.O. —— (1913b) Money Trust Investigation. Hearings Before the Committes Appointed Pursuant to House Resolutions 429 and 504 to Investigate the Concentration of Control of Money and Credit, Washington, DC: G.P.O. Ramirez, C. (1992) ‘Did J.P.Morgan’s men add liquidity? Corporate investment, Cash flow and financial structure at the turn of the twentieth century’. Working Paper, Harvard University. Rasmusen, E. (1989) Games and Information: An Introduction to Game Theory, Basil Blackwell, Oxford and Cambridge. Riesser, J. (1911) The Great German Banks and their Concentration, Washington D.C.: US Government Printing Office. Ross, S. (1973) ‘The economic theory of agency: the principal’s problem’, American Economic Review, May, 3: 363–367. —— (1977) ‘The determination of financial structure: the incentive signalling approach’, Bell Journal of Economics, 8, 2: 23–40. —— (1978) ‘Some notes on financial incentive signalling models, activity choice and risk preferences’, Journal of Finance, XXXIII, 43: 777–794. Rothschild, M. and J.Stiglitz (1976) ‘Equilibrium in competitive insurance market: an essay in the economics of incomplete information’, Quarterly Journal of Economics, 90: 331–359. Salinger, M. (1984) ‘Tobin’s q, unionization, and the concentration-profits relationship’, Rand Journal of Economics, 15: 159–170. Sayous, A.E. (1899) ‘Les banques allemandes en cas de crise ou de guerre’, Revue d’Economie Politique, 142–166. —— (1907) Lex Banques de Dépôts, les Banques de Crédit et les sociétés Financières, Paris. Scott, K.E. (1991) ‘Banking and antitrust: limiting industrial ownership by bank? Comment’, Journal of Institutional and Theoretical Economics, 147, 1: 202–206. Shavell, S. (1979) ‘Risk sharing and incentives in the principal and agent relationship’, Bell Journal of Economics, 10: 55–73.
172
BIBLIOGRAPHY
Shiller, R. (1981) ‘Do stock prices move too much to be justified by subsequent changes in dividends?’, American Economic Review, 71, 2: 421–36. —— (1989) Market Volatility, Cambridge, MIT. Smirlock, M., Gilligan T. and W.Marshall (1984) ‘Tobin’s q and the structure performance relationship’, American Economic Review, 74, 4: 1051–1060. Solow, R.M. (1982) ‘On the lender of last resort’, in C.P.Kindleberger and J.P. Laffarge (eds) Financial Crises: Theory, History and Policy.’ Cambridge University Press. Stiglitz, J.E. (1985) ‘Credit markets and the control of capital’, Journal of Money, Credit and Banking, 17: 133–152. —— (1982) ‘The inefficiency of the stock market equilibrium’, Review of Economic Studies, 49(2), April, 241–261. Stiglitz, J.E. and A.Weiss (1981) ‘Credit rationing in markets with imperfect information’, American Economic Review, 71: 393–410. Stole, L.A. (1990) ‘Mechanism design under common agency’, Mimeo, MIT, November (unpublished). Teichova, A. (1988) ‘Rivals and partners: banking and industry in Europe in the first decades of the twentieth century’, Uppsala Working Papers in Economic History, Report from the Vienna Banking Industry Symposium. Thornton, H. (1802) An Enquiry into the Nature and Effects of the Paper Credit of Great Britain, edited by F.A.Hayek, Fairfield Augustus: M.Kelley. Tilly, R. (1966) Financial Institutions and Industrialization in the Rhineland 1815– 70, London: University of Wisconsin Press. —— (1986a) ‘German banking 1850–1914: development assistance for the strong’, Journal of European Economic History, 15, 1: 113–152. —— (1986b) ‘Germany 1815–1870’, in R.Cameron, et al. (eds) Banking in Early Stages of Industrialisation: a study in comparative economic history New York, London, Toronto: Oxford University Press. Tobin, J. and W.R.Brainard (1977) ‘Asset markets and the cost of the capital’, in Balassa B. and Nelson R. (eds) Economic Progress, Private Values and Public Policy: Essays in Honor of William Fellner, North Holland. Webb, D. (1987) ‘Conflict of interest, bond rationale and the value maximisation rule’, Economica, 54, 3: 455–463. Weber, A. (1903) Die rheinisch-westfälischen Provinzialbanken und die Krisis, Leipzig. Weston, J.F. and T.E.Coppeland (1987) Financial theory and corporate policy, 3rd edn, Addison Wesley Publishing Company. Whale, B.P. (1930) Joint Stock Banking in Germany: a study of the German Credit Banks before and after the War, London: Macmillan and Co. Wilson, C. (1977) ‘A model of insurance markets with incomplete information’, Journal of Economic Theory, 16: 167–207. Wrück, K.H. (1990) ‘Financial distress, reorganization and organizational efficiency’, Journal of Financial Economics, 27, 2: 419–444 Yoshikawa, H. (1980) ‘On the q theory of investment’, American Economic Review, 3: 739–743.
173
INDEX
adverse selection 9–10 affiliated companies 42–4; data on 142–5; depreciation rate and investment 92–5, 98–104; effect of bankruptcy on 53–65; investment and cash flow 80–90; and non-affiliated companies, comparisons 71–3; share price fluctuations 60–4; share prices, monthly (1866–1868) 148–51; summary statistics on 105 agency costs 9, 12 agency theory 1, 7; banks, supervisory role of 15–24; basic concepts 8–12 Allais (company) 44, 67; bankruptcy, impact of 68–70; data on 147 Aoki, M. 117 asset diversification 11–12 asymmetric information see information asymmetry
banks: association with firms, analysis 117–20; financial intermediation by see financial intermediation; monitoring by see monitoring by banks; and stock market 51–2; supervisory role of see supervisory role of banks. See also central banks Banque de France 29, 38; and bankruptcy of Crédit Mobilier 112 Banque de l’Union Parisienne 28, 48, 51 Banque Transatlantique 36 Becht, M. 27, 31, 34, 40, 42, 45–6, 51, 86, 93, 95, 110, 116–17 Bernanke, B. 114 Bonnet, V. 109 Born, K.E. 52 Bouvier, J. 7, 30, 112, 121, 124 Bovykin, V.I. 76
Bagehot, W. 38 bank finance 16–19; and market finance 19–20; and supervisory role of banks 16–19 bank-firm relations: analysis of 117–20; freedom from 24; Japan 23–4 banking panic and bankruptcy of Crédit Mobilier 114–16 bankruptcy of Crédit Mobilier 3, 35–7; banking panic in 114–16; and collateral 33; effect on affiliated companies 53–65, 73–4; effect on non-affiliated companies 65–71, 73–4; financial crisis, creating 111– 14; general influence of 54–8; highrisk projects 13; shareholders’ attempts to prevent 128 9; and stock market prices 41–2
Cameron, R.E. 43, 76, 132 capital scarcity 18–19 capital stock and investment 82, 84–6; and depreciation rate 93–5 Carosso, V.P. 51–2 cartels 23 cash flow and investment expenditure 77; and Crédit Mobilier 78–90; dataon 153–61; and depreciation 103; in Japan 124; regression on 83–7 central banks: as lender of last resort 38–9; role in Germany 125 Chatillon (company) 44, 45, 67; bankruptcy, impact of 68–70; data on 148; investment and cash flow analysis 80–90 collateral 33 commitment by banks 23–4
174
INDEX
common agency 12–15; assumptions 14 15; compensation 15; contract terms 14; risk-neutrality in 14 Compagnie des Chemins de Fer de l’Est 43; asset prices 54; bankruptcy of Crédit Mobilier, impact of 55–9; data on 144; investment and cash flow 80–90; share price fluctuations 60–4; share prices, monthly (1866–1868) 149–51 Compagnie des Chemins de Fer du Midi 32, 41, 43; asset prices 54; bankruptcy of Crédit Mobilier, impact of 55–9; data on 143; investment and cash flow 80–90; share price fluctuations 60–4; share prices, monthly (1866–1868) 149–51 Compagnie des Chemins de Fer du Nord 41; asset prices 54; bankruptcy of Crédit Mobilier, impact of 55–9; data on 143; investment and cash flow 80–90; share price fluctuations 60–4; share prices, monthly (1866– 1868) 149–51 Compagnie des Chemins de Fer d’Orleans 41, 43; asset prices 54; bankruptcy of Crédit Mobilier, impact of 55–9; data on 143; investment and cash flow 80–90; share price fluctuations 60–4; share prices, monthly (1866–1868) 149–51 Compagnie des Chemins de Fer de l’Ouest 43; asset prices 54; bankruptcy of Crédit Mobilier, impact of 55–9; data on 143; investment and cash flow 80–90; share price fluctuations 60–4; share prices, monthly (1866–1868) 149–51 Compagnie des Chemins de Fer PLM 41, 43; asset prices 54; bankruptcy of Crédit Mobilier, impact of 55–9; data on 143; investment and cash flow 80–90; share price fluctuations 60–4; share prices, monthly (1866–1868) 149–51 Compagnie d’Eclairage par le Gaz 41; asset prices 54; bankruptcy of Crédit Mobilier, impact of 55–9; data on 143 Compagnie Générale des Omnibus (Paris) 31, 43; bankruptcy of Crédit Mobilier, impact of 55–9; data on
143; share price fluctuations 60–4 share prices, monthly (1866–1868) 149–51 Compagnie Générale Transatlantique 31, 41 Compagnie Immobilière 31, 32, 35, 43, 97; asset prices 54; and bankruptcy of Crédit Mobilier 113, 118; bankruptcy of Crédit Mobilier, impact of 55–9; data on 142, 143, 144; investment and cash flow 80– 90; investment regression 98–101; share price fluctuations 60–4; share prices, monthly (1866–1868) 149–51 Compagnie Parisienne de Distribution d’Électricité 43 Compagnie Parisienne du Gaz 32 Compagnie Transatlantique 32, 36, 43; asset prices 54; and bankruptcy of Crédit Mobilier 112; bankruptcy of Crédit Mobilier, impact of 55–9; data on 143; share price fluctuations 60–4; share prices, monthly (1866– 1868) 149–51 compensation in common agency 15 contracts: and banks, supervisory role of 21–3; competition for 21; terms in common agency 14 Crédit Lyonnais 120–3; credit policy of 123; definition 121; dividends, distribution of 123; financial participation 122; specific activities 121–2 Crédit Mobilier: advances to companies (1856–1866) 32–3; advantages/ disadvantages of association with 116–24; affiliated companies 42–4; bankruptcy of see bankruptcy; collateral, not asked for 33; control of firms 3; credit policy 123; criticisms of 28; data (1852–1867) 136, 137–8; data (1903–1914) 136, 139; definition 121; dividends, distribution of 123; evaluation of association with 60–4; excess volatility tests on stock market 45–53; fall of 34–5; historical background 29–30; liquidation account of 36; monthly share prices (1866–1868) 149–51; notaffiliatedcompanies 44–5; projects 31; reconstitution of 35–7;
175
THE ROLE OF BANKS IN MONITORING FIRMS
representation on boards 31; specific activities 121–2; supervisory role, limited 22; supervisory role of 111–16 Crédit Mobilier Français 28, 37–40 credit policy: Crédit Mobilier and Crédit Lyonnais 123; in supervisory role of banks 110 credit rationing 10 Darmstädter Bank 27, 31, 110 De Long, J.B. 26–7, 28, 31, 34, 40, 42, 44–6, 51, 65, 74, 76, 86, 117, 132 debt: defined 12; inside 16–17 Denain (company) 44, 67; bankruptcy, impact of 68–70; data on 147 Denuc, J. 79 depreciation rate and investment 91–2; regression on 92–5, 98–104 d’Erlanger, E. 35 Deshons, M. 10 Dessirier, J. 46, 47 Deutsche Bank 26 Devinney, T.M. 16, 17 Diamond, D. 10, 20 Diskonte Gesellschaft Bank 26 Dollfus Mieg et Cie 29 Dresdner Bank 26, 126 Dupont Ferrier, P. 30, 42 Electrometallurgie de Dives 41, 43 Engberg, H.L. 52 enterprises, financing of 30–4 excess volatility tests on stock market 45–53 external finance and banks, supervisory role of 16–19 Fama, E.F. 17 Fazzari, S. 77, 124 Feldenkirchen, W. 91 Filature (company) 67; bankruptcy, impact of 68–70; data on 147 financial crises: and bankruptcy of Crédit Mobilier 111–14; and real output 112 financial intermediation by banks 20; and commitment 23–4; and investment 21; role of 113 Firminy (company) 44, 45, 67; bankruptcy, impact of 68–70; data on 148; investment and cash flow 80–90
firms: banks’ association with, analysis 117–20; control of by Crédit Mobilier 3; investment performance of 23. See also bank-firm relations fixed effect model of investment 87–90; and depreciation rate 95–7 Fourchambault (company) 44, 45, 67; bankruptcy, impact of 68–70; data on 148; investment and cash flow analysis 80–90 France: banking crash (1882) 48; economy (1853–1914) 40–2; interest rate rises (1870–1872) 42; monitoring role of banks 51; share price indices 140–2; stock markets, volatility tests on 49–50; supervisory role of banks, comparisons with Germany 124–8 Franche Comté (company) 44, 67; bankruptcy, impact of 68–70; data on 148 Freixas, X. 10, 13 Friedman, M. 112 Furet, F. 30 Galai, D. and Masulis, R. 12 Gale, D. 10 Germain, H. 121, 123, 128, 129, 134 Germany: banks, caution of 49; banks, short-term policy 22; capital scarcity in 18–19; financial participation 125–6; investment policy 125; monitoring role of banks 51; stock markets, volatility tests on 49–50; supervisory role of banks, comparisons with France 124–8 Gerschenkron, A. 18–19, 20, 22, 23, 120, 130, 132 Gille, B. 44, 48, 79 Gillet, M. 30 Givors (company) 44, 67; bankruptcy, impact of 68–70; data on 147 Gjesdal, F. 9 Goodhart, C. 39 Grossman, S.J. 11 Hart, O.D. 11 Haussmann, Baron 35 Hellwig, M. 10 Hilferding, R. 130 L’Home (company) 44, 67; data on 148 Hoshi, T. 23, 24, 77, 79, 80, 115, 133
176
INDEX
Majluf, N. 24
Houillière (company) 67; bankruptcy, impact of 68–70; data on 147 Hubbard, R.G. 77, 124 Humphrey, T. 39 information asymmetry 9, 10; and banking panics 115; and bankruptcy of Crédit Mobilier 111; and investment 77; and role of banks 17 inside debt 16–17 internal finance: and agency costs 24; and banks, supervisory role of 16–19 investment: and capital stock, regression on 82, 84–6; and cash flow, regression on 83–7, 152–60; and Crédit Mobilier 78–90; and depreciation rate, regression on 91–5, 98–104; and financial intermediation by banks 21; fixed effect model 87–90; and information asymmetry 77; performance of firms 23 Japan: bank lending relationships 118; bank-firm relations 23–4; investment and cash flow 124; investment expenditure and cash flow 77 Jeidels, D. 78, 126 Jensen, M.C. 9, 11, 12, 24 joint-stock banks 27 Jorgenson, H.T. 77 J.P.Morgan Company 26, 28, 51, 77 Kalay, A. 10 Kashyap, A. 77, 79, 80, 115, 133 Kindleberger, C.P. 111, 113 Koechlin, A. 29 Kose, J. 10 Laffont, J.J. 10, 13 Leland, H.E. 11 lender of last resort, central banks as 38–9 Lenoir, M. 45, 46 Levy-Leboyer, M. 47, 134 L’Home (company): bankruptcy, impact of 68–70 liquidation account, Crédit Mobilier 36 liquidity: central banks, providing 39; and Crédit Mobilier 86; Germany and France, comparisons 125; and investment 77 Loutchitch 41
Marine (company) 44, 45, 67; bankruptcy, impact of 68–70; data on 148; investment and cash flow analysis 80–90 market finance and bank finance 19–20 Mathias, P. 132 Maubeuge (company) 44, 67; bankruptcy, impact of 68–70; data on 148 Mayer, C. 16, 21, 22, 23, 119 Meckling, W.C. 9, 11, 12, 24 Milde, H. 16, 17 Miller, M. 3, 77 Mine Firminy 80 Mine Saint Etienne 80 Mines Grand Combes 80 Mines Montrambert la Beraudière 80 Miskin, F. 111 Modigliani, F. 3, 77 monitoring by banks: and contracts 21–3; as financial intermediation 20; model of 17 moral hazard 9–10 Myers, S. 24 Myers, S.C. 10, 12 Napoleon III 35 non-affiliated companies 44–5; and affiliated companies, comparisons 71–3; data on 146–7; depreciation rate and investment 92–5, 98–104; effect of bankruptcy on 65–71; investment and cash flow 80–90; share price fluctuations 67–71; summary statistics on 105 Nouette-Delorme, E. 53 Ohio Life and Trust Company 111 Oppenheim, A. 27 Paris, K. 129 Péreire, Emile 29 Péreire, Eugène 36 Péreire, Isaac 29, 121 performance-related compensation in common agency 15 Petersen, B. 77, 124 Plenge, J. 43 Pujo Committee (1913) 86 Pyle, D.H. 11
177
THE ROLE OF BANKS IN MONITORING FIRMS
Ramirez, C. 79, 86, 93, 95, 110, 116–17 Riesser, J. 78 risk-neutrality in common agency 14 Ross, S. 12 Saint Simon, Henri, comte de 119 Sayous, A.E. 122, 125, 126 Schaaffhausen’schen Bankverein 27 Scharfstein, D. 77, 79, 80, 115, 133 Schartz, A. 112 Schneider (company) 44, 67; bankruptcy, impact of 68–70; data on 147 Seillière, Baron 29 share prices: affiliated companies 60–4; and bankruptcy of Crédit Mobilier 112; fluctuations 34; indices, France 140–1; monthly, affiliated companies 148–51; non-affiliated companies 67–71 shareholders on bankruptcy of Crédit Mobilier 128–9 Shiller, R. 45, 47 signalling theory 1; basic concepts 8–12; model of 11 Société de l’Hôtel et des Immeubles de la Rue de Rivoli 31, 32 Société d’Eclairage Électrique 43 Société des Machines Outils 44 Société des Omnibus de Paris 43 Société Immobilière 34 Société Metallurgique de MontbardAulnaye 43 Société Parisienne d’Eclairage 41 Société Parisienne d’Eclairage et de Chauffage par le Gaz 43 Société Raffineries Say 43 Solow, R.M. 39 Stiglitz, J.E. 112
stock market and banks 51–2 Stole, L.A. 13 supervisory role of banks 15–24, 109–31; association with, advantages and disadvantages 116–24; bank finance and market finance 19–20; banking panic in bankruptcy 114–16; and cartels 23; and contracts 21–3; and Crédit Lyonnais, comparisons 120–3; credit policy in 110; finance 16–19; financial crisis in bankruptcy 111–14; France and Germany, comparisons 124–8; intermediation and commitment 23–4; and investment performance of firms 23; model of 17–18 Sylla, R. 51–2 Terrenoire (company) 44, 67; bankruptcy, impact of 68–70; data on 147 Thalmann and Co. 37 Thornton, H. 38–9 Tilly, R. 22, 125 Tobin, J. 82, 91 Union Fire 55–9; data on 143 l’Union (insurance company) 43; bankruptcy of Crédit Mobilier, impact of 55–9; data on 143 United States: monitoring role of banks 51; stock markets, volatility tests on 49–50 volatility ratios 49–52 Wallut, C. 36 Whale, B.P. 127 Wrück, K.H. 118
178