INTERNATIONAL CO OPERATIVE BANKING ASSOCIATION SEMINAR Co operative logic versus

INTERNATIONAL CO-OPERATIVE BANKING ASSOCIATION • SEMINAR • Co-operative logic versus market logic applied to capitalization of co-operative banks and to their growth strategies Beijing Pyongyang South Korea Seoul Tokyo INTERNATIONAL CO-OPERATIVE BANKING ASSOCIATION JOURNAL NO. 13, 2001 ISO 9001 Certified Business solutions through information technology TM Commitment is... Forging lasting partnerships Empowered by past successes, CGI intends to continue forging strong alliances with credit unions worldwide. In North America alone, it has already established partnerships with credit unions serving more than 10,000,000 members. CGI is particularly proud of its association with the Mouvement des caisses Desjardins, a leading provider of financial services in Québec, Canada. Through a strategic 10-year alliance valued at more than US$800 million, Desjardins is delegating the management of its data processing operations to CGI while retaining control of its technological strategy. www.cgi.ca cooperative logic versus market logic applied to capitalization of cooperative banks and to their growth strategies INTERNATIONAL CO-OPERATIVE BANKING ASSOCIATION No. 13, 2001 CONTENTS Page INTRODUCTION : Claude Béland A word from the President SPEAKERS : Benoît Tremblay Cooperative banks and the mobilization of capital: To what end, with which partners and with what consequences for members? Klaus P. Fischer Do Agencies rate cooperative bank-issued bonds fairly? 24 5 3 ICBA CONTACTS 37 ICBA MISSION The International Co-operative Banking Association (ICBA) is a specialised organisation of the International co-operative Alliance (ICA). The ICBA was set up by national co-operative banks and financial organisations: • • • To exchange information; To promote co-operation among co-operative banks; To promote the development of new co-operative banks through advice and assistance; To research and study projects of common interest, i.e. capital formation, co-operative values as applied to banking, etc. • Membership is open to all banks and central thrift and credit organisations through one of several Regional Committees or by direct membership of the Central Executive Committee. Our membership brings together representatives of co-operative banks and financial institutions worldwide, including the International Raiffeisen Union, European Association of Co-operative Banks, and the World Council of Credit Unions. The Association is, therefore, a true global and democratically-elected specialized organisation of the ICA, representing the views of co-operative financial institutions all over the world. This Journal contains the presentations made on the occasion of the International Co-operative Banking Association Seminar held in Seoul, South Korea, October 15, 2001. INTRODUCTION TO THE Seoul SEMINAR Mr. claude béland President of the INTERNATIONAL CO-OPERATIVE BANKING ASSOCIATION A word from the president The two presentations featured at this ICBA Seminar provide very interesting insights into the thematic of "cooperative logic versus market logic applied to capitalization of cooperative banks and to their growth strategies". These presentations shed some light on the rationale of various stakeholders, namely members, investors and financial agencies, with regards to their respective participation and appreciation of the cooperative banks’ capitalization strategies. Traditionally, most cooperative banks were found in a weak competitive environment in which cooperative identity received strong expression. For various structural reasons, we are witnessing a shift where cooperative identity loses part of its meaning while the environment becomes highly competitive. This in turn has a strong influence on financial management and the management of the capital structure of the cooperative banks. One of the answer to those challenges has been demutualization. However, the majority of cooperative banks have chosen to remain faithful to their cooperative roots. They have sought alternative ways to improve their competitiveness while preserving and even strengthening their cooperative identity. This proves to be quite a challenge as their capital requirements are ever increasing in the face of the globalization of world economies and because of the need, for some, to expand outside their domestic territory. 3 The first presentation unveils the results of the recent study done by the HEC. It will stress that the cooperative identity, when viewed as an integral part of the business strategy, can not only drive the capitalization strategy but can also be used as a competitive tool which can greatly enhance the perceived value of the cooperative bank and in turn increase its ability to raise capital. The presentation will briefly explore various financial models borrowed by some known mutual banks to compete in an increasingly global market environment. Results of the study are based on a broad review of literature and case studies of the Rabobank, Co-op Bank, Crédit Mutuel de France, Crédit Agricole, CERA and Desjardins as well as on interviews of a few African and one Indian mutual banks. The second presentation highlights the main conclusions of another recent study done by Laval University’s finance deparment. It tries to answer this question : "Do rating agencies such as Fitch, Moody’s, Standard and Poor’s and others calibrate the rating of bonds issued by cooperative banks and stock banks to the true risk of insolvency of the issues? The approach taken is to compare consistency between investors and agencies’ appreciation of the default risk of the bond issues of a wide sample of cooperative banks and of joint stock banks of Europe and North America. I would like to thank Professors Benoît Tremblay and Daniel Côté from the HEC and Professor Klaus Fischer from Laval University for their valuable research work and for having chosen the ICBA Seminar as first platform to disclose their findings. We can certainly appreciate from these presentations the tangible benefits that the cooperative financial organizations can reap from working more closely with the university researchers of their community. I also thank all participating mutual banks for sharing their experience. Lastly, these studies would not have been possible without the important financial and technical contribution of Développement International Desjardins. DID chose to associate ICBA to the project, thus ensuring that a wider number of co-operators could benefit from the lessons to be learned. Good reading ! 4 Mr. Benoît Tremblay Mr. Tremblay has a Ph.D. from L’ École Pratique des Hautes Études, Paris, France. He is Associate professor at l’Ecole des Hautes Études Commerciales (HEC-Montréal, Canada) where he teaches mainly management courses since 1976. He conducts research on managerial issues related to various cooperative sectors and recently on cooperative banking in the world. He has been deputy minister for cooperative development, board member of many cooperative organisations and was cofounder of “le Centre de gestion des coopératives” at HEC in 1975. Currently, he is much involved with the development of the Centre Desjardins on cooperative financial institutions management at HEC-Montréal. Cooperative banks and the Mobilization of Capital: To what end, with which partners and with what consequences for members? By Benoît Tremblay, PhD., In collaboration with: Daniel Côté, PhD., HEC, Montréal, Canada Introduction Over the last few years the ICBA has become an important forum for discussion about the issues raised by the capitalization of cooperative banks. I would like to thank the Association’s executive members, and the Secretary, Mr. Ghislain Paradis, for his interest in research and for his always stimulating comments. My presentation is a part of the research that I am conducting with Daniel Côté at HECMontréal, so I will frequently use “we” as expression of both understanding. My presentation focuses on the cooperative banks in Western countries whose strategies illustrate major trends of the cooperative banking sector. The major objectives of this research are: • To inventory innovative capital mobilization practices; • To establish a typology of observed capitalization strategies and structures; • To draw out capital mobilization strategies likely to foster and enlighten discussions by directors of cooperative banks. 5 I will present the practices of some of the cooperative banks that we have observed, with a view to putting forward and illustrating a framework for analyzing strategies which simultaneously take into account the evolution of their cooperative identity and their competitive environment. 1. The evolution of strategies…and capitalization needs Historically, cooperative banks have developed on the fringes of the banking sector, for the most part with their own laws and their own regulations. Over the past 30 years, in Western countries, there has been an increasing harmonization of the rules applied to cooperative banks with those applying to the banking sector as whole. Moreover, given their size and the diversity of their activities, cooperative banks are fully experiencing the effects of the new competitive framework. These two factors are major contributors to the evolution of their capital requirements. In the 1980s, the traditional regulatory framework has been replaced with new regulations aimed essentially at reinforcing the security of transactions and the whole system. The most constraining manifestation of which has been the application of widely accepted capitalization ratios of the Basle Agreement, 1988. Moreover, the deregulation led to the introduction of competition into a sector containing all financial services. This shift led banks to try to meet the needs of different segments of the population, all the while offering an increasingly wide range of financial products. So, the heightened pressure on the need for capital can also be associated to the shift towards the universal bank. The 1990s were characterized by strategies which were major consumers of capital, such as the high cost of information and telecommunications technologies and the rising costs of increasingly sophisticated product development, which increase the importance of volume and critical size. On the other hand, the possibilities enterprises have for decomposing their value chain to concentrate on elements at the core of their expertise may reduce capital needs. Although fundamental, these factors do little to excite the imagination compared to the sector’s spectacular restructuring strategies by way of mergers and acquisitions. We know that many managers of cooperatives also dream about these strategies for which they feel poorly equipped. I would like to point out that specialists and analysts are not unanimous about the relevance of these merger and acquisition strategies. For example, Pfeffer in Looking for Success in all the Wrong Places (1988) and Coureil in Valeur Ajoutée, Construire et Développper les compétences de l’Entreprise (1997) clearly demonstrate that size has little to do with the success of businesses. For practitioners, I would like to cite a recent study by the firm of consultants Ernst & Young, entitled Measuring the Future, the Value Creation Index, 2000, the conclusion of which maintains that “The wave of consolidations and globalisation that have been transforming the financial services landscape, have, alas, done little to energise the valuations of financial ser- 6 vices companies…Growing the asset base, a traditional driver of value, has become less and less important, not just in financial services, but across industries…According to the Value Creation Index Study, the top non financial performance drivers for financial services are alliances, human capital and quality of management.” In light of our own analyses, we could add the following to these key factors: relations with the community. Here is something which opens the door to creativity; size is perhaps not the main success factor. This argument finds an echo at Standards & Poors, quoted in The Tribune (29/11/2000): “Indeed, be they domestic or transnational, mergers and acquisitions remain a complex equation. The choice of a consolidation model, the integration of different cultures, technological and organizational constraints, risk assessment, and management of execution risks are as many obstacles explaining why mergers and acquisitions have, until now, rarely met preset objectives. To the point where, to optimize the return on these operations and to satisfy the value creation dogma, acquirers often aggressively manage their own funds via share buy backs or payment of exceptional dividends. The combination of these elements explains the cautious approach adopted by Standard and Poor’s in its assessment of external growth operations.” Although industry is immersed in restructuring activity, as much with regard to corporate strategies (mergers, acquisitions, alliances, etc.) as with respect to the reorganization of profes- sions, company success is certainly less “reliant on the environment” than it would appear at first glance, as we can see in light of the analyses presented above. We can thus hypothesize that there is a margin of manoeuver for different strategies without sacrificing excellence. Moreover, even though empirical findings do not appear totally to support the paradigm of critical size, rapid internationalization and restructuring as the royal road to success, several of the strategies proposed are major consumers of capital and bring into question the capitalization strategies of cooperative banks. Before developing the presentation of these strategies, it is worthwhile to take a brief look at the development of cooperative banks’ capitalization strategies over the past 20 years. 2. The capitalization strategies of cooperative banks Generally speaking, for decades, the equity funds of cooperative banks have essentially been accumulated on the basis of reinvested surpluses. The pace of their development was thus highly dependent upon their capacity to generate these funds and to convince their members to reinvest them. The more rapid evolution of the financial sector over the past 20 years has led many cooperative banks to diversify their capitalization practices. In the 1980s, cooperative banks innovated in responding to the new rules of capitalization by creating category B shares, which offered a limited return, subject to the annual financial results and without vote, according to various terms. Among the cases studied, Crédit 7 Mutuel, the Caisses Desjardins and CERA, the Belgian cooperative bank, created category B shares in different contexts. Crédit Mutuel used this innovation within the framework of the application of the credit deregulation policy. To have access to credit, members were required to borrow 5% more than their credit needs and to reinvest this part in category B shares. In the case of the Caisses Desjardins, the distribution of B shares was facilitated by a fiscal incentive given by the State for a few years. In the case of CERA, B shares were gradually released according to the corporation’s needs and carried a substantial interest rate. In the 1990s, several cooperative banks became interested in using debt securities as financial levers. Generally speaking, the cooperative banks never used these securities at the same proportion as that was the case for joint stock banks. This tendency to assume lower risks is consistent with the cooperative model, which takes into account collective equity and the value maintaining the user-relation. In the same spirit, the relatively moderate performance of cooperative banks lowers possible recourse to certain financial instruments, the real cost of which is high and which carry high risks (leveraged buyout, junk bonds, etc.). Among the cases studied we have chosen three examples which bear witness to change in the relationship to financial markets. Firstly, the Mouvement Desjardins, which took its inspiration from securitization models already existing on the market and created Capital Desjardins Inc., a structure with the sole objective of buying debt securities issued individually by the credit unions and grouping them together and issuing a bond with the same term and yield. This structure permitted the development of an original relation with markets corresponding to the network’s decentralized nature and which offered the best possible value for their shares on the market. Another case which raises a special attention is that of an important subordinated loan by cooperative Bank of Great Britain which was negotiated with a major institutional lender and for which the assessment of the value of the Bank’s name was a significant part of the actual value. This case confirms the growing importance of the evaluation of intangible assets in the global evaluation of a given corporation. The large-scale use of hybrid securities is a relatively recent phenomenon in financial markets and little used by cooperatives. Rabobank took advantage of the exceptional rating it was given by major rating agencies by recently issuing two series of hybrid securities. First, in November 1999, a 650 million eurodollar issue of a Tier one capital, non-voting, perpetual preferential security with a fixed return of 7%. Then in June 2000, it issued one billion eurodollars worth of Rabobank Member Certificates intended only for members of local credit unions and Rabobank employees, which had a variable cumulative dividend 1% higher than ten-year bonds issued by the government of Holland. The success of these issues could clearly be a source of inspiration for other cooperative banks. 8 Among the cases we studied, Crédit Agricole (CA) and its various components are without doubt the institutions which have made the most use out of mechanisms implemented by the French government in the late 1980s and early 1990s to make it easier for cooperatives to gain access to additional equity funds. “Cooperative investment certificates” were used by several Caisses Régionales to obtain capital needed for the CNCA’s mutualization in 1988 and for other investment projects. The irony of this situation is that these are the same securities which will soon be used in the listing on the stock market of a structure which includes CNCA in its entirety, all its wholly-owned subsidiaries, and 25% of the Caisses Régionales’ capital. The Caisses Régionales will hold 70% of the new-created holding while the balance will be offered to the public (up to 20%) and employees (up to 10%). In order to acquire a 25% participation in the Caisses Régionales, the CNCA had to resolve the problem of how they were valued. Two approaches were retained. The first approach was a multicriteria one based on the PER (Price Earning Ratio), the ROE (Return on Investment) and net assets. The first two criteria are each worth 25%, and the third 50%. If the Caisse is not listed on the stock exchange, the CNCA plans to use a theoretical PER by comparing it with equivalents. This approach will enable the calculation of the Associates’ cooperative certificate issue (ACC). For already listed Caisses (18), the CNCA will use the average stock market value of the cooperative investment certificates (CIC) as at July 20, 2001. This entry on the stock market of a part of CA’s activities is in addition to activities by the Caisses Régionales, 18 (out 48) of them use the financial market through the issuance of cooperative investment certificates listed on the stock exchange. Collectively, they have issued Cooperative Certificates worth around 3 billion euros. The new capitalization strategy is closely linked to an internalization strategy via acquisitions, which necessitates the cohabitation of the cooperative logic with the capital valorization logic, a subject we discuss below. Among the innovations, it is also worth noting the merger of Belgium’s CERA cooperative bank and Kredietbank, a Belgian joint stock bank. This merger seem arising less from an immediate need for capital than a strategic positioning in light of the consolidation of Belgium’s banking sector. Without being privy to the decisions made, it would appear that when the Dutch ING took control of the Banque Bruxelles Lambert and when the Dutch Fortis Group bought all of Générale de Banque’s capital, the signal was perhaps given for the CERA-Kreietbank merger. This genetically curious merger can be viewed as serving the higher interests of Belgium in creating Belgium’s largest bank with 25% of its banking market and 10% of the insurance market. This case enables us to situate the stakes involved in the reorganization of the banking sector by taking into account national issues and not only cooperative or market stakes. 9 Moreover, the evolution of this cooperative bank warrants closer examination. In this regard, the cooperative has become a financial holding, the equity and revenue of which depend on the performance of KBC, a joint stock corporation. On one hand, CERA appears to its members as a mutual investment fund tied to a set of particular institutions (KBC) and, on the other hand, as a charitable foundation managing the indivisible collective equity of several generations of cooperators. This foundation’s mission is among other things to promote the development of cooperatives in sectors other than the banking sector. For the time being, the cooperative holding is mute with respect to the management of KBC, but a future critical episode could come from an outside purchase offer which would enable the substantial valuation of the cooperative holding’s equity, and which would provide even more resources for cooperative promotion. In this light, what would be the position of the cooperators? 3. A model to represent the forces and trends characterizing the evolution of cooperative banks Heightened pressure due to economic, regulatory and technological shakeups has led to a period of profound questioning by everyone. These changes affect all organizations involved, forcing them to reassess their competitive advantages and thus their capacity for competing in what has become a hostile environment. In such a difficult environment, managers and directors of cooperative banks are questioning the organization’s ability to compete. Some are advocating a transformation into a joint stock company; others are reaffirming their belief in the relevance of cooperatives in this kind of environment; and many are in a period of profound questioning. The changes in the competitive environment are not the only ones influencing the development of cooperatives. Various structural changes also have an impact in varying degrees on the intensity of application of cooperative rules and contribute to this identity crisis. It should be noted that in a world in which individuals are defined by their interests, it is not obvious that the cooperative distinction can be recognized and appreciated. The size of local cooperatives also raises the issue of collective action, to the extent that it is necessary to take into account the quasi-public nature of cooperatives. To this, we must add that the specific mission of cooperative banks in Western countries is no longer defined by the situations of abuse or absence of service which characterized the initial conditions. Members now experience their relationship with their cooperative in a highly evolved competitive environment offering choices as much at the individual level as at the level of collective action. For the purposes of this discussion, we retain two poles enabling us to represent the evolution of this dual tension characteristic of cooperative organizations. A first pole represents the heart of cooperative identity: values and legitimacy, democratic cooperative practices, characteristic features, etc. This first pole serves to illustrate the evolution of cooperative identity, the dynamic of which can range from a strengthening of this 10 identity to a complete loss of identity in the event of demutualization. The second pole represents the evolution of the market rules in which cooperatives develop. It serves to illustrate the intensity of competitive forces, deregulation, internationalization, solvency of demand, etc. The evolution of a cooperative’s activities can move towards a hyper-competitive universe or towards a non-market universe in which there is no longer a real solvency of demand. By positioning a cooperative in terms of these poles, we can characterize its situation and draw out the main factors influencing its transformation (Figure 1). Figure 1 To enable us to illustrate the diversity of cooperative profiles, we have broken down each axis into three possibilities. The vertical axis (cooperative identity) is broken down into: a strong intensity of application (+) of cooperative values, rules and principles; a weak intensity of application (-); and demutualization when the cooperative adopts a corporate status. The horizontal axis is also divided into: a high competitive intensity (+); a low competitive intensity (-); and a nonmarket environment. As such, this model enables us to construct a typology in which we will be able to find six different typical cases. 11 Quadrants I, II, III and IV and two cases outside the quadrants • I: strong cooperative intensity and low competitive intensity; • II: weak cooperative intensity and low competitive intensity; • III: weak cooperative intensity and high competitive intensity; • IV: strong cooperative intensity and high competitive intensity; • V: A cooperative in a non-market universe; • VI: Demutualization. The arrows allow us to illustrate an important hypothesis, namely: we are witnessing a shift from position A (1st quadrant) to position B (3rd quadrant). This hypothesis merely takes up the analysis of the main trends we noted above. On the basis of this hypothesis, we can expect a weakening in the intensity of application of cooperative rules whenever market rules are intensified. This scenario is based on the hypothesis that at the moment of the cooperative’s creation, the entrepreneurial function requires a high mobilization of members when the initial market context is typically associated with a weak competitive environment (monopoly, absence of a solvent market, information asymmetry, etc.), leading to situations of abuse. This double situation places this cooperative in the 1st quadrant. In the 3rd quadrant, we can see the decrease in intensity of the application of cooperative rules and the strengthening of market rules as the increasingly unique basis of decision arbitrages. The position of managers can be comfortable for a certain period of time, but the gen- eralization of arbitrage as a function of market rules is likely to lead to a loss of distinction and lead actors to reconsider the “meaning and legitimacy” of cooperative status in a competitive context. One can imagine that these cooperatives would experience increasingly strong internal tensions or, to the contrary, a loss of interest by members and the development of behavior limited to a market logic. This kind of situation could lead to eventual demutualization or to a revitalization of the cooperative, depending on the orientations adopted in light of new situations. The positioning of cooperatives in the 4th quadrant (position C) requires the conditions for the implementation of reinforced cooperative rules and principles. This position is thus necessarily based on the reinforcement of associative structures as well as on the search for and management of a continually renewed cooperative coherence. 4. Four cases of cooperative banks: where are they to be situated on the matrix? The objective of this research focuses primarily on innovative capitalization practices of cooperative banks. We presented above the evolution of capitalization practices observed over the past 20 years in cooperative banks. Throughout this research, however, we have adopted the view that capital needs (and the vehicles developed to satisfy them) must be associated with the strategies of the banks. Indeed, the orientations and strategic choices lie at the heart of each cooperative bank’s response to the dual logic illustrated above. 12 This model enables us to situate these decisions in the various contexts in which cooperative banks operate “more than a century later,” when the initial situation (the 1st quadrant) is nothing but a memory. Several factors can characterize this evolution and explain the degree of vitality that we observe at present. We do not claim to be able to invoke all of these factors, nor to present an adequate assessment of each of the cases we present. However, we believe that it is very useful to reconsider the cases studied in this research in light of this matrix. Positioning the different cooperative banks while illustrating both their strategic orientations and the adopted capitalization vehicles will bring insight to the discussion of various scenarios. The cases that I will present are observable reference points with regard to cooperative rules, adopted strategies and capitalization strategies. Given the objectives of our research, I have retained cooperative banks which strike me as particularly dynamic in this regard: CERA, Crédit Agricole, Rabobank and cooperative Bank. (a) CERA/KBC: Mutation towards a capitalist status and trivialization of the Cooperative Form CERA, a cooperative bank, merged with Kredietbank in 1998 to become KBC, a joint stock company, a substantial portion of the capital of which (around 40% at the time of the merger) is held by CERA Holding, which has a cooperative status. A lawsuit contesting the treatment offered to its cooperative members was only settled in 2000, and was finally adopted in early 2001. The litigation focused on the proportion of the accumulated surplus to be distributed to current members, with revenues from the non-distributed part to be used for promoting the cooperative formula and for funding social works in Belgium. Cooperative rules and their application. Before its merger with Kredietbank, CERA’s profile was very similar to that of the vast majority of cooperative banks. Arising out of the agricultural class in the last century, CERA had 207 local credit unions with 938 banking outlets and 410,000 members among its 1.5 million clients in 1997. In light of M. Norman’s excellent analysis of the application of cooperative rules at CERA, we can say somewhat laconically that over the past several years, its cooperative practices have satisfied the legal requirements without, however, being embodied in stimulating organizational practices. Norman concludes by saying that cooperative principles at CERA have become nothing more than the basis for corporate discourse. The loss of intensity in the application of cooperative principles and rules probably accounts for the fact that CERA’s demutualization has not engendered a negative reaction among its members, except for a demand for a better sharing of the corporation’s real value in light of the fact that the initial decision was to accord members only the nominal value of their shares. Market rules and adopted strategies. During the 1980s, CERA confirmed its shift towards a universal bank. 13 Agricultural lending went from 80% in 1967 to 26% in 1992, even though at the time farmers still accounted for 54% of its members. This shift was accompanied in the 1990s by the adoption of practices similar to those of private banks. These practices, well detailed in the 1997 annual report, are based on gaining control of the profitability of clients, products and activities, as a function of the performance objectives for the equity funds. These practices account for CERA’s financial success close to the time of the merger, when its return on equity (ROE) went from 7.58% in 1995 to 11.82% in 1996 and to 15.89% in 1997. This performance occurred in a market undergoing rapid change, in which competition was stronger and in which all are on the lookout for acquisitions and alliances. The merger of CERA and Kredietbank enabled the creation of the Belgium’s largest bank, with assets of 160 billion euros. KBC holds 25% of the banking market and 10% of the insurance market in Belgium. The newly created group rapidly displayed its international ambitions by implementing an aggressive external growth strategy. It invested 1.5 billion euros in banking and insurance acquisitions in eastern Europe (Czech Republic, Hungary, Poland), which it views as a second domestic market. In 1999, KBC’s return on own funds was 20.5%, compared to 16.5% in 1998. It undertook a rationalization exercise which saw the closing of 600 of its 1,500 outlets in Belgium. Capitalization practices: recourse to the stock market. Given that it is listed on the stock exchange, the KBC Group has access to capital that it can raise via public offerings. In this connection, its intention to issue 500 to 700 million euros in shares was checked due to the drop in their value of more than 35% in 1999. The year 2000, however, saw its return to an above-average profitability (24.70%), and the settlement with its dissident members is such that it will allow for a more favourable assessment by the market. Situating CERA on our matrix. The case of CERA at the end point of its trajectory can be easily situated on the matrix (Figure 2). Clearly, it is located at position D, outside the cooperative universe. However, it is important to reconstruct the trajectory of this cooperative bank over the course of the preceding decades. As presented above, it seems clear that over the years, what we call the vitality of cooperative rules was gradually vitiated to the point where it essentially became symbolic. This conclusion is also based on the kind of “business arbitrages” observed at CERA in the years preceding the merger. The goal was radically changed to the pursuit of a maximum return on invested capital. The choice of “clients” was made on the basis of their contribution to the cooperative’s equity funds performance, and only 30% of them were members. This presentation of the evolution of CERA allows us to note that initially there was an internal demutualization, followed by an external demutualization, likely accelerated by consolidation constraints on the sector at the national level. It is not surprising, then, that so 14 few members manifested opposition to this project. It is more interesting to note that the “cooperative arbitrage” function was ultimately completely shifted onto a foundation that pursued a general interest mission while remaining owner of a substantial portion of KBC. Does this mean that it is no longer possible to associate a cooperative logic with the logic of banking activity? This would appear to be the conclusion reached by CERA’s directors. To the extent that cooperative logic becomes almost completely absent from the practices and arbitrages of banking activity, it would seem that this transformation can occur without members’ acting as a collectivity and who only seek to benefit on an individual level, as was the case with CERA’s members. (b) Crédit Agricole: towards the stock market References with regard to cooperative rules. Crédit Agricole (CA) has 2,672 local credit unions grouped into 48 regional credit unions, themselves grouped in the Fédération Nationale du Crédit Agricole (FNCA) and holding 90% of the Caisse Nationale du Crédit Agricole’s (CNCA) capital. They make up the base of Crédit Agricole. The local credit unions are representative structures bringing together more than 5.5 million members who name the 33,413 administrators. Financially, their activities are integrated with those of the regional credit unions, which also have a network of 7,679 outlets and 10,100 “points verts” or outlets installed in retail businesses, serving 15.5 million clients. Figure 2 15 The FNCA is the deliberative body for the regional credit unions and the Group. It plays the role of the professional chamber. Since the adoption of the January 1988 mutualization law, the CNCA has been a public limited company whose capital is held by the regional credit unions (90%), the administrators, the employees and former employees of Crédit Agricole, and, since 1996, by Groupama SA. As of December 31, 2000, Crédit Agricole managed assets of 535 billion eurodollars (3,513 billion francs). Its own funds amounted to 25,792 billion eurodollars (169 billion francs). CA has a 22.3% share of the savings market and 17.6% of the loan market. It has a solvency ratio of 11%, ROE of 10.1% (11.6% in 1999) and an ROA of 0.49%, while its operating coefficient is 64%. The ratio of own funds to weighted assets amounts to 9.5% (May 2000). References with regard to market rules. CA has three strategic orientations: a) strengthening the local bank; b) development of the investment bank and assets management; and c) increasing its presence in foreign retail banks (Italy, Argentina, Portugal, etc.). CA has been multiplying its external growth operations in recent years, and it is estimated that its diversification strategies have required 30 billion francs of its own funds during this period. The Caisses Régionales have provided nearly 15 billion francs in capital to the CNCA. Due to its entry on the stock market, CA has sought to finance its development without demutualizing. Capital raised in this manner will enable it to undertake major external growth operations. This external growth strategy has already led it to take control of the Indosuez bank, 16% of Banco Inteso’s (Italy) capital, 10% of Crédit Lyonnais shares, 6.7% of the Commercial bank of Greece, and 31% of Impériale de Lyon. Crédit Agricole also holds 9% of BES’s (Banco Esperanto Santo, Portugal) capital. It is also active in Poland through the acquisition of two-thirds of ELF’s capital (real estate leasing), evaluated at 400 million eurodollars. In addition to its internationalization strategy, there have also been inroads in consumer credit (Sofinco), life insurance (Prédica) and damage insurance (Pacifica). All CA’s acquisitions have been integrated into the network of subsidiaries controlled by CNCA. In recent years, CA has also consolidated its position as leader in the French market in terms of results, financial base and market shares alike. Crédit Agricole’s entry into the stock market coincided with the group’s aspirations with regard to the control of Crédit Lyonnais at a time when the French government would likely sell its current 10% percent holding. The latter gave Merril Lynch the mandate to privatize Crédit Lyonnais. The association with Crédit Agricole will permit the creation of a major French banking pole of “global proportions,” and able to become one. Crédit Lyonnais is thus viewed as a strategic pawn in the current European and world game where CA try to be a player. According to its directors, there is no question of abandoning the cooperative status. A genuine demutualization of CA would have people up in arms. Given the 16 listing scheme, the demutualization issue will not become a problem in the short term. In the medium and long term, however, this mixed mutualism-capitalism formula will depend on the pace of increases in capital and the capacity of the Caisses Régionales to keep pace with the raising of funds. If CA were completely demutualized and the surplus redistributed, members (5.6 million) could expect to each receive 8,000 euros from the operation. CA’s value is estimated at around 50 billion euros (twice the value of its own capital), from which must be subtracted the 3 billion in cooperative certificates issued by the regional credit unions. In all, the stock market entry of 20% of the new CNCA’s (the listed vehicle) capital should represent a 4 billion euro operation. Questions…the search for a new coherence. The partial demutualization of CA raises several questions. First, with regard to the medium-term interest of eventual investors, since it is the assets of the Caisses Régionales which are the most solid, but they only account for a small part of the vehicle and the regional credit unions control the majority of the listed vehicle’s capital. The relevance of the strategic operations with regard to the share valorization will be at the heart of the concerns of external investors, observers and perhaps mutlualists. The rights of minority shareholders and the legitimacy of their concerns may well significantly complicate internal arbitrages with regard to the distribution of costs and the setting of prices among the various entities, and make it more difficult to coordinate the setup as a whole. Situating Crédit Agricole on our matrix. As was the case with CERA in Belgium, national concerns are clearly not absent from the strategies adopted by Crédit Agricole. However, Crédit Agricole’s size and very broad base in France enable the mutualist structure to maintain its identity and, for the moment, to retain its leadership position. Several factors lead us to situate it in the 3rd quadrant, on the demutualization path. Firstly, the ratio of nonmember clients its three to one. As well, there is the number of agencies (7,679) and service outlets (10,000) compared to the number of local credit unions (2,672). Ultimately, these two relations will only weaken cooperative practices. The second consideration in our classification of Crédit Agricole involves its adopted strategies. Clearly there is a desire to “strengthen its position on the French market,” but there is also an external growth strategy of a major international bank. How is this to be associated with the activities of its cooperative? It is thus not surprising that Crédit Agricole is turning to the stock market to finance its strategy. Funds raised on the stock market, both by the Caisses Régionales and the vehicle eventually set in place by the CNCA, are remunerated on the basis of the return on capital, and the value of these shares will fluctuate as a function of results and stock market activity. The increasingly pronounced rupture between the activities of Crédit Agricole and those of its members, coupled with an external growth strategy based on acquisitions, can only strengthen the domination of the logic of capital to the detriment of the cooperative logic, which is convergent with the entry onto the stock of a substantial portion (30%) of its value. The extent of its international operations and the inter- 17 vention methods are thus leading to a mutation, the form of which remains to be determined, but which threatens cooperative revitalization. (c) Rabobank: a return to the membership Cooperative rules. Rabobank heads up a network of around 400 local cooperative banks grouping together more than 700,000 members and serving nearly 9 million individual and business clients. Created on the Raiffeisen formula, local credit unions did not require an initial share purchase, but members had to assume unlimited liability. All surpluses were thus put into the reserve, a practice still operative today. In terms of representation, local credit unions are grouped into 43 regional districts where policy matters and orientations are discussed. In the 1990s, Rabobank’s cooperative status came under intense review, and the 1995 congress concluded that its cooperative status had contributed to its success and needed to be revitalized. The decision to open its membership to all clients and to propose incentives arose from the policies determined at this congress. Market rules, a highly competitive environment. Rabobank enjoys a very strong position in its domestic market comforting two giants, ABN-Amro and ING. Rabobank holds 85% of the Dutch agriculture and agribusiness market, 40% of small business lending, 25% of mortgage loans, and 15% of lending to large businesses. It re-centered its international activities in the agribusiness niche, in which it is aiming for 4% of the world market. With a balance sheet of 340 billion euros at the end of 2000, Rabobank has 140 outlets in 37 countries. Having remained outside the concentration movement observed in Europe over the past few years, it has recently sought to develop alliances and partnerships with a view to “opening the way for a European cooperative bank.” Even though the failure of the planned merger with the German DG Bank (central bank for the German cooperative network) dampened its aspirations, Rabobank’s subsidiaries have been particularly active over the last two years in alliances and partnerships, as can be seen in the growth of participation of minority interests in the balance sheet. In any event, its strength in its own market and its skill in certain niches make it a very interesting partner. Capitalization practices…a return to the membership. Rabobank’s decision to maintain and strengthen its cooperative status for its Dutch activities resulted, in 2000, in the issuing of investment certificates (25 euros each), offered only to members of local credit unions and employees, and having a variable cumulative dividend 1% higher than 10-year Dutch government bonds. Although it expected to raise 400 million euros in this way, the real demand amounted to 892 million. This capitalization strategy is part of a larger strategy of maintaining closer links with its members and of recruiting new ones (the membership has been declining for the past ten years). Indeed, Rabobank announced this year that it had recruited 210,000 new members. In this connection, Hans Smits announced,“No, I'm not blinded by the figures. It is only the beginning of what we promised each other: en route to more than a million members in 2003. And it is also 18 only the beginning in terms of the bond with members and member involvement. We will have to put a lot of energy in the coming period into the revitalization of our trusted, but equally modern organizational form…It is my firm belief that the 21st century will not only be about selling products and services. No, it will also be about giving meaning to the relationships with customers and members…Our cooperative background is not a barrier, even though some people occasionally try to convince us otherwise. On the contrary, it is an asset, a major competitive advantage.” Situating Rabobank on our matrix. Even though Rabobank’s historical evolution has some similarities with those of CERA and Crédit Agricole, the 1995 congress was characterized by a desire to break with the trend and to undertake cooperative revitalization. Even though it seeks to become an international reference in agribusiness financing, Rabobank’s overall activities strike us as being in line with its members’ activities. The recently advocated approach with a view to “opening the way for a European cooperative bank” was based on alliances and partnerships with other cooperative banks. The failure of this initiative does nothing to diminish the value of the energies invested in opening the way for new cooperative perspectives. The reconsideration of its cooperative status in the mid 1990s and the clear resulting choice no doubt provide Rabobank with a head start in the revitalization of the trust of its membership. In this light, we situate it in the upper part of the 3rd quadrant, and we feel that it is moving towards the 4th quadrant. (d) Cooperative Bank: The effect of loyalty and the legitimacy of promoting distinctive values Cooperative rules. In contrast to the preceding cases, Cooperative Bank does not have a cooperative status; it is a whollyowned subsidiary of the Consumer Wholesale Society (CWS), a consumer cooperative directly descendent from the Rochdale pioneers. CWS board members are also administrators of the cooperative Bank. It promotes principles inherent to cooperation: quality and excellence, participation, freedom of association, education, quality of life, and integrity. It regularly consults its members in order to develop an ethic in line with the values of its adherents. It respects their ecological and moral values, and refuses to invest in certain activities or organizations. Three percent of its before tax profits are used for donations to local communities. Market rules…a strategy basic on ethics. Cooperative Bank has a distinctive position in the market given its ethical and ecological policies. Its practices integrate a “stakeholder” approach, which it rigorously applies. It systematically measures the value given to its clients, its shareholder (CWS), its employees, its suppliers, the local communities in which it operates, the environment, and past and future generations. This systematic measuring is carried out by outside experts and widely diffused to all its partners and the public at large. With regard to products and services, cooperation bank is innovative in its use of new technologies. It is renowned for the excellence of its service. Considerable energy is devoted to ensuring that its employees understand the bank’s values and speci- 19 ficity. It provides them with training courses on its heritage, values and what distinguishes it from its competitors. Capitalization Practices…the loyalty and legitimacy effect. Cooperative Bank has experienced strong growth since the mid-1990s. Its cost/revenue ratio (efficacy ratio) went from 75% in 1994 to around 60% in 2000. The return on equity has experienced phenomenal growth, going from 12% in 1994 to an average of 26.5% from 1996 to 2000. With regard to capitalization, the solvency ratio (BIS) remained above 12% during this period of strong growth, and cooperative bank generated a cash flow sufficient to ensure the great majority of its own funds needs. Lastly, note must be made of the “Cooperative Bank” brand value, which contributed significantly to the evaluation of the corporation’s overall value in the negotiation of a loan from an institution investor on the NewYork market. Situating Cooperative Bank on our matrix. A member of the family of consumer cooperatives, Cooperative Bank has developed considerable coherence in its consulting, business and capitalization practices. It consults its clients and seeks to develop a strategic positioning and commercial practices in line with their values. Its stakeholder approach, with the systematic measuring of the value given to each of its partners and its widely diffused external verification, constitutes a remarkable effort at transparency which considerably increases the legitimacy of its actions. The involvement of its employees and the particular training in the bank’s values and specificity reinforce the coherence of its strategy. Its competitive positioning is systematically demonstrated by its excellent results. Its productivity is on a par with the best in the industry, and at the same time, it constantly innovates. These performances incline us to feel that Cooperative Bank has a profile consistent with the 4th quadrant. Figure 3 5. Scenarios and issues related to capitalization Can cooperatives maintain their status, their goals and the resulting arbitrage logic while resolving the thorny capitalization problem? Has any progress been made with regard to the important issue of the capitalization of cooperative banks? We have identified a certain number of innovative practices without, however, finding clear paths to defining cooperative capitalization in the 21st century. To the contrary, it would appear that over the past 20 years, the owner-user duality has more or less been drained of substance—members have become clients. Since members no longer know how to value their particular relationship to their cooperative, any solution to the 20 capitalization problem can only be a hybrid one setting off a mutation of some sort. To meet the imperatives of competition, cooperative banks have vigorously developed the business side of things, without asking themselves whether cooperative specificity can constitute a competitive advantage. A recent study by Lévesque and Côté, involving executives of 34 local Desjardins credit unions reveals that the management staff at 19 credit unions who see advantages related to a cooperative status also feel that they have a favourable competitive position on the market, whereas the management of 15 credit unions who have a negative view of their status also feel that they have an unfavourable competitive position in their market. Forthcoming case studies will add to our understanding of these observations. Can cooperative banks be content with being a good bank? The quadrant three position strikes us as being less and less comfortable. The many reorganizations shaking up the financial services industry force all actors to clarify their position, going from their mission to arbitrage, practices, and to be sure their strategic orientations. The intensification of competition has required organizational alignment to become clearer. “Confused” organizations will not be able to resist this additional pressure. Cooperatives are organizations which have certain advantages, but which also have certain constraints. We must recognize the limits specific to the model, i.e., constraints on use (member activity tied to corporate activity), constraints on capital (social share, slow accumulation via surpluses, redistribution linked to member activities with his/her cooperative), and constraints on the decision-making process (democratic structure). Why should these constraints be tolerated when the market demands clarity and flexibility? Above all, why maintain them if we do know how to (or cannot) exploit the advantages they procure? We believe that third-quadrant cooperative banks are likely to see increased pressure coming not only from the market, but also from managers, employees and, lastly, members. At the base of the different scenarios which inspired the cases we observed and the research we conducted, we need to make a fundamental distinction between cooperative banks which seek and elaborate strategies to strengthen cooperative identity and those centered only on market considerations. It strikes us that without a significant effort to give “meaning and legitimacy” to each of these cooperative banks, pressure for demutualization coming as much from the market as from managers and ultimately members can only increase. The interest for strengthening cooperative identity must first show the reinforcement of the competitive position of cooperatives wishing to move into quadrant four. We must show that this kind of position would also enable the significant resolution of the problem of capitalization. In our view, the most promising pathway is to rethink the link with members by demonstrating the very significant proximity between cooperative identity and loyalty. Loyalty is based on a set of practices articulated around fundamental values 21 and practices which are completely integrated into strategic orientations, and cannot be achieved in the margins. The stakes and challenges of loyalty also involve the organization’s capacity for “earning” the loyalty of its employees; for it is they who will merit the loyalty of the members. The association between loyalty and competitive advantages is well documented in the literature, which draws out the fact that loyalty enables both an increase in productivity and a reduction in costs. It also leads to improvements in service and greater member satisfaction. Productivity, costs, service quality, and greater satisfaction: key elements of competitive advantage, which does nothing to exclude the need for continually delivering a very good value “business proposal.” Figure 4 The most astonishing result of loyalty is the realization that corporations that know how to manage (and earn) this kind of trust have little need to rely on the stock market to support their growth. Their profitability is such that self-generated funds are sufficient to capitalize the enterprise. If we accept this reasoning, the issue is to know how to redefine the associative structure and cooperative identity as the lever of choice for developing loyalty. cooperative Bank offers a very interesting means of doing so. It has taken the idea of transparency to great lengths in seeking a balance among its various partners approach. Here, over and above mere words, the identification of precise indicators and the systematic measuring of 22 the value delivered to all ensure the legitimacy essential to the process. The shift announced by Rabobank also strikes us as moving in the same direction. Here, we do not have sufficient information to judge the undertaking in greater detail. Rabobank is a cooperative bank that has assets 30 times greater than those of Co-op Bank, has considerable penetration in its own market, and has a significant international presence. It is thus a very different field of experimentation that the other large cooperative banks would do well to heed. Conclusion I am well aware that this discussion is only a beginning in the development of a relevant analytic framework for proposing an approach to renewal which could be of interest both to members and to senior management. Over the past few years, several cooperative sectors have been seriously weakened by profound changes in the society as well as in the industries in which they operate. Strengthening of cooperative identity can only be achieved if it leads to a reinforcement of competitive position and an improvement in cooperative banks’ performances. I can only hope that the ICBA continues to be a privileged venue for exchanges of innovative experiences. Note: For a more in-depth discussion of the relationship between loyalty and capitalization, see F. Reichheld’s work, published by the Harvard Business Press. 23 Klaus P. Fischer M. Fischer has a Ph.D in finance from Texas A&M University. He is Professor of finance at the Faculty of Administrative Sciences at Laval University (Québec, Canada). His research and teaching work focuses on corporate finance and financial intermediation (especially of mutual type) with special interest in microfinance issues. His current research is centered around the governance, regulation and supervision of mutual intermediaries in developing countries. M. Fischer has published numerous articles in research-oriented journals but also likes to publish articles of vulgarisation on his research work. Do Agencies Rate Cooperative bank-issued bonds fairly? By Klaus P. Fischer, Ph.D. In collaboration with: Ridha Mahfoudhi, MBA Selma Shaker, MBA Laval University, Québec, Canada Note: We received useful comments from Professors Guy Charest, Bernard Lamont and Patrick Savaria of the Faculty of Administrative Sciences, Laval University. We are grateful for their reading, but all remaining errors are ours. The views expressed in this document are those of the authors and in no way represent the official position of Developpement International Desjardins (DID) or any other institution involved in the organization of this event. The problem How would you feel if somebody, endowed of a lot of authority, tells your customers that your products are not “quite” as good as those of the guy down 24 the street. You know that this is not true. You have seen their product failing all around while yours stood. Then, that somebody says that the higher price customers are willing to pay for your product is most likely too much. Lastly, to top it off, that somebody turns around and tells you in the face that “you are doing great.” Wouldn’t you feel that you are being patronized? Unfortunately, that is exactly what rating agencies (RAs) are doing to cooperative banks (CBs). To the question of whether RAs rate CBs bonds fairly, my answer is “no, they don’t.” Incidentally RAs and investors do not seem to agree on the value they assign to bonds. Investors are willing to pay a higher price for these bonds than what the rating would suggest they should pay. I do not believe that there is malice on the side of the RAs. They have nothing to gain from being the “bad guys.” But that doesn’t change the situation. To their credit I will add that they have improved in the “fairness” of their rating. Ten years ago, the situation was much worse. Today, we even observe a few cooperative banks enjoying a “triple A” rating (are they the ones with the most joint-stock bank-like appearance?). But this does not change the fact that on average they appear not to be fair, even today. How did I arrive at this conclusion? I will answer with two groups of arguments. The first group is of a theoretical nature. I explain why investors should expect bonds issued by CBs to be exposed to lower default risk (and thus should, other things being equal, display a higher rating) than bonds issued by joint-stock banks (JSBs), and not the other way around. There are indeed several arguments that support this hypothesis that, I suspect, RAs genuinely ignore or chose to ignore. I will also note some of the “peculiarities” RAs themselves identified about CBs. The second group of arguments is of a statistical nature. I will present results of a statistical analysis performed on a matched sample of CB and JSB bonds. But, before, let us briefly define what rating means and point to the consequences of having one’s bonds rated “unfairly.” What exactly does a rating measure? When a firm obtains a rating on a bond it issues, this rating is an assessment of : 1. The probability of default or delayed payment, and 2. the expected severity of loss from default of holding the bond. Thus, when the average rating of a CB bond is inferior (however small) than that of a JSB, the rating agencies are in effect stating that the average probability of default or delayed payment is higher, and the expected losses from holding a bond, are more severe for a CB bond than for a JSB bond. The consequences to CBs of “unfair” ratings are not trivial. They potentially increase the cost of financing available to them through the market. This, in turn, will have a two-fold effect. On the one hand, it will reduce the competitiveness of CBs in financial markets. On the other, it will negatively affect members’ welfare through higher rates on their credit, reduce benefits owed to them and increase risk of troubles in the system. However small this effect may be, it is only fair to confront RAs about their rating practices. 25 Professor Tremblay just spoke to us about the practices and sources of capital available to cooperative banks. Issuing bonds is indeed one of them (using the Basle standards, as subordinated capital, for example), but it is difficult to recommend making use of them unless the issues are indeed fairly evaluated by market participants. And in this, RAs have a particularly heavy responsibility. Why CB bonds should present a lower default risk A number of differences between stock banks and cooperative banks suggest important differences in the risk exposure of CBs and JSBs and the bonds they issue. Perhaps the most important ones, two theoretical and one empirical, are: 1. Moral risks between shareholders and debt-holders 2. Managerial conservatism in mutual intermediaries 3. The role of deposit insurance I will now briefly present each of these arguments. 1. Moral risks: Joint stock banks take more risk This is due to the well-known conflicts of interest between shareholders and debtholders in a stock bank. As has been amply demonstrated in the banking literature, in JSBs, bondholders and depositors are subject to moral risks exercised by shareholders. The heart of the argument is that shareholders are subject to incentives to assume risk beyond what is good for debtholders (through on- and off-balance sheet positions). This will lead to a transfer of wealth from the first (bondholder and depositor) to the second (stockholder). In the presence of deposit insurance, shareholders are subject to incentives to also expropriate the deposit insurance fund. The most transparent theoretical explanation of these incentives resides in the option nature of stock (as a long position in a call on the bank assets bought from the liability holders) and the deposit insurance contract (as a long position on a put option bought by the shareholder from the deposit insurance fund). I will not enter into the details of this theory since it is well documented (and its existence widely confirmed) in the banking literature. In the case of CBs, shares are not subject to fluctuations in market value and the beneficiaries of residual cash flows are both debtors and non-bond liability holders. Thus, the incentive to transfer wealth from liability holders to shareholders (debtors and depositors) is either inexistent or strongly diminished. This lowers incentives for risk-taking in CBs reducing default risk on liabilities. As an empirical support for this line of argument we note the work of Chou and Cebula [1996]. They found that during the 1980s United States Savings and Loan (S&L) crisis, states with a greater proportion of mutual associations suffered a lower failure rate than states with a larger proportion of stock owned S&L. A similar result is obtained by Desrocher and Fischer [2001] who found that the main cause of joint stock bank failure is risk-taking, while in financial cooperatives risk-taking plays little or no role in the failure of mutual institutions. 26 2. Managerial conservatism: Cooperative banks take less risk This is due to the fact that managers (as we all know, quite “almighty” individuals in a CBs) are fundamentally conservative. Managers, like everybody else in this world, are considered risk-averse and are intent on keeping their jobs. Thus they prefer to hold safe assets that will ensure solvency of the institution. Some authors have argued for, and tested the hypothesis that managerial conservatism may in fact offset risk-taking that stems from moral risks (our first argument) in JSBs with diffuse ownership. Results tend to support this hypothesis. For example, Anderson and Fraser [2000] found that managerial shareholding is positively related to total and bank-specific risk. This implies that when control (under management) and ownership (by shareholders) are separated or ownership is very diffuse, the bank tends to assume less risk. Diffuse ownership is, of course, one of the main characteristics of CBs. You may think that large, publicly held JSBs (such City Bank) should then be very safe. This may or may not be so. Management of JSBs is subject to efficient disciplinary mechanisms by stockholders through: i) the stock market; ii) the market for mergers and acquisitions; and iii) incentive schemes such as stock options. As a result, incentives of managers tend to be aligned with the interests of shareholders. Thus, they are likely to engage in risktaking policies that transfer wealth from liability holders (depositors and bondholders) to shareholders despite their fundamental conservatism and even when the JSBs are of diffuse ownership. These control mechanisms are absent in mutual intermediaries. In a CB no control of managers by “the market” exists because: 1. shares are not traded in the market 2. there is no market for mergers and acquisitions of CBs, and 3. stock options on the bank are not possible. Further, there is wide agreement that corporate governance and control of managers by “shareholders” in mutual intermediaries is definitively weak. This allows managers to pursue their own interests and in particular engage in expense preference behaviour and avoid risk-taking. This should result in a less risky management style for mutual intermediaries than in JSBs. 3. Mutual banks do not take advantage of deposit insurance (while stock banks do) Stock banks almost everywhere in the world have often exercised the right to transfer the liabilities of the institution to deposit insurance (DI) schemes (i.e. exercise the put option bought from the deposit insurer). But when was the last time that you saw a major CB-networkrelated institution making use of insurance coverage? Failing institutions have been absorbed in to the system through a process of mutualization of failing cooperative liabilities, be this through internal institution insurance schemes or through a plain transfer of assets and liabilities to the rest of the system. Kane and Hendershott [31] in a suggestively titled article, “The Federal Deposit Insurance Fund That Didn’t Put a Bite on U.S. Taxpayers,” explained the judicious use that the United 27 States credit unions have made of their DI scheme, in stark contrast to the stockowned components of the financial system that caused huge losses to taxpayers. Further, most (but not all) cooperative bank networks have, on top of the publicly sponsored insurance scheme, built their own “insurance or contingency funds.” They are ready to perform rescue operations to save institution leaving liability holders untouched. In some countries, DI funds charge CBs a lower premium in recognition of lower risk exposure they present to the fund. Is it that everybody else sees that CBs represent a lower risk except RAs? Even more importantly, standard DI schemes for JSB cover deposits only. Bondholders are left to fight for their money in case of failure. In contrast, most internal insurance schemes for CBs do not cover deposits but the institution itself, and implicitly cover bondholders. This means that JSB insurance schemes are of no use to bondholders while CB insurance schemes do cover all liability holders, including bondholders. Even if we ignore all other arguments presented before, this reason alone should make CB bonds subject to a lower default risk than JSB bonds and thus benefit from a better—not worse—rating. If that is not enough, let us drive the nail in with a last stroke: it is widely known (at least in academic circles) that the presence of deposit insurance encourages JSB shareholders to take even more risk, and this incentive increases as the financial situation of the bank deteriorates in what is knowing as “gambling for survival”. Again, this gambling for survival is related to the “option-like” property of DI and shares. This incentive is almost absent in CBs since managers hold a fixed-payment contract. Furthermore, the mutual nature of CB insurance schemes encourages cross-supervision and prudent behaviour. As you can see, all three arguments suggest that bonds issued by CBs should be subject to a lower risk of default, and on average benefit from a higher rating than bonds issued by JSBs. I know of no argument that would suggest that JSB bond should, a priori, be less subject to default risk. There is no evidence either that CBs use less sophisticated risk management techniques than JSB (such as VAR and RAROC). But that is, of course, all before the fact. What happens in practice is the focus of attention of the last part of this presentation. The calibration of rating Rating is essentially a “subjective” assessment. Thus we cannot speak of objective rating because by definition this does not exist. No RA would say it does, because it could be made accountable of its “objectivity” or lack thereof. Researchers prefer to use the words “consistency” or “calibration”. It is in this very sense that I speak of fairness. The question of calibration (or fairness, to use our word), has been the object of some debate. Some specific forms of bias in calibration (or lack of fairness) have been reported. 28 Kuhner [2001] provides a short catalogue of criticisms of the rating industry. The biases take five forms: • power without accountability (not my words!) • conformity bias (i.e. agencies tend to herd, rarely diverging from each other’s rating or adapt to stock market analysts’ evaluations) • sociocultural bias (in favour of AngloAmerican management styles) • punishment attitude (toward ”disobedient” firms) • pro-cyclical bias (e.g. in the Asian crisis, ratings did not provide any warning, only after the crisis did ratings change—for the worse) Clearly, none of these “known” biases apply exactly to CBs. However, they show that CBs would not be the first to be discriminated against. But what do the RAs themselves say about CBs? I have reviewed documentation produced by Moody’s, S&P and Fitch about the rating of CBs. I present a brief summary of these opinions because they are very indicative of what they consider, and also what they ignore. Interestingly, none of the three arguments presented above that suggest a priori a lower default risk for CB bonds than for JSB bonds is ever mentioned. Ironically, RAs say they give CBs “a very good rating”! To be frank, that statement sounds somewhat condescending. They also say that they have a “neutral view of the merits of stock vs. mutual forms of ownership”. Assuming they mean by that that they are not favourably biased towards JSBs, then let me tell you that the facts contradict this statement. There are a number of aspects that RAs consider. I note four that appear most consistently, more or less in order of importance: 1. Degree of centralization, with a clear preference for a high level of centralization in the decision-making process over decentralized ones. 2. The existence of cross-guarantees or mutual guarantees, with a preference for explicit cross-guarantees over coverage based on mutual insurance funds. 3. The number of tiers in the structure, with a preference for lower tiers (say two rather than three) that makes centralized decision-making more efficient. 4. They prefer networks with small and homogeneously sized base units over networks with large units and heterogeneously sized. Other features they appreciate as typically present in most CBs are: 1. A high level of capitalization. 2. Closeness to and understanding of retail customer base. 3. Excellent capacity to mobilize deposits (they mean “core deposits”, the cheapest and lowest risk form of bank liability). 4. Relatively low bad loan rate, which they attribute to “customer closeness” (no mention is made of lower fundamental incentives for taking risk). So, for example, why does Rabobank enjoy an AAA rating? Because it seems to have tailored itself to please the RAs – of course, I doubt that that was exactly their intention. In the first four criteria they always come to the preferred choice: a cen- 29 tralized, two-tier structure, with an explicit cross-guarantee support system and rather small and homogeneous-sized network of base units. On top of that, they present a good standing in the other three criteria (capitalization, closeness to clients and capacity to mobilize core deposits). If that were not enough, Rabo is a “too big to fail” bank, meaning the Netherlands’s government would simply not allow that Rabo to fail. On the cynical side, I suspect that if a JSB presented all these attributes, they would probably create an additional rating category on top of AAA to accommodate it. Rabo being a CB, it just gets a triple A. I cannot avoid noting some contradictions and uncomfortable bias in their criteria. I will only comment on two. First, no matter how much you may argue, centralization in decision-making on one side, and closeness to customers on the other, are to a considerable extent contradictory goals. While this may perhaps be feasible in a small economy like that of Netherlands (i.e. Rabo), it will be much more difficult to achieve in the case of a much larger German Raiffeisen system (which is perhaps being penalized for its three-tier structure and more decentralized decision-making process). Second, and this is closer to our main concern: fairness in rating. RA preferences for centralized and two-tier structures smacks a clear bias for JSB-like structures. The closer a CB is structured to a vertically integrated JSB—as is clearly the case of Rabo—the better rating you get. Now, that is precisely the point. Why does a centralized structure represent a lower risk to bondholders...? In the documents they produce they never explain, only state their preference. This flies in the face of the historical reality: the huge success of mutual banks in Europe and other places in the world can be attributed to the decentralized nature of federated networks of individually autonomous base units. For many reasons, I do not have the time to discuss this here. One of the key features of these structures is that they allow a high level of independence in decision-making at its local level and closeness to its members, in stark contrast to centrally managed JSB branch networks. Methodology and statistical model Let us now stop speculating and focus on numbers. The RAs (and some in the audience) may argue that all said is just “ivory tower theorizing”. So let the data talk. I will quickly present the data used to pass to the more interesting part: statistical results. Of course, the quality of data influences the validity of results. Data We took much care to extract all information available in the commercial database. We used, in order of importance: • Datastream. This is an on-line database used by institutional investors worldwide. From this database we extracted all data relevant to hundreds of CB and JSB bonds. • Daniel Côté, Martine Vézina and Micheline Tétrault. Profil : des institutions bancaires coopératives dans le monde. Centre de gestion des coopératives/DID. 1998. Here we obtained issuer-related information for CBs to construct the samples. • Euromoney. Various issues. Also used to obtain issuer-related information for JSBs (and some CBs) to construct the samples. 30 We built two pairs of matched samples of CB and JSB bonds. This follows a wellestablished practice in bond-related research. By “matched sample”, we mean we took for each CB bond in a sample, another bond issued by a JSB that presents similar characteristics to the CB. This yields pairs of samples with an equal number of CB and JSB bonds. The purpose of matching is to eliminate some sources of variation that are considered particularly important. One example is the country of origin. It is well known that RA will not give an institution a higher rating than that allotted to the country of origin. This will also help to control variations due to the macroeconomic conditions at the time of issue. The characteristics used in building the two pairs of samples, the countries represented and the number of pairs obtained are as follows. Statistical analysis The methodology adopted in this study is strongly influenced by the potential validity of results and availability of certain data. One approach to evaluate the fairness of rating would be to establish the relationship that exists between specific financial data of the issuer to the rating. This would have been an interesting approach and by no means superfluous. However, it would have had two weaknesses: i) RAs could ignore results with the legitimate claim that whatever they may be, they mean very little since their assessment is subjective and “forward looking” while we just use historical accounting data; ii) there are considerable difficulties in collecting all financial data for an internationally representative sample of CBs and JSBs over a period of, say, 10 years. A second approach often used to address calibration of ratings is to compare ratings with ex-post default. This approach is not applicable to our situation, the reason being that default rates in the sample of banks studied here (CBs and JSBs) is close to zero! Matched sample “S” §1 Country of residence §2 Size of issuer §3 Date of issue Austria (RaiffeisenV.) Belgium (Bacob; Cr. Prof.) Canada (Desjardins) Finland (Okobank) France (Cr. Agr. ; Cr. Mutuel ; Bque Pop.) Germany (DG Bank, SGZ Bank) Netherlands (Rabobank) 113 pairs Sample Matching criteria Matched sample “R” §1 Country of residence §2 Size of issuer §3 Date of issue §4 Size of issue Austria (RaiffeisenV.) Canada (Desjardins) Finland (Okobank) France (Cr. Agr.; Bque Pop) Germany (DG Bank) Netherlands (Rabobank) Countries/ Associations Sample size (No of pairs) 105 pairs 31 A third approach, the one taken in this study, is to compare consistency between investors and agencies’ perception of default risk of the issues. The first is reflected in the yield demanded on the bonds and the second reflected in the rating. Thus, to test whether rating agencies, when they rate CB bonds do so fairly, we will make the following assumption: even though rating agencies may not be fair (i.e. may fail to calibrate their rating to the true risk of default), investors are. What this means is that, although rating agencies may not incorporate into the rating all relevant differences in the risk exposure of bonds issued by JSBs and CBs, investors on average, are aware of these differences and thus will price bonds accordingly. If investors and RAs agree on their assessment, then the rating would contain all the relevant information to explain the risk premium required by investors on bond yields. If that is not the case, the fact that the bond is issued by a CB or a JSB (represented by a dummy variable) would also have an explanatory power. I explain this in more detail below. Please note also that investors, just as RAs are “forward looking”. Thus, on that account, both measures (required return and rating) are comparable. This comparability underlines the appropriateness of the approach. We performed several tests to check and countercheck results. We found for example that on average, CB bonds are likely to be rated lower than JSB bonds. We also found that investors do not let rating by agencies influence the yield they require from the bonds, at least not after having controlled for readily available issue-specific characteristics. Finally and most importantly, we found that investors require a lower yield for CB bonds than for JSB bonds despite the likelihood that CB bonds may be rated lower than JSB bonds! The last two results taken together imply that while investors value CB bonds higher than JSB bonds, RAs do the opposite! This result is at the heart of our general conclusion that RAs do not rate CB bonds fairly. Let us start with some simple statistics. In the following table, we present the means of ratings and yields for CBs and JSBs. There you can see that in both samples the mean yield is lower for CB bonds than for JSB bonds and that the difference is statistically significant. This means that investors require a lower risk premium for CB bonds than for JSB bonds. The rating is however lower for CB bonds than for JSB bonds. In one sample the difference is significant and in the other, not. Table 1: Key variables means for CB bonds and JSB bonds Sample “R” Variable Yield (y) Rating (Z) Yield (y) Rating (Z) CB bonds 5.84% 0.64 (AA/AA-) 6.11% 0.78 (A+) JSB bonds 6.6% 0.66 (AA/AA-) 6.73% 0.87 (AA-) Stat. Sign. Diff. ( =0.10) Yes No Yes Yes “S” 32 Of course, we didn’t just stop at the computation of some means. I am not going to tire you with the details of the analysis we did. However, you can get a pretty good idea of our testing procedure by looking at the following regression equation: Among the “control variables” we include market-wide data and issue-specific and issuer-specific characteristics. These variables were selected after a review of earlier works on bonds suggested their pertinence. As market-wide data we used: 1. Market-wide risk premium defined as the difference between the yield of a 10-year government bond minus the yield of a 3-month treasury bill. This premium is a proxy of the risk aversion observed in the market and a common measure used in bond and stock yield analysis. 2. Risk-free rate measuring the level of interest rates at the date of issue. For the issue-specific characteristics, we used: 1. (Log of the) size of the issue: its use can be justified on several grounds, as proxy of financial leverage. A larger issue will have, other things being equal, a larger effect of engaging the bank's cash flow than a smaller issue 2. Periodicity of coupons: defined as one over the length of period between coupon payments 3. Duration: a measure of bond volatility 4. Size of the coupon 5. Type of the coupon in the form of a dummy variable that takes a value of 0 for floating coupons and 1 for fixed (or zero) rate coupons 6. Place of issue, a dummy variable that takes the value of 0 for domestic issues and 1 for international issues 7. Options attached to the issue. A dummy variable that takes the value of 0 for ordinary issues and 1 for callable issues 8. Collateral, a dummy variable that takes the value of 0 for subordinated non-collateralized issues and of 1 for issues with collateral Issuer-specific characteristics: Considering our approach, we have focused only on variables that are readily available in the data sources listed above such as country of origin and size of the issuer. You may recall that the matched samples were built around two issuerrelated variables: country of origin and size of the institution. Now, I know that this is not very business-like, but is very “academic-like”. A warning is necessary: failing to find a statistically significant coefficient could mean one of two things: that rating agencies rate fairly (or even ”overrate” in case of a positive and significant coefficient) that they do not rate fairly but investors do not know better. 33 And, with the data available, if the coefficient is not different from zero there is simply no way to distinguish between these two possible conclusions. Fortunately (for us, as researchers but not for CBs and RAs), we do not face this dilemma. Results are consistent with our hypothesis: they do not rate fairly (i.e. the coefficient for the dummy is negative and statistically significant). Below, we have reproduced the same equation where we substituted the symbols with numerical values of the coefficients estimated with the samples “R” and “S”. Below is the coefficient value and in parenthesis, I provide the t-statistic. ignore the rating or that the rating contains no information beyond the information already incorporated in the (publicly available) variables obtained from Datastream. You might say: “but rating is supposed to already incorporate all this issue and issuer-specific information.” Thus, either they or the rating are superfluous in the regression. Indeed, this would be so if we were able to uncover all the relevant information that goes into the rating. The impressive teams RAs work with should mean that they can uncover information that we, as common mortals, just peeking into a commercial database, would not be able to see. But this does not seem to be the case for the sample under consideration. After controlling for the various market, issue and issuer-characteristics, the rating gives no additional information. Now, what do we learn from this regression? A negative and statistical significant coefficient for the dummy confirms that rating agencies do not rate fairly (underrate) but that investors know better. If investors and RAs agree on their assessment, then the rating represented by the variable Zi (the rating) would contain all pertinent information and the coefficient for the dummy variable Di would be zero (meaning not significantly different from zero). The value of this coefficient and its test statistics (the t-value) are thus the central pieces of information that can be extracted from this analysis. Validity of results Let us discuss a little more the comparisons we are making. Numbers suggest that investors disregard the rating by RAs We note that in the sample “R”, the coefficient for the dummy is negative and statistically significant. The only coefficient shown that is statistically significant at the 5% level is the one of the dummy in the sample “R” (3.604>1.96). In the sample “S” the coefficient is still negative and of similar value but not significant. For both samples the coefficient for the rating is non significant. What this means is: 1. Investors require a lower yield on CB bonds than on JSB bonds. 2. Rating has no effect on the yield. The interpretation is either that investors 34 and demand from CB bonds a lower risk premium than from JSB bonds. They would, of course, not do that if they perceived that the default risk on these bonds is higher. No need to go and ask investors in laborious interviews how they assess the default risk of CB and JSB bonds. The proof of the pudding is in the eating, and in this case it means that investors (those guys who put their money on the line) are prepared to pay a higher price for CB bonds than for JSB bonds! Now, there are good reasons to believe that markets (the aggregation of all investors) know better than RAs. Bankissued bonds are held by thousands of investors, big and small, individuals and institutions, unsophisticated and highly sophisticated, and none of them is ready to pay a cent more than what bonds are worth according to risk of losses. And investors, just like RAs, are “forward looking.” RAs do no more than just a technical analysis and do not engage a cent behind their assessment. In fact, as far as the market is concerned, the “tougher” they appear the better (do you remember the argument of their power without accountability?). Thus, in case of doubt they prefer to be tough than soft. It is my contention that RAs do not quite understand CBs and confuse quality with semblance to a JSB, doing the former a great disservice. We cannot exclude the possibility that, although investors “know better”, they are still being influenced “negatively” by the ratings that agencies give CBs. If this is the case, a fairer rating could result in even lower risk premiums (higher bond prices) than what CB bonds are enjoying now. For this reason, however happy we may be to hear that investors seem to assess CB bonds better than RAs, the situation should not be accepted without a fight. Have RAs learned over the years? Numbers suggest that yes. When we perform the same regression over time (i.e. over and over again for sub-samples in a moving time window), we obtain the following result. The spotlight is still on the coefficient for the dummy Di: Figure 1: Recursive regression: Sample “R” As you can see, for the same sample “R”, the 95% confidence interval on both sides of the coefficient value plot remains below zero. This means that for all sub-samples the coefficient remains negative and statistically significant throughout the 1990-2000 period. This is a quite strong result! One can observe however, that the upper confidence band approaches zero progressively. Figure 1: Recursive regression: Sample “S” 35 For the sample “S” the confidence interval starts negative and stabilizes around zero but with the coefficient remaining generally on the negative side. Taken together, these results suggest that RAs have learned and are becoming “fairer”, but an additional learning effort is warranted. Conclusions 1. Due to corporate ownership and governance it is expected that cooperative banks (CBs) assume less risk than joint stock banks (JSB)s. Thus bonds issued by CBs should, other things being equal, display a lower ex-ante default risk than those issued by JSBs. This in turn should be reflected in higher ratings for CB than for JSB bonds. 2. In practice, rating agencies (RA) are likely to assign a similar or lower ratings to CB bonds than to JSB bonds, countering existing governance theories. This may increase the cost of financing to members (already the most disadvantaged social sectors of the economy) and increases instability in the system, both unjustifiably. 3. Investors, however, consistent with theory, perceive that CB bonds represent a lower risk than JSB bonds, demanding a lower risk premium despite the likelihood of a lower rating assigned by an RA (in other words, investors know better than RAs). 4. RA agencies have learned over the years, with ratings becoming “fairer” but that are still biased. One now sees some CBs being rated AAA. 5. The claims by the rating agencies that they rate CBs “very favourably” appear condescending and contradicted by facts. 6. It behoves them that the CB movement should confront RAs about their rating practices, encouraging them to re-examine the frame of reference of CB bond rating. No one is asking RAs to become “unjustifiably biased” in favour of CBs. But if CB bonds present, on average, a lower default risk as every bit of theory and data seem to suggest, then they have the moral obligation to rate them accordingly. Claims of “favourable ratings”, being untrue, should be abandoned. Maybe RAs should, for a moment, relinquish their “power without accountability”, and be prepared to listen. Thank you! References Ronald C. Anderson and Donald R. Fraser. Corporate control, bank risk-taking and the health of the banking industry. Journal of Banking and Finance, 24:1383–1389, 2000. Ray Y. Chou and Richard J. Cebula. Determinants of geographic differentials in the savings and loan failure rate: A heteroschedastic obit estimation. Journal of Financial Services Research, 10:5–25, 1996. Martin Desrocher and Klaus P. Fischer Corporate governance and depository institutions failure: the case of an emerging market economy. CREFA Working Paper 0111. August, 2001. Christoph Kuhner. Financial rating agencies: Are they credible? insights into the reporting incentives of rating agencies in times of enhanced risk. Schmalen-bach Business Review, 53:2–26, January, 2001. 36 INTERNATIONAL CO-OPERATIVE BANKING ASSOCIATION ADDRESSES Mr. Claude BÉLAND PRESIDENT Association Internationale des Banques Coopératives 100, avenue des Commandeurs Lévis (Québec), CANADA G6V 7N5 Telephone : 1.418.835.2444 Fax : 1.418.833.4769 Telex : 051.3533 Mr. Ghislain PARADIS SECRETARY Association Internationale des Banques Coopératives 100, avenue des Commandeurs Lévis (Québec), CANADA G6V 7N5 Telephone : 1.418.835.2444 Fax : 1.418.833.4769 Telex : 051.3533 E-mail : gparadis@did.qc.ca Mr. Étienne PFLIMLIN VICE PRESIDENT AND REGIONAL CHAIRMAN EUROPE Président Confédération Nationale du Crédit Mutuel 88-90, rue Cardinet 75017 Paris, FRANCE Telephone : 33.1.44.01.10.01 Fax : 33.1.44.01.11.63 E-mail :etienne.pflimlin@creditmutuel3d.com Mr. Gideon MURIUKI VICE PRESIDENT AND REGIONAL CHAIRMAN EAST, CENTRAL & SOUTH AFRICA Managing Director Co-operative Bank of Kenya Ltd. P.O. Box 48231 Nairobi, KENYA Telephone : 254.2.249707 Fax : 254.2.227747/336846 Telex : 22938 E-mail : CoopbankMD@form-net.com Mr. Miguel CARDOZO VICE PRESIDENT AND REGIONAL CHAIRMAN LATIN AMERICA Asesor Institucional Cooperativa Nacional de Ahorro y Credito (COFAC) Casa Central, Sarandi 402 C.P. 11000 Montevideo, URUGUAY Telephone : 598.2.916.08.26 Fax : 598.2.916.00.31 E-mail : cofac@multi.com.uy Mr. P.A. KIRIWANDENIYA VICE PRESIDENT AND REGIONAL CHAIRMAN ASIA AND THE PACIFIC Chairman of the Regional Co-op Banking Association For Asia and the Pacific E-4, Defence Colony, 3rd Floor New Delhi-110024, INDIA Telephone : 91.11.469-4989 Fax : 91.11.469-4975/469-4964 E-mail : sanasafe@isplanka.lk Mr. Noureddine OMARY VICE PRESIDENT AND REGIONAL CHAIRMAN NORTH AFRICA AND MIDDLE EAST Président Directeur Général Banque Centrale Populaire 101, boulevard Mohamed Zerktouni B.P. 10 622 21100 Casablanca, MAROC Telephone : 212.22.46.91.64 Fax : 212.22.20.19.32 Telex : Bancepo 21723 - 23078 E-mail : nomary@cpm.co.ma B