Thesis Reference

Thesis

The mexican debt crisis of 1982 Redux: domestic banks, international interbank markets and debt renegotiation

ALVAREZ, Sebastian

Abstract

Cette thèse étudie l'implication des banques mexicaines dans la finance internationale pendant la décennie précédant la crise de la dette de 1982. L'analyse se concentre sur un mécanisme de transmission reliant la banque internationale, l'économie domestique et le financement souverain, à savoir les dépôts interbancaires. Elle montre que l'incorporation à grande échelle des dépôts interbancaires internationaux dans les stratégies de financement des banques commerciales mexicaines a: (a) rendu les banques plus vulnérables aux fluctuations financières mondiales, (b) été à l'origine de risques plus élevés pris par le système bancaire, et (c) contraint la stratégie de négociation du gouvernement mexicain à la suite de son propre défaut. En utilisant à la fois des méthodes historiques et quantitatives et une analyse tant micro que macroéconomique, la thèse propose une nouvelle dimension pour comprendre les origines et le développement de la crise internationale de la dette des années 1980.

Reference

ALVAREZ, Sebastian. The mexican debt crisis of 1982 Redux: domestic banks, international interbank markets and debt renegotiation. Thèse de doctorat : Univ. Genève, 2016, no. SdS 50

URN : urn:nbn:ch:unige-912555 DOI : 10.13097/archive-ouverte/unige:91255

Available at: http://archive-ouverte.unige.ch/unige:91255

Disclaimer: layout of this document may differ from the published version.

1 / 1 The Mexican Debt Crisis of 1982 Redux: Domestic Banks, International Interbank Markets and Debt Renegotiation

THÈSE présentée à la Faculté des sciences de la société de l’Université de Genève par Sebastian Alvarez

sous la direction de Prof. Mary O'Sullivan et Prof. Juan Flores Zendejas

pour l’obtention du grade de Docteur ès sciences de la société mention histoire économique et sociale

Membres du jury de thèse:

Prof. Lucio BACCARO, président du jury, Université de Genève Prof. Juan FLORES ZENDEJAS, co-directeur de thèse, Université de Genève Prof. Mary O'SULLIVAN, co-directrice de thèse, Université de Genève Prof. Ugo PANIZZA, The Graduate Institute, Geneva Prof. Jérôme SGARD, Sciences-Po Paris Prof. Gail TRINER, Rutgers University, USA

Thèse no 50 Genève, 6 octobre 2016

Abstract

This dissertation investigates the involvement of Mexican banks in foreign finance during the decade leading up to the 1982 debt crisis. More specifically, it focuses on a transmission mechanism that links international banking, domestic economy and sovereign finance, which has been overlooked so far: interbank deposits. The dissertation demonstrates that the large-scale incorporation of interbank deposits from international banks into the funding strategies of Mexican commercial banks made: (a) the banks more susceptible to world financial fluctuations, (b) was at the root of greater risks in the banking system and (c) constrained the negotiating strategy of the Mexican state in the wake of its own default. Through a combination of macro and micro-analysis and the use of both historical and quantitative methods, I propose a novel dimension to understand the origins and development of the international debt crisis of the 1980s based on the study of the Mexican banking sector.

------

Résumé

ACKNOWLEDGEMENTS

I am particularly grateful to my advisors Mary A. O'Sullivan and Juan H. Flores Zendejas for careful guidance, engagement and support. My special thanks to Gail Triner and Jérôme Sgard for valuable feedback provided all throughout my Phd. I am also indebted to Lucio Baccaro, Stefano Battilossi, Gustavo del Angel, Marc Flandreau, Carlos Marichal, Aldo Musacchio, Sergio Negrete Cárdenas, Ugo Panizza, Huges Pirottes, Ricardo Solis, Richard Sylla and Eugene White for critical comments and helpful suggestions at different stages of my research. Participants at workshops, seminars and conferences at the University of Sussex, Oesterreichische Nationalbank, Université de Neuchatel, Universidad del Pacífico and Pontíficia Universidad Católica del Perú, University of Pisa, Universidad de Buenos Aires, Université Libre de Bruxèlles, Banamex, University of Geneva, El Colegio de Mexico, Rutgers University, New York University, SSHES/SGWSG, Universidad de Cantabria, Freie Universität Berlin, CLADHE III, Universidad Carlos III de Madrid and Université de Fribourg are also thanked for comments. Finally, I would like to thank archivists at Banamex, Banco de Mexico, Federal Reserve Bank of New York, Bank of England, Bank for International Settlements, International Monetary Fund, World Bank, Lloyds Bank and Columbia University for assistance, as well as Lee Buchheit, Angel Gurría, Carlos Tello and William Rhodes for sharing their insightful experiences.

CONTENTS

Introduction ...... 1 Chapter One. Analytical Framework ...... 3 1.1. The underpinnings of debt crisis...... 4 1.2. The economics and politics of international lending ...... 8 1.3. The management of sovereign debt crisis ...... 11 1.4. A new perspective from the role of Mexican banks ...... 15 Chapter Two. Historical Background ...... 18 2.1. Mexico, commercial banks and the international debt crisis of the 1980s ...... 18 2.2. Euromarkets, international finance and external Indebtedness ...... 22 2.3. Interbank markets, domestic banks and banking fragility ...... 29 2.4. Foreign agencies, debt renegotiation and crisis management ...... 35 2.5. Domestic policies, macroeconomic imbalances and financial crisis ...... 39 2.6. Sources, data and methodology ...... 44 Chapter Three. Internationalization and Condition of Mexican Banks Prior to the 1982 Crisis ...... 48 3.1. Introduction ...... 49 3.2. The historical context of the Mexican banking industry ...... 50 3.3. Incursion into foreign finance, 1972-1976 ...... 53 3.4. From domestic to foreign funding, 1977-1982 ...... 58 3.5. The underlying weaknesses of the banking industry ...... 61 3.6. Inspection at the bank level ...... 64 3.7. An econometric interlude ...... 68 3.8. Conclusion ...... 72 Appendix ...... 74 Chapter Four. International Interbank Markets and Financial Crisis ...... 77 4.1. Introduction ...... 78 4.2. Foreign intermediation by Mexican banks ...... 80 4.3. An interbank market based business model ...... 82 4.4. The risks behind Mexican foreign agencies ...... 85 4.5. The making of a distressed banking system ...... 88 4.6. Banking difficulties in the face of crisis ...... 91 4.7. Interbank funding during debt renegotiations ...... 94 4.8. Conclusion ...... 97

Chapter Five. Domestic Banks and the Mexican External Debt Renegotiating Position ...... 98 5.1. Introduction ...... 99 5.2. The puzzle behind Mexican debt deals ...... 101 5.3. Excluded debt, bargaining tools, and sovereign debt theory ...... 107 5.4. Default and implications for the banking systems ...... 111 5.5. Mexican foreign agencies and the interbank situation ...... 115 5.6. Who bail them out? ...... 119 5.7. Conclusion ...... 124 Final Remarks ...... 126 Insights for the literature on debt crises ...... 128 Towards a new understanding of the crisis...... 130 A novel perspective on bank nationalization ...... 133 References ...... 137 Archives ...... 137 Financial Press ...... 138 Bibliography...... 139

"Put simply, the potential impact of Mexico's crisis is greater than the collapse of the Penn Square Bank and Drysdale Securities in the U.S., the affair of Banco Ambrosiano in Italy and the problems of AEG-Telefunken in West Germany combined"

The Financial Times, August 20, 1982

Introduction

On Friday August 20, 1982 at 10 a.m., Mexican officials met with a group of creditor bankers at the Federal Reserve Bank of New York. There, they announced to the international financial community that they could not meet their external debt obligations. The Mexicans requested the banks grant a moratorium on principal repayments that were coming due on its public sector foreign debt during the next three months, and asked the bankers present at the meeting to assist them in handling these problems and help them secure acceptance from all creditor banks for their refinancing proposal. By that time, Mexico was in conversations with the U.S. government and the monetary authorities of major industrial countries for emergency financial assistance and bridge loans to keep the country afloat. Formal discussions had also started with the International Monetary Fund (IMF), which aimed to reach an economic adjustment agreement that would allow the country to borrow from the Fund.

Mexico's moratorium was the spark that unleashed what has become known as the international debt crisis of the 1980s. Although, as IMF historian James Boughton pointed out, "Mexico was not the first indebted economy to erupt, nor the largest, nor the one with the most serious economic or financial problems, the 1982 Mexican crisis was the one that alerted the IMF and the world to the possibility of a systemic collapse." 1 During the next few months that followed Mexico's announcement, a number of other large international debtor countries that faced similar difficulties, such as Argentina and Brazil, pursued the same path, spreading the crisis regionally and all over the world. Latin America and Africa were the most affected regions, accounting for the lion's share of defaults and debt crises, but numerous developing nations in the Middle East, Asia and Eastern Europe were confronted with debt payment problems as well.

The scale and extent of defaults represented a major shock to the international capital markets and the world economy. For the industrial world, the heavy indebtedness of developing countries and the related exposure of the commercial banking system was a serious threat that placed the stability of the international financial system in jeopardy. As for debtor countries, both the crisis and the way it was managed entailed an enormous financial burden with devastating effects on their economies. Rising unemployment levels were accompanied by steep declines in real income per capita and in standards of living for the majority of the population, causing political and social unrest throughout the region. In Latin America, the financial debacle turned into a major development crisis unprecedented since the Great Depression, with the region suffering its worst recession of the twentieth century, in what has been called the 'lost decade'.

The debt crisis of the 1980s motivated an expanding research on international debt markets and sovereign debt crises that grew out of that event. The literature questioned the fundamentals of sovereign debt markets in the absence of legal enforcement mechanisms, the political and economic underpinning of international lending and external debt, and the ways to manage and resolve debt crises. These questions were first formulated and addressed during the period following the crisis, and today are at the base of an extensive corpus of research in economics and economic history. The experiences of the 1980s have made clear the enormous costs associated with default and debt

1 Boughton (2001, p. 281).

1 crisis, pushing scholars to investigate the factors behind these problems and discuss measures to improve the performance of sovereign debt markets and prevent future crises.

This thesis revisits the dynamic of the international debt crisis of the 1980s and underpins the debt management strategy from the standpoint of debtor countries. It analyses the case of Mexico and from this empirical experience, attempts to draw wider implications for other Latin American countries. In particular, it explores the involvement of Mexican banks in international lending and foreign borrowing and their potential impact on the debt crisis. Until now, the historical literature has remained mostly silent about the Mexican and broader Latin American debt crisis, largely because of a lack of primary sources that have only recently become available. The main purpose of this study is to thoroughly explore the experience of Mexico based upon archival records and derive new insights on the creation of the crisis, as well as on the country's renegotiating position in the aftermath of the moratorium. The 1980s debt crisis was one of the major international financial meltdowns of the twentieth century, and, as such, it deserves to be reconsidered with new historical evidence.

The study is organized in five chapters. Chapter One reviews the theoretical and empirical literature on sovereign debt crisis and sets out the analytical framework. Chapter Two provides the historical and methodological background and discusses the Mexican case and the role of its domestic banks in the broader Latin American and international context. Chapter Three focuses on the international expansion of Mexican banks during the 1970s and the condition of the domestic banking system in the lead-up to the crisis. Chapter Four examines the vulnerabilities the banks created while intermediating foreign finance with domestic borrowers and their financial position in the aftermath of default. Chapter Five analyses how the situation of the banks affected the renegotiating position of Mexican officials with international creditors during the debt rescheduling process. The last section draws overarching conclusions, linking the weakness of the banking sector with sovereign debt-payment problems to provide a new understanding of the crisis.

Chapter One

Analytical Framework

The international debt crisis of the 1980s was the first global financial meltdown of the postwar period. This worldwide wave of defaults and restructurings, along with a new generation of financial crises and external indebtedness problems in emerging market economies during the 1990s and early 2000s, gave rise to a large volume of research on the causes and consequences of sovereign debt crises. The crises were not only limited to debt repayment difficulties: they usually involved severe disruptions to the foreign exchange markets and failures in the banking sector that threatened both troubled countries as well as the international financial system.

The literature has developed a variety of related issues ranging from how to govern sovereign debt markets to how to manage and contain international financial crises in the aftermath of a default. How do these crises develop? Are they only problems for sovereigns, or do they extend to currency and banking problems? If they are connected, what are the transmission channels? How does causality work? In terms of the policy debate, who is responsible for the crisis? Undisciplined borrowing countries, careless lenders, or both? How should the costs of sovereign debt workouts be shared among debtor and creditors? At an international level, can a sovereign debt restructuring process help prevent and resolve crises?

These questions permeate my study. As such, this thesis deals fundamentally with the literature on sovereign debt, crisis management and international finance. It examines the case of Mexico, a major international borrower and the country responsible for unleashing the crisis at the international level, focusing particularly on the country's domestic banking sector. Although largely neglected in the literature, leading domestic commercial banks from debtor countries were generally involved in sovereign debt and international finance, and they were also called upon to manage the crisis after the default. Thus, analyzing the banking sector and the international activities of leading domestic banks helps us understand the depth of the crisis and the complexity of the problems these countries faced.

This chapter reviews the literature on these three main branches of research. The first branch focuses on the economic theory of sovereign debt and its implications for interpreting defaults and debt crises. The second examines the analytical framework of financial crises and the boom and bust cycles that underlie defaults, along with the economics and politics of international lending and capital markets in the lead up to the 1982 debt crisis. The third branch surveys the research that deals with the debt restructuring procedure of the 1980s and international financial crisis management in general. A final discussion places this study at the intersection of these three corpuses of research and argues that looking at Mexican domestic banks not only provides a better understanding of Mexico's debt crisis in 1982, but also allows for new insights into existing literature.

1.1. The underpinnings of debt crisis

The great wave of defaults in developing countries during the 1980s was a windfall for the literature on debt crisis. The scale and extent of the international debt crisis encouraged the development of a growing body of research on the underpinnings of international debt markets and sovereign defaults. Questions such as how sovereign debt can exist in the absence of legal enforcement mechanisms, what motivates sovereigns to repay their debts, and under what conditions countries default, which now lie at the heart of the economic theory of sovereign debt, were first formulated during those years.

Indeed, much of the economic research on sovereign debt is based on a seminal paper by Eaton and Gersovitz (1981). Motivated by the surge of large-scale foreign borrowing by developing countries and increasing episodes of default, the authors question the underlying fundamentals of sovereign debt markets and international lending. As they observe, a distinguishing feature of sovereign finance is that unlike with domestic private debt contracts, there is no institutional framework, analogous bankruptcy law, or direct authority to regulate and enforce international loan repayment. Alternately, they explain that the possibility of exclusion from capital markets and future credit could function as an incentive, enticing international borrowers to maintain repayments on outstanding foreign loans and avoid defaults.

The underlying premise of Eaton and Gertsovitzs' approach is that a government decision to stop servicing its debts entails negative economic consequences. An extensive amount of research has developed on the basis of this idea, adding other reasons why sovereigns are motivated to honor their foreign debt obligations.2 The notion of costs of default became, indeed, the fundamental piece of the analytical and methodological framework of the economic literature on sovereign debt markets.3 In fact, as Panizza, Sturzenegger, and Zettelmeyer (2009) point out, in the classic theory of sovereign debt defaults must be costly for international lending and borrowing to exist.4 Otherwise, it is argued, sovereigns would have strong incentives to default on their foreign obligations and creditors would likewise not be motivated to lend. Hence, both theoretical and empirical work has focused on identifying what the costs of defaults are as well as determining their magnitude.

A first group of papers builds upon the idea of access to capital markets and includes aspects other than exclusion that could concern borrowing countries and push them to repay. Bulow and Rogoff (1989b), for instance, bring up the possibility that a default may affect the reputation of the borrower, downgrading its credit standing, and therefore damaging the terms and financial conditions for future financing.5 In this regard, in his study of sovereign borrowing across three centuries, Tomz (2007) argues that the Argentine government's decision not to default during the 1930s was largely motivated by reputational considerations. The same book analyzes the cost of borrowing over the long term and finds that defaulting countries usually paid higher rates than non- defaulting ones. Similar patterns are observed by Flandreau and Zumer (2004) for the 1880-1914

2 As posed by the literature, the interest is in whether these reasons are sufficient to enforce debt repayment, and thus serve as arguments to explain international borrowing in a context of the limited rights of creditors. 3 See Kaletsky (1985). 4 As Panizza, Sturzenegger, and Zettelmeyer (2009, p. 674) stress "models of sovereign debt distinguish themselves primarily in terms of which cost they emphasize." 5 This idea about reputational concerns is formally developed in Cole and Kehoe (1995, 1998).

4 period, as well as by Sturzenegger and Zettelmeyer (2007) and Borensztein and Panizza (2008) for the 1990s and 2000s.6 Conversely, Eichengreen (1989) finds no evidence of different ability to borrow between defaulting and not-defaulting countries when looking at U.S. lending during the 1945-55 period, and Lindert and Morton (1989) show that defaulters paid no extra premium when borrowing in 1976-79.7

Another line of research has focused on the effects of defaults on international trade and commercial transactions. As first noted by Bulow and Rogoff (1989), a potential reduction of import and export levels, either as a result of trade sanctions, retrenchment of trade credit, or other possibly adverse effects of default can incentivize repayment of sovereign debt. Andrew Rose (2005) shows, for instance, that debt reschedulings during the 1980s were associated with a long-lasting decline of bilateral trade between defaulting and creditor countries. However, Borensztein and Panizza (2010) finds that, although statistically significant and economically sizeable, the effect of default on trade is only transitory. Furthermore, Borensztein and Panizza (2009) are not able to identify the mechanism or channel through which defaults affect trade: direct import sanctions or restrictions, and damage to the creditworthiness of exporters or trade finance.

In addition to the effects on international commercial and financial transactions, a recent strand of literature has looked at the consequence of defaults for the domestic economy, in particular the banking sector. Gennaioli, Martin and Rossi (2014), for instance, extend the previous works on sovereign debt by emphasizing the role of domestic banks and financial markets in reducing the incentives to default. In this respect, Borensztein and Panizza (2009) find that defaults tend to increase the probability of a domestic banking crisis. Moreover, Reinhart and Rogoff (2011) show that there is a close association between banking crisis and default in both advanced and developing countries' histories alike. But according to their study, a banking crisis usually precedes or coincides with a sovereign debt crisis, although the interactions can run either way.8 Overall, although the theory of sovereign debt gives a central role to the notion of cost of default, empirical research has found only ambiguous evidence of their existence and importance.9 In fact, while there is a large consensus that defaults matter and that they entail some costs, at least temporally, their role in deterring a default is much more controversial.

6 Panizza, Sturzenegger, and Zettelmeyer (2009) review the bulk of empirical studies that deal with this issue and conclude that defaults seem to upwardly affect the cost of borrowing in the short term, but that the effect eventually disappears over time. 7 In analyzing capital market access during the 1960s, Cardoso and Dornbusch (1989) and Jorgensen and Sachs (1988) find no evidence that Argentina, faithful repayer in the 1930s, enjoyed better conditions than defaulting Latin American countries. On the other hand, Özler (1993) finds that countries with a default history were charged higher rates for bank lending during the 1970s. Taken together, as Tomz ans Wright (2013) sum up, the literature suggests that defaulters temporarily lose access to the international capital markets and pay higher interest rates for a short period after they resume borrowing. 8 The current European crises are a vivid illustration of this phenomenon: in virtually all troubled countries, sovereign debt payment difficulties are narrowly intertwined with problems in the banking sector (Shambaugh, 2012; Attinasi, Checherita, and Nickel, 2009; CGFS, 2011). See Correa and Sapriza (2014) for a detail discussion on the links between sovereign debt crisis and banking crisis. 9 Along the same lines, another group of papers has analyzed the impact of defaults on output and economic growth but there is no conclusive evidence either (Borensztein and Panizza, 2009; De Paoli, Hoggarth, and Saporta, 2006; Levy-Yeyati and Panizza, 2011; Sturzenegger, 2004)

A corollary of the approach of the classic literature of sovereign debt is that the decision to service external obligations is the result of a cost-benefit analysis: a government will stop reimbursing its debts if it finds that the benefits of not paying are larger than the costs associated with default. In the words of Reinhart and Rogoff, while "a bankrupt corporation may simply not be able to service its debts in full as a going concern, a country defaulter (…) has typically made a strategic decision that (full) repayment is not worth the necessary sacrifice."10 Defaults and debt renegotiations are, therefore, the result of a situation where the borrower makes the decision not to repay based on a consideration of the relative costs and benefits of default, even though it is a complex calculation that may include both economic and financial factors as well as political and social ones.11

A main implication of this analysis is that sovereign debt crises have been narrowly interpreted as a consequence of an unwillingness to pay on the part of the borrower. Reinhart and Rogoff (2009) stress that willingness to pay problems, in contrast to the incapacity or inability to pay, have been the main determinant of country defaults over the last five to six centuries. To make this case for the 1970-2008 period, they look at the burden of external debt at the time of default and find that about half of the defaults or restructurings by middle income countries occurred at levels of external debt that were less than 60 per cent of GDP. They argue that such defaults occurred even "when, under normal circumstances, real interest payments of only a few per cent of income would be required to maintain a constant level of debt relative to GDP, an ability that is usually viewed as an important indicator of sustainability."12 According to them, although as percentage of exports or government revenues, payment would have been much higher, in most of the cases reimbursements could have been worked out.

The 1982 Mexican and Latin American debt crises bring the limitations of this approach into sharp focus. At the time of the defaults, total external debt was about 50 per cent of the GDP in Mexico, and below Reinhart and Rogoff's threshold of 60 per cent for most countries in the region.13 However, as Latin American financial historian Carlos F. Diaz-Alejandro observes, a salient feature of the government repayment decisions during the 1980s was the continued servicing of the public external debt, even with delays and arrears, and the provision of extraordinary facilities to service private external debt, which in some cases was explicitly socialized. This behavior not only shows commitment and determination to pay, but also sharply contrasts with the decision that these same countries made to suspend debt repayment during the 1930s.14 Even when compared to the previous waves of debt crisis in Latin America in the 1820s and 1870s, it seems that during the 1980s, countries made strong efforts to reimburse their debts and avoid outright default.15

Indeed, contrary to the classic theory of sovereign debt, the literature specifically dealing with the international debt crisis of the 1980s has largely adopted a capacity to pay perspective. This research primarily focused on the international and domestic factors that led borrowing countries into situations where they were no longer able to service foreign debt. In this respect, while one group of

10 Reinhart and Rogoff (2009, p. 59). They add that, "since most country defaults happen long before a nation literally runs out of resources, in most instances, with enough pain and suffering, a determined debtor country can usually repay foreign creditors." 11 See Obstfeld and Rogoff (1990) for a detail discussion of willingness to pay and the cost of default. 12 Reinhart and Rogoff (2009, p. 54). 13 See Reinhart and Rogoff (2009, p. 23), Table 2.1. 14 Diaz-Alejandro (1984, p. 356). 15 See Marichal (1989).

6 scholars has interpreted the crises to be mainly the result of policy mistakes and economic mismanagement in debtor countries, another group has stressed the role played by external macroeconomic shocks and circumstances beyond the debtors' control. The first group has mainly focused on macroeconomic imbalances and expansionary economic policies, typically resulting from a lack of fiscal and monetary discipline, which led to large public sector deficits, overvalued exchange rates, and an unsustainable accumulation of debt throughout the 1970s.16 The second group has emphasized exogenous global shocks and the increasing difficulties in the world economy, such as high international interest rates, prolonged recession, trade policies of industrial countries, and the deteriorating terms of trade for developing countries.17

A large body of empirical research investigates the determinants and causes of sovereign debt crisis. The most common approach has been regression analysis between a variable defining a debt crisis and a number of economic, political, and institutional variables. This procedure allows scholars to include proxies for willingness and capacity to repay, as well as liquidity and solvency indicators. Macroeconomic variables reflecting insolvency, such as the ratios of debt to GDP, or debt to total exports, economic growth, and real exchange rate levels, tend to be significant predictors of debt crisis. A worsening liquidity position, as measured by indicators such as the ratio of short-term debt to total debt, and the level of international reserves compared to foreign short-term debt, among others, normally increase the likelihood of a sovereign default. Finally, several studies cite de role of debt history, institutions, and other political factors that capture a country's willingness to pay.18 Such a mix of results can only be reconciled by treating the different arguments as integral parts of a complex phenomenon instead of competing explanations.

Another perspective on debt crisis relies on the idea that in some occasions, debtor countries may have borrowed excessively. That is, as Panizza, Sturzenegger, and Zettelmeyer (2009) explain, a situation where a borrowing country has accumulated an amount of debt that is unsustainable from an ex ante welfare perspective. In contrast to the theories put forth above, in these cases, the debt burden becomes so excessive that even a small external shock or change in the domestic economic conditions could eventually undermine the capacity to pay and trigger a default. For instance, the IMF (2003) estimates the stock of debt that would be sustainable according to the historic performance of a country's primary surpluses, and finds evidence of significant overborrowing in Latin American and other developing regions for the 1985 and 2002 period.19 McKinnon and Pill (1996) stress the fact that a number of countries, such as Uruguay, Argentina, and Chile in the late 1970s, as well as Turkey and Sri Lanka during the early 1980s, suffered episodes of 'boom-time overborrowing' before entering into debt crisis. They name this phenomenon the "overborrowing syndrome" and use it as the analytical framework to study the East Asian crises of 1997-98.

16 See, for instance, Sjaastad (1983) and Wiesner (1985). 17 See Cline (1984, 1985), Dornbusch (1989) and Sachs (1989). 18 See Tomz and Wright (2013) for a survey of the empirical research undertaken by economists, historians, and political scientists on sovereign debt and default. 19 They find that the typical developing country with a default history has overborrowed 2.5-3 times more than non-defaulters, which indeed usually do not overborrow.

1.2. The economics and politics of international lending

The approaches developed in the previous section focus on the incentives and behavior of debtors and creditors, but place more limited attention on systemic factors or global shocks. Conversely, another stream of economic literature has emphasized the role of the supply of external finance rather than domestic factors when evaluating the causes of debt crisis. Sachs (1983), for instance, explains how a debt crisis can be triggered by the fear of a government default and the break out of a liquidity crisis or panic in international lending (self-fulfilling runs on debt).20 A sovereign debt crisis can also erupt as a consequence of a currency crisis provoked by a sudden-stop in capital flows as described in Calvo (1998). Along the same lines, Reinhart and Reinhart (2009) shows that sovereign default episodes are systematically preceded by capital flow bonanzas, which are normally associated with more volatile macroeconomic environment and external accounts.

More broadly, credit supply and external shocks are main elements in the analytical framework of financial crises outlined by Minsky (1972) and Kindleberger (1978). According to these authors, a typical financial crisis follows a boom-and-bust cycle, in which an unstable expansion of credit fuels a bubble (the mania phase of the cycle) that eventually bursts with the emergence of a shock or disrupting event, thus precipitating the crisis (the panic and crash phase of the cycle). Lending is at the heart of their analysis, since they hypothesize that the cycles of manias and crashes result from rapid expansions and sharp contractions of the supply of credit, respectively. Although not directly applied to explain sovereign debt crises, the post mortem 2005 edition of Kindleberger's Manias, Panics, and Crashes points out that the surge in banks loans to Mexico and other developing countries in the 1970s was one of the top ten financial bubbles in the history of financial crises.21 According to him, while in the early 1970s, lending to Mexico was based on a realistic assessment of its credit standing, in the years preceding the crisis, during the mania phase of the expansion, the country increasingly relied on cash from new foreign loans to pay the interest on outstanding foreign debts, and ultimately imploded with the collapse of syndicated bank loans in 1982.22

In fact, other important features that are neglected in the explanations of the previous section relate precisely to the cyclical and global character of sovereign debt crisis. Although isolated episodes have, of course, occurred across countries at different moments in history, an aggregate perspective on defaults and debt restructurings over the long run shows that in many cases, they emerged in waves and affected a large number of countries around the world. Certainly that was the case for international debt crisis of the 1980s, but, as a number of scholars have stressed, similar historical patterns can be observed during the 1820s after the Napoleonic wars, between the 1860s and the 1880s, and in the 1930s during the Great Depression.23 The fact that defaults periodically broke out in many different countries almost simultaneously and under an array of economic, political, and institutional situations points to the presence of a systemic factor that goes beyond the situation of individual lenders and borrowing countries.

20 See also Cole and Kehoe (2000). 21 Kindleberger and Aliber (2005, p. 8). 22 Idem, p. 39, 109 and 243. 23 For an overview of the historical cycles of sovereign defaults on external debt see Reinhart and Rogoff (2009), Chapter 5; and Sturzenegger and Zettelmeyer (2006), Chapter 1.

In line with the Kindleberger-Minsky approach, it is noteworthy that these historical waves of defaults correspond to boom and bust cycles in international lending. In his study of debt crises in Latin American between 1823 and 1930, Carlos Marichal characterizes these crises as loan cycles or recurrent stages of loan boom and debt-default. In the same vein, Suter (1992) builds on Marichal's work and extends the analysis to countries from other regions and over a longer period of time, finding a similar pattern between global debt crises and boom periods in capital exports and external borrowing. He typifies the international debt cycles as made up of three consecutive phases: the expansion of international lending, the outbreak of widespread defaults, and the negotiation of debt settlements. Flores Zendejas (2015) takes a different perspective and underlines the importance of lending booms in sovereign debt crisis by focusing on debtor countries with a record of defaulting that underwent boom-bust cycles without failing to repay. This was the case of Brazil in the 1820s, Mexico in the 1880s, Argentina in the 1870s and 1930s, and Colombia during the 1980s, all of which, he argues, were unexpected non-defaulters. He makes an interesting point: a pattern among these countries, which distinguishes them from defaulters, is that none of them took part in the lending booms that preceded the crises.

The previous discussion begs the question what brought about the international lending booms. Eichengreen (2003) sheds lights on this issue by looking at the periods of large-scale overseas lending that took place during the twentieth century.24 He observes that a similar setting underlies cyclical bursts of international lending and that major episodes of default played a key role.25 These bursts of lending all tend to occur during the upswing of the global business cycle, in periods of expanding international trade, under supportive political conditions, and have been associated with financial innovation.26 As for the boom of the 1973-82 period, he stresses that the economic setting in which the flows to Mexico and developing countries took place was one of increasing growth and declining trade barriers, where world trade expanded nearly twice as fast as world income, and the Eurodollar market grew and syndicated loans emerged as the main international lending device. Regarding the political dimension, although he states that the post 1973 surge of syndicated bank lending was not based upon active encouragement by the U.S. government, other scholars have argued that European governments supported the petrodollar recycling process as explained below.

A number of studies have analyzed supplier behavior and the reasons for the massive expansion in commercial bank lending to developing countries during the 1970s. At an international level, the 1973-74 historical oil price hikes have been identified as a decisive factor that pushed up the supply of credits; large international banks became major depositaries of oil exporting countries' trade surpluses, which increased liquidity in the system and provided banks with new loanable funds (Friedman, 1983; Kane, 1983). At a national level, during the 1970s, the slowdown in the European postwar economic boom and the recession in the U.S. economy diminished the domestic demand for credits in advanced industrial countries and led the banks to look for new lending opportunities abroad (Friedman, 1977; Hayes, 1977). Applying this framework, Beek (1977) and others have stressed the role of developing countries' strong income and export growth rates over the past decades, which were especially extraordinary among the higher income countries in Latin America

24 He identifies four lending booms: 1880 to 1913, the decade following WWI, the years of petrodollar recycling that ended with the Mexican crisis of 1982, and the 1990s. 25 See Eichengreen (1991) for his analysis of international lending and debt-defaults. 26 See Eichengreen (2003, pp. 14-18).

9 and East Asia. These were the main drivers of lending in the context of the banking industry's ongoing process of international expansion (Devlin, 1978; Kapur, 1977).

Within this perspective, a branch of the literature has attempted to explain why banks had over- extended their lending to high-risk developing countries.27 One line of research accounts for excessive lending by emphasizing the role of information, particularly the lack of data and the information asymmetries that existed between debtors and creditors, as well as between banks themselves (Kletzer, 1984). A second line of research has stressed that commercial banks underestimated country exposure and the risk of default, either due to "disaster myopia" (Guttentag and Herring, 1985) or eager competition for market share and profits (Devlin, 1989), by relaxing the lending limits and thereby extended credits above the level they would have normally granted. Finally, a number of scholars have raised the issue of moral hazard and wondered whether bank lending was not further encouraged by the possibility of lender of last resort protections afforded by their home country governments or the IMF in the case things went wrong (Bacha and Diaz- Alejandro, 1982; de Vries, 1983; Vaubel, 1983).28 This idea connects with accounts by political scientists about the politics and institutional foundations of international finance during the 1970s.

In fact, an important element neglected by economists' explanations is the political component behind commercial banks' lending behavior. Yet, as Eric Helleiner and others have stressed, the expansion of international banks and foreign finance during the post Bretton Woods era not only responded to technological development and market forces, but was framed by the rules that were either dictated or implicitly tolerated by advanced industrial states.29 Helleiner (1995) demonstrates that the initial U.S. and British support of the Euromarkets, which was the institutional framework where the banks recycled petrodollars to developing countries, initiated a competitive process of increasing liberalization and deregulation of the financial markets among governments. These policy tools which, he argues, were chosen by policymakers to deal with employment, foreign exchange earnings, and funding for current accounts and fiscal deficits in their national economies, proved to be determining factors in the creation of a new regime for international finance and global debt.

Subsequent contribution in a similar vein highlight the importance of considering government behavior and political choices in defining international financial outcomes generally, and in the surge of international lending and borrowing during the 1970s in particular. Wellons (1987), for instance, analyzes how major industrial countries reacted to the new configuration of global financial and external imbalances caused by the 1973 oil shock and the large balance of trade deficits they were confronted. He argues that the government of Japan, along with France, Germany and the United Kingdom, the three largest European countries, relied on their international credit lines to drive up exports, and also drew on their banks as policy tools to accommodate to the oil shock and improve their trade balances. The rationale behind this argument is that, instead of adjusting their domestic economies, policymakers decided to increase the foreign demand for their products by providing

27 The fact that banks had incurred excessive lending to developing countries was a widespread opinion. See, for instance, Cline (1984), Teeters (1983), Weintraub (1983) and Lessard (1983). 28 Empirical research has tried to find evidence of moral hazard behavior by analyzing the pricing of the loans to developing countries, but no conclusive evidence exists in the literature. See Flores (2012). 29 See, in particular, Helleiner (1996), Frieden (1987), Pauly (1988), Strange (1988), and Abdelal (2007).

10 client countries with the financing they need to grow and import. The commercial banks were especially well-suited to fulfill the role.30

Along those lines, Kapstein (1994) claims that the recycling of petrodollars to developing countries was not due to pure market forces. Instead of acting on their own, banks were supported, and indeed encouraged in their international expansion, by national governments. He shows that a major policy response of OECD governments to the oil shocks was to encourage developing countries to accommodate the price increases by borrowing from banks.31 The inflationary policies of industrial countries would be another example of this, since these policies generated negative real interest rates that made borrowing so attractive. But, as Kapstein acknowledged, officially encouraging developing countries to lend does not necessarily explain bank policy, unless the banks themselves had an interest in pursuing these new lending opportunities. In this regard, Lissakers (1991) argues that governments offered explicit incentives for banks to lend to developing countries through the tax code. She explains how taxation methods in the borrowing country combined with the treatment of those taxes in the U.S. (similar to other industrial countries) allowed the banks to turn the tax credit from their loans into profit. Thus, it acted as a subsidy for bank lending to developing countries that pushed the supply for credits.

At the international level, moreover, recent works have also linked the international lending activities of private commercial banks to the IMF and its role in the international monetary and financial system. Sgard (2016) documents that, as the banks regained access to international liquidity and played a greater role in financing developing countries, the IMF began to meet with them regularly to discuss monetary and debt policy issues.32 Altamura (2015) shows that some voices within the Fund emphasized the need to work with commercial banks in assisting Fund- member countries that faced balance of payment difficulties. He notes that the Fund leaked an increasing amount of confidential information to the banks, which the latter required to better evaluate the situation of borrowing countries and asses lending risks.33 The main implication of these works is that the international financial intermediation by private commercial banks may not only have been supported by governments but facilitated, and even encouraged, by the IMF itself.

1.3. The management of sovereign debt crisis

Similarly to the causes and explanations for the crisis, a great deal of research in economics and political science was also conducted on the way sovereign debt problems were managed in the 1980s. Among economists, the interest was in how to address debtor country repayment difficulties and the process through which sovereign debt crises could be resolved. The main policy challenges were to renegotiate debts with creditor banks, avoid outright defaults, and preserve the stability of the international banking and financial system. As for political scientists, the focus was on international coordination and the political economy of the debt rescheduling arrangements. In

30 Gisselquist (1981) Lipson (1981), Wellons (1987a), Spindler (1984) and Cohen and Basagni (1981) share a similar approach. 31 The inflationary policies of industrial countries generated negative real interest rates of the 1970s that made borrowing so attractive. 32 See also Lipson (1979) and Cohen (1981). 33 Altamura (2015, p. 228).

11 particular, there has been great interest in the question of how such a variety of official and private actors, especially among creditors, managed to find a collective solution for the crises.

In fact, the Mexican default of 1982 represents an important turning point in the treatment of sovereign debt and broader financial crises. Looking to the past, Suter and Stamm (1992) highlight the fundamental difference between the 1980s and previous waves of debt crises in the 1820s, 1870-90s, and the 1930s, and show that the 1980s defaults were largely forestalled by the multilateral debt workout mechanism.34 Looking ahead, many of the debt restructuring techniques and financial instruments developed during that time period became commonly used in the management of subsequent debt crises until the present (Rhodes, 2011). Moreover, Sgard (2002) shows that the international responses and stabilization strategies adopted to handle the surge of financial crises in East Asia and other emerging economies of the 1990s and 2000s were, to a large extent, rooted in the financial architecture and institutional arrangements adopted during the international debt crisis of the 1980s.35

The basic strategy to deal with the Mexican and other developing countries debt crises consisted in rescheduling and conditional lending.36 On one hand, creditor institutions, which included creditor governments central banks, and international organizations, as well as private commercial banks and official credit institutions, engaged in cooperative efforts to provide financial assistance to troubled countries and guarantee that debt service would be maintained. On the other hand, debtor countries approached the IMF to negotiate and take part in an adjustment program and austerity measures, as part of a Stand-By Agreement (SBA) or Extended Fund Facility (EFF). In other words, debtor countries had to commit to adjusting their economies as much as possible, while creditors committed to jointly provide the additional funds to meet the balance of payments needs. The rationale behind this approach was to avoid a major disruption in the reimbursement of external debt, given the large exposure of the banking sector to developing countries debt and to allow for structural changes in debtor countries to restore their ability to repay in the medium and long term.

The active involvement of official creditors was perhaps the most salient feature of the debt management strategy. Sachs (1986) argues that, unlike during the 1930s, the United States came to take on the role of hegemonic power in the 1980s and led an interventionist international effort to tackle the crisis and preserve international financial stability, as discussed by Kindleberger (1973). Krugman's (1994) analyzes U.S. policy on developing country debt and emphasizes the central role of the U.S. government in rescuing Mexico and, more importantly, in setting the international debt workout procedure for dealing with other countries as they entered into debt crisis.37 In the same vein, Marichal (2008) points to the increasing role of the U.S. government in financial negotiations in Latin America in the postwar period, and, in particular, during the debt reschedulings of the 1980s. Eichengreen and Portes (1989) also underline political involvement during the 1980s, and argue that

34 See also Diaz-Alejandro (1987). 35 See also Boughton (2000). 36 A description of the debt strategy and how it evolved between 1982 and 1989 can be found in Cline (1995), James (1996), and Boughton (2001). 37 See also United States General Accounting Office (1997) and Kraft (1984) for an account on the role played by the U.S. government in the management of the Mexican debt crisis of 1982, and Buchheit (1990) for a more general overview of the role played by the U.S. government in sovereign debt negotiations in the 1980s.

12 the reason why governments were less inclined to press for favorable settlements on behalf of private creditors in the 1930s was that their banking systems were not at risk.

The participation and engagement of the IMF in the management of debt crises is also a distinctive element of the 1980s. Boughton (2001) focuses on the role played by the Fund in the debt rescheduling processes and explains how it emerged and consolidated itself into the institution responsible for the creditor's strategy in dealing with troubled countries. Its long experience with balance of payments problems and policy surveillance was instrumental in handling the external liquidity problems of indebted countries (Pauly, 2008). This experience, along with its recognized role in conditional lending, were the foundations of the support received by the U.S. and European creditor governments (Lipson, 1987; Sachs, 1988). Indeed, an agreement with the IMF was a precondition for debtor countries to approach creditor government for financial assistance and commercial banks for rescheduling or refinancing outstanding debt (Mentré, 1984).38 Moreover, from that point on, IMF conditionality and adjustment programs were to become the central piece in the management of future international financial crisis (Boughton, 2000).

The economic rationale behind official involvement and the strategy adopted to deal with the crisis can be found in the liquidity and solvency debate. An important, though not straightforward, distinction that should be made by a policymaker when considering to intervene and financially assist a distress borrower is whether it faces a liquidity shortfall or, if the problem is insolvency.39 The influential work of William Cline (1984) diagnosed debt crisis as founded upon liquidity rather than solvency problems, and thus, recommended that troubled countries be supported until the situation was alleviated. Conversely, this diagnosis implies that debt forgiveness or write-down was not a recommended policy alternative for dealing with developing countries' debt problems (Cline, 1985). With an improvement in the international commercial and financial conditions and some internal economic reforms, debtor countries were expected to be again in a position to repay (Cline, 1983; Feldstein, 1986). Fiscal discipline and adjustment was the necessary counterpart of financial assistance so that creditworthiness and market confidence could be restored.

At an international level, unlike the case of domestic liquidity crises, a major problem is the absence of a central bank or other jurisdictional authority to provide lender of last resort assistance (Guttentag and Herring, 1983). For Wellons (1987b), the debt management strategy of the 1980s was actually a complex device that provided international lender of last resort assistance, though in a diffuse and indirect way. According to him, G-5 governments and central banks, along with their big banks, worked together to allocate responsibilities and furnish liquidity to the international system. Kindleberger (1988) offers a similar interpretation and underlines the function of the strategy in "providing lender of last resort support for these sovereign-risk loans on top of the marked steadiness of money-market banks in continuing to lend to these countries."40 Paul Volcker, president of the U.S. Federal Reserve at the time of the crisis, explicitly defined the strategy in these

38 Although the Fund and industrial countries governments had a long history of working together in the context of the Paris Club reschedulings (Hardy, 1982), this relationship became more intense during the 1980s (Trichet, 1989; Noyer, 1994). 39 See Eaton, Gersovitz, and Stiglitz (1986) and Solberg (1992) for a discussion about liquidity vs. solvency debt crisis of a sovereign borrower. 40 Kindleberger (1988, p. 116).

13 terms and stressed that on an international scale, they were all collectively providing lender of last resort coverage.41

The alternative interpretation is that while it rescued debtor countries, the debt management strategy was also an indirect bail-out for the banks. Having a multilateral organization and public resources bail out private agents was, however, a potential institutional and political problem. This is one of the reasons why the IMF and the governments promoted a burden-sharing approach in which the banks were "bailed-in," as framed in Roubini and Setser (2004). All creditor banks contributed and did their share. To address the sovereign debt crisis, banks were expected to take part in a multilateral strategy: rescheduling outstanding debt, stretching principal payments, and extending new money facilities conditioned on IMF agreements with debtor countries. In particular, the provision of new fresh money by the banks became a central component of the refinancing programs with debtor countries (Bergsten, Cline, and Williamson, 1985; Cline, 1987).

A large body of research developed among economists and political scientists about the complex organizational challenges and conflicts of interest involved in the management of the sovereign debt crisis. The central question they address was how differences among creditors were worked out and how the restructuring procedures would be enforced. The nature of the collective action problem was twofold. On one hand, securing the participation of the banks as a group in a concerted strategy was challenging because, although it was in their interest to avoid defaults, the individual incentive was to reduce exposure and to halt lending to troubled countries (Cline, 1983). Banks could thus free ride on the IMF and creditor governments' lending in seeking reimbursement for their claims. On the other hand, banking itself was a large international sector with heterogeneous institutions facing different stakes and divergent interests regarding the debt crisis.42 Thus, cooperation among banks themselves was inherently difficult and often strained with individual opportunism proving to be a major problem (Lipson, 1985).43

Jérôme Sgard (2016) argues that enforcing the collective debt management strategy of the 1980s relied on a decision-making procedure based on the mutual veto power of the parties involved in the negotiations. According to him, the key element behind commercial bank cooperation was the not- lending-into-arrears policy of the Fund.44 Rather, what finally cajoled the banks into collective arrangements was the decision of the Managing Director to require a formal commitment on new loans. Had the banks failed to commit their share, he wouldn't have sent the SBA or EFF to the Board for approval and debtor countries would have had little choice but to default. In contrast, by refusing to lend, banks could veto an agreement between the IMF and a country if, for example, the adjustment program was believed to be insufficient to guarantee debt service repayments. Debtor countries —the third party involved— had to agree to both arrangements.

41 See Volcker and Gyohten (1992, p. 203). 42 This was typically the case of small and regional banks that were only marginally exposed to the crisis and thus had the incentive to take the losses instead of increasing exposure. See, for instance, Gibbs (1984). 43 Formal analyses of the free rider problem and disruptive behavior by commercial banks during debt renegotiations can be found in Sachs (1983), Krugman (1985), Caskey (1989), Fernandez and Kaaret (1992). 44 The rationale behind was that "any assumption regarding arrears, hence debt more generally, inevitably affected how the ‘financing gap’ of the country (i.e. current balance plus amortizations) would be covered over the duration of the SBA. And this arrangement had to be stated ex ante in the Letter of Intent, if the Executive Board was to extend a stand-by loan." (Sgard, 2016, pp. 110–111).

Creditor country governments and financial regulators also played an important role in the enforcement of the collective strategy. Eichengreen (2003b) and others have pointed out that regulatory incentives were used to leverage the power of the IMF and persuade the banks to reach agreements on refinancing troubled countries. In the case of United States, for instance, the President of the Federal Reserve, Paul Volcker, publicly announced that banks who agreed to the Fund's demands and provided new loans that facilitated the adjustment process of debtor countries, helping them to service their international debt, would not be subject to 'supervisory criticism'.45 Financial authorities from other creditor countries also showed flexibility in the application of banking regulations and accounting policies: they allowed the banks not to register losses on their international loans or reveal the possible insolvency behind their balance sheets (Lipson, 1985). And, in cases where such incentives proved to be insufficient to induce additional lending, policymakers applied moral suasion or more direct pressure on banks through telex or direct phone calls.46

Finally, an important issue in this literature is the coordination among banks. Commercial banks organized themselves into Bank Advisory Committees (BAC) that assembled representatives of the major 10 to 15 international banks, with the mandate to act on behalf of 500 to 1,000 creditor banks during renegotiation with debtor countries (Lomax, 1987; Rieffel, 2003). Within this institutional arrangement, large money-center banks managed to exert pressure and coerce small and regional to participate in the rescheduling deals. They could do so given their market power and the reliance of regional banks on interbank funding, as well as other services like correspondent banking and loan syndication provided by the larger banks (Lipson, 1985). Information about debtor countries was largely centralized and controlled by the money-center banks, which made smaller banks dependent on them for dealing with foreign debt issues (Gibbs, 1984). Still government officials, federal regulators, and central bank authorities sometimes had to step in when large banks proved unable to coerce the smaller ones in certain cases (Stallings, 1983).

1.4. A new perspective from the role of Mexican banks

It should be clear by now that since sovereign debt crises are a multidimensional phenomenon, there is no single way of approaching them. The discussion in the previous sections suggests that a thorough account of a debt crisis should take into consideration the systemic and cyclical forces underlying the functioning of world capital markets, the economics and politics of international borrowing and lending, as well as the domestic and external factors behind these financial transactions. In addition, although there are common features and patterns among crises and defaults across time and space, the historical circumstances and country specificities play a significant role in the dynamics and resolution of each particular case.

Standing on the shoulders of these extensive bodies of research I have just surveyed, as well as on additional historiographical work and primary sources, this thesis revisits the 1982 Mexican debt crisis. From an analytical point of view, the study is framed by the questions and the analytic tools discussed above, with the aim of providing a more complete picture of the event in question. It does so by focusing on the role and the activities of a sector that has been thus far neglected in the

45 See Kraft (1984, p. 49). 46 See, for example, James (1996, pp. 381-385) for debt negotiations with Brazil in 1982-83, Boughton (2001, p. 301) for the management of the interbank run against Mexican banks in September 1982, and Aggarwal (1996, p. 366) for the Mexican refinancing plan of 1987.

15 accounts of the crisis: domestic Mexican banks. Because the Mexican commercial banks participated in international borrowing and sovereign lending, they became directly involved in the creation of the crisis and in external debt renegotiations. The analysis developed in this dissertation lies at the intersection of the literature on sovereign debt, international finance, and crisis management with a twofold aim: to shed new light both on our understanding of the crisis and on the interpretations advanced in previous works.

The importance of the domestic banking sector in the study of debt crisis has been stressed in the recent economic literature on sovereign debt and financial crisis. However, despite the fact that Mexico, along with other debtor countries, had undergone serious problems in the lead up to the default, the role of domestic banks has been largely neglected in the explanation of 1980s debt crisis. Moreover, as the current European debt crisis has shown, the banking sector and domestic bank activities in the international capital markets can play a central role in the debt repayment difficulties faced by the sovereign. This case study of Mexico supports the connection between weak domestic banking sectors and sovereign defaults found by Diaz-Alejandro (1985) in the 1983 Chilean crisis, and demonstrates that, similar to Europe today, the international activities of leading domestic banks were at the heart of the storm.

The involvement of Mexican banks in the international capital markets brings an interesting element into the discussion between supply-oriented and demand-oriented explanations of sovereign debt crises. On one hand, together with the Federal government, the public agencies, and the non- banking private sector, as borrowers in the international wholesale markets, Mexican commercial banks were part of the country's demand for foreign capital. On the other hand, because they relent these funds back home or to final borrowers in other developing countries, they were also on the supply side, as providers of syndicated or direct foreign loans. Mexican banks, therefore, likely exacerbated the dynamic of external debt accumulation and over-lending to Mexico, by simultaneously pushing both the demand and supply of credit upwards. This shows how difficult, and indeed inappropriate, it can be to consider lender or borrower incentives and behavior as a macroeconomic aggregate without differentiating between actors. Along the same lines, explanations of debt crisis cannot be based on one of the two sides of the market as if they were independent from each other.

This study provides also new insight on the international political economy of foreign finance during the decade leading up to the 1982 debt crisis. In looking for the politics behind international bank lending and the petrodollar recycling process, political scientists have largely focused on the role played by the governments of advanced industrial countries and how they accommodated their goals and interests. The fact that leading commercial banks from Mexico, as well as from other major borrowing countries, were based in London and New York and worked along with major U.S. and European banks, brings the governments of developing countries into the story. The internationalization of Mexican banks could hardly be accomplished, let alone started, without the consent and support of the Federal government and financial regulators. As in the case of industrial countries, the international expansion of Mexican banks responded to market forces and individual bank behavior, but it was also framed by domestic economic policy objectives of the national government and financial authorities.

The political economy of Mexican banks internationalization had implications that go beyond the national level. As Helleiner and others have demonstrated, the rules that governed the world capital markets in the post Bretton Woods era were directly or implicitly defined by the policy choices of advanced industrial states. Although a large economy in the developing world, Mexico was a small international financial player and thus had only limited influence in the establishment of global financial rules. Instead, Mexican policymakers faced a predetermined international setting and external financial forces that had to be accommodated into national economic and political objectives. This doesn't mean, however, that the financial policy and regulatory choices adopted by the Mexican government did not affect international capital markets. On the contrary, the destabilizing behavior that Mexican banks caused in the foreign capital markets shows that in international finance and global debt, it is important to take the debtor-country side of the story into account as well.

Finally, bringing Mexican banks into the analysis contributes to the literature on sovereign debt crisis management. Leaving domestic banks out of the story implies that the banking sector situation at the time of the crisis was not important in terms of the debt management strategy. On the contrary, because of their involvement with foreign finance, the debt rescheduling procedure set up in the wake of the Mexican crisis had to handle both the repayment problems of sovereigns as well as the external indebtedness and exposure of leading domestic banks. In fact, Mexican banks faced a serious conflict of interests in this respect and were in an awkward position during debt negotiations. On one hand, they sat on the debtor side of the table along with the Mexican government to renegotiate its external obligations. On the other hand, they were also on the side of creditors, along with the rest of the international commercial banks that had outstanding claims on Mexican public and private borrowers.

This study also contributes to the literature on the enforcement of the collective debt management strategy of the 1980s. While much of the discussion has focused on collective action problems among creditors and how these disputes were resolved, the way that debtor countries were dragged into the debt rescheduling process has received much less attention. This is surprising, given that managing of the crisis proved to be extremely costly for debtor countries, and that successfully avoiding a collapse of the international banking and financial system was largely dependent on forcing debtors to participate in the creditors' collective strategy. This study argues that a main instrument the creditors used to coerce Mexico into accepting their renegotiating demands was the reliance on foreign capital by Mexican policymakers to keep domestic banks alive.

Chapter Two

Historical Background

2.1. Mexico, commercial banks and the international debt crisis of the 1980s

Although the financial crisis of the 1980s was global in scale, Latin America was the epicenter, and Mexico was the country in the eye of the storm. The sudden end to international lending that followed the outbreak of the Mexican crisis in August 1982 caused other heavily indebted Latin American countries to confront debt-payment problems as well. Brazil and Argentina, which along with Mexico were the largest economies in the region, approached their international creditors during the last quarter of 1982 and began to negotiate refinancing plans.47 One by one, developing countries entered into multilateral debt renegotiations.48 By the beginning of 1983, virtually all Latin American countries, with the exception of Colombia, were in discussions with the IMF, developed countries' governments, and private creditor banks to negotiate adjustment programs and rescheduling plans.49 By November of that year, the IMF reported that 20 cases of debt restructuring by developing countries had been completed and 7 were still under negotiation: Latin America and the Caribbean accounted for 12 and 4 of these, respectively.50

Figure 1 shows the spread and extent of external debt payment problems in the developing world during the following years. The number of countries in default on foreign currency bonds or bank debt went from 25 in 1982 to 39 in 1983 and over 43 in 1986. Latin America and Africa were the most affected regions, accounting for about 85 per cent of defaults in the period, with developing countries in the Middle East, Asia and Eastern Europe making up the remainder.51 In Africa, defaulting countries spent, on average, 5.53 years in default between 1982 and 1989, while Latin American countries spent 6.83 years in default during the same period. The great number of countries that signed IMF-agreements and enacted adjustment programs is also an indicator of the scale of the problem. Almost all defaulting countries subscribed to an IMF agreement at some point since creditor banks required adherence to these agreements as a necessary condition for debt restructuring.52 In Latin America, major economies like Mexico, Brazil and Argentina were in default

47 For an account on the origins of the crisis in Argentina and Brazil, see Boughton (2001, pp. 327-345). 48 External debt payment problems were not new in Latin America. Peru and Bolivia, among others, had already undergone debt restructurings prior to 1982. See Hardy (1982). 49 See Altimir and Devlin (1992) and Devlin (1989, pp. 181-196). 50 IMF SM/83/227, Table 9, p. 38, and Table 1 in the Appendix. 51 For a historical perspective on the number and importance of defaults in thes 1980s, see Reinhart and Rogoff (2009), Chapter 5, in particular Figure 5.1, p. 72. 52 There were cases of countries that adopted IMF programs even though they never defaulted, such as China, Korea and Portugal.

18 during almost the entire period and, indeed, signed IMF or debt rescheduling agreements several times during those eight years.53

Figure 1. Default in developing countries, 1982-1989

Source: Standard and Poor's (2002). Latin America was relevant not only because of the scope and intensity of its crises, but also because the bulk of international debt was concentrated there. Table 1 shows the level of total outstanding external debt for the largest borrowing countries of the developing world: Brazil, Mexico, Argentina and Venezuela were the top four debtors in 1982. Moreover, as can be clearly seen, their foreign indebtedness levels were considerably higher than for the largest developing debtor countries in Eastern Europe, Africa, Asia and the Middle East. For instance, Brazil's external debt was about 3.5 times higher than Egypt's, the largest non-Latin American defaulting country. In comparison with African countries, the other region that was most impacted by debt crises, the Brazilian debt represented 7.6 times that of Morocco, Africa's largest defaulter.

Even more important than debt levels is the fact that within the developing world, Latin America as a region held the majority of commercial bank claims. In June 1982, before the outbreak of the crisis, commercial banks from the Group of Ten (G-10) and Switzerland reported to the Bank for International Settlements (BIS) total claims on Latin American countries to be US$ 191.5 billion, an amount representing as much as 58.7 per cent of their assets with developing countries (see Table 1). Mexico was the country where banks' exposure was the largest, and along with Brazil, Venezuela and Argentina, accounted for almost 85 per cent of their Latin American assets. Table 1 also makes evident the extent that Africa represented a much lesser problem than Latin America for creditor banks in this respect. Although largely affected by defaults, this region represented only 8.7 per cent of the assets of developed countries' banks. After Latin America, the regions where banks' exposure was the highest were Eastern Europe and Asia, but the shares were considerably lower and defaults and debt crises were much less frequent.

53 For a review of the renegotiations round of Latin American countries, see Altimir and Devlin (1994), Devlin and Ffrench-Davis (1995) and Devlin (1989, pp. 183-190).

Developing countries with large indebtedness levels on the brink of the crisis were the result of a vigorous borrowing-lending process that had been established during the previous decade. Table 1 shows that between 1972 and 1982, heavily indebted defaulting countries increased their external debt, on average, at annual rates ranging from 11.7 to 31.8 per cent. Foreign financing was mainly conducted through syndicated or direct international loans granted by private commercial banks operating in the Euromarkets. After the oil shock of 1973, the Euromarkets became the dominant institutional mechanism for recycling the oil-exporting countries' large revenues, which were deposited in the international banking system, to the private and public sectors of borrowing countries.54 As a result of this process, commercial banks evolved into the most important source of international financing and the main creditors of developing countries, surpassing the prior predominant positions of international organizations and governments from the industrial world.55

Because of its large oil wealth, Mexico became a preferred destination for international lenders and, along with other countries in the region, it attracted the lion's share of Eurolending. By 1981, Latin America had absorbed almost two-thirds of the loans extended to the developing world and U.S. banks took a prominent role as main suppliers of portfolio flows to the region.56 For Mexican borrowers, and particularly the public sector, Eurocredits proved to be a better source of funds than Eurobonds and other financial instruments available in the international capital markets. As Negrete notes, as early as 1974 and "in just six months, with two syndicated loans, Mexico had borrowed virtually the same nominal amount accumulated through bond offerings in the 1963-72 decade."57 By 1982, bank lending represented about 90 per cent of Mexico's total outstanding liabilities to non- official creditors, while the balance consisted of publicly-issued bonds that were in the hands of individual investors and other credit facilities from private non-banking institutions.58

However, international lending to developing countries through the Euromarkets was neither limited nor even exclusive to banks from industrial countries. Since the early 1970s, commercial banks from borrowing countries had been expanding their international activities and becoming involved in both the Euromarkets and the petrodollar recycling process. Countries such as Mexico and Brazil were both major borrowers and lenders in the world capital markets, and in fact, promoted their government-controlled domestic banks or private sector banks to reach the world capital markets.59 While it has been evoked by national historiographies, especially by Latin American scholars as will be shown later, the participation of debtors' banks in Eurolending has been given very little attention.60 This issue is largely absent in the extensive corpus of literature on international finance and global debt during the lead up to the crisis of the 1980s.

54 See Wellons (1977). 55 Moffitt (1984). 56 See Devlin (1987) and Stallings (1987, pp. 94-104). 57 Negrete (1999, p. 154). It was through Eurocurrency syndicated lending that developing countries reached the Euromarkets, because, after all, "LDCs rarely qualifi[ed] for access to the Eurobond and foreign bond markets" and therefore had "only limited access to [them]." "Eurobond Survey," The Banker, September 1977, p. 75 and p. 87, respectively. 58 World Debt Tables, 1984-1985 Edition. 59 "Consortium banks on course," The Banker, February 1976, pp. 170-171. 60 For an exception, see Quijano (1987, pp. 241-258), Chapter 6.

Table 1. External borrowing of the largest developing debtor countries by region

Total External debt Total Ext. Bank Debt*

Years in Default 1972 1982 June 1982 Annual growth 1982-89 US$ Mil. US$ Mil. rate US$ Mil. Share

Latin America 191,490 58.7% Brazil 11,864 94,429 23.1% 1983-89 50,460 15.5% Mexico 8,352 86,275 26.3% 1982-89 62,405 19.1% Argentina 6,894 43,787 20.3% 1982-89 23,627 7.2% Venezuela 2,614 32,182 28.5% 1983-88 22,805 7.0% Chile 2,963 13,959 16.8% 1983-89 10,888 3.3% Peru 3,585 10,871 11.7% 1983-89 5,134 1.6% Colombia 2,965 10,520 13.5% 5,002 1.5% Ecuador 538 7,808 30.7% 1982-89 4,343 1.3% Uruguay 440 1,907 15.8% 1983-85, 1987 1,045 0.3% Eastern Europe 43,311 13.3% Yugoslavia 3,438 16,077 16.7% 1983-89 9,243 2.8% Hungarya n.a. 6,739 6,777 2.1% Greecea 1,339 6,719 17.5% 8,795 2.7% Poland n.a. n.a. 1982-89 13,643 4.2% Romania n.a. n.a. 1982-83, 1986 4,375 1.3% Asia 40,522 12.4% India 10,029 27,810 10.7% 1,341 0.4% Indonesia 5,863 25,133 15.7% 4,963 1.5% The Philippines 2,671 24,413 24.8% 1983-89 8,125 2.5% Korea 3,088 21,499 21.4% 16,591 5.1% Malaysia 940 13,354 30.4% 3,777 1.2% Thailand 1,229 12,235 25.8% 2,826 0.9% Pakistan 4,055 11,527 11.0% 759 0.2% Africa 28,527 8.7% Algeria 1,550 17,639 27.5% 6,465 2.0% Morocco 1,186 12,401 26.5% 1983, 1986-89 3,352 1.0% Nigeria 1,082 11,992 27.2% 1982-89 5,732 1.8% Ivory Coast 580 8,961 31.5% 1983-89 2,929 0.9% Sudan 452 7,169 31.8% 1982-89 959 0.3% Tanzania 1,396 6,130 15.9% 1984-89 257 0.1% Zairea 573 4,049 21.6% 1982-89 984 0.3% Tunisia 753 3,777 17.5% 1982 933 0.3% Middle East 22,245 6.8% Egypt 1,952 27,323 30.2% 1984 4,726 1.4% Turkey 3,555 19,716 18.7% 1982 2,912 0.9% Israela 3,585 14,900 15.3% 1989 5,832 1.8% Syriaa 337 2,616 22.8% 595 0.2% Jordana 171 1,685 25.7% 1989 516 0.2%

* Total bank claims of BIS reporting countries with developing countries as classified by the World Bank (offshore centers are excluded). a Total External debt includes only public / publicly guaranteed external debt from World Bank's World Debt Tables. Source: World Bank's World Development Indicators and Standard and Poor's (2002).

This study fills that gap by exploring the involvement of Mexican banks in foreign finance and their potential role in the 1982 debt crisis. Mexican banks first reached the international capital markets between 1972 and 1974, through the creation of London-based consortium banks in partnership with banks from developed and developing countries. They engaged more deeply in international businesses during the following years, by opening their own branches and agencies in London and the U.S., the major world financial centers at that time. Through their associated consortium banks and their networks of foreign offices, Mexican banks took on an important role in international lending to both Mexico and other Latin American countries.

The thesis investigates three different aspects of the Mexican bank internationalization process. Chapter Three explores the factors driving the international expansion of Mexican commercial banks, as well as its repercussions for the health and condition of the domestic banking system. In Chapter Four, I reconstruct the business model behind the banks' international activities, and examine the risks and vulnerabilities it created by intermediating foreign liquidity for final borrowers in Mexico. Finally, Chapter Five analyses the impact of the exposure of the domestic banking system to the 1982 debt crisis and, specifically, how the banks' situation affected the renegotiating position of Mexican officials with their international creditors. The general argument is that domestic commercial banks, through their involvement in international finance and Eurocurrency markets, contributed to substantial instability in the Mexican economy that made it extremely vulnerable to internal dynamics and external shocks.

2.2. Euromarkets, international finance and external Indebtedness

The 1980s international debt crisis was not an isolated event, but rather part of a chain of recurrent surges in foreign lending followed by defaults throughout financial history. During the previous two centuries, waves of sovereign debt crises occurred from the 1820s through the late 1840s during the aftermath of the Napoleonic Wars, over the early 1870s up to the 1890s, and in the Great Depression of the 1930s that extended until the early 1950s.61 Latin America was part of this historical trend, with newly-independent states defaulting on their foreign debts during the 1820s, and a succession of new lending booms followed by debt crises in 1873, 1890, and 1931. Large-scale inflows of foreign capital were integral to these boom and bust cycles.62

The expansive phase of the 1982 debt cycle was founded upon the enthusiastic wave of foreign bank loans to developing countries in previous years. After several decades of operations largely concentrated on retail banking inside national boundaries, U.S. and European banks started to develop businesses abroad and to expand their international financial activities during the late 1960s and early 1970s.63 The internationalization of the banking industry and the re-opening of the international capital markets that followed the end of Bretton Woods were accompanied by an increasing penetration in the developing world and a boom of cross-border bank lending. Total outstanding bank claims on developing countries rose from approximately US$ 22.3 billion to US$ 140.4 billion between 1973 and 1979, representing an average annual growth of 35.8 per cent.

61 See Reinhart and Rogoff (2009), Chapter 5. 62 See Marichal (1989) for an historical account of lending booms and debt crisis in Latin America. 63 Altamura (2015) shows the profound transformation and internal reorganization undertaken by British and French banks during those years to be able to carry out international activities.

Annual growth of outstanding claims then began to slow, and was 24 per cent in the 1979 to 1980 period, 18 per cent in 1981, and only 7 per cent in 1982.64

The Euromarkets were the basic platform used by private commercial banks to build up their international lending activities. What had originated as a pool of dollars held outside the U.S. banking system during the postwar period, was then expanded by European countries, primarily in London, during the 1960s and 1970s, and became a much larger and active market of dollar- denominated foreign currency deposits and Eurocurrency operations.65 In the beginning, these operations consisted essentially of placing or borrowing funds in the Eurocurrency interbank markets, but the banks would later enlarge their business through the creation of new instruments, notably the Eurobonds and Euroloans, intended to finance non-bank customers.66 Particularly important was the syndicated Euroloan market, whose access was largely restricted to all but the most creditworthy borrowers prior to the oil shock in 1973, but went on to become the main lending instrument for public and private sector borrowers from developing countries.67

The historical rise in oil prices was a decisive factor in the evolution of foreign lending to developing countries in the lead-up to the 1982 debt crisis. Eurocurrency deposits, which had grown almost three-fold over the 1970-1973 period, became the largest single depository for the substantial trade surplus of oil exporting countries from 1974 on.68 As large amounts of U.S. dollar liquidity streamed into international private banks in London, they became available to the rest of the banking system through the Eurocurrency wholesale interbank markets, and thus provided the banks with considerable new loanable funds.69 A mechanism, known as the petrodollar recycling process, was set into motion, where dollars flowing to OPEC countries as result of the increase in oil exports were recycled and flowed back to the rest of the world.70 Within the international banking and financial system, while the petrodollar surpluses boosted the syndicated Eurocredit market, they also overcome the Eurobonds and other traditional types of private finance as source of financing.

The counterpart to the rush of international bank lending was a large-scale demand for external finance. At the aggregate level, the increasing surpluses accumulated by oil exporting countries were, after all, a mirror image of the deteriorating current account of oil importers. This included both industrial countries, which consumed the largest share of global energy production, and developing economies, which had structurally negative trade balances and were dependent on imported oil for growth.71 Within this context, the international banking sector had the resources

64 For summary descriptions of developing country indebtedness and growth see Aronson (1979), IMF (1981), and World Bank (1981). 65 An important stimulus to the development of the Euromarkets and London as its main marketplace was the introduction of the "interest equalization tax" (IET) by the U.S. administration in 1963. See Aliber (2000). 66 Together the Eurodollar, Eurobond and Euroloan markets set the foundations of international finance and global debt during the decade leading up to the outbreak of the crisis of 1982. See Schenk (1998) for an account of the origins of the Euromarkets in London. 67 A loan syndicate can be defined as a group of financial institutions, formed by a manager or a group of co- managers, which jointly lends funds on common conditions to a borrower. See Wionczek (1979) for an analysis of developing countries external indebtedness and the Euromarkets. 68 See Kane (1983, pp. 110-111) and Roberts (2001, p. 94), Table 6.2. 69 For an explanation on the mechanics of the Eurodollar market see Giddy (1994, pp. 35-37). 70 The second oil shock of 1979 gave more impetus to the process, as once again banks became major depositories of OPEC funds. 71 See World Bank (1981, pp. 49-63).

23 and was particularly well-situated to intermediate financial surpluses and deficits between countries on a worldwide scale. To the extent that the private banking institutions allocated international liquidity to countries where foreign capital was needed, they helped accommodate the shift of global external imbalances that the rise in energy costs provoked in the 1970s.72

The reconfiguration of world trade flows and balances was accompanied by a change in the balance of payment financing patterns in the developing world. In contrast to industrial countries, which mainly attempted to adjust oil-related deficits through an expansion of exports,73 developing countries increasingly relied on external borrowing based on recycled petrodollars, primarily bank lending, to bridge the wider financial gap.74 However, there were important differences in the development of foreign indebtedness across regional groupings of developing countries, with middle-income economies borrowing relatively more from private financial markets than their low- income counterparts. In Africa, for instance, where external debt expanded more rapidly than in any other region, foreign borrowing relied almost exclusively on official sources of funds.75

Therefore, the escalation of bank indebtedness was a predominantly middle-income economy problem. In fact, according to the World Bank, "commercial bank lending to developing countries was concentrated almost totally in the middle-income countries."76 Among the largest borrowers in middle-income countries, were oil exporters, such as Mexico, Venezuela and Algeria, and upper middle-income countries like Brazil, Spain, Argentina, Yugoslavia and South Korea. By the beginning of the 1980s, these eight countries accounted for about two thirds of total outstanding bank debt.77 Most of these countries had participated in the postwar economic boom, and were perceived to be relatively prosperous emerging economies at the time of the oil shock, qualifying for bank loans according to the usual country risk criteria.78 In contrast, access to the Euroloan market by lower- income economies was much more limited and they could only borrow meager amounts from international private commercial banks.

The increased participation of middle-income countries in bank lending was accompanied by a growing concentration on Latin American borrowers. The period from the end of World War II to 1980, as Bértola and Ocampo (2012) states, "was marked by the highest economic growth rates [that had been] attained by Latin America in its entire history."79 Economic performance during the 1966-1973 period was particularly outstanding, with large Latin American countries displaying not only strong economic growth, but also improved purchasing power of their exports, along with

72 See Cohen and Basagni (1981) for an analysis of the role of international banks in the financing of balance of payment deficits. 73 See Wellons (1987). 74 Other channels that petrodollars were recycled through were direct aid by oil producers and remittances from migrant workers in the Middle East. Additionally, two oil facilities were arranged by the IMF to channel surpluses accumulated by OPEC to developing countries in 1974 and 1975. See World Bank (1981, pp. 50-53). 75 Overall, between 1973 and 1982, official creditors accounted for as much as 85% of external financing of low income oil-importing countries and about 25% in the case of higher income countries; the respective balances of 15 and 75% corresponded to private financial institutions and other private creditors. See Sachs (1982, p. 13), Table 4. 76 World Bank (1981), Chapter 5. 77 Ibid, p. 52. 78 See Friedman (1983) for an account of country risk practices by international banks during the 1970s and 1980s. 79 Bértola and Ocampo (2012b, p. 139).

24 greater diversification in the commodities exported and destination markets.80 Thus, by the time of the first oil shock, Latin American countries had become a preferred place to invest banks' increasing loan funds: they were creditworthy and considered "good" borrowers in the private international capital markets 81

Lending to Latin America not only appeared justifiable to international banks, but appealed to industrial countries policymakers who sought to ease the impact of oil shocks. Faced with the big balance-of-trade deficits of the mid-1970s, as Wellons (1987) explains, the governments of major industrial countries sought to expand their exports to developing countries, in particular to Latin America. As these economies underwent industrialization processes, they became more reliant on imports of capital goods and equipment from industrial countries. Furthermore, imports of goods and services were also increasing as a consequence of Latin American countries' high rates of economic growth, their population explosion, and their rapid urbanization process.82 In such a context, the multiplying effects of the provision of trade credits and financing for investment projects on industrial countries' exports to the region were significant.83

One outstanding, though neglected, feature of international lending to Latin America was the participation of domestic commercial banks as foreign lenders. In the case of Mexico, for instance, Quijano (1987) shows that Mexican banks participated in approximately one third of the total credit raised by Mexican public and private sector borrowers in the syndicated Euroloan markets between 1974 and 1978. The author shows that Brazilian banks led eight syndicated loans to Brazil and participated in another 29 syndicated credits between October 1978 and December 1979, of which 20, or 81 per cent of the total amount, went to home country borrowers, while the remainder went mainly to other Latin American countries. Although to a much lesser extent, Argentine banks were also involved in the Euromarkets and participated in international lending, as did banks from Venezuela and other smaller Latin American countries as Colombia, Chile, and Peru.84

Latin American banks first reached the international capital market as shareholders of London-based consortium banks in the early 1970s. Consortium banks, also called Eurobanks, were independent banks collectively owned by several different banks, mainly established in London and conceived for Eurocurrency operations in any of its forms.85 After the establishment of Midland and International Banks (Maibl) in 1964 - the first of its type, the development of banking consortium went hand in hand with the evolution of the Euromarkets. As the size and scale of the Euromarkets grew

80 See Diaz-Alejandro (1984). 81 Compared to other developing countries in expansion, Eichengreen and Fishlow (1996, p. 39) observed that the weighted average annual growth rate in East Asia in the 1965 to 1973 period was lower than the levels attained in Latin American. Among the largest Latin American borrowers, Brazil experienced the miracle of the late 1960s and early 1970s, Mexico displayed high rates of expansion of the region, and Argentina achieved peak growth rates during these years. 82 See Bértola and Ocampo (2012b), Chapter 4 for an analysis of economic performance and the industrialization process in Latin America from the postwar period to the 1980s. 83 In this regard, a number of scholars have argued that the boom in international lending and borrowing by developing countries was actually encouraged by industrial country governments. See, for instance, Kapstein (1994), Helleiner (1996) and Altamura (2015). 84 See Quijano (1987, pp. 241-258), Chapter 6. Large commercial banks from Brazil, Mexico, and Argentina ranked in the top 50 of syndicated lead managers to Latin American in 1979 and 1980. See Latin American survey, Euromoney, April 1981, pp. 50-53.

25 exponentially in the wake of the financial fallout from the oil crises of the 1970s, the number and scale of operations of consortium banks in London have also surged.86

During the heyday of the 1970s and 1980s, when the Eurocurrency markets experienced its most dynamic expansion, virtually every major international bank participated in at least one consortium bank. However, ownership was not limited to the world's largest international banks and it often included small banks from many countries or regions that sought to combine their resources and get involved in international finance. This group of London-based international consortium banks included a large variety of institutions which focused on a diversity of businesses ranging from short- term trade finance to longer term bonds and credits to multinationals and foreign governments all around the world or from particular geographical areas or industries. As they expanded, consortium banks became active players in the development of international banking, particularly through their participation in sovereign and syndicated lending to developing countries.

Between 1972 and 1974, four consortium banks were created in London with the participation of Latin American shareholders (see Table 2). The involvement of Latin American banks in consortium banking was part of a trend among a handful of developing country financial institutions that got involved in the international financial community and the Euromarkets at that time. As reported in The Banker, client nations of the international money and capital markets had been increasingly promoting, through government-controlled domestic banks or private sector banks, the creation in London of consortium banks, in partnership with European and North American banks, specialized in international banking to local customers. Thus, in addition to those owned by large international banks, there were a significant number of consortium banks that had participation of Middle Eastern and Latin American banks.

Among the group of Latin American consortium banks, the Euro-Latin American Bank, also called the Eulabank, was the largest with assets of £ 1,518.2 million in 1982. It was 50 per cent European, with the other 50 per cent belonging to 10 Latin American banks from 9 different countries, to which Banco del Estado de Bolivia would be added in 1979. It was founded to strengthen the economic ties between Latin America and Europe and was mainly focused on medium and long-term Eurocurrency loans, project finance, and Latin American trade finance. 87 The other three Latin American consortium banks were the European Brazilian Bank, which was majority owned by Banco do Brasil with 31.9 per cent of the shares, the Libra Bank, of which Brazil Banco Itaú and Mexico Bancomer owned 8 per cent each, and the International Mexican Bank (Intermex), which was majority owned by Banco Nacional de Mexico (Banamex) with 38 per cent of the shares.

As Philip Wellons explains, the function of these consortium banks was "to act as a go-between for domestic borrowers, including their home office, and to raise money (...) in world markets for their home countries."88 Intermex, the main Mexican consortium bank operating in London, provides the most clear and representative example of the general pattern. The interest of the group was to operate as international commercial bank and to specialize in assembling international finance for

85 Bank of England defined consortium banks as "banks which are owned by other banks but in which no one bank has more than 50% ownership and in which at least one shareholder is an overseas bank." See "Consortium banks on course," The Banker, February 1976, p. 167. 86 See Roberts (2001) for an account of the rise and fall of consortium banking in London. 87 Ibid, pp. 252-253. 88 Wellons (1977, p. 77).

Mexico through the participation in loans and bond issues. In the words of Hector Reyes, the Mexico City manager of Intermex, they "went to London to borrow Eurodollars to lend to Mexico."89 This bank would, indeed, play a leading role in syndicated bank lending to Mexico during the years preceding the debt crisis of 1982.90

Table 2. London-based consortium banks

Founder Banks Assets in Consortium Founded 1982 Status until Latin America Other (Mil. £.)

Latin American C. Banks 3,358.0 Euro-Latinamerican Bank 1972 Banca Serfin, Banco de Algemene Bank, Banca 1972-1990 1,518.2 Colombia, Banco de la Naziolale del Lavoro, Nacion, Banco de la Banque Bruxelles Nacion Argentina, Lambert, Banque Banco de la Republica Nationale de Paris, Oriental del Uruguay, Barclays Bank Banco del Estado de International, Chile, Banco do Brasil, Bayerische Banco Industrial de Hypotheken, Dresner Venezuela, Banco de Bank, Osterreichische Pichincha, Banco Landerbank, Banco Mercantil de Sao Paulo, Central, UBS, Deutsche (5% each) Sudamerikanische Bank (each less than 5%) European Brazilian Bank 1972 Banco do Brasil (31.9%) Deutche Bank (13.7%), 1972-1988 766.0 UBS (13.7%), Dai-Ichi Kangyo Bank (8.8%)

Libra Bank 1974 Banco Itau (8%), Chase (23.6%), 1974-1990 692.6 Bancomer (8%) Mitsubishi Bank (10.6%), Royal Bank of Canada (10.6%), Westdeutsche Landesbank (10.6%), Credito Italiano (7.1%), National Westminster (5%), Swiss Bank Corporation (10.6%) Espirito Santo (5.9%) International Mexican Bank 1973 Banamex (38%) Inlat (13%), Bank of 1973-1992 381.2 America (20%), Paribas, Dai-Ichi Kangyo, Deutsche Bank and UBS (7.25% each)

Middle Eastern C. Banks 5,203.0

Japanese C. Banks 1,190.4

U.S. and European C. Banks 11,406.8

Note: 35 of the 39 United Kingdom (U.K.) Consortium Banks, as defined by the Bank of England 1975-1987.

Source: Based on Roberts (2002).

89 "World Push by Mexican Banks Irks Rivals," The New York Times, April 18, 1981. 90 See Negrete Cárdenas (1999, p. 400-404), Table B17.

In parallel to the creation of consortium banks, during the 1970s, Latin American banks opened their own branches and agencies in London, as well as in other international financial centers. Between 1973 and 1982, the number of branches and agencies of Latin American banks in London increased from 6 to 18, with the majority of the new offices opening after 1977 (see Table 3). In terms of nationality, while in 1973 only four Latin American countries – Argentina, Brazil, Mexico, and Chile – had banking representation in London, by 1982 the number had increased to nine. Similarly, as shown in Table 3, during this period, Latin American banks also opened branches and agencies in the United States, and their expansion there was even more dramatic. While in 1973 there were only 7 branches of Latin American banks in the U.S. – 5 of those being Brazilian – this number jumped to 48 by the end of 1982. Like in London, the opening of U.S. banking offices by Latin American banks was largely concentrated in the 1977 to 1982 period, and New York was the main destination.

Table 3. Foreign agencies and branches of Latin American banks

London United States 1973 1977 1982 1973 1977 1982 Latin America 6 9 18 7 16 48 Argentina 2 2 2 1 1 8 Brazil 2 4 6 5 9 18 Chile 1 1 1 0 0 1 Colombia 0 0 2 0 1 3 Mexico 1 1 4 1 3 10 Peru 0 0 1 0 0 0 Uruguay 0 0 1 0 0 1 Venezuela 0 1 1 0 2 5 Panama 0 0 0 0 0 1 Paraguay 0 0 0 0 0 1

Source: Based on The Banker and FFIEC 002.

In fact, the U.S. experienced impressive growth in its foreign banking community during the 1970s.91 The abolition of the Interest Equalization Tax (IET), along with the expiration of exchange controls in January 1974, made the U.S. capital market once again relevant for foreign borrowers and international banks. Apart from New York, London's rival international financial center, California and Chicago also welcomed an increasing number of foreign banks, though on a much smaller scale.92 By the end of 1982, total assets of Latin American banks in the U.S. reached nearly US$ 10 billion, of which the agencies and branches in New York accounted for 63.1 per cent, Los Angeles 17.7 per cent, San Francisco 12.7, and Miami, Chicago and Washington the remainder. The main reason that Latin American banks, like other foreign banks, established locations in the U.S., and in particular in New York, was to access its money market and open a dollar-base funding channel.

91 See Terrell (1993), United States General Accounting Office (1979), Terrell and Key (1977) and Damanpour (1990). 92 Up until and including 1970, 73 foreign banks were directly represented in New York, while by the end of the decade, this figure had more than tripled. By 1980, foreign banking assets in New York accounted for approximately 70% of U.S. total foreign banking assets, compared with the 23% of California and 3% of Chicago. The Banker, February 1980, p. 87.

The presence of Latin American financial institutions abroad shows that debtor countries, like their creditor counterparts, also underwent a process of internationalization in the banking sectors. In the case of Mexico, as Chapter Three explains, domestic banks increasingly turned to the international capital markets in the years that preceded the 1982 international debt crisis. Confronted with domestic fundraising problems and increasing competition from foreign banks in the early 1970s, leading commercial banks found the international wholesale markets to be a rich alternative source of funding. As in many other Latin American countries, this was a period of high inflation and heavy financial regulation in Mexico, or domestic financial repression as defined by McKinnon (1973) and Shaw (1973), where interest rates were not proportionally adjusted by the monetary authorities.93 In this context, the international capital markets offered banks a new, unregulated, more flexible space to develop their lending activities and overcome domestic constraints.

Foreign finance and international loans were the cornerstone of borrowing countries' economic goals, and the involvement of domestic banks was fundamental. During the 1970s, administrations throughout Latin American sought to promote economic growth through a developmentalist and nationalist strategy.94 At this time, they were undertaking the final stages of the import-substituting industrialization process. Financing for the government and public enterprises, as well as for the private sector, was needed to support industrial and economic growth.95 Additionally, international interbank lending was an instrumental source of foreign exchange, as it helped financial authorities to manage their countries' monetary and exchange rate policies. In most Latin American countries, including Mexico, Brazil and Argentina, governments not only encouraged the internationalization of private banks, but also directed state-owned commercial banks to internationalize.

2.3. Interbank markets, domestic banks and banking fragility

A central feature of the expansion of international bank lending during the 1970s was the large interbank market. In 1982, the Bank for International Settlements (BIS) study group on the international interbank market estimated that up to three-quarters of the international lending boom that had taken place in previous years consisted of interbank positions (see Figure 2.1).96 Although the interbank market had always been at the core of the international banking system, it took on a larger role after the first oil shock, becoming, in Michael Moffitt's words, "the mainstay of the Euromarkets."97 But despite its central function in the banking system and the significance of its size and scale, up to now the Euromarkets interbank issue has eluded serious scholarly attention in most of the accounts of international finance and external indebtedness in the wake of the debt crisis of the 1980s.

International interbank transactions consisted of fast, informal transfers of short term funds between banks. They not only enabled individual banks to lend without being tightly constrained by deposits they attracted from the non-banking sector, but, like other money market transactions,

93 This refers to government regulations and non-market restrictions that constraint the functioning of financial intermediaries, such as high bank reserve requirements, interest rates ceilings, directives on the allocation of credit and capital controls. As developed in further detail in Chapter 3, expect for capital controls, all these restrictions were part of the Mexican banking policy during the 1970s. 94 See Bulmer-Thomas (2003, pp. 313-352), Chapter 10. 95 See Frieden (1991) and Bértola and Ocampo (2012b, pp. 162-189). 96 BIS Archive, Box 1/3A(3)M Vol. 1. 97 Moffitt (1984, p. 69).

29 they also provided a way to adjust the volume and nature of their assets and liabilities. At an international level, the interbank market acted as a channel from banks with a domestic dollar base or an excess of deposits to direct lending toward banks where direct lending exceeded deposits. Indeed, as the Eurocurrency deposits expanded, the interbank market provided a flexible device that made large amounts of loanable funds quickly available, at a low cost.98 This both allowed the banks to meet new lending opportunities and encouraged the recycling of petrodollars, and also let them control the interest and exchange rate risks that arose from their foreign business.99

Figure 2. The international interbank market in the 1970s

Figure 2.1. Euromarkets and interbank market Figure 2.2. U.S. and U.K. interbank market

1000

900

800

700

600

500

400 billion of dollars 300

200

100

I-73

I-74

I-75

I-76

I-77

I-78

I-79

I-80

III-73

III-74

III-75

III-76

III-77

III-78

III-79 III-80

BIS Banks total cross-border liabilities in foreign currencies BIS banks interbank cross-border liabilities in foreign currencies

Source: Bank of England's Quarterly Bulletin and Financial Accounts of the United States.

Consortium banks, which were major players in the Euromarkets and recycling process largely relied on the international interbank market to develop their business. Unlike commercial banks, they had no branch network or territorial presence, and, so as financial historian Richard Roberts put it, they "were born naked, with neither deposits nor clients."100 Unable to cultivate a strong retail deposit base, these banks were completely dependent on the purchase of deposits in the wholesale money markets. Thus, their liability structure was largely dependent on obligations to the banking sector. Their main funding source was short-term borrowing from other banks, but they also drew on negotiable London dollar certificate of deposits, floating-rate loan notes, Eurobonds, loans from other branches, and domestic funds from parent banks.101 Conversely, the asset side of the balance sheet was mainly made up of Euroloans, and to a much lesser extent, liquid assets and reserve balances in the form of informal credit facilities with other banks.102 Finally, since they borrowed more than they placed, consortium banks were net takers of funds in the interbank market.

Indeed, consortium banks were more dependent on the Eurocurrency interbank market than any other type of bank operating in London. As the Bank of England reported in 1981, as much as 80 per cent of the foreign liabilities of consortium banks during the 1978-80 period were due to the

98 See Bell et al. (1982a, pp. 15-17). 99 See BIS (1983) for a study of the international interbank market during this period. 100 Roberts (2001, p. 26). 101 See Dufey and Giddy (1994, pp. 216-232). 102 See Davis (1980, pp. 39-40).

30 interbank market, followed by Japanese and other overseas banks (65-69%), British banks (58%), and U.S. banks (46%).103 Within the total, U.S. banks were net suppliers and the main contributors of funds to the interbank market, while the consortium banks were, relative to their size, the largest net takers of these funds.104 The net position and maturity composition of consortium banks' balance sheets, exhibited in Table 4, show the extent to which their funding was concentrated on short-term borrowing from banks and claims on long term loans to non-bank borrowers. The figures also show the substantial increase in borrowing at less than one year and loans made at maturities of more than one year that took place between 1973 and 1982.105

Table 4. Net liabilities and claims of London consortium banks Millions of pounds sterling

1973 1982 Sector \ Maturity 1 mo. 1 yr. 1 mo. 1 yr. < 1 Total < 1 Total breakdown to < 1 to < 3 > 3 yr. to < 1 to < 3 > 3 yr. mo. Sector mo. Sector yr. yr. yr. yr. U.K. interbank market -105 -462 -14 2 -579 -1,179 -4,901 190 194 -5,696 Other U.K. residents 6 53 54 90 203 -90 17 31 120 78 Banks abroad -273 -741 18 -8 -1,004 -2,040 -2,361 964 1,393 -2,044 Other non-residents 107 438 335 559 1,439 -655 1,052 2,409 4,723 7,529 Net CD & CP held* 162 328 -194 -330 -34 49 39 35 35 158

Total -103 -384 199 313 25 -3,915 -6,154 3,629 6,465 25

* Net certificates of deposits issues and commercial paper held.

Source: Bank of England's Quarterly Bulletin.

While initially limited to the world's major international banks, the interbank market increasingly became used by a larger number and wider variety of banks as it expanded. The market grew from a few hundred institutions in the mid-1970s to well over 1,000 banks from more than 50 countries by the early 1980s.106 This expansion entailed both greater volumes of operations between banks in the same financial center, as well as more cross-border business, including inter-office positions and genuine interbank activities among banks throughout the world.107 Many smaller institutions, ranging from regional banks in advanced industrial countries to commercial banks in less developed countries, became active participants in the Eurocurrency interbank market. For these institutions, as in the case of consortium banks, this interbank market provided attractive wholesale source of funding to develop their domestic and international activities. In the case of large Latin American banks, the bulk of these operations were undertaken from their foreign agencies and branches, the platform used by parent banks to become involved in the Euromarkets.

Table 5 presents the 1982 balance sheet of the London agencies and branches of commercial banks from the three larger Latin American debtor countries. Together, they had consolidated assets and

103 Ellis (1981). 104 Ibid, pp. 359-360. 105 Aronson (1976) shows that consortium banks relied more on short borrowing and long lending than other types of banks operating in London. 106 Giddy (1981). 107 See BIS (1983).

31 liabilities of about US$ 8.5 billion, of which 61.1 per cent corresponded to six Brazilian agencies, 23.2 per cent to four Mexican agencies, and the balance of 15.7 per cent to two Argentine agencies.108 The large part of liabilities to banking institutions, which accounted for as much as 85 per cent of the total obligations of the Brazilian and Mexican agencies, makes the interbank nature of these agencies' funding strategies clear. The fact that the net position (assets less liabilities) vis-à-vis banks in the U.K. and abroad were negative implies that they were net borrowers in the interbank market. Conversely, on the asset side, claims on their own offices were larger than the liabilities to them, as was also the case for the remainder of the assets. This indicates that the foreign liquidity these agencies raised in the international wholesale markets was channeled to parent banks or to non- bank borrowers outside the U.K.

Table 5. Branches and agencies of Latin American banks in London End-June 1982, millions of dollars

Brazil Mexico Argentina* Net Net Net Assets Liabilities Assets Liabilities Assets Liabilities Position Position Position (1) (2) (1)-(2) (1) (2) (1)-(2) (1) (2) (1)-(2) Banks 3,921 4,520 -599 1,068 1,824 -756 1,308 750 558 In the U.K. 404 1,435 -1,031 260 1,147 -887 350 194 156 Outside the U.K. 570 1,938 -1,368 331 429 -98 176 498 -322 Own offices abroad 2,947 1,147 1,800 477 248 229 782 58 724 Non-Banks 1,083 369 714 984 52 932 In the U.K. 90 40 50 0 0 0 Outside the U.K. 993 329 664 984 52 932 Other 178 292 -114 70 246 -176 20 578 -558 Negotiable papers 4 199 -195 3 243 -240 0 555 -555 Other 174 93 81 67 3 64 20 23 -3 Total 5,182 5,181 0 2,122 2,122 0 1,328 1,328 0

* Data for End-March 1982

Source: Bank of England's archives.

Access to international wholesale liquidity was not limited to the London Eurocurrency interbank market. The re-opening of the U.S. capital market in the early 1970s made it possible for foreign banks who had established locations in New York, California and Chicago, the country's major money market centers, to raise dollars directly in the domestic interbank market. In particular, New York, which was by far the largest money market in the world, offered a wide range of instruments, high liquidity, and extensive international operations. In fact, interbank liabilities of foreign banking offices in the U.S. underwent an impressive expansion during this period, beginning at close to US$ 2 billion in 1970 and reaching almost US$ 50 billion by the beginning of the 1980s (see Figure 2.2).109 For the agencies and branches of Latin American banks, along with most foreign banking offices

108 The six Brazilian banks were Banco do Estado de Sao Paulo, Banco do Brasil, Banco Real and Banco Mercantil do Sao Paulo; the four Mexicans were Banamex, Bacomer, Banca Serfin and Multibanco Comermex; and the two Argentine were Banco de la Nacion Argentina and Banco Galicia y de Buenos Aires. 109 Z.1 Financial Accounts of the United States, Historical Annual Tables 1965-1974 and 1975-1984.

32 operating in the U.S., money market instruments were the single and most important source of funding.110

As Table 6 shows, the balance sheet of the U.S. agencies and branches of Brazilian, Mexican and Argentine commercial banks were structured similarly to their London counterparts. As of June 1982, they had consolidated assets and liabilities of about US$ 9.5 billion (one billion dollar more than in London), of which US$ 8 billion, or about 85 per cent was owed to banking institutions. Of this amount, 67.2 per cent were liabilities to banks in the U.S., while the remaining 14.1 and 18.7 per cent were liabilities to banks outside the U.S. and head offices respectively. In fact, although levels varied among the countries, banks in the U.S. were agencies' main creditors, accounting for between 40 and 60 per cent of their total liabilities. The large part of these obligations consisted of deposits and trade balances that the banks held with the agencies, but there were also substantial amounts of federal funds, borrowed money, and interbank credit lines. As in the case of London, the agencies' net position was negative, which meant that the banks in the U.S. were their net suppliers of funding. Obligations with the non-bank sector, both in the U.S. and abroad, and other liabilities represented an average of 15 per cent of the balance sheet liabilities.

Table 6. Branches and agencies of Latin American banks in the U.S. End-June 1982, millions of dollars

Brazil Mexico Argentina Net Net Net Assets Liabilities Assets Liabilities Assets Liabilities Position Position Position (1) (2) (1)-(2) (1) (2) (1)-(2) (1) (2) (1)-(2) Banks 2,942 3,507 -565 924 2,551 -1,627 631 1,899 -1,268 In the U.S. 727 2,707 -1,979 300 1,854 -1,554 114 784 -670 Outside the U.S. 496 277 219 264 488 -224 430 358 72 Own offices abroad 1,719 523 1,196 360 210 150 87 757 -670 Non-Banks 1,326 827 500 1,904 10 1,894 1,354 156 1,198 In the U.S. 217 734 -517 8 1 7 6 110 -104 Outside the U.S. 1,109 93 1,016 1,896 10 1,886 1,348 46 1,302 Other 224 159 66 81 347 -266 143 73 70 Negotiable papers 1 81 -80 2 293 -291 27 28 -1 Other 223 77 146 79 54 24 116 45 71 Total 4,492 4,492 0 2,909 2,909 0 2,127 2,127 0

Source: FFIEC 002.

In contrast, the net position of the agencies compared to the non-banking sector was largely positive. In fact, the private and public non-bank sector was the agencies' largest group of debtors, accounting for about 40 per cent of consolidated total assets (see Table 6). These claims primarily consisted of lending facilities to final borrowers outside the U.S. Loans and advances granted by Argentine and Mexican agencies to non-U.S. residents reached US$ 1.3 billion and US$ 1.6 billion, respectively, amounts that represented as much as 99 per cent of their loan portfolio. As for

110 As stated by Serge Bellanger, vice-president of the Institute of Foreign Bankers and Crédit Industriel et Commercial's New York branch manager, when looking at the liabilities side of overall foreign banks, "interbank borrowings from the domestic and Eurodollar markets still remain a major component of the funding strategy." "The foreign challenge to U.S. banks," The Banker, October 1978, p. 40.

Brazilian banking offices, the corresponding figures were US$ 1 billion and 84 per cent. These were either direct loans granted from the U.S. agency to a foreign borrower or syndicated credits that their parent banks participated in, along with other international banks, though their foreign agencies. Thus, in the U.S. as well as London, the function of agencies of commercial banks from Brazil, Mexico and Argentina was to raise interbank money that they then used to relend abroad.

The intermediation between foreign capital and final borrowers performed by Latin American banks meant greater risk for the foreign agencies as well as for their domestic parent banks. In the case of Mexico, as Chapter Four expands upon, the increasing trend of leading domestic banks to rely on the international interbank market to fund cross-border lending created new risks and vulnerabilities. Increased reliance on foreign interbank borrowing meant a higher weight of foreign currency external debt in the liabilities, while dollar-denominated claims were largely concentrated in Mexican borrowers, who mainly operated in the domestic currency. The result was an accumulation of currency mismatches on the balance sheets of the agencies and parent banks, which would become a major problem when the prospect of devaluation loomed in early 1982.

In addition, the use of interbank money as the basis for Mexican banks' loans created significant maturity and interest risks. Although banks incur maturity mismatches between their borrowing and lending activities, the pattern of maturity transformation in the agencies' foreign businesses was presumably unsustainable. Interbank placements or credit lines had very short maturities, normally between overnight and six months, but the loans they extended were arranged at much longer term. Additionally, while interbank funding lines were set at floating rates, the bulk of the loans that Latin American banks made were at pre-determined fixed interest rates.111 These mismatches were not a major problem in normal times, as agencies could easily roll over their interbank credits when they came due, but it certainly became more problematic in periods of financial distress and rising interest rates. As bad times emerged, agencies could only find new interbank credit lines at shorter maturities and higher costs, making them vulnerable to eventual creditor bank lending limits and their refusal to roll over deposits as they became due once the crisis broke out.

As a matter of fact, the size of the international interbank market shrank as soon as the international debt crisis broke out. In November 1982, BIS official Alexandre Lamfalussy reported to the members of the Eurocurrency standing committee that "there has been a shrinkage of interbank position and a halt in the cross-border interbank market."112 Many of the largest U.S. money center banks, which were the net suppliers of funds within the market, started to reduce market in-and-out trading where risk was greater. Banks relying on the interbank market for funding, therefore, faced a liquidity squeeze and had to find substitutes for the interbank market to cover their funding needs. That proved particularly problematic for the international banking agencies and branches of troubled debtors, such as Mexico, Brazil and Argentina, which were heavily dependent on foreign wholesale interbank liquidity, largely exposed in their home countries, and had only very limited alternative source of funding.

111 Interbank placements or credit lines were typically arranged at LIBOR plus a premium, depending on the risk associated with the borrowing bank. 112 FRBNY Archive, Box 108406, Notes on the G-10 Governors' Meeting held at the BIS, November 8, 1982.

2.4. Foreign agencies, debt renegotiation and crisis management

Although overlooked in the international debt crisis literature, interbank deposits became an issue of major concern during debt renegotiations in the aftermath of the Mexican moratorium. The outbreak of the crisis in Mexico disturbed the normal interbank funding transactions and rollovers of existing money market lines with the agencies and branches of Mexican banks and other Latin America countries as well. Bankers' acceptances, pre-export financings, straight Eurodollar and prime based advances all became increasingly difficult in a nervous and uncertain market atmosphere. As the crisis spread throughout the region, the increased risk perception on operations with banks from troubled countries prompted creditor banks to reduce their business and involvement with them, stop renewals of interbank claims, and demand to be paid off at maturity.

Figure 3. Interbank liabilities of Latin American Bank's foreign agencies

Source: Bank of England's archives and FFIEC 002.

Figure 3 shows the evolution of the liabilities to banking institutions of the U.S. and London agencies and branches of Brazilian, Mexican and Argentine banks in 1982. It provides a clear perspective on the erosion of interbank funding that these agencies suffered around the outbreak of the international debt crisis. Between September and December of 1982, Brazilian agencies lost US$ 1.7 billion dollars in the interbank market, an amount representing a drop of 25 per cent of their outstanding obligations in only three months. In the case of Mexico, withdrawals by creditor banks seem to have preceded the outbreak of the crisis, with the agencies losing US$ 886 million, or 30 per cent of interbank funding between June and September of 1982. As for the Argentine banks, the decline of interbank liabilities was largely concentrated in the London agencies, which was certainly related to the Falklands War that began in April 1982.

Problems in the interbank deposit businesses were not confined to Latin American banking offices. A Bank of England internal report stresses that during 1982, a number of other British and foreign

35 banks, whose liabilities were owed to the interbank market and whose assets were not liquid, also encountered funding difficulties in the London Eurocurrency market. According to this report, for instance, the balance sheets of Portuguese and South Korean branches "continued to contract in the second quarter… [and] this was largely associated with reduced inter-bank activities on both sides of their balance sheets, as their agencies ran down business with banks in London and elsewhere."113 In the eyes of London supervisors, these agencies appeared particularly vulnerable to changes in market sentiment, partly because they were dependent on market funding and had large country risk exposure.114 Although to a lesser extent, consortium banks also faced hard times funding themselves in the interbank market for similar reasons. Indeed, some of them were forced to request substantial support from their shareholders.115

As their interbank funding lines dried up, these institutions, especially Latin American banks, faced funding pressures. Foreign banks were their single most important suppliers of immediate liquidity and, as previously discussed, they had only limited capacity to adjust their assets since most of them were long term loans or illiquid claims. Most of them did not have a liquidity cushion and were unable to fund themselves outside the international interbank market. They had no dollar retail deposit base, were not secured by the Bank of England in London or the Federal Deposits Insurance Corporation (FDIC) in the U.S., and could not borrow from the U.S. Federal Reserve discount window. Recourse to foreign exchange from head offices and parent central banks in a context of dwindling international reserves and balance of payment crises in their home countries was not a realistic possibility either. The contractions in interbank liabilities could, therefore, seriously damage the liquidity, and even the solvency, position of these agencies, creating the very real risk that they would no longer be able to reimburse their creditors on time.

However, the possibility of a disruption of interbank payments had implications that went far beyond the financial situation of individual banking institutions. If a bank or an agency became insolvent and, as a result, was unable to repay its liabilities, then the solvency of those banks which had provided it with interbank funds would also be called into question, as the Herstatt-Franklin crisis in 1974 had shown. Under such circumstances, there was a danger that the failure of one bank or agency, which in the absence of the interbank market would have been an isolated incident, would spread throughout the banking system and cause a domino effect collapse. The Mexican agencies had US$ 5.2 billion of outstanding interbank liabilities spread throughout 1,000 banks across different international financial centers. Focusing on Latin American, according to Paul Mentre's computations, total interbank liabilities reached about US$ 15 billion by the beginning of 1983.116 Given the high volume, the uncollateralized nature, and the pyramid structure of these interbank transactions, the potential systemic danger resulting from a payment disruption by Latin American banks was considerable.

113 Bank of England Archive, Task Force, File 13A195.1, Review of international financial development 1982- 1983. 114 The report stresses that even "French banks have given some concern in the market as they continue to be very heavy takers of almost any maturity available for large amounts." Ibid. 115 Bank of England Archive, Task Force, File 13A195.1, Note for Record of the Banking Supervision Division, February 3, 1983. 116 Brazil US$ 6 billion, Mexico US$ 5.2 billion, Argentina US$ 1.4 billion, Chile US$ 1.2 billion and Peru US$ 1.2 billion. FRBNY Archive, Box 108403, The International Interbank Market and International Banking Lending, June 28, 1985.

Indeed, the situation of the agencies was a matter of grave concern for policymakers and financial authorities on both sides of the table. From the beginning of debt renegotiations, Latin American central bankers and secretaries of finance insisted on the need to maintain interbank deposits at the offshore agencies of their domestic banks. Still, although spreads were revised upward and yields were increased substantially to encourage the banks to maintain or increase deposit levels, the leakage continued. Agencies looked to cover the leaks through asset reduction, the use of headquarters' internal liquidity, emergency support from home country central banks, and the use of overnight credit lines but these solutions were insufficient. Eventually, creditor countries' supervisory authorities, especially the Federal Reserve and the Bank of England, intervened to persuade international banks to limit the reduction of their deposits to ensure that foreign branches of Latin American banks did not default on their obligations and maintain confidence in the system.

Different approaches were used in order to secure the renewals of interbank deposits at maturity. As sovereign debt lawyer Lee Buchheit explains, some countries, such as Argentina and Brazil, asked their creditors to sign formal agreements whereby creditor banks agreed to maintain their interbank deposit liabilities at their levels on the date of the moratorium declaration. Other countries, like Mexico or the Philippines, agreed to a clause in their restructuring documents that stated that a default event would be triggered if the aggregate level of interbank liabilities placed with the offshore agencies and branches of their domestic banks were to drop below the pre-moratorium levels.117 For their part, the agencies agreed to continue to pay interest on these liabilities when they came due, and their home governments and central banks agreed to make the necessary foreign exchanges available to do so. The principle underlying these approaches was to avoid restructuring interbank debts, which would have seriously disturbed the international financial market, and caused problems for proposed rollovers.

These schemes were an integral part of the broader financial packages and restructuring programs implemented to manage debt-payment crises and guarantee the stability of the international financial system. When Mexico communicated its difficulties to pay foreign debt obligations to the U.S. on August 12, 1982, the Federal Reserve and Treasury Department promptly developed a rescue strategy to provide the country with direct emergency funding. Presumably out of concern with the potential implications for its banking system, U.S. officials led an international cooperative effort for greater financial assistance to prevent an interruption of repayments to international commercial banks.118 The extent of Mexico's debt-service obligations relative to creditors' individual resources and financial capacities eventually required the participation of all creditors to keep the country afloat. Debtor governments, European and Japanese central bankers and finance ministries, the IMF, and commercial banks in both the U.S. and abroad were all mobilized in this regard. An international policy response to confront the Mexico's debt payment problems and handle the broader international debt crisis of 1980s was developed among through the cooperation of all these actors.

As noted earlier, the IMF emerged as the institution responsible for the management and the execution of the creditors' collective strategy. The Fund's implication in the international debt crisis was strongly supported by the governments of industrial countries, especially the United States.119 Once they realized the importance of keeping debtors afloat with enough liquidity, the Fed and the

117 Buchheit (1991, pp. 15-16). 118 See United States General Accounting Office (1997, pp. 19-34). 119 See Cohen (1986), Cohen (1982), Frieden (1987), Krugman (1994), and Sachs and Williamson (1986).

U.S. Treasury turned to the institution with the most experience in dealing with developing countries' external problems. On the one hand, the Fund's long trajectory with balance of payment crises gave it the know how to treat the Third World's debt troubles, major advantage compared with most creditors. On the other hand, its conditionality policy served as a valuable element in making adjustment programs a precondition for borrowing countries to reschedule their debts and access to creditors' financial assistance.120 As for the Fund itself, the crisis provided it the opportunity to assume a new function as crisis manager in the international financial system, and recover the legitimacy it had lost after the end of Bretton Woods.121

Commercial banks, which were debtor countries' main creditors, were the final component of the creditors' debt management strategy. By that time, though banks had developed a systematic procedure to approach developing countries for joint lending purposes, they had only sporadically confronted repayment problems of debtor countries and, therefore, lacked the experience to handle a sovereign debt crisis on such an international scale.122 Given the large number of lending institutions involved, the heterogeneity among them and their different levels of exposure to the crisis, the collective engagement of commercial banks in the management of the crisis was a critical but major challenge.123 The IMF played a central role in securing their participation, by making commercial bank financing a condition for approval of stand-by and extended agreements with troubled countries.124 At the same time, adjustment programs became a condition for developing countries to have their debts restructured by commercial banks and to receive new lending.125 In the broader debt management strategy, banks took on the role of rescheduling outstanding debt, stretching principal payments, and extending new money facilities, all of which were required under IMF agreements with debtor countries.

Involving officials from creditor countries and international agencies was key to coping with the crisis and averting a collapse of the banking system. After all, as Paul Volcker concedes, the debt management strategy provided lender of last resort assistance to the national and international banking system through the involvement and contribution of a multiplicity of actors.126 The increase in official lending from creditor governments and international organizations, along with the savings that austerity programs generated from debtor economies, financed an outward transfer of resources to the benefit of commercial banks. Latin American countries, which had previously been net recipients of resources from abroad became net exporters of capital beginning in 1982.127 International commercial banks provided new loans as part of the restructuring agreements, but they received a much larger amount from borrowing countries in debt-service payments. In the end,

120 See Pauly (2008). In a broader perspective, creditor governments and the IMF had been in discussions about sovereign debt issues in the context of the Paris Club since its creation in 1956. Only when the number of countries facing debt problems increased at the end of 1970s and the beginning of the 1980s, however, did these institutions begin to meet more frequently and start to collaborate on debt matters in a consistent way. See Rieffel (1985, 2003) and Trichet (1989). 121 See Boughton (2000, 2001). 122 See Hardy (1982) and Rhodes (1994, 2011). 123 See Fernandez and Kaaret (1992) and Lipson (1985). 124 Lipson (1987) highlights the level to which IMF agreements conditioned and determined banks' private lending and rescheduling activities as one of the most salient aspects of the new debst policy regime. 125 Sgard (2012). 126 See Volcker and Gyohten (1992, p. 203). See also and Wellons (1987). 127 In terms of GDP, the amount of resources the region transferred abroad between 1982 and 1989 was between two and three times what it had received during the 1970s. See Bértola and Ocampo (2012a).

38 the way in which the crisis was managed allowed the creditor banks to rebuild their capital base and increase their reserves and by the end of the decade, they were in a position where they could accept losses without compromising the confidence in the international banking system.128

Although bank failures and a disruption of the international financial system were successfully avoided, the management of the crisis entailed substantial economic and social costs to Latin America. Between 1982 and 1989, Latin American countries rescheduled their debts several times and participated in multiple IMF-adjustment programs that proved to be unsuccessful in generating lasting solutions to crisis. As a result of the extent of the debt burden and the belt tightening to reduce it, economic growth was deeply restricted. GDP per capita, employment, and other social and living standard indicators deteriorated as the years passed and debt problems remained unsolved. As early as the mid-1980s, Diaz-Alejandro argued that "what could have been a serious but manageable recession [had] turned into a major development crisis unprecedented since the early 1930s."129 The unequal distribution of the burden of the crisis raises the important question of why Latin American countries played along with creditors in the international debt negotiations? In the case of Mexico, as Chapter Five explains, the exposure of the domestic banking sector to its own debt crisis and the implications of its dependence on foreign funding for the stability of the Mexican financial system seem to have been crucial factors in accounting for the weak bargaining position of Mexican policymakers during renegotiations with international creditors.

2.5. Domestic policies, macroeconomic imbalances and financial crisis

Embedded in the international framework I have just described, economic policy in debtor countries also played an important role in their increasing involvement with foreign finance and external debt. In the case of Mexico, macroeconomic management, and in particular fiscal and monetary policy, contributed to the creation of the financial crisis of 1982 and influenced the subsequent debt renegotiations. The combination of loose fiscal control with a fixed foreign exchange rate regime created a series of macroeconomic problems and imbalances that made the Mexican domestic economy vulnerable to changes in the international capital markets and the external shocks of the late 1970s and early 1980s.

Following the devaluation of August 1976, after 22 years of fixed exchange rates, the Mexican national currency stabilized at the new parity of about 22.5 pesos per dollar by early 1977. This balance of payment crisis was the consequence of fundamental disequilibria, namely inflation, increasing fiscal deficits and negative current account balances, which had been accumulating since the end of the stabilizing development program.130 As part of the financial stabilization strategy, Mexican authorities looked for financing from the U.S. and subscribed to a three-year standby agreement with the IMF that called for the usual monetary and fiscal austerity.131 Although Mexico was under the tutelage of the IMF at the beginning, it abandoned the adjustment program in June

128 Discussions about bank's capital throughout the text refer to total capital (equity, surplus and undivided profits) unless otherwise specified. 129 Diaz-Alejandro (1984, p. 335). 130 See Buffie and Sanginés-Krause (1989, pp. 141-147). 131 See Negrete Cárdenas (1999, pp. 209-214) for an explanation of how the crisis was managed and the financial system stabilized.

1978, and macroeconomic management dispensed with the targets of fiscal deficit, external indebtedness and wage increases in favor of more expansionary economic policies.132

Moreover, as the oil reserves of the country expanded, the Mexican government became even more deeply engaged in a growth-led strategy based on strong fiscal stimulus.133 While current and capital public expenditures in terms of GDP rapidly increased between 1978 and 1982, revenues expanded at a much more modest pace so that the government's fiscal deficit soared (see Table 7).134 Foreign borrowing came to finance the bulk of the large budget deficits, but the government also used of domestic financing generated by the sale of Treasury Certificates (Cetes) to the banking system and the public. The surge in public spending and the crucial role of the government as a producer led to an expansion in aggregate demand and boosted economic growth to very high levels (see Table 7).

However, in contrast to the stabilizing development period, high economic growth was no longer coupled with price stability. After many years of low inflation in Mexico, inflation started to rise in the early 1970s, increasing to double-digits levels by 1973 and becoming a real problem for the rest of the decade. Mexico's inflationary process was part of a worldwide phenomenon and the consequence of external factors as well as domestic economic policy mismanagement. The strong increase in nominal and real wages between 1970 and 1975, the devaluation of 1976 and the monetary impact of public sector deficits exacerbated the inflationary pressures. Between 1977 and 1982, rises in nominal wages and inflation occurred at a very similar pace, and the expansion of the monetary base and money supply was between one a half and two times as large as the increase in the price level during most of the period (see Table 7).

Within a currency board regime, high and rising inflation led to a considerably overvalued foreign exchange rate. Between 1977 and 1982, the Mexican peso appreciated in real terms against the U.S. dollar at an average annual rate of 6.5 per cent, which represented a cumulative appreciation of 37.5 per cent over the entire period. By the time of the devaluation of February 1982, the real exchange rate had attained similar levels to the peak reached in 1977 in the wake of the currency crisis. The appreciation of the real foreign exchange along with a deterioration of the terms of trade and a strong demand for imports driven by the country's large economic growth in the late 1970s, had a negative impact on Mexico's trade balance. The rise of international interest rates also contributed to an increasing current account deficit, which almost tripled from 2.4 to 6.2 per cent of GDP between 1977 and 1981. Surpluses on the capital account, largely based on external bank borrowing, allowed for a balancing of the country's external position even as the current account steadily deteriorated.

Under such circumstances, the decision of the Mexican government and financial authorities to defend the parity of the national currency with the dollar fueled expectations of future depreciation. In a context of free convertibility of the peso with the dollar and negative domestic real interest rates, these expectations translated into increasing amounts of capital flights since the second half of 1981. From 1981 to mid-1982, the Mexican private sector fled the national currency and took out

132 Ibid, 55-56. 133 See Bazdresch and Levy (1991, pp. 246-252). 134 Zedillo (1985).

40 of the country about US$ 16.5 billion according to Negrete Cárdenas' estimates (see Table 7).135 Within the domestic financial system, depositors moved their savings from pesos to mexdollars and then even started to withdraw their balances from these dollar deposit accounts. 136 These developments were made possible by the fact that Banco de Mexico was engaged in selling dollars at the parity rate while international reserves were replenished by increasing recourse to foreign capital and external debt.

Indeed, as Jeffrey Sachs (1989) explains, capital flight was closely intertwined with the external indebtedness processes and the macroeconomic imbalances mentioned above. As the government increased spending and the central bank expanded the money supply, the higher money balances in the hands of the private sector led to a weakening of the exchange rate as it converted cash into foreign currency. The trend created inflationary pressures that the central bank unsuccessfully attempted to control by keeping the exchange rate from depreciating by selling dollars, thereby leading to a rise in the foreign assets holdings of the private sector that were largely transferred abroad. With a negative trade balance, an adequate level of central banking reserves was maintained only by foreign-exchange inflows from oil exports and private and public sector borrowing in the international capital markets.137

As capital flight continued and fiscal and current account deficits grew, Mexico experienced the largest run-up in foreign debt of the period. Between the beginning of 1980 to the end of 1982, total external debt more than doubled from US$ 40.2 to 84.8 million (see Table 7). While increasing by 12.6 and 20.3 per cent in 1978 and 1979, total external indebtedness accelerated with growth rates of 26 and 47.6 per cent in 1980 and 1981 respectively. The lion's share was due by the state, but the private sector was also increasingly borrowing abroad. In fact, private foreign debt increased from 22.9 per cent of total external debt in 1977 to 33.3 per cent in 1980 and 30.6 per cent by the end of 1982. Large private industrial firms and commercial banks had direct access to overseas financing, and they could therefore benefit from a wide range of foreign bank loans at more attractive rates and financial terms than scarce and expensive domestic funding despite the currency risk.138

Faced with such a deteriorating economic situation and the tightening of international credit, the government adopted a number of measures by mid-1981 aimed at improving the balance of payments and reducing the fiscal deficit. To ameliorate the external situation, the annual depreciation rate of the peso against the dollar was slightly raised, import controls were restored while export subsidies increased, and interest rates revised upwards more regularly. As for the

135 The phenomenon of capital flight against the Mexican currency has originated a literature on the "peso problem" and speculative attacks in unsustainable fixed foreign exchange regimes. See, for instance, Krasker (1980) and Lizondo (1983). 136 Mexdollars were dollar-denominated checking and time accounts held in the Mexican banking system. 137 In the words of President Lopez Portillo, "I implicitly accepted to increase the external debt to keep dollar reserves in the central bank and the free convertibility of our currency […] In a few words, I accepted implicitly to finance capital flight." López Portillo (1988, pp. 1151-1152). Quoted in Negrete Cárdenas (1999, p. 239). See also Quijano (1983, pp. 349-353). 138 As long as the foreign exchange is expected to be steady or to move below the gap between domestic and external interest rate, borrowers would find convenient to go abroad and take advantage of the differential. In this regard, it was believed in Mexico back then that the country's oil surplus and increasing exports would allow for putting an end to the period of frequent balance of payment crisis or "external strangulation." For an analysis of the factors accounting for the rise of external private debt see Cardero and Quijano (1983, pp. 240- 255), and Molina Warner (1981) for sectorial account of the private sector's external indebtedness process.

41 public finances, the government announced a reduction in the annual budget of the public sector and introduced a number of rules and regulations to avoid future budget increases.139 The adjustment program proved, however, unsuccessful in addressing the underlying macroeconomic imbalances, and by the end of the year the flight of capital and the deterioration of the balance of payments had accentuated. By the beginning of 1982, Banco de Mexico decided to withdraw from the currency market and let the peso float freely.

After the devaluation of February 1982, the government announced new stabilization programs. They consisted of austerity measures similar to the ones implemented in the previous program, aiming to curb aggregate demand though restrictive fiscal and monetary policies.140 However, inflation continued to escalate and, fueled by political uncertainty in an electoral year, capital flight started again in mid-March. As a result, after almost five months of stability for the peso, at a parity of around 45 pesos to the dollar, the Mexican currency depreciated again in August to a maximum of 114.7 pesos per dollar. It eventually declined to 70 pesos per dollar in September after the introduction of capital and exchange controls by the Mexican central bank.

With the wave of currency devaluations and the rise of international interest rates, the peso value of outstanding dollar obligations and external debt payment services soared. Indeed, the liabilities of the countries' largest private firms and economic groups such as Alfa, Visa, Vitro and Celanese, significantly increased between February and August 1982. In the case of Alfa, for instance, liabilities doubled within this period and came to represent 1.5 times of its assets, meaning that the company virtually lacked equity.141 As for the banking sector, although banks indebted abroad - usually part of the same business conglomerates - had generally passed the exchange risk to final borrowers, they were not protected from the devaluations because their foreign currency assets were not overseas but largely in Mexico. As I discuss in the following chapters, the payment difficulties of some of its big customers and the outbreak of the debt crisis brought the domestic banking system under serious financial distress.

The balance of payments crisis of 1982 was, therefore, initially transmitted through the foreign exchange markets, eventually leading to a sovereign debt and a banking crisis. By the middle of April 1982, the Alfa Industrial Group, Mexico's main economic conglomerate and the country's largest private international debtor, suspended principal payments on its US$ 2.3 billion foreign debt. The government's first reaction to Alfa's debt payment problems, as well as the difficulties of other major economic groups and private companies, was to intervene and provide financial assistance to avoid their bankruptcy. In October 1981, the government's public works bank (Banobras) had approved a large loan for US$ 480 million to allow Alfa to meet its dollar debt obligations with international creditor banks.142 But the Mexican government was not in good shape either and had its own debt repayment issues. As the dollars need to reimburse external debt rose, foreign reserves fell and the rollover of international credit proved more difficult and the government finally went into default in August 1982.

139 See Tello (1984, pp. 77-81). 140 See Looney (1985, pp. 111-113) and Tello (1984, pp. 91-84) for a description of the February and April Stabilization Programs. 141 See Cardero and Quijano (1983, p. 281). 142 On May 14, 1982 Banco de Mexico communicated its decision to back firms with foreign debt by absorbing up to 28% of the exchange risk. Comercio Exterior, Vol. 32, N° 11, p. 1181.

Table 7. Macroeconomic indicators, 1972-82 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 Real Sector GDP (Bn. US$) 45.4 55.8 73.1 89.7 90.8 84.2 106.1 139.7 194.8 263.8 182.1 Growth rate (%) 8.5 8.4 6.1 5.6 4.2 3.4 8.3 9.2 8.3 8.5 -0.5 Public Sector Expenditure / GDP 22.9 25.8 27.0 31.9 32.0 30.0 31.4 33.0 33.5 39.7 44.5 Revenues / GDP 18.7 20.2 21.1 23.2 23.8 24.6 28.9 26.7 26.9 26.7 28.9 Financial deficita / GDP (%) 4.5 6.3 6.7 9.3 9.1 6.1 6.0 6.8 7.5 14.1 16.9 Monetary variables Monetary base (M1)b (annual growth, %) 21.2 26.7 20.1 21.1 35.7 26.3 31.6 33.7 33.4 33.3 54.1 Money supply (M4)c (annual growth, %) 17.9 14.1 18.1 26.8 14.2 31.9 35.2 38.1 43.7 48.4 75.8 Inflation (annual, %) 5.6 21.4 20.7 11.3 27.1 20.7 16.2 20.0 29.8 28.7 98.8 Interest rate Nominal (year average, %) 7.9 8.3 10.4 11.0 10.0 10.7 10.5 15.0 22.6 30.8 45.8 Real (year average, %) 2.3 -10.4 -8.0 -0.2 -12.1 -8.0 -4.9 -3.8 -5.0 1.8 -25.1 Exchange rate Nominal (year average) 12.5 12.5 12.5 12.5 15.4 22.6 22.8 22.8 23.0 24.5 57.2 Real (1980 = 100) 103.2 101.3 111.3 114.3 104.4 83.3 86.9 89.5 100.0 115.7 86.3 Foreign reserves (End year, Bn. US$) 1.42 1.52 1.53 1.67 1.45 1.96 2.30 3.09 4.0 5.03 1.83 External Sector Trade Balance / GDP (%) -2.4 -3.3 -4.5 -4.1 -2.9 -1.3 -1.7 -2.3 -1.6 -1.5 3.9 Current account / GDP -2.9 -3.2 -4.7 -5.0 -4.0 -2.4 -3.0 -3.4 -5.4 -6.2 -3.2 Capital account / GDP -0.3 0.9 2.1 1.8 1.6 -2.0 -0.8 0.6 5.8 10.1 5.5 Capital flightd (Bn. US$) -0.60 0.69 0.81 0.86 2.97 0.72 0.20 0.32 0.08 12.41 7.33

Terms of trade (PX/PM , 1972 = 100) 100 114.7 90.2 77.1 86.5 81.1 69.1 70.9 82.2 87.8 85.3 Foreign indebtedness Public sector (Bn. US$) 4.8 6.8 9.7 14.6 20.8 23.8 26.4 29.8 33.9 52.2 58.1 Private non-banking sector (Bn. US$) 2.6 3.2 4.5 5.6 4.9 5.0 5.2 7.9 11.8 14.9 18.0 Commercial banks (Bn. US$) 0.2 1.6 1.8 2.0 2.6 5.1 7.0 8.0 Total External Debt (Bn. US$) 7.4 10.0 14.2 20.3 27.3 30.6 33.6 40.3 50.8 74.1 84.1

a Financial deficit includes also “financial intermediation” expenditures, so that it is not equal to the difference between total revenues and total expenditures (fiscal deficit). b Coins and banknotes in hands of the public plus cheque accounts in domestic currency c M1 plus cheque accounts in foreign currency, short-term, up to three-month, saving instruments, medium and long term, over three months, saving instruments. d Calculated as a “residual” of the balance of payments. Source: Negrete Cárdenas (1999), Solís and Zedillo (1985) and Banco de Mexico. 43

In the months following the outbreak of the debt crisis the Mexican government announced the nationalization of the banking system. Already in March 1982, President Lopez Portillo had asked a group of advisors, which included structuralist economists Carlos Tello and José Andrés de Oteyza, to quietly work on policy options to confront the financial crisis.143 They proposed the nationalization of the banking system and the introduction of formal foreign exchange controls to stem the rising tide of capital flight and financial speculation as well as to line up the financial sector behind the economic policy goals of the government. The economic and financial authorities, Secretary of Finance Jesús Silva-Herzog and Governor of Banco de Mexico Miguel Mancera Aguayo, and the incoming President Miguel de la Madrid were opposed to these measures. Indeed, as is emphasized in the literature and their own testimonial accounts, the nationalization decision took them by surprise and they were not even consulted about it.

The expropriation of private commercial banks is a major event in Mexican economic history and, as such, is still a matter of debate and controversy. Although traditionally seen as an ideological and opportunistic political decision, prominent Mexican banking scholars, such as Carlos Marichal and Gustavo del Angel, have suggested the possibility that the nationalization was a mechanism to rescue a system on the brink of collapse.144 The presence of underlying economic and financial problems for the banking sector that are emphasised in this thesis, especially among leading domestic banking institutions, provides compelling support for that suggestion. Although addressing this issue in depth is beyond the scope of my study, in the concluding chapter I discuss how the weakness and fragility of the banking system at the time of the crisis may have contributed to the political reasons behind the nationalization decision.

2.6. Sources, data and methodology

This study draws on a wide range of secondary sources as well as first-hand historical documents from a variety of institutions that include private banks, regulatory agencies and international organizations in both debtor and creditor countries. On the debtor side, the historical archives of Banco de Mexico, Mexico's central bank and financial authority, along with those of Banco Nacional de Mexico (Banamex), one of the two largest domestic private commercial banks and the only one whose archives are open to the public, were the two most important primary sources. On the creditor side, the archives of the Federal Reserve Bank of New York, the Bank of England and the Bank for International Settlements have provided the bulk of the records and documents on which this research relies on. Most of these institutions have only recently opened their archival collections from the end of the 1970s and the 1980s, and thus they have not previously been used for this purpose. Although they are all used throughout the study, these sources' relative weight differed across chapters, as they each played different roles in addressing the various research questions.

Chapter Three draws on quantitative and qualitative records from Mexican sources. On the one hand, the Annual Reports of Banco de Mexico are used to assess the financial condition of the Mexican banking sector. They are complemented with data from the Monthly Bulletin of the Comisión Nacional Bancaria y de Seguros (CNBS) available at the Library of Banco de Mexico. They record the balance sheets of commercial banks and were used to develop the financial analysis, including principal component and regression analyses. The yearbooks of the Mexican Stock

143 See Tello (1984) and Tello (2014, pp. 261-303). 144 Marichal (2011) and del Angel (2002).

Exchange and the compendium of historical financial and monetary statistics of the Banco de Mexico have provided important additional data. In addition, the minutes of the meetings of the Governing Board of Banco de Mexico, available at its historical archive, proved central for understanding the financial challenges that Mexican monetary authorities confronted during the 1970s. These documents offer a detailed record of the discussions among board members from 1970 to 1982. The minutes of the Executive Committee of Banamex, a pioneer institution in international banking in Mexico, provided internal insights into the situation of domestic commercial banks. Together they allow for an examination of the interests of the actors involved in the internationalization process of the Mexican banking sector.

Although extremely rich and valuable for the decade leading to the 1982 crisis, these archival sources are more limited for the following years. The quality of the records and the information contained therein significantly diminished, especially since August 1982. In the case of Banamex, the last minute registered in the book of the Executive Committee's meetings dates from 27 August 1982, just before the nationalization of the bank, and the book is blank after that. Tough less dramatic, the quality of the minutes of the Governing Board of Banco de Mexico also deteriorates after the nationalization: while the Board met between 7 and 10 times per year between 1972 and 1982, records show only 4-5 meetings per year from there on. As far as the nationalization decision is concerned, there is no trace of discussions about it within these primary sources. This is not surprising given that the decision was relatively sudden and came from outside the banking sector; any historical study on the political economy of Mexican bank nationalization would likely require drawing on archives and records other than banks'.

Chapter Four is largely drawn from the archives of the Federal Reserve Bank of New York. The FRBNY played a central role during the external debt renegotiating process with Latin American countries. FRBNY officials were directly involved in the negotiations and they kept a close eye on the evolving situation of the agencies in the U.S. in the aftermath of the crisis. Many meetings between parties about the debt crisis took place at the FRBNY building. Thus, a large number of documents, consisting mainly of meeting memoranda, telex messages and, to a lesser extent, internal reports, are held in the central records of its historical archives. The collections of Antony Salomon and, in particular, of Sam Y. Cross, the contact person for Mexican negotiations and the Bank Advisory Committee, have been fundamental for developing a picture of the Mexican agencies' business and the difficulties they confronted in the aftermath of the crisis. Finally, the Report of Assets and Liabilities of U.S. Branches and Agencies of Foreign Banks of the Federal Financial Institutions Examination Council (FFIEC 002), available online in the historical commercial bank database of the Federal Reserve Bank of Chicago, provided the data to reconstruct the business models and assess the imbalances that Mexican agencies accumulated in their balance sheet through their banking activities in the U.S.

Chapter Five relies mainly on documents from the archives of the Bank of England and the Bank for International Settlements. The "Apocalypse Now" collection contains the background documentation for a study of the possible consequences of a default by a major borrowing country that the Bank of England began preparing in 1977. A number of files and discussions held with other central bankers on these issues have been crucial for identifying the importance of the interbank market as a channel for transmitting default throughout the banking system. The files of the task force that was set up in 1982 by the Bank of England to consider all aspects of the Mexican debt

45 problems and possible repercussions for banks in London and the international banking markets were also very important. These files contain information on the balance sheets of the Mexican agencies in London that complement the information from the U.S. reported in the FFIEC 002. As for the Bank for International Settlements, the documents consulted correspond to the collection of the Study Group on the international interbank market, which was also established in 1982. The Mexican agency situation and international banking stability issues were also discussed in the Standing Committee on the Eurocurrency Market and the Basel Committee on Banking Supervision in Basle. Though these files have not been disclosed by the BIS yet, a number of memorandums and notes on the meeting from several G-10 central bankers were available in the archives of the FRBNY and the Bank of England.

A number of other archives have been consulted in the development of this research though their contribution is less significant. Among these, the archives of the International Monetary Fund are the most important. The personal files of Managing Director Jacques de Larosière, Paul Mentre's papers, and the country files on Mexico of the Western Hemisphere Department also had significant references to the interbank market situation and the renegotiation of Mexican interbank debt. Internal reports of the IMF staff and minutes of the meetings of the Board, as well as office memoranda available online have also been consulted and are referred to throughout the study. It should be noted that at the time of exploring the archives, priority was given to other institutions involved in the crisis, as the records of the IMF had been already used by other scholars for similar purposes and their work used extensively in this research.145 The two last institutions that complete the list of archives explored for this research are the World Bank, in particular Eugene H. Rotberg's files and the World Debt Table Reports, and the historical records of the Latin American Division of Lloyds Bank International. The personal papers of economic historian Carlos F. Diaz-Alejandro, who had an insightful view of the Latin American financial crises of the 1980s, were also consulted in the Columbia University Library.

In addition, secondary sources were reviewed to complement these primary historical records. They include the financial and economic sections of The New York Times and the Financial Times, two newspapers which closely followed the events related to the international debt crisis. The collections of Euromoney and The Banker, who specialized in banking and international finance and were the two most important publications in the financial press at that time, as well as The Economist have been also important as sources of comparative data. These magazines have also been used to track down the activities and businesses of Mexican banks in the U.S. and London, as well as their involvement in the Euromarkets. Other financial publications consulted include Salomon Brothers' Reviews on Bank Performance, Latin American Weekly Report and the Bank of London & South American Review. An extensive number of scientific articles and books regarding the debt crisis, most of which were authored at the time of the crisis, have also been a fundamental part of this research. All these publications were either physically or electronically accessible from the library of the University of Geneva and, in particular, the Elmer Holmes Bobst Library at New York University.

Finally, interviews were held with key actors in the Mexican debt renegotiation process. The four persons interviewed were: Angel Gurría Treviño, who was Mexico's Director of Public Credit at the Secretary of Finance at the time of the crisis, and the leading external debt negotiator in the

145 See, in particular, James (1996) and Sgard (2016).

46 upcoming years; Carlos Tello Macías, the head of Secretaría de Programación y Presupuesto in 1976- 77 and the Governor of Banco de Mexico from the nationalization of Mexican banking system in September 1, 1982 until the end of that year; William R. Rhodes, a former senior international officer and senior vice chairman at Citibank, who was the leading external debt negotiator for the banks in renegotiations with Latin America and chairman of Mexico's Bank Advisory Committee; and Lee C. Buchheit, a lawyer at Cleary Gottlieb Steen & Hamilton, LLP, who participated in the drafting of sovereign debt restructuring contracts as part of that firm's representation of Latin American countries. These four men were all actively involved in the Mexican crisis from different positions and institutions; their testimonies have provided invaluable insights from insiders in developing the research for this thesis.

Chapter Three

Internationalization and Condition of Mexican Banks Prior to the 1982 Crisis*

Abstract

This chapter explores the international expansion of Mexican banks and its implications for the domestic banking system during the decade leading up to the 1982 debt crisis. In contrast to the prevalent focus in the literature on profitability and performance, I examine the asset and liability structure of the banking sector and show that there were clear signs of deterioration in its financial condition well before the onset of the crisis. Financial statement analysis reveals that the banks engaged in international lending and foreign funding were the ones with the greatest propensity to be adversely affected by this problem. The international expansion of Mexican banks emerged and developed as an exit option to domestic funding problems and increasing competition from foreign bank loans, in a context of growing needs for financing and foreign exchange in Mexico. The chapter provides new insights on the domestic and international political economy of a lending-borrowing mechanism that led to what became the largest global financial crisis since the Great Depression.

* A shorter version of this chapter is published as Sebastian Alvarez, "Venturing Abroad: The Internationalisation of Mexican Banks Prior to the 1982 Crisis," Journal of Latin American Studies, forthcoming.

3.1. Introduction

The decade that preceded the debt crisis of 1982 witnessed a sharp expansion of international banking and the overseas lending and borrowing activities of commercial banks. After the oil shock of 1973, as large amounts of petrodollars streamed into the international banking system, private financial institutions increasingly engaged in the Euromarkets and the recycling of wholesale market liquidity in the form of syndicated or direct loans to developing countries.146 Mexican banks were no exception:147 cross-border claims and obligations went from representing about 0.7 per cent of total assets and liabilities of the banking sector in 1975 to 20 per cent in 1982.148 Involvement with foreign funding and international lending created a series of new vulnerabilities for the domestic banking system, in particular related to currency risks and world financial fluctuations.

Until now, scholars have paid little attention to the internationalization of Mexican and Latin American banks during the lead up to the 1982 debt crisis.149 This is unfortunate because, as the debt and banking crises in Europe have recently illustrated, the condition of domestic banks is typically, and indeed significantly, affected by the activities they may undertake in the international capital markets.150 It is all the more surprising given that banking crises broke out in Argentina in 1980, Chile and Uruguay in 1981, Colombia and Ecuador in 1982, Peru in 1983 and, sometime later, Bolivia in 1986.151 Other countries, such as Brazil and Venezuela, although not affected by systemic banking meltdowns, still experienced banking failures in the 1980s. Mexico, unlike most of its neighbors, did not suffer from a banking crisis or bankruptcies, but this does not mean that the situation of the banking sector was positive. On the contrary, leading domestic banks accumulated serious mismatches on their books while running their international activities as I develop in further detail in the following chapter.

This chapter is concerned with the health of the Mexican banking system in connection with the international expansion of leading domestic banks between 1972 and 1982: How involvement with foreign finance and the funding strategy affected the financial position of the banks? Why Mexican banks increasingly turned to the international capital markets in the years preceding the outbreak of the crisis? To address these questions, I draw on a number of historical sources that have not been used for these purposes before. To evaluate the role and condition of Mexican banks, I rely on data from their balance sheets, as published in the Monthly Bulletin of the Financial Analysis Unit of the Comisión Nacional Bancaria y de Seguros (CNBS). The bulletin, first issued in December 1978, provides the most detailed information available on the financial condition of Mexican commercial banks. At this time, the collection is housed in its entirety in the Library of the Banco de Mexico, the country's central bank. Data from annual reports and a compendium of historical financial statistics from the Banco de Mexico are also used to complement the analysis. My account is also based on documents and records from the historical archives of Banamex - one of the two largest private commercial banks in Mexico – and, in particular, minutes from the meetings of its Executive

146 For accounts with a special focus on Latin American countries see Devlin (1989), Griffith-Jones and Sunkel (1986, pp. 57-80), Griffith-Jones (1984), Stallings (1987) and Bacha and Diaz-Alejandro (1982). 147 See especially the work of Quijano (1981, pp. 241-258) and Cardero et al. (1983). 148 Banco de Mexico, Informe Anual (Mexico City, 1975-1983). 149 Major exceptions are Diaz-Alejandro (1985) and Frieden (1987b). 150 See, for instance, Special Investigation Commission (2010) and Honohan et al. (2010). 151 See Laeven and Valencia (2008) and Sundararajan and Baliño (1991).

Committee. Finally, and with a view to complementing my analysis with the regulators' perspective, I draw on the minutes of the Banco de Mexico Governing Board, which are available at its historical archive.

Scholarly views on Mexican banks prior to the 1982 debt crisis has usually examined the performance at the industry level, and give the sense of a system operating under normal returns and low risks with limited penetration in the national economy.152 In this chapter I argue, in contrast, that the Mexican banking sector showed clear signs of a deterioration in its risk and financial position well before the onset of the crisis. Between 1977 and the early 1980s, commercial banks became twice more leveraged and much more reliant on debt instead of equity to finance the expansion of its assets. At the bank level, financial statement analysis shows that banks engaged in international lending and foreign funding were the most severely affected by these problems, displaying worse capitalization levels, a weaker liquidity position and a more instable funding base than those that operated at only a national level.

This decade of Mexican economic history is seen as one of dramatic growth of public expenditure and fiscal deficit accompanied by aggressive recourse to external indebtedness, which has been usually related to underlying political and populist purposes.153 This chapter adds the role of domestic financial institutions to the story and argues that the large scale incorporation of international liquidity by private and mixed commercial banks was at the heart of increasing risk in the banking system. Within a context of poor domestic funding and rising competition from foreign bank lending, internationalization provided Mexican banks with the possibility to secure their financing needs and protect market share. Yet, it helped the Mexican government and the central bank to accommodate the increasing financial and foreign exchange needs of the developmentalist strategy and fixed exchange rate regime.

This chapter stresses the importance of considering the situation of domestic banks during the run up to the sovereign debt crisis of the 1980s. Section 3.2 provides a historical background on the development of Mexican banks, while Section 3.3 describes the consolidation and internationalization process that began in the early 1970s. The expansion of the banking industry and its deeper involvement with foreign finance, as developed in Section 3.4, resulted in increasing fragility and new risk, as discussed in Section 3.5. Section 3.6 shows the extent to which the weakness of the system was based on asset liability mismanagement by large banks involved in foreign businesses, which, as argued in Section 3.7, limited their ability to fund themselves with less risky domestic liquidity. Despite increasingly leveraging foreign funding for expansion, I suggest that while these banks were not necessarily fueling the crisis through bad lending, they became a dangerous force of instability because of their greater risk positions.

3.2. The historical context of the Mexican banking industry

From the postwar years up to 1982, the Mexican banking sector experienced a stable and continuous enlargement of its intermediating activities. After the collapse and destruction of the financial system during the Mexican Revolution at the beginning of the twentieth century and its subsequent reorganization and reconstruction during the 1920s,154 Mexico's banking system would

152 See del Angel (2002, pp. 18-62), del Angel (2005), and Haber and Musacchio (2014). 153 See, in particular, Negrete Cárdenas (forthcoming) and Bazdresch and Levy (1991). 154 See Anaya Merchant (2002) and Goldsmith (1966) respectively.

50 enjoy an unprecedented level of growth during the following decades. As documented by del Angel (2002), both in terms of the number of the institutions established and in the volume of transactions and lending activities they carried out, the domestic banking sector underwent a powerful expansion from the 1930s on.

The foundation for the Mexican banking system was laid in the two decades that followed the revolution. One of the most important measures was the creation of a central bank in 1925, Banco de Mexico, which would become the institution at the center of the financial system. The Comisión Nacional Bancaria, later renamed as the CNBS, was created the previous year as the supervisory agency for the banking sector. In terms of the legal framework, new legislation and amendments were passed during the 1920s and 1930s, and the legal rules for Mexican banking were further elaborated with the adoption of the Banking Law of 1941. These legal innovations established a system of specialized banking in Mexico that ruled the banking industry until 1975, when the legislation went through its first substantial modification.155 Around the same time, the creation of development banks, such as Banco Nacional de Crédito Agrícola in 1926 and Nacional Financiera in 1934, were also major events in the development of the modern Mexican financial system.156

Figure 1. Evolution of the Mexican banking sector, 1925-1982

40%

30% Total banking assets / GDP

20%

10%

1925

1928

1931

1934

1937

1940

1943

1946

1949

1952

1955

1958

1961

1964

1967

1970

1973

1976 1979 1982

Source: Estadísticas Históricas de México, INEGI (1985).

The Mexican banking sector grew and gradually consolidated during the postwar era (del Angel, 2010). The Banking Law of 1941 divided Mexican banks into three main types of financial intermediaries: mortgage banks, financieras (similar to industrial banks), and deposits and savings banks, otherwise known as commercial banks.157 Each of these had a specific set of operational boundaries, their own balance sheet, and was a separate legal entity. Despite strict regulatory constraints in terms of their funding instruments, reserve requirements and lending activities,

155 See del Angel (2002), Chapter 4.4. 156 Development banks were state-owned financial institutions that differed legally from the private intermediaries and were not regulated by the Banking Law. 157 The other, less important, financial firms were thrift institution, trust organizations and residential mortgage banks. See Acosta Romero (1978) for a detailed account of the different financial entities contained in the Code.

51 commercial banks managed to enlarge the scale of their operations beyond the scope of their original grant of authority. As Gustavo del Angel explains, they were able to "bypass these restrictions by taking over other intermediaries and using them as their legal mandates to channel funds with greater flexibility."158 By the 1960s, commercial banks and their associated firms had formed financial conglomerates or holdings - legally recognized with the banking law amendment of December 1970 - which became, in general, the financial units of larger economic groups.159

However, the consolidation of the banking sector was not confined to the amalgamation among financial intermediaries. It also strengthened their position within the domestic economy. The development and enlargement of banking activities during this period took place in a time when the Mexican economy experienced a remarkable expansion, with growth rates that often exceeded those observed in the banking sector. Particularly between the late 1950s and early 1970s, the banking sector deeply penetrated the national economy and strengthened its role in the economic development of the country. While in 1958, total assets of private financial intermediaries represented approximately 15 per cent of the GDP or US$ 1.5 billion, by the end of 1972 they were as much as 34.1 per cent or US$ 15.3 billion (see Figure 1).160 This rapid pace of expansion came along with an increased concentration of the banking industry and consolidation among the country' leading banks.161

Two private banks stood out from the rest of the financial institutions in the sector. The first of these was Banco Nacional de Mexico, known as Banamex. Founded in 1884 as the result of a merger between Banco Mercantil Mexicano and Banco Nacional Mexicano, Banamex was one of the oldest banks in the country and a leading player in the domestic commercial banking market. The other leading player was Banco de Comercio, known as Bancomer. Created during the early 1930s in Mexico City, Bancomer would quickly expand its activities to the national level, becoming a major actor in the banking industry and the main competitor of Banamex. Whereas originally established as commercial banks, these two institutions progressively acquired, created, or grouped with other financial firms, such as financieras, mortgage banks, brokerage houses, leasing agencies, insurance companies, among others, and covered a vast network of financial services, becoming the two biggest banking groups in the country.

The position of Banamex and Bancomer, as well as of other leading banks, and the concentration of the industry were further strengthened by the reorganization of the Mexican banking system during the mid-1970s. With the passage and enactment of the Multiple Bank Law in 1975-1976, banking groups were legally permitted to consolidate their different financial institutions into one single banking entity, called a multiple bank. The reform was aimed at transforming the previous arrangement, which had included a great number of specialized financial companies. The goal of this law was to create a universal banking system composed of a small number of large banks. By the beginning of the 1980s, multiple banks represented as much as US$ 54.6 billion or 91 per cent of the total assets and liabilities of the entire banking sector, and Banamex and Bancomer alone accounted for almost a half of this, both of them having similar shares (see Figure 2).

158 del Angel (2002, p. 95). 159 See Hamilton (1983) for a description of economic and financial groups in Mexico around that time and del Angel (2006) for detailed account of the connections between commercial banks and business groups. 160 See del Angel (2010, p. 637), Gráfica 15.1. When compared to other similarly sized economies, such as those of Argentina and Brazil, penetration of the Mexican financial system was relatively modest. 161 See del Angel (2002), Chapter 6.

Figure 2. Composition of the banking market in the early 1980s Total Assets in December 1980, in millions of dollars

Source: Bulletin of the CNBS and Banco de Mexico's Annual Report

Mexico's third largest private bank was Banca Serfin, which controlled nearly one tenth of the country's total banking market. Serfin was created in July 1977, after the huge conglomerate Valores Industriales SA (VISA group) acquired Financiera Aceptaciones, Mexico's largest industrial bank, and merged it with Banco de Londres y México and three regional banks, forming Banca Serfin. Multibanco Comermex was the fourth largest bank, followed by Banco Mexicano-Somex and Banco International, both of which were majority-owned by the Mexican government. As for the remainder of the banking sector, it was mainly made up of multiple banks, but there were also a number of deposits and saving banks, financieras and mortgage banks which had not evolved into multiple banks and kept their previous legal standing; together they accounted for over a quarter of the Mexican banking system's total assets (see Figure 2).

A special remark needs to be made with respect to foreign banks and their participation in the Mexican domestic banking market. Unlike during Porfirian times, the financial system reconstructed in the aftermath of the Revolution was almost entirely national. With the Law of Banking Institution of June 1932, the operations of foreign banks in Mexico were legally forbidden. Citibank was the main exception with full branch facilities and commercial banking activities in the country. The other foreign banks were only allowed to open representative offices that, although not permitted to collect savings from the public, could associate with national banks and leverage this network of correspondent banks to conduct business. The presence and activities of foreign banks nevertheless remained relatively discreet until the late 1960s and 1970, when they began to open offices and engage in lending activities in Mexico.162

3.3. Incursion into foreign finance, 1972-1976

A salient feature in the historical evolution of the Mexican banking industry during the post-World War II period is the major disruption it suffered during the 1970s. As evident in Figure 1, after decades of stability, expansion and increasing penetration in the national economy, the rise of the commercial banking sector came to a halt around 1972. Following the historic peak of 34.1 per cent

162 See Sánchez Aguilar (1973) and del Angel (2002), Chapter 5.4.

53 that year, the banking system's total assets relative to the GDP progressively diminished until 1977, when it reached 25.8 per cent, the lowest value of the decade. This trend, which Mexican scholars have named the 'financial disintermediation process',163 would be reversed in coming years and by the early 1980s, commercial banks had regained what they had lost.

The underlying factor behind the contraction of banks' total assets was a slowdown in bank fundraising. Estimates from Quijano (1987, p. 177) show that during the period from 1956 to 1960 and 1964-1970, while commercial banks' domestic funding increased, respectively, by 10.2 and 18.1 per cent annually in real terms, the average annual growth rate between 1971 and 1978 was a much more modest 1.7 per cent.164 By 1971-72, about 93 per cent of the banks' funding base consisted of local deposits and savings from the private and public sector, 4 per cent were transactions between domestic financial institutions and the remaining 3 per cent was made up by other domestic liabilities.165 This funding structure and weak performance in terms of domestic funding, in the context where the Mexican economy expanded at an annual rate of 5.6 per cent, implied a relative decline in the financial resources available to banks166

The minutes of the Executive Committee of Banamex witness the serious difficulties the banking sector experienced in terms of domestic financial resources. During their weekly Wednesday meeting, Banamex's General Director Agustin Legorreta would periodically talk about the scarcity of cash resources and the slow evolution that deposit accounts showed during 1971 and 1972. Funding difficulties went beyond the case of Banamex, since, as he stressed, "the situation was not exclusive to [their] bank, but general to all the credit institutions of the country, particularly affecting small institutions, which have been failing to meet their legal deposits."167 Bank fundraising difficulties were also a matter of concern for national financial authorities who, in looking for a solution, introduced a reform to the Banking Law in December 1973, which authorized Banco the Mexico to "equip the Mexican banking system with more and more flexible fundraising instruments."168

The causes of the banks' domestic fundraising problems were related to inflation and the interest rate policy followed by the central bank. For many years, until the late 1960s, inflation in Mexico was low (below 4 per cent annually), but it had started to rise in the early 1970s, with rates oscillating between 4 and 6 per cent, and became a real problem beginning in 1973, when prices rose by 21.4 per cent over the course of the year. High inflation was a problem for banks because of the difficulties in adjusting interest rates on deposits and savings instruments upwards. As a result, 1972 was the last year with positive real interest rates. As shown in Table 1, from 1973 onwards, nominal interest rates were persistently below inflation rates. In Mexico, as in many other Latin American countries, this was a period of heavy financial regulation, or financial repression, and interest rates were not market-determined, but instead, established by the monetary authority. In this respect, as

163 See, in particular, del Angel (2010, p. 637) and Quijano (1987, pp. 170-180). 164 Buffie and Krause (1989) estimate that the total stock of real bank funds fell 13.3% from 1973 to 1976. 165 Banco de Mexico, Annual Report of 1972, Table 20, p. 73. 166 For the 1956-1960 and 1964-1970 periods, the average annual growth rates of the GDP were 6 and 6.4% respectively. 167 Banamex Archive, Libro No. 2 de Actas de la Comisión Ejecutiva, August 11, 1971 Meeting. 168 Acta No. 2406, February 1974. New saving regimes were then introduced by Banco de Mexico, that, in the words of Governor Fernández Hurtado, aimed to "provide domestic savers with a wider range of investment opportunities, in terms of timing and performance, and tried to encourage fundraising by putting emphasis, not so much on important increases in the return on investment, but on a more adequate timing structure, covering the gap that was currently observed in that structure and giving the Mexican banking system a more competitive position, without representing excessive costs for the institutions."

Buffie and Krause (1989, p. 146) observed, "whereas real deposit rates were maintained at positive levels throughout [the era of Stabilizing Development],169 after 1972 this policy was allowed to lapse."170

Table 1. Macroeconomic and banking indicators, 1972-1976 In percentages

1972 1973 1974 1975 1976 Macroeconomic variables

GDP real growth rate 8.5 8.4 6.1 5.6 4.2

Inflationa 4.9 12.1 23.8 14.9 15.8

Devaluationa 0 0 0 0 23.2

Fiscal deficit/GDPb 4.5 6.3 6.7 9.3 9.1

Current account/GDP -2.9 -3.2 -4.7 -5.0 -4.0

External indebtedness/GDPc 16.4 18.1 19.8 23.0 29.3

Banking sector

Interest rated 7.9 8.8 11.5 12.0 12.1

Total lending/GDP 19.3 17.4 14.9 14.7 14.0

Domestic funding/GDP 32.1 30.0 26.9 27.4 25.7 Foreign borrowing/Total liabilities 0 0 0 0.7 2.7

a Annual rate. b Public sector borrowing requirement (includes public financial sector) from Negrete Cárdenas (1999). c Total external debt from World Bank's World Development Indicators. d Average annual interest rate of bank's domestic fundraising instruments from Banco de Mexico' Historical financial statistics.

Source: Banco de Mexico and INEGI.

The resulting funding difficulties gnawed away at the lending capacities of the banks. Bank lending was further affected by the decision made by financial authorities to increase reserve requirements, which, according to Buffie and Krause (1989), rose from about 31.3 per cent in 1970 to 51.1 per ent in 1976.171 As a result of these events, between 1972 and 1976, the bank loan portfolio declined

169 Stabilizing Development - desarrollo estabilizador - is the name given to the economic program that prevailed in Mexico between the beginning of the Lopez Mateos Administration in 1958 and the end of the Diaz Ordaz Administration in 1970, in which the country experienced high levels of growth and low inflation. See Izquierdo (1995) or Ortiz Mena (1998). 170 In fact, the minutes of the Governing Board show that determining interest rates was pretty much made based on the evolution of interest rates in international capital markets, in particular in the U.S. and the Eurodollar, in order to keep domestic instruments at competitive levels and avoid outward flows of capital. Although inflation was a matter of discussion, it didn’t come up during the talks on interest rate policy. See, for instance, Actas No. 2401 (July 1973) and 2407 (April 1974) and 2420 (January 1976). This is consistent with the position of the central bank, which addressed banks' fundraising instruments through the introduction of new saving instruments and not by raising interest rates, as previously mentioned. 171 The Reserve requirement system was a very complex network of ratios, in which different specialized institutions, such as mortgage banks, financieras, deposit banks, etc., and each of the fundraising instruments was subjected to specifics coefficients. There were, indeed, several increases in reserve requirements between 1970 and 1976, which were motivated as anti-inflationary measures. These increases also provided additional resources to finance the public sector. See, in particular, Acta No.2410, October 1974.

55 from 19.3 to 14 per cent of the GDP (see Table 1).172 This contraction of domestic bank lending was ill-timed, as the economy and demand for credit continued to grow in Mexico, stimulated by an aggressive expansion of public spending and large fiscal deficits. 1972 was, indeed, the beginning of a decade in Mexican economic history underpinned by populist macroeconomic programs and dramatic expansionary fiscal policy.173 Along with other Latin American economies, Mexico was undertaking the final stages of the import-substituting industrialization (ISI) process and bank lending was a major piece of the funding strategy.174 For the Mexican banks, as Banamex' General Director pointed out, the lack of funding was exacerbated by the rising demand for credit that banks faced from both the private and, especially the public sector, which, as he stressed "was exerting strong pressures on the credit institution for financing."175

In a context of rising demand for credit and insufficient domestic supply, the financing that domestic banks were not able to provide was supplied by foreign banks. The country had been borrowing in the international capital markets for some time prior to the early 1970s, but the boom in international lending began in earnest with the oil shock of 1973 and the rise of the petrodollar recycling process in which Mexico, and Latin American generally, participated.176 Green (1998) shows how Mexico's external debt, both from the public and private sector, started to accelerate vigorously beginning in 1973. For instance, while between 1970 and 1972 the Mexican public sector had raised US$ 2.08 billion in the international capital markets, in 1973 alone, it borrowed US$ 2 billion, and during the following three years, it borrowed a total of US$ 12.5 billion.177 To put it differently: in each successive year between 1973 and 1976, the Mexican public sector took on five times more foreign loans than during the entire 1970-73 period. Of these amounts, about four fifth came from international banks operating in the Euromarkets.

In the eyes of leading Mexican banks, the increasing role of foreign banks in lending to Mexico was a major threat. Their position was that, as Legorreta explained to Committee members, if they were not involved in the international capital markets "they would be condemned to be a mere supplement to foreign banks," which were ready to provide the funding the country was demanding to finance its economic development at very attractive interest rates.178 It is within this context of poor domestic funding, rising competition, and the loss of market share to foreign institutions in terms of lending, that Mexican banks turned to the international capital markets. In fact, as the minutes of Banamex's Executive Committee show, "the impossibility of meeting the domestic demand for credit with domestic resources" was a crucial factor in the decision of the bank to look at the Euromarkets and to get involved in international banking.179

172 The level of 1971-72 was a historical peak after years of vigorous expansion since the mid-1950s when the ratio of banks' loan portfolio to GDP was around 7.5%. See Del Angel (2002), Chart 2.2, p. 24. 173 See Negrete Cárdenas (1999) and Cárdenas Sánchez (1996, pp. 86-119). 174 See Bértola and Ocampo (2012). 175 Banamex Archive, Libro No. 2 de Actas de la Comisión Ejecutiva, August 11, 1971 Meeting. By way of an example, in the August 29, 1973 meeting, Agustin Legorreta informed the Committee of a conversation he had had with Ernesto Hernández Hurtado, Governor of the Bank of Mexico, concerning the financial needs of the Federal government. Hernández Hurtado had told him that the government would need 2 billion pesos (approximately US$ 160 million) to cover public spending until the end of the year and that he was expecting leading private banks to provide the financing. See Acta No. 2399, April 1973. 176 See Devlin (1989). 177 See Green (1998), Table I.12, p. 42. 178 Banamex Archive, Libro No. 3 de Actas de la Comisión Ejecutiva, February 9, 1971 Meeting. 179 Ibid.

The case of Banamex, a pioneer and leading financial institution in Mexican international finance, illustrates the path Mexican banks followed to reach the Euromarkets. During early 1972, the Executive Committee started to discuss alternatives to enter the international capital markets and, eventually, decided the best way to do it was through the creation a London-based multinational bank associated with large foreign banks.180 The International Mexican Bank, known as Intermex, was finally opened in April 1, 1974, with Banamex as the main shareholder holding 38 per cent of the bank's initial capital.181 Around that same time, Bancomer and Banco de Londres y Mexico (which later became Banca Serfin), the two other largest Mexican banks, also founded their own consortium banks, the Libra Bank and the Euro-Latinamerican Bank (Eulabank) respectively.182 These banks, particularly Intermex, would take a predominant role in international lending to Mexico during the upcoming years.183 But the internationalization of Mexican banks could not have been accomplished, let alone started, without the consent and support of the federal government and the country's financial regulators. In 1974, an important amendment to the banking law was passed that permitted Mexican banks to access international capital markets.184 More specifically, the amendment explicitly empowered domestic banks to participate in the capital stock of foreign financial institutions and to open agencies and branches upon receiving authorization from the Secretary of Finance. For Mexican banks willing to get involved in the Euromarkets, this was a major reform since until that point, domestic bank participation in international finance was not contemplated in the national banking law. These changes aligned with other measures taken by the Mexican financial authorities at that time that tended to soften controls on banks and liberate the domestic financial sector.185 For Mexican policymakers, the involvement of domestic banks in foreign finance was desirable for at least two reasons. First, the public sector was confronting a steep increase in the fiscal deficit - it rose from 4.5 per cent of the GDP in 1972 to 9.1 per cent in 1976, see Table 1, and thus was in need of financing. As previously noted, the government had been putting pressure on the banks for financing that banks were not able to afford with domestic financial resources. Second, given the fixed exchange rate and growing inflation, the Mexican economy was also suffering from capital flight and a worsening current account, which both led to a deterioration in the country's balance of payment position. In this context, letting banks access international money markets promised to be useful, not only because it would provide them with additional funds to finance the government (and the ISI policies), but also because foreign exchange could help to bridge the peso gap. A

180 The alternatives under consideration were: a) to open a branch in London, b) to organize a society in London controlled by the bank, and c) to associate with other European and American banks to create an institution in which Banamex held an important share. In the opinion of the Director, the last option "was probably the most convenient solution because of the advantages it represent[ed], in terms of prestige and appropriate contacts among other." Banamex Archive, Libro No. 3 de Actas de la Comisión Ejecutiva, February 2, 1972 Meeting. 181 Bank of America (20%), Inlat (13%), Deutsche Bank (7.25%), Paribas International (7.25%), The Dai-Ichi Kangyo Bank (7.25%) and Union Bank of Switzerland (7.25%) were the other founders and shareholders of Intermex. 182 The arrival of Mexican banks in London was part of a general trend among a handful of developing countries' banks that established consortium banks there between 1972 and 1974. See Consortium Banks on Course, The Banker, February 1976, pp. 170-171. 183 Negrete Cárdenas (1999), Table B17: Leaders in Bank Syndicated Loans to Mexico, 1973-82, pp. 400-404. 184 Borja Martínez (1978, pp. 431-342). 185 See Kaminsky and Schmukler (2003).

57 currency crisis would nevertheless break out in 1976, leading to the first devaluation of the peso after 22 years of stable parity with the dollar.186

3.4. From domestic to foreign funding, 1977-1982

1977 marks a turning point in the evolution of the Mexican banking system during the 1970s. After half a decade of contracting activities and losing presence in the national economy, the banking industry started to recover. Total assets, which represented 25.8 per cent of Mexican GDP in 1977, grew to 32.4 per cent in 1979, and 35.9 per cent by 1981 (see Figure 1). In terms of lending, the loan portfolio of domestic banks rose from 12.8 to 19.2 per cent of the GDP between 1977 and 1981 (see Table 2). By the beginning of 1982, the domestic banking sector had not only regained the ground it has lost during the years of financial disintermediation, but its weight in the national economy was even greater than the historic highs it had reached in 1972.

The revival of the domestic banking sector was possible because banks succeeded in reversing the trend of previous years and increased their funding base. In December 1976, Gustavo Romero Kolbeck was named Governor of Banco de Mexico, and from the beginning of his mandate, he emphasized the need to strengthen the funding and financial capacity of the domestic banking system. 187 In order to so, Banco de Mexico proceeded to create and restructure financial instruments available to attract savings from the public, with an emphasis on encouraging long term funding in the national currency. Nominal interest rates were also revised upwards, and in 1977, "for the first time since 1972, the interest rate of longer term deposits became positive in real terms."188 Interest rates on domestic fundraising instruments would be periodically revised, especially beginning in late 1979, and adapted to inflation levels. As a result, by the end of 1977, domestic funding of the banking sector has increased by 18.6 per cent in real terms, and would continue to expand at a rate of 8.8 per cent per year between 1978 and 1982.

At that time, expanding the lending capacities of the banks was important both because it was believed necessary to support economic growth and also because it helped companies with liquidity problems. With the devaluation of 1976, a number of firms indebted in dollars abroad experienced payment difficulties and relied on financing from domestic banks to cover their financial needs. Romero Kolbeck explained to Banco de Mexico board members that the mechanism "observed in the business sector consisted of obtaining credits in national currency and using those amounts to pay liabilities in foreign currency."189 During the last four months of 1976, Banco of Mexico had already been supporting the banking system to assist companies with cash flow problems, and in

186 Negrete Cárdenas (forthcoming) argues that the crisis of 1976-77 was, in reality, a short-lived debt crisis that passed mostly unnoticed. 187 See Actas No. 2430 (March 1977) and 2438 (July 1977) 188 Banco de Mexico, 1977 Annual Report, p. 49. This was not, however, the case for all savings instruments, since, on average, interest rates continued to be lower than the annual rate of inflation (see Table 2). Two factors are worth mentioning with respect to the new interest rate policy adopted in 1977. A clear separation was introduced between savings instruments in foreign and domestic currency. Since March 22, 1977, interest rates for dollar saving instruments were determined daily at one point above the interest rate of its equivalent instruments in the Euromarkets. As for savings instruments in pesos, interest rates were to be revised more frequently, and determined as ceilings and not as fixed rates, like before. 189 Acta No. 2432, July 1977.

January 1977, it made up to 1.7 billion of pesos (US$ 770 million) available to the banks in order to maintain their credit capacity.190

A step forward in improving bank lending capacity was the reform of the reserve requirement regime. On April 1, 1977, a unified reserve requirement system replaced the complex existing structure of multiple requirements with one single reserve ratio for all liabilities in national currency. The process of homogenization of reserve requirements came along with a general reduction of their levels, which was more important for multiple banks than for specialized financial institutions. Up until March 1977, the average reserve requirement ratio had been about 50 per cent; that April it was unified and reduced to 39.5 per cent for multiple bank institutions and later reduced to 37.5 per cent in August of the same year.191 With these measures, the financial authorities aimed to both simplify a system that has become very complex and also expected to release considerable amounts of resources that would become available to banks for lending.192

Though domestic funding was important, the recovery of the banking sector after 1977 was also based on increasing use of foreign finance by Mexican banks. In 1975, obligations of the commercial banking system with foreign creditors represented US$ 176.9 million or 0.7 per cent of the sector's total liabilities (see Table 1). Although clearly limited in scale, reliance on external borrowing as source of funding escalated considerably thereafter. While in 1977, foreign obligations accounted for US$ 630 million or 3.1 per cent of the funds raised by the banking sector, they progressively increased up to 1982, when they reached US$ 8.53 million or 20.2 per cent of the sector's total liabilities (see Table 2). This consisted of wholesale interbank funding that leading Mexican banks accessed through their expanding network of foreign agencies and branches in the main international financial centres as developed in the next chapter.

The example of Banamex is representative of the general pattern in this regard. After its first incursion to the Euromarkets as shareholder of Intermex, Banamex eventually decided to become involved in the international capital markets on its own. In April 1974, after reporting his presence at the opening of Intermex in London, Banamex's Director brought a proposal to open a branch in Los Angeles, California to the attention of the Executive Committee, as part of the development of the international operations of the bank. 193 The project consisted of replacing the existing representative office with a branch, and the reasoning given was that this upgrade made it "eligible for loans from American banks that could be invested in the United States, Mexico or in another country."194 The Los Angeles branch opened in February 1975, and soon after, the representative office of the bank in New York was also converted to branch status and in 1978 a new branch facility was created in London. Like Banamex, other leading Mexican banks developed a similar network of foreign agencies and branches, including Bancomer and Banca Serfin and, to a lesser extent, Multibanco Comermex, Banca Somex and Banco International. By doing so, these banks gained

190 Banco de Mexico, 1977 Annual Report, p. 45. See also Acta No. 2428, January 1977. 191 The process of conversion into a universal banking system began in December 1976 and by the end of 1977, multiple banks accounted for 71% of the resources of the banking system. Ibid, p. 42. 192 In contrast, while also unified, the reserve requirement for liabilities denominated in dollars was raised from 30 to 75% cent as a measure to prevent the dollarization of the banking system. See Acta No. 2432, July 1977. 193 Banamex Archive, Libro No. 6 de Actas de la Comisión Ejecutiva, April 17, 1974 Meeting. 194 Ibid. The other advantages mentioned in support of branch status were the capacity to intermediate between U.S. residents doing businesses in Mexico or Mexican residents that needed to make payments in the U.S., as well as in handling collections in both countries. They were not authorized, however, to take deposits from the public.

59 access to source of funds overseas and became able to borrow in the international wholesale interbank markets.195

Table 2. Macroeconomic and banking indicators, 1977-1982 In percentages

1977 1978 1979 1980 1981 1982 Macroeconomic variables

GDP real growth rate 3.4 8.2 9.2 8.3 7.9 -0.5

Inflationa 29.1 17.5 18.2 26.4 27.9 58.9

Devaluationa 46.5 0.6 0.3 0.8 6.7 121.9

Fiscal deficit/GDPb 6.1 6.0 6.8 7.5 14.1 16.9

Current account/GDP -2.4 -3.0 -3.4 -5.4 -6.2 -3.2

External indebtedness/GDPc 35.9 32.6 29.6 26.5 32.1 52.9

Banking sector

Interest rated 14.0 15.9 17.5 24.3 31.8 46.1

Total lending/GDP 12.8 13.3 17.8 17.9 19.2 23.1

Domestic funding/GDP 24.1 27.2 28.7 28.7 30.5 34.5 Foreign borrowing/Total liabilities 3.1 8.7 9.1 12.0 13.0 20.2

Initially permitted by the amendment of the Banking Law in 1974, the participation of Mexican banks in foreign financial markets was further favored by the reforms embodied in the Multiple Bank Law. Since the law implied the consolidation of assets of many financial institutions into a larger one, as Cardero et al. (1983, p. 96) explain, multiple banks were in much better shape to place themselves in the international capital markets as bigger players with an impressive overall volume of business. Size and name seemed to have been determining factors in the ability of banks to raise fund in the international money markets.196

Financing was a fundamental underlying component of Mexico's economic growth between 1977 and 1982. The development of the economy relied on recently discovered oil reserves and great amounts of investments were needed to exploit this wealth.197 Increasing demand for financing

195 Chapter Four describes the international business model that Mexican banks developed through their foreign agencies and branches. 196 The words of Marquis Gilmore, agent and senior Vice President of Banco International's agency in New York, are illustrative: "since the parent bank is the smallest of the six Mexican banks operating in the U.S. and since it is not well known, the agency is unable to raise funds in the money markets directly, but has been able to obtain the needed funds through its international bank correspondents." FRBNY Archive, File: Mexican Government 1917-1984, Office Memorandum, August 25, 1982. See BIS (1983) for a description of the lending and borrowing practices in the international interbank market. 197 See Cárdenas Sánchez (1996).

60 came from both the public sector, whose public deficits were reaching extremely high levels (see Table 2), and also from the private sector, which had considerably expanded its borrowing towards the end of the decade.198 For Mexican banks, domestic sources proved insufficient to satisfy the loan needs of both the public and private sectors, which encouraged them to expand fundraising abroad. The words of Banamex' General Director during a meeting of the Executive Committee in March 1980, illustrate the situation: "[the bank] could not serve nor meet the needs of their big clients if [they] could not count on resources coming from abroad."199

3.5. The underlying weaknesses of the banking industry

Recourse to foreign borrowing to overcome domestic financing limitations created new vulnerabilities for the Mexican banking system. As previously mentioned, external fundraising instruments were made up of credit lines or interbank placement from foreign banks operating in the Euromarkets or the U.S. international money market and were mainly conducted through overseas banking offices. Maturities were usually very short, between overnight and six months, and consisted of dollar denominated instruments, arranged at variable interest rates. Increasing reliance on foreign interbank funding exposed the domestic banking system not only to currency risk but also to funding strains that could develop in the international financial markets. The larger the weight of foreign interbank liabilities in the funding structure, the larger the exposure to currency risks and liquidity or funding strains in the international money markets.

The involvement of Mexican banks with international finance through London-based consortium banks did not create such vulnerabilities. Although these banks relied heavily on international wholesale liquidity for funding,200 they were separate institutions with their own capital base and legally independent from shareholder banks. Unlike foreign agencies and branches, their borrowing and lending activities were not consolidated in the balance sheets of parent banks. Only the shares were exhibited on the books of owner banks (asset) and standby facilities that the Bank of England normally asked them to commit as support in case of emergency (liability).201 But since this was not a formal requirement, these back-up lines of credit were not legally binding. Had the consortium bank become distressed and wanted to draw on these lines, shareholder banks could refuse to grant them, if for some reason they found it inconvenient.202 Thus, for the Mexican banking system, the international operations of its consortium banks were not a major source of vulnerability. In

198 Private sector foreign debt increased from US$ 7.1 billion in 1978 to US$ 15 billion in 1980, while in the same period, the public sector external debt rose from US$ 26.2 billion to US$ 33.8 billion. 199 Banamex Archive, Libro No. 12 de Actas de la Comisión Ejecutiva, March 12, 1980 Meeting. 200 For instance, in 1975, 15% of Intermex's liabilities consisted of primary deposits from clients, while the remaining 85% were funds raised in international money markets. See Banamex Archive, Libro No. 8 de Actas de la Comisión Ejecutiva, February 18, 1976 Meeting. 201 For instance, in November 1974, Banamex had committed US$ 9 million in favor of Intermex as stand-by credit line, which had been authorized by Banco de Mexico, and was to be registered "as a credit at a term of two years, with the possibility to renew it one or more times, and that would be used only in case of emergency to strengthen the liquidity position of Intermex." Banamex Archive, Libro No. 7 de Actas de la Comisión Ejecutiva, November 6, 1974 Meeting. This stand-by credit line was later increased to US$ 15 million in 1977 and to US$ 15.8 million in 1979, but was never used by Intermex. 202 As stressed in a Bank of England report, it was possible that when distressed banks "most wanted to draw on these lines, the banks granting the standby lines might decide to cut back or cancel them, in order to safeguard their own position." Bank of England Archive, Apocalypse Now – 3A143/1, June 1980.

61 contrast, the direct involvement of domestic banks in international banking through their headquarters in Mexico or their agencies and branches overseas was of growing concern.203

Adding to the vulnerabilities, the Mexican banking system experienced a deterioration of the domestic funding base. On one hand, between 1978 and 1982, liquid saving instruments from the non-banking private sector persistently reduced their share as source of funding. More specifically, as evident in Figure 3.1, the decline of checking account deposits explains this contraction. While in 1977, sight deposits represented over 20 per cent of the funds of the sector, by the early 1982 their weight has fallen to just under 10 per cent. The rest of the liquid fundraising instruments, made up of savings accounts and a variety of term deposits with short maturities (one month or less), were maintained at around 17-18 per cent of total liabilities. In terms of the funding structure of the banking system, domestic liquidity was substituted with international liquidity: declining sight deposits, which were the least expensive non-equity source of funding for commercial banks, were replaced by larger interbank foreign borrowing that entailed higher risks (see Figure 3.1).

Figure 3. Funding structure of the Mexican banking system, 1978-1982 Figure 3.1. Liquid funding instruments Figure 3.2. Long Term Deposits Percentage of Total Liabilities Percentage of Total Liabilities

Source: Banco de Mexico's Annual Reports (several issues)

On the other hand, although stable in terms of the volume of liabilities, the internal composition of long-term funding also suffered important transformations (see Figure 3.2). While time deposits of one year and over accounted for 70 per cent of total illiquid liabilities in December 1978, they represented 26 per cent by end-1981 and around 18 per cent in 1982. Conversely, time deposits with maturity of less than one year accounted for 30 per cent in 1978, but represented as much as 82 per cent by 1982. Within this category, shorter term deposits (three-month deposits) were the more dynamic component, increasing their share from 12 to 47 per cent over the period. The concentration of three-month term deposits was the result of a yield structure in which nominal interest rates of saving instruments at different maturities were similar, and even converged

203 This makes a big difference in the exposure of the banking system to the 1976 and 1982 crises. In the 1976 crisis, when most of the Mexican bank involvement in the international capital markets had been done through consortium banks — as the network of foreign agencies and branches had not yet been developed, the exposure of the banking system was much more reduced than in 1982, when the banks were directly involved in international finance and foreign debt weighted heavier on their books.

62 throughout the period.204 It thus seems logic that, in a context of rising inflation and diminishing spreads between long and short term deposits, people would prefer the saving instrument with shorter maturity. For the banking system, this shortening of the maturity structure of term deposits implied a weakened and less stable funding base.

Apart from these changes in funding, the build-up of risks in the banking system was also associated with a more general deterioration of the balance sheet structure. Figure 4 displays the evolution of the leverage ratio and the debt-to-equity ratio for the Mexican domestic banking system: they both show increasing levels of risks between 1977 and 1982. The debt-to-equity ratio, calculated as the ratio between total liabilities and paid-in capital, indicates the extent that the banking sector relied on debt instead of equity to finance the expansion of its assets. The ratio passed from 33.4 per cent in the early 1977 to around 60 per cent in 1980, and to 82.8 per cent by the beginning of 1982: a 2.5 increase in the 5-year period of exchange rate stability. Because of the growing share of dollar liabilities with the foreign sector, ratio values would become much higher after the devaluation of the peso in February 1982.

Figure 4. Risk indicators for the Mexican banking system, 1977-82

120 4.5

Leverage ratio (right scale) 4.0 Debt-to-equity ratio (left scale) 3.5

80 3.0

2.5

2.0 percentage

40 1.5

1.0

0.5

0 0.0

I-77

I-78

I-79

I-80

I-81

I-82

II-77

II-78

II-79

II-80

II-81

II-82

III-77

III-78

III-79

III-80

III-81

IV-77

IV-78

IV-79

IV-80 IV-81

Source: Banco de Mexico's Annual Report (several issues)

Although these figures give a clear sense of the deteriorating trend they say less about how dangerous the values were. The case of Intermex, which operated in London under the regulation of Bank of England, provides an interesting benchmark. At the end of 1974, the year of its creation, Intermex had a capital base of £ 2.5 million and liabilities that represented 10 times its capital. By 1976, its liabilities had increased up to 19 times the level of paid-in capital, which pushed shareholder banks to increase its capital to £ 5 million, diminishing the debt-to-equity ratio to 14.205 In 1977, Intermex's liabilities had climbed to around £ 130 billion "giving a debt to equity ratio of 26

204 In 1978, the average interest rate of 3-month deposits was 11.2% and 15.5% for 18-month deposits. These rates were, respectively, 13.2 and 15.7% in 1979, 20.6 and 22.8% in 1980, 29.6 and 32% in 1981, and 43.6% in 1982 for both instruments. 205 Banamex Archive, Libro No. 8 de Actas de la Comisión Ejecutiva, February 8, 1976 Meeting.

63 to 1, which was outside the policy of the Bank of England, who consider[ed] a ratio of 20 to 1 between debts and capital manageable."206 With these figures in the background, the numbers for the Mexican banking system look quite worrisome: debt to equity was already 50 per cent higher than the level the Bank of England considered prudent by early 1977.

The evolution of the leverage ratio also displays increasing risk levels in the banking system. Between 1977 and the beginning of 1982, the ratio of equity and reserves to total assets was reduced by half, passing from 4 to 2 per cent (see Figure 4). This ratio complements the debt-to- equity perspective. It shows that not only was the banking system relying more on debt than equity to finance the development of its activities, but that reserves were not improving either. In other words, the banking sector was taking on more debt to increase its assets without proportional increments to its capital base. Naturally, lending was the banking sector's main activity, with a loan portfolio representing between 50 and 55 per cent of the assets during the period. While in 1977, dollar lending accounted for about 20 per cent of total lending (the balance consisted of credits denominated in pesos), by the early 1982, however, its weight has increased to 30 per cent and to 42 per cent after the devaluation of February.

3.6. Inspection at the bank level

This section analyzes whether higher risk in the Mexican banking sector was a homogeneous phenomenon. Did this risk affect some banks or groups of banks differently than others? To investigate banks' risk position, I use quarterly bank-level information from the Comisión Nacional Bancaria y de Seguros' (CNBS) Monthly Bulletin and analyze a number of financial ratios reconstructed from the balance sheets of the banks.207 The period I cover begins in June 1979, the sixth month after the first Bulletin was published, up to June 1982, before the announcement of the Mexican government's temporary debt moratorium and the later nationalization of the banking system.208 The analysis includes 23 multiple commercial banks, for which there is complete and consistent information for the entire period.

As of early 1982, the 23 banks considered here accounted for US$ 71.9 billion or 95.8 per cent of the assets of multiple banks, and as much as 90 per cent of the assets of the Mexican commercial banking system. In Table 3, the banks are classified as internationally or domestic oriented, depending on whether they were substantially involved in foreign finance. The group of internationally oriented banks contains Mexico's largest institutions: Bancomer, Banamex, Serfin, Comermex, Mexicano-Somex and Banco Internacional (Banca Promex was part of the banking group Mexicano-Somex). As previously discussed, these were the banks that were most actively engaged in international banking activities. In contrast, the group of domestic-oriented banks is made up of smaller institutions that operated only at a national level. It was made up of 16 of the 23 banks in the sample but held only 18 per cent of their assets, with Banco del Atlantico, Banpais and Banco B.C.H. among the biggest. At an industry level, the segment of internationally oriented banks

206 Banamex Archive, Libro No. 10 de Actas de la Comisión Ejecutiva, January 5, 1977 Meeting. 207 They were published by the Financial Analyses Unit (Unidad de Análisis Financiero) under the name of "Boletín mensual de indicadores y estados financieros de las instituciones de créditos" from 1978 to 1979 and "Boletín de indicadores financieros de la banca multiple privada y mixta" from 1980 onwards. Publications are available at the Banco de Mexico Library. 208 The two first quarters are discarded because the bulletins do not include as many banks as the following ones, which would necessitate a considerable reduction of the sample in order to have a complete time series.

64 represented three-quarters of the commercial banking system, and domestic-oriented institutions accounted for the remaining quarter.

Table 3. Total assets of 23 multiple banks January 1982, millions of dollars

Domestic-oriented banks 13'096.2 Internationally-oriented banks 58'849.7 Banco del Atlantico 1'618.3 Bancomer 18'624.2 Banpais 1'519.4 Banamex 17'145.2 Banco B.C.H. 1'470.1 Banca Serfin 7'355.5 Banco de Credito y Servicio 1'237.0 Multibanco Comermex 5'841.4 Banca Cremi 1'127.4 Banco Mexicano-Somex 6'161.8 Multibanco Mercantil de Mexico 1'037.6 Banco Internacional 3'119.5 Banca Confia 979.9 Banca Promex 602.2 Credito Mexicano 573.6 Banco Regional del Norte 552.2 Actibanco Guadalajara 542.7 Unibanco 529.9 Banco Continental 477.0 Banco Mercantil de Monterrey 472.1 Banco del Noroeste 437.4 Banco Sofimex 332.1 Banco Occidental de Mexico 189.6

Source: CNBV, Monthly Bulletin.

Distinguishing between these two groups of banks allows for an evaluation of the role of international activities in the deteriorating health of the Mexican banking system. On one hand, access to international wholesale liquidity, which was identified as a potential source of vulnerability in the previous section, was available to internationally oriented banks, but not to those operating at a national level. On the other hand, as Chapter Four shows, Mexico's six largest banks incurred serious maturity, interest rate and currency mismatches on the balance sheet of their foreign agencies and branches while developing their international banking activities. Had international financial intermediation been a significant source of additional vulnerability, one would expect the international bank group to exhibit higher risk levels than domestic-oriented banks.

Since there is no information available on financial terms or the composition of the loan portfolio and funding instruments, I evaluate the risk position of the banks through a comparative analysis of their balance sheet ratios. Figure 5.1 and 5.2 plots the mean weighted by size of bank (total liabilities) and the 50 per cent central distribution of the debt-to-equity ratio and the quick ratio respectively for the group of international and domestic banks. Both ratios show considerably higher risk levels for internationally oriented banks than for those operating only in the domestic market. The first figure shows that the former had been much more aggressive than the latter in financing their growth with debt instead of shareholders' equity, and that they became significantly more leveraged toward the end of the period. Access to the international capital markets gave Mexican leading banks the possibility of finding new sources of funding to further expand their activities. The second figure also displays higher levels of risk for the internationally oriented bank group: the lower

levels of the quick ratio indicate a more limited ability to meet their short-term obligations with liquid assets, and thus a worse short-term liquidity position.

Figure 5. Risk levels of internationally vs. domestic oriented banks, 1979-1982 Figure 5.1. Debt-to-equity ratio Figure 5.2. Quick ratio

150 3

2.5

100

2 Quick Ratio Quick

Debt-to-equity Ratio Debt-to-equity 1.5

50 1

1979-Q2 1979-Q4 1980-Q2 1980-Q4 1981-Q2 1981-Q4 1982-Q2 1979-Q2 1979-Q4 1980-Q2 1980-Q4 1981-Q2 1981-Q4 1982-Q2

International banks Domestic banks International banks Domestic banks

Note: Debt-to-equity ratio = Total Liabilities / Total Equity Note: Quick ratio = Current Assets / Current Liabilities

Source: CNVB Multibank Bulletin (several issues).

Greater risk levels for internationally oriented banks are further confirmed when considering a larger number of balance sheet ratios. The appendix develops a Principal Component Analysis (PCA) based on ten financial ratios (see Table A1) that are usually used in finance to conduct financial statements and banking analysis, and have been identified in the literature as important indicators of bank risk management.209 PCA proves to be a useful methodology here, because it allows for extracting the most relevant information from the series of financial ratios and to identify the sources of bank vulnerabilities, as well as risk factors behind their balance sheet structures. The analysis shows that on average, internationally oriented banks displayed slightly worse capitalization levels than domestic-oriented banks, and suffered from a much more pronounced deterioration of capital ratios toward the end of the period (Figure A1.a). Significant differences also exist when comparing the liquidity position of the two groups of banks. While domestic-oriented banks' ability to meet their financial obligations with liquid assets improved from 1981 up until the end of the period, a deterioration and subsequent modest recovery are observed in the asset liquidity position of internationally-oriented banks (Figure A1.b). Discrepancies are also observed with respect to funding structures. The maturity composition indicator of the banks' funding base shows a relative larger concentration in short-term instruments for the internationally oriented bank group. Higher reliance on liquid fundraising instruments cannot be explained by broadening sight deposit base, but instead, because of increasing interbank borrowing. Indeed, sight deposits had actually decreased their share as a source of funding all through the period from 1978 to 1982.210 As previously posited and clearly shown in Figure 6,

209 These ratios were chosen and defined in order to have a comprehensive representation of the banks' assets and liabilities structures and the risks underlying their balance sheets. Similar indicators have been also employed by historians to assess the situation of banking institutions in times of financial distress or crisis. See, for instance, Calomiris and Mason (1997) and White (1984). 210 Of the remaining 16 domestic commercial banks considered here, only Multibanco Mercantil de Mexico and Banca Continental followed a similar trend. As for the rest, sight deposits either consistently increased their

66 international wholesale liquidity seems to have been providing these banks with the resources they were losing in their checking accounts.211 In contrast, domestic-oriented banks that were also suffering from significant declines in sight deposits, such as Banco Occidental de Mexico, Multibanco Mercantil de Mexico, Banco Mercantil de Monterrey and Banco Continental, did not increase interbank borrowing. Instead, their balance sheets exhibit offsetting changes on time deposit accounts, which indicates a change in the composition of retail deposits that did not necessarily imply a less stable and riskier funding base, as in the case of banks operating abroad.

Figure 6. Change in the share of deposits and bank borrowings to total liabilities, 1978-81

Note: Computations for Banco Mexicano-Somex, Banca Internacional, and Banco del Noroeste correspond to 1979 and 1981. Source: CNVB Multibank Bulletin (several issues).

As for the condition of the banks in terms of the quality of their assets, no clear distinction can be made between the two groups. If anything, the indicator of banks' asset quality shows a slightly better performance by internationally oriented banks, with the two groups displaying worsening levels during the last two quarters of the period (Figure A1.d). T-tests performed on the principal component scores show that the differences between the means are not statistically significant in any of the quarters of the investigation period. Similar results are obtained when evaluating the ratios of troubled assets to total assets and return on total assets separately, as well as when comparing the banks in terms of their values. Thus, it seems that no differences existed in the quality of the assets of both groups of banks. This finding suggests that the risks associated with their asset portfolio were similar. It appears that the devaluation of 1982 is the factor behind the deterioration of banks' asset quality, and that it affected all banks —not just those involved in international lending.

share as a source of funding, such as in the case of Credito Mexicano and Banca Cremi, or alternated years of increasing or declining share in the banks' liabilities. 211 Recall that Banca Promex belonged to the Mexicano-Somex Group.

3.7. An econometric interlude

The declining share of sight deposits as source of funding, as mentioned above, deserves closer scrutiny. Difficulties in deposit-taking seem to have been, indeed, a point of concern for banks. At the Mexican bankers' annual convention in Acapulco in June 1980, they underscored the hard time banks were having in capturing and absorbing new deposits. Bankers argued that this problem was due to the inflationary process affecting the country, and also to the direct competition for savings that came from government papers such as the certificados de tesoreria (Cetes) and, to a lesser extent, from the stock exchange. For their part, the reply from officials to the delegates' complaints hinted at troubles in the banking sector that could potentially be linked to reluctant depositors. At the meeting, Gustavo Romero Kolbeck, the central bank president, pointed out that Banco de Mexico "has been assisting banks that have over-committed themselves in the extension of credit to customers."212 It is thus worth investigating the potential link between the performance of sight deposit accounts and bank health.213

Although a complete model is not developed here, the single-equation regressions in this section may clarify whether the difficulties in obtaining new deposits were related to the higher risks observed on banks' balance sheets. The basic idea is to relate the evolution of sight deposits with risk indicators and test whether bank risk level is associated with differential changes in the accounts. The dependent variable is the quarterly percentage change of sight deposits from the domestic non-banking sector. The independent variables are the first four principal components in the prior quarter, namely capital adequacy (PC1), liquidity position (PC2), maturity structure of the funding base (PC3), and asset quality (PC4). As already noted and developed in much more detail in the appendix, these components can be interpreted as indicators of bank risk level and financial fragility.

A number of reasons favor the use of the first four principal components instead of the original financial ratios as explanatory variables. First of all, the principal components provide much clearer and more robust indicators of a bank's risk level, than the balance sheet ratios they were derived from. Second, many of the original financial ratios are very similar measures; it is not clear which of these should be included in the regression and which should be left out. That is the case, for instance, in the ratios such as equity to total assets, total capital to total assets and reserves and equity to total liabilities, which are all indicators of bank capitalization levels. Finally, based on the above-mentioned reasons, as well as the fact that financial ratios vary between zero and one, a high degree of multicollinearity is observed when running panel data regressions on the original variables. In short, not only is the risk perception given by principal components clearer, but they also provide uncorrelated predictors that incorporate most of the information contained in the original financial ratios.214

Table 4 reports the fixed effect panel data regressions for the 23 Mexican commercial banks between 1979 and 1982. Overall, the results show consistent evidence of a negative association

212 In fact, by that time Bank of Mexico was providing significant amounts of funds in special "auctions" to help commercial banks meet reserve requirements. "Bankers' hard-luck stories fail to move government," Latin American Weekly Report, WR-80-25, 27 June 1980, p. 7. 213 It is worth mentioning here that there was not a deposit insurance system in Mexico at that time. The first one was the FONAPRE, which was established in 1986. 214 Recall that these four principal component account for as much as 81.6% of the variation of the financial ratios data set (See Table A1.).

68 between the risk level of the banks and the performance of sight deposit accounts. The coefficients indicate that a worsening in the liquidity position of the bank (decrease in PC2), a shortening on the maturity structure of the funding base (increase in PC3) and deterioration in the quality of the asset portfolio (increase in PC4) have significant negative impacts on the growth rate of sight deposits across the four different specifications. Unlike them, the coefficient on the first principal component (PC1) is negative, although not statistically significant in any of the models: at best, no link is found between bank capitalization levels and deposit growth.215

Table 4. Regression analysis: domestic funding capacity and risk level (1) (2) (3) (4) VARIABLES Model 1 Model 2 Model 3 Model 4 Dependent variable is sight deposits growth rate Capital adequacya -0.00502 -0.0108 -0.0181 -0.0131 (0.00434) (0.0142) (0.0164) (0.0117) Liquidity positionb 0.0737*** 0.0587*** 0.0559*** 0.0394*** (0.0133) (0.0133) (0.0136) (0.0133) Funding maturityc -0.0871*** -0.0795*** -0.0738*** -0.0488** (0.0196) (0.0201) (0.0196) (0.0186) Asset qualityd -0.0153 -0.0212* -0.0257** -0.0229*** (0.00982) (0.0114) (0.0112) (0.00793) Inflatione 0.361*** -0.701** -1.645*** (0.114) (0.334) (0.400) Cetese -0.00288 -0.0848 -0.192** (0.0721) (0.0794) (0.0771) Total assets (ln) -0.247*** -0.245*** -0.146 (0.0685) (0.0717) (0.0905) Time 0.0666*** 0.118*** (0.0235) (0.0236) Sight deposits growth rate (t-1) -0.291*** (0.0693) Constant 0.0803*** 0.709 6.114*** 10.18*** (0.00105) (0.527) (2.154) (1.930)

Observations 270 270 270 268 R-squared 0.21 0.24 0.26 0.38 F-statistics 24.7 18.3 17.9 15.1 Fixed-Effects YES YES YES YES Number of panels (banks) 23 23 23 23

Cluster-robust standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 a PC1: Higher values indicate better capitalization levels. b PC2: Higher values indicate a better liquidity position. c PC3: Higher values indicate a shorter term funding structure. d PC4: Higher values indicate worse asset quality levels. e Inflation and Cetes are calculated as the natural logarithm of the CPI and of the ration of Cetes yield to the average of banks' domestic cost of borrowing respectively.

215 The Hausman test has been pursued to evaluate whether fixed effects or random effects should be used, with results indicating that the fix effect specification is the preferred model in all the cases.

The regressions of Columns 2 to 4 also assess the effect of other variables that potentially influence banks deposit performance.216 The coefficient of the variable Cetes, calculated as the ratio between the interest rate of government papers and bank deposits, shows a negative impact on the relative yield of alternative saving instruments on bank domestic funding performance. The effect of inflation on sight deposits is more unclear than might be expected, and has a significantly negative impact in the models that include a time-control and lag variables. 217 Finally, the size of the bank, measured by total assets, appears to be negatively associated with the growth rates of liquid deposits. Overall, these regression results are consistent with the previous discussion of internationally oriented banks as being in a weaker financial position and the most adversely affected by domestic funding difficulties.

It is now interesting to analyze whether differences existed in the performance of sight deposits accounts in internationally and domestic-oriented banks due to factors other than those included in the regressions. Since time-invariant factors cannot be directly considered in fixed effect models, interactions of a zero-one variable for domestic-internationally oriented banks with quarter dummies are included. The estimates of this interaction's terms were calculated under the four models of Table 4 with no clear results. Coefficients alternate positive and negative signs between quarters and between the different regressions, and in the bulk of the cases they are not statistically significant. This means that, everything else being equal, there is no evidence related to differences in deposit growth between both groups of banks. In other words, given the same level of risk and of the other explanatory variables, the performance of sight deposit accounts at internationally oriented banks was not greater or less than for domestic-oriented banks.

These results confirm the association between deposit-taking and bank risk levels, but the interpretation is more ambiguous that it might first appear. It could be that many of the explanatory variables, and in particular risk indicators, depend on the evolution and performance of bank deposits. As shown in the previous sections, recourse to foreign finance, which was at the root of increasing risk levels in the Mexican banking sector, was accompanied by a contraction of sight deposits. It thus seems likely that a negative feedback mechanism might have been in place: banks borrowed abroad to make up the shortfall of domestic liquidity, which, in turn, raised worries among customers about the health of the banks, further affecting deposit funding.218 However, because the concern here is whether banking risk can be viewed as a symptom of domestic funding difficulties rather than determining the causality relationship, endogeneity is neither a major nor fundamental issue.219

216 Given the presence of aggregate variables as predictors, standard errors have been cluster by time period in these regressions. 217 Yet, the switching sign for the inflation variable once a trend is introduced is intuitive. The trend captures the long-run average effect of inflation on nominal variables, so it is logically positive. Once purged from that influence, the inflation proxy captures the remaining short-run yearly fluctuations of the inflation rate, which has the usual negative effect on money holdings. 218 For the banks, the alternative to foreign borrowing would have been to attempt to improve their long term deposit base. However, the cost of raising dollars abroad was significantly lower than interest rates on domestic fundraising instruments and, additionally, foreign borrowing was not subject to reserve requirement while domestic deposits were at 37%. 219 Simple and partial correlation analysis has been performed between dependent and independent variables to check for the positive or negative association observed in regressions. In all cases, the endogeneity bias that may result in the regression analysis does not change the nature of the relationship between the variables or the significance, particularly regarding deposit growth rate and banks risk levels. Although a procedure

Although addressing this issue is well beyond the scope of this section, it is noteworthy to explore the potential mechanism linking deposit account performance and bank risk levels. One possibility is that, as previously mentioned, depositors withdraw their money from risky banks. An implicit assumption behind this reasoning is that customers were aware or informed of the financial condition of the banks, and that their deposit decisions depended on the perception of banking as risky. One way to check for this hypothesis is to compare the evolution of banks' sight deposits accounts and their stock prices. Since stock prices were common knowledge because they were published in the financial press, one would expect a close relationship between banks' sight deposit accounts and share prices if depositors considered the condition of the banks before placing their deposits.

Figure 7. Evolution of sight deposits accounts and banks' stock prices, 1978-1982 Figure 7.a. Bancomer Figure 7.b. Banamex

Figure 7.c. Banca Serfin Figure 7.d. Banca B.C.H.

Source: CNVB Multibank Bulletin and Anuario bursátil (several issues).

properly designed to deal with endogeneity and reverse causality issues would require instrumental variable analysis, this is well beyond the scope of this study.

Figure 7 displays the evolution of sight deposits as a source of financial resources and share prices for four large Mexican banks. In the case of Bancomer, Banamex and Banca Serfin, the three largest internationally oriented banks, the declining portion of sight deposits went almost hand-in-hand with falling share prices. Banca B.C.H, one of the largest banks in the group of domestic-oriented banking institutions, provides a counter example that seems to confirm the argument: The ratio of sight deposits to total liabilities does not fall and stock prices still fit the overall trend. In line with the regression results, this would suggest that depositors' behavior cannot be explained only by macroeconomic factors such as inflation or relative yield of saving instruments but that depositors were rationally reacting to changes in the individual condition of the banks.220

However, there might have been other factors underlying the fall in deposits. In a context of high economic growth and increasing domestic saving rate in Mexico, the fact that there were no liabilities of the banking system offsetting the declining share of sight deposits with internationally oriented banks raises the question of where was all this money going.221 This is an important concern in connection with the wave of capital flights affecting the country during those years, which was believed to be intermediated through the banking system. Given the lack of an international network for small domestic banks and the much limited presence of foreign banks in Mexico, internationally oriented banks were in a good position to carry the outflow of dollars. It is therefore possible that the money these banks were losing in their deposit accounts was actually being transferred abroad, but this requires further investigation.

3.8. Conclusion

This chapter shows the importance of considering the condition of the domestic banking system of debtor countries in the lead-up to the international debt crisis of the 1980s. It is revealing that during the years preceding the crisis, the position of the banking sector weakened, and the banks involved in international finance incurred relatively larger asset liability mismanagement. These facts suggest that the banking system was in bad shape by the time the crisis broke out, and that Mexican banks' increasing involvement in international banking was at the core of the problem.

The deteriorating health of the Mexican banking sector was twofold. On one hand, liabilities gradually became more concentrated in short-term obligations, which made the funding base of the system less stable. In particular, interbank financing from foreign banks considerably increased in importance as a source of funding, and thereby exposed the system to currency risk and shifting of conditions in the international financial markets. On other hand, the general balance sheet structure of the banking industry also weakened. Between 1977 and 1982, Mexican banks recovered from half a decade of financial disintermediation, but their expansion was attained by an increasing reliance on debt, in particular foreign debt, instead of equity, with no improvement in reserves. The worsening of capitalization levels and liquidity position, as well as the shortening maturity of the funding structure affected much more seriously the large Mexican banks involved in international finance than those operating only at a national level.

220 Unfortunately, it is not possible to generalize this analysis to the whole sample of the 23 commercial multiple banks, as not all of them issued shares, or were listed on the stock markets on a monthly basis. In addition to the four banks considered above, Multibanco Comermex and Banca Confia also have a complete series of stock prices for the analyzed period. Similarly, both banks performed in the same pattern. 221 Domestic private savings relative to Mexican GDP passed from 13.5% in 1977-78 to 14.6% in 1980 and 18.4% by the beginning of 1982. See Negrete Cárdenas (1999, p. 309), Table A2.

The chapter raises important issues regarding the Mexican banking system's exposure to the 1980s international debt crisis. Through their involvement in foreign capital markets, Mexican banks actively participated in syndicated and direct international lending for their customers in Mexico and other borrowing countries. This implies that the 1982 Mexican default may have not only risked the solvency of commercial banks from creditor countries, but also put its own domestic financial institutions at peril as well. In the Chapter Five I analyze the exposure of large internationally oriented banks to debt in Mexico and other defaulting countries to understand the extent to which this represented a threat to the domestic banking system.

A final issue that this chapter raises is related to the economics and political underpinning of international finance in the decade leading up to the debt crisis. The fact that the Mexican government and national financial authorities supported the internationalization of Mexican banks is revealing. They show that not only were international and domestic economic forces pushing the banks to go abroad, but that national political leaders and regulators saw it as an opportunity for meeting their economic policy goals. Finally, the arrival of Mexican banks in major international financial centres, however, raises the question about the position of host countries' financial authorities, particularly given the risky businesses they engaged in. Although this requires further research, it is imaginable that allowing Mexican financial institutions to participate in the Eurolending business was part of a strategy by creditor governments and major international banks to develop lending deals with Mexican borrowers.

Appendix

Financial ratios, risk level and principal component This appendix examines the differences in risk levels between internationally and domestic oriented banks though Principal Component Analysis (PCA). By reducing the dimensionality of the dataset, the main factors driving the variability of the data can be identified. PCA is used here to extract the most relevant information from a series of financial ratios and to identify the risk factors underlying the balance sheets of banking institutions. In what follows, I explore the correlation (factor loadings) between principal components and the original financial ratios. This allows me to interpret their meaning and look for patterns or distinctive features among banks, in comparison with bank component scores.

One single application of the principal component method is performed on the entire set of panel data that includes ten financial ratio series on each of 23 commercial banks for the 13 quarters from June 1979 to June 1982. Four components are found to have eigenvalues larger than one and therefore, will be kept to perform subsequent analysis. They account for 81.5 per cent of the variation in the data set, of which 32 percentage points correspond to the first principal component, 22.3 to the second, and the remaining 27 are equally distributed among the last two components. The variation explained by each component, as well as the factor loadings are reported in Table A1.

Table A1. Principal components, financial ratios and factor loadings

Comp1 Comp2 Comp3 Comp4

Troubled Assets / Total Assets 0.1139 0.0745 0.1878 0.6578 Loans / Total Assets 0.0005 -0.6772 -0.0115 -0.1158 Liquid Assets / Total Assets -0.0208 0.6898 -0.0586 -0.0644 Equity / Total Assets 0.5027 -0.0887 -0.2374 0.2229 Total Capital / Total Assets 0.5135 0.0565 0.1606 -0.1217 Equity and Reserves / Total Liabilities 0.5844 -0.0249 -0.0428 0.0348 Returns on Assets 0.0741 0.0206 0.1062 -0.6304 Interbank Borrowing / Total Liabilities -0.2281 -0.0427 0.6184 0.1995 Time Deposits / Total Liabilities -0.1357 0.1183 -0.5487 0.0363 Sight Deposits / Total Liabilities 0.2295 0.1791 0.4272 -0.2133

Per cent of variation explained 32.04 22.31 13.77 13.46

Note: Factor loadings represent the correlation coefficients between the principal component and financial ratios.

Source: Based on CNBS Bulletin (several issues).

As observed in Table A1, three of the ten original financial ratios, those referring to the bank's capitalization levels, are the most strongly correlated with the first principal component: equity to total assets, total capital to total assets, and equity and reserves to total liabilities.222 The higher the value of the ratio, the more the component increases, which suggests that it can be viewed as a measurement of the capital adequacy of the banks, with lower values indicating worse capitalization

222 The threshold is set at 0.4 and, therefore, only factor loadings larger than 0.4 in absolute terms (in bold in Table A1) are retained for analysis.

74 levels. Component scores show that no bank or group of banks follows a clear pattern or has consistently higher or lower values than the others.223 Instead, what can be observed is a general deterioration of commercial banks' capital adequacy throughout the years. Figure A1 plots the weight average and the component scores 50 per cent central distribution for the internationally and domestic oriented bank groups. There is clear downward trend in the scores of the first principal component for both groups of banks (see Figure A1.a), with internationally oriented banks displaying a slightly lower average and suffering from a more pronounced decline by the end of the period.

The greatest differences between both bank groups can be observed in the following two components. Factor loadings for the second principal component show a strong association with the variable loans to total assets (negative) and liquid assets to total assets (positive). Higher values in this component are, therefore, associated with a bank being better able to meet its obligations with most liquid assets and/or with lower exposure to relatively illiquid assets (loans). This suggests that the component can be considered a measure of banks' assets liquidity. Figure A1.b shows no clear differences in banks' asset liquidity until mid-1981, with both groups of banks moving away from each other from at point on. While there was a clear improvement in the component scores of domestic-oriented banks, values for internationally oriented banks decreased until the end of 1981 and stagnate thereafter. These lower scores of internationally oriented banks in comparison to domestic banks indicate a relatively weaker situation in terms of their liquidity position.

The third component displays high loadings in the following ratios: time deposits to total liabilities, sight deposits to total liabilities, and borrowing from banks to total liabilities, all of which are indicators of the banks' funding structure. Correlation is negative with the former and positive with the latter two, meaning that banks with higher values of this component tend to have relatively lower ratios of time deposits and, conversely, higher shares of sight deposits and bank borrowings. This contrast between short-term instruments with long-term ones could suggest that the component provides a measure of the maturity composition of the financing of banks, with high scores indicating a funding base more reliant on short-term funding. Figure A1.c illustrates internationally oriented banks as a distinctive group in terms of funding structure, with higher average component scores during the entire period. This result shouldn't be surprising since, as previously discussed, during these years, these banks had increased their reliance on short-term interbank borrowing in the international capital markets as recourse against domestic fundraising difficulties.

Finally, the forth principal contrast the variable return on assets (ROA) with the ratio of troubled assets to total assets. Since the correlation is negative with the former and positive with the latter, it appears that this component represents characteristics accounting for the quality of the bank assets. Banks with higher scores in the component could be associated with lower relative levels of returns and higher shares of troubled assets, and thus poor asset quality. Component scores show no evidence of significant differences between the two groups of banks. It is worth noting that component scores rise for both groups of banks toward the end of the period, indicating a general deterioration of bank asset quality as the crisis approached (see Figure A1.d).

223 Banco Occidental de Mexico, which is among the smallest banks in the sample, is the only exception, with components scores higher than the rest for most of the period.

Figure A1. Principal component scores by group of banks, II-1979 / II-1982 Figure A1.a. Capital adequacy Figure A1.b. Liquidity position

2 3

0 1

0 -2

-1

Scores for principal component 2 component principal for Scores

Scores for principal component 1 component principal for Scores

-4 -2 1979-Q2 1979-Q4 1980-Q2 1980-Q4 1981-Q2 1981-Q4 1982-Q2 1979-Q2 1979-Q4 1980-Q2 1980-Q4 1981-Q2 1981-Q4 1982-Q2

International banks Domestic banks International banks Domestic banks

Higher values indicate better capitalization levels. Higher values indicate better liquidity position.

Figure A1.c. Funding maturity Figure A1.d. Asset quality

3 2

2 1

-1 -1

Scores for principal component 4 component principal for Scores

Scores for principal component 3 component principal for Scores

-2 -2 1979-Q2 1979-Q4 1980-Q2 1980-Q4 1981-Q2 1981-Q4 1982-Q2 1979-Q2 1979-Q4 1980-Q2 1980-Q4 1981-Q2 1981-Q4 1982-Q2

International banks Domestic banks International banks Domestic banks

Higher values indicate a shorter term funding structure. Higher values indicate worse asset quality levels.

Source: CNVB Multibank Bulletin (several issues).

Chapter Four

International Interbank Markets and Financial Crisis*

Abstract

The international banking crisis that began in 2007 has brought the relationship between international banking activities and financial crises to the forefront. The growing reliance on foreign interbank funding by domestic banks has been recognized as a crucial factor in explaining the banking and sovereign debt crisis currently affecting several peripheral European countries. This chapter shows that the link between financial crisis and international interbank lending is not a new phenomenon; a similar trend can be observed in the Mexican banking sector during the run-up to its 1982 debt crisis. I explore the international activities of Mexican commercial banks in the years preceding the country's default and demonstrate that they became involved in international lending which was funded largely through heavy short-term interbank foreign borrowing. I provide new archival evidence which shows that in intermediating foreign finance with local public and private borrowers, Mexican banks incurred maturity, interest rate and currency mismatches and dangerously increased their risk position. My analysis provides insights for understanding the Mexican debt crisis as closely intertwined with problems in the domestic banking sector, which were, in turn, linked to its involvement in the international financial system.

* A slightly different version of this chapter is published as Sebastian Alvarez, "The Mexican Debt Crisis Redux: International Interbank Markets and Financial Crisis, 1977-1982," Financial History Review, 2015, vol. 22, no. 01, pp. 79–105.

4.1. Introduction

The global banking meltdown that began in August 2007 has led to a growing body of research on the relationship between banks' international activities and financial crises. While some of these studies have focused on the role of foreign off-balance sheet exposures in the banking sector, scholars have also highlighted the vulnerability to crisis created by banks' wholesale external borrowing in the international interbank market. Specifically, researchers have argued that a growing reliance by banks in certain countries on short-term cross-border funding (Fender and McGuire 2010; Merk Martel, Van Rixtel, and Gonzalez Mota 2012), along with currency and maturity imbalances between their assets and liabilities (CGFS 2010), made their national banking sectors dangerously vulnerable to shifts in U.S. money markets and international interbank lending.

Indeed, discussions concerning international interbank markets have moved into the spotlight during the current European crises. Since the intensification of the financial crisis in the second half of 2008, Euro area banks have been confronted with major dislocations in international wholesale markets and have consequentially experienced severe funding strains (Caruana and Van Rixtel, 2013; Van Rixtel and Gasperini, 2013). The heightened dependence of domestic banks on foreign interbank funding has been recognized as a primary factor in explaining the vulnerability of certain banking sectors, such as Ireland and Iceland, to the crisis (Honohan et al. 2010; SIC 2010). As for other troubled peripheral countries, such as Spain and Greece, problems in the banking sector have become increasingly intertwined with sovereign debt crises, which have put further pressures on banks, both in terms of access to funding and its cost (Shambaugh 2012; CGFS 2011).

Similarly, although to a lesser extent, the scale and scope of the interbank market expanded exponentially during the run-up to the 1980s international debt crisis. In 1982, the Bank for International Settlements (BIS), the institution responsible for much of the recent research, established a study group on the international interbank market. The size and the speed with which the market has grown in the wake of the financial fallout from the oil crises of the 1970s have captured the attention of supervisory authorities. By early 1982, interbank positions worth several hundred billion dollars and represented, as the BIS reported, up to three-quarters of the international lending boom that had taken place in previous years.224 There is reason to think, therefore, that there was an important link between bank's wholesale foreign activities and the financial crisis of the 1980s, in what had threatened to be the largest international banking meltdown since the Great Depression.

This chapter focuses on the link between foreign interbank market operations and financial crisis in Mexico, from 1977 up to the outbreak of the debt crisis in 1982. Mexico is a valuable case study to address these issues. First, Mexico was not only one of the biggest international debtors at this time, but also the country whose moratorium in August 1982 triggered the international debt crisis that brought the whole global banking and financial system on the brink of collapse.225 Second, the country's leading banking institutions were actively involved in foreign finance as both major borrowers and lenders in the international capital markets. Finally, there is good reason to assume that the experience of Mexico is echoed by that of other Latin American countries, such as Brazil and

224 BIS Archive, Box 1/3A(3)M Vol. 1., Study Group on the International Interbank Market. See also BIS (1983) pp. 17-19. 225 Boughton (2001, pp. 281-317).

Argentina, whose governments and financial institutions displayed similar dynamics to their Mexican counterparts and also suffered from serious financial crises.226

During the lead-up to the debt crisis, when considerable amounts of foreign finance flowed into Mexico, leading domestic commercial banks became highly involved in intermediating foreign finance with domestic final borrowers (Quijano 1987, pp. 241-258). However, despite the importance of their international borrowing and lending activities, neither the literature on Mexico's external debt nor the research on the sovereign debt crisis and the subsequent debt renegotiation process have given much attention to the role of the domestic banking sector.227 This chapter shows how Mexican commercial banks were crucially entangled in the country's external indebtedness process through the channelling of international wholesale liquidity back home, and that in doing so, they dangerously increased their risk position.

Until now, international interbank markets and developing countries' commercial banks have been largely absent in the extensive literature on the debt crisis of the 1980s.228 In this chapter, I reconstruct the essential elements of the international business model of Mexican banks by drawing on a variety of archival sources that have recently become available. I show how, through their branches and associated banks overseas, Mexican commercial banks raised large amounts of foreign capital in the U.S. and British international interbank markets, which they used to relend either directly, or through off-shore centres, to public and private borrowers at home. I provide new evidence which shows that in running their international activities, they accumulated serious maturity, interest rate and (indirect) currency mismatches on their balance sheets. My account is largely based on Federal Reserve Bank of New York archives and from forms that Mexican banks in the U.S. were required to file with the Federal Financial Institutions Examinations Council (FFIEC), known as FFIEC 002 Report Forms.229 I also draw on reports and historical statistics from the Banco de Mexico as well as documents and records from archives of the IMF and the BIS.

In the run-up to the crisis, a number of events contributed to a worsening of banks' financial mismatches and led to a deterioration of their financial position. The sharp increase in international interest rates during the late 1970s and early 1980s exacerbated the interest rate mismatch between the banks' liabilities that had been contracted at floating rates and their foreign loans, which had been largely arranged at predetermined interest rates. At the international level, moreover, beginning in mid-1981, a general retrenchment in the U.S. interbank market endangered Mexican banks' single most important source of foreign funding. Furthermore, the February-March 1982 devaluations made it more difficult for large private Mexican companies to reimburse their dollar debt to domestic banks and for Mexican banks, in turn, to service their foreign creditors. Finally, the Mexican government's moratorium declaration in August 1982 delivered the coup de grace to Mexican banks' wholesale foreign borrowing.

Although overlooked in the literature, Mexican banks' interbank foreign liabilities played a key role during the renegotiations and in the stabilization programs that followed Mexico's debt moratorium.

226 See Altimir and Devlin (1994), Devlin and Ffrench-Davis (1995), Diaz-Alejandro (1984), and Sachs and Williamson (1985). 227 Main references on the subject are Dornbusch (1990), Green (1988, 1998), Kraft (1984), Marichal (2011), Negrete Cárdenas (forthcoming), Solís and Zedillo (1985), and Zedillo (1985). 228 See, for instance, Cline (1984, 1995) and Devlin (1989), three of the most influential works on the accounts of the international debt crisis of the 1980s. 229 Federal Financial Institutions Examinations Council (FFIEC) Report of Assets and Liabilities of U.S. Branches and Agencies of Foreign Banks (FFIEC 002).

Mexico's negotiators, along with its official international creditors, sought to have creditor banks maintain access to interbank credit lines for Mexican banks. After the nationalization of the Mexican banking sector, and as part of the first rescheduling program for Mexican debt, an agreement was reached to freeze interbank deposits with the foreign branches and agencies of Mexican banks at the August 1982 pre-moratorium levels. This agreement would be extended in subsequent rescheduling programs. The outstanding interbank loans would remain frozen for almost ten years, until a definitive market-oriented solution was finally proposed in 1991, as part of the banks re- privatization program.

The rest of the chapter is organized as follows. Section 4.2 presents evidence showing the extent to which Mexican commercial banks relied on foreign interbank borrowing to fund their international loans prior to 1982. Section 4.3 describes the business model and the economic rationale behind their foreign activities. In Section 4.4, I analyse the rising fragility of the Mexican banking system, emphasising especially the exposures and imbalances incurred by large international banks. Section 4.5 explores the financial difficulties that Mexican banks faced in the run-up to the crisis, and the solution finally adopted to secure their interbank funding needs in Section 4.6. In the final section, based on the analysis of of the experience in Mexico, I draw more general conclusions about the link between the international activities of banks and financial crises during the 1980s and highlight avenues for future research on the topic.

4.2. Foreign intermediation by Mexican banks

In the decade-long run-up to the 1982 crisis, Mexico experienced the most notable growth of its external debt in the history of the country (Marichal, 2000; Solís & Zedillo, 1985). Mexicans' ability to borrow in foreign capital markets reflected a broader enthusiasm among international creditors in providing financing to developing economies during the 1970s. Following the oil shock of 1973, the petrodollar recycling process flourishing in the Euromarkets eventually led to an international lending boom to the developing world, where Mexico, along with other Latin American economies, attracted the lion's share (Devlin, 1989). At that time, Latin American economies were undertaking the final stages of the import-substituting industrialization (ISI) process and foreign capital was a major piece of the funding strategy (Bértola and Ocampo, 2012b). In Mexico, as in other Latin American countries suffering from credit booms followed by financial crises and busts in the early 1980s, foreign borrowing was partly intermediated by the domestic banking system.230

Between 1977 and 1982, when substantial amounts of capital flowed into Mexico, the commercial banking sector significantly increased its foreign liabilities. While the external debt of the Mexican public sector increased by one and a half times during that period, commercial banks' foreign indebtedness more than tripled.231 Indeed, the rise of banks' external debt is even more striking when considered in terms of their balance sheets. According to Banco de Mexico's annual reports, obligations to the external sector went from only 3.1 per cent of the total liabilities of the commercial banking sector in 1977 to 20.2 per cent in 1982, a bit less than a sevenfold increase in five years. These foreign liabilities essentially consisted of loans from international banks - in

230 See Sundararajan and Baliño (1991) and Mendoza and Terrones (2008) for a review on credit booms and financial crises during the 1970s and the 1980s in Latin America. 231 Based on data from Negrete Cárdenas (1999) and Solís and Zedillo (1985).

80 particular interbank facilities - operating in main international financial centres.232 Such figures give a clear sense of the extent and the intensity of commercial banks' borrowing in the international capital markets during the years preceding the crisis.233

The rise in banks foreign liabilities came along with an improvement in their lending capacities. Total financing provided by the commercial banking sector rose from US$ 12.6 billion in 1977 to US$ 46 billion at the beginning of 1982, expanding at the rate of 38 per cent per annum on average. Interestingly, dollar financing was the more dynamic component, with the dollar loan portfolio increasing by 5.5 times compared to a tripling in loans in pesos during the same time period. Banks' higher dollar lending capacities relied, to a large extent, on the foreign currencies that they were able to borrow on international capital markets. Intermediation of foreign capital with Mexican borrowers was mainly made through syndicated Eurocredits, but also by direct lending from London and other main international financial centres.

Syndicated deals were brought to London and promoted by a large international commercial bank, usually from the U.S. or another developed country, under a mandate granted by the borrower. Once there, the leading bank formed a lending syndicate with other banks and worked together to provide funds. Although largely dominated by banks from industrialized countries, Mexican commercial banks actively took part in these operations. According to Quijano (1987, pp. 244-247), of the US$ 16 billion raised by the 322 Eurocredits granted to Mexico between 1970 and 1979, US$ 5.1 billion or 32 per cent included participation by Mexican commercial banks, both as leaders or as associated syndicate members. Banamex and Bancomer, the two largest Mexican commercial banks, along with their London-based consortium banks, the International Mexican Bank (Intermex) and the Libra Bank, accounted for the vast majority of these operations. Indeed, as shown in Negrete Cárdenas (1999), in terms of the number of Eurocredits granted to the Mexican public sector between 1973 and 1982, these banks ranked alongside major foreign lenders such as Bank of America, Chase Manhattan and Citibank in their importance.234

Mexican commercial banks gained a foothold in the major international financial centres during the 1970s. They first arrived in London through the creation of consortium banks in joint ventures with other international banks from developed and developing countries. The Libra Bank, founded in 1972, along with Intermex and the Euro-Latinamerican Bank (Eulabank), both established in 1974, were the three consortium banks with Mexican ownership.235 The 1974 Mexican banking reforms empowered domestic banks to participate in the capital stock of foreign financial institutions and to open agencies and branches abroad.236 In effect, in addition to their involvement in London-based consortium banks, late in the decade, leading Mexican commercial banks would eventually establish

232 Banco de Mexico's annual report of 1978 details the composition of commercial banks' obligation with the external sector, which reached US$ 620 million in 1977 and US$ 823 million in 1978 and consisted entirely of all loans from foreign banks (pp. 109-110). 233 At that time, the Mexican banking system was made up by commercial banks, state-owned development banks and Banco de Mexico. For a description of the Mexican banking system and its place within the financial system as well as the interaction and network relationship with the rest of the financial actors see del Angel (2002), in particular, Chapters 4 and 5 as well as appendix 1 and 2. A broader characterization of the Mexican financial system during those years can be found in Solís (1997). 234 Intermex was the largest Mexican lender. It participated in 24 Eurocredits and ranked 5th among the leaders in bank syndicated loans to Mexico between 1973 and 1982 (Negrete Cárdenas 1999, Table B.17, p. 400). 235 They were owned by Bancomer, Banamex and Banca Serfin with 8, 38 and 6% of the shares respectively. 236 See Borja Martínez (1978, pp. 431-432).

81 their own international offices. By 1982, the six largest commercial banks Bancomer, Banamex, Banca Serfin, Multibanco Comermex, Banco Mexicano-Somex and Banca Internacional, which represented up to three-quarters of the commercial banking market share in Mexico, were running their own branches and agencies in the main international financial centres, with London and New York as their primary destinations.237

The internationalization of Mexican banking took place in a context of financial and banking deregulation. During the mid-1970s, Mexico, along with other Latin American countries with highly repressed financial systems, introduced a number of that attempted to liberate the financial sector. Kaminsky and Schmukler (2003) found that between 1973 and 1974, both capital account controls and regulations on the domestic financial system were relaxed in Mexico. As far as commercial banks were concerned, financial private institutions were now allowed to engage in offshore borrowing and to issue certificate of deposits at market-determined interest rates. Deregulation continued in the following years, through reductions and unification of the reserve requirement, softening interest rate controls, and increases on dollar borrowing and lending limits. The introduction of multi-purpose banking in 1975, which replaced the former restricted regime of specialized banking with a universal banking system, was a step forward. The reform lifted regulations that had previously pushed specialized financial institutions to operate in a single financial market and provided banks with greater flexibility in their intermediation activities.238

4.3. An interbank market based business model

During the 1970s, Mexican banks created consortium banks and set up overseas offices, in large part, to involve themselves in Euromarkets and engage in Eurocurrency businesses. Their presence in London and New York allowed the banks to access the two biggest international money markets and open a dollar-based funding channel. The fact that the vast majority of commercial banks' foreign offices were branches or agencies, as opposed to subsidiaries, reveals that their parent banks were not interested in developing regular retail banking businesses in the host country, which would have required bank subsidiary status.239 As agencies or branches, these banks were forbidden from taking conventional direct deposits. The focus was thus on wholesale banking instruments, like federal funds and interbank credit lines, which were available in the marketplace.

Most of the funding to meet the demand for international loans by Mexican banks came from the international interbank market. As Paul Mentre emphasized in a report for the Institute of International Finance, it was by accessing the interbank market that “LDC commercial banks typically borrowed on the U.S. domestic market or on the London dollar market to relend directly, or through offshore centres, to final borrowers.”240 In practice, interbank markets acted as channels from banks with a domestic dollar base or an excess of deposits towards banks where direct lending exceeded

237 There were in total 21 foreign branches and agencies of Mexican commercial banks in six foreign cities: Bancomer, Banamex and Serfin with four offices each, Comermex with five and Somex and Banca Internacional with two each (CIEN-A13/E-68/Agosto de 1982). 238 For an explanation of the multi-bank reform and the implications for the Mexican banking and financial system see del Angel (2002) and Seijas Román (1991). 239 Branches were not technically defined as banks, which raises an important difference in terms of regulation. Since they were not legally separate from their parent banks, they were not separately capitalized and were primarily supervised by their home authorities. As such, they were not subject to the host country's reserve requirements. 240 Mentre P. (1985), "The International Interbank Market and International Bank Lending," FRBNY Archive, Box 108403.

82 deposits. Such was the situation of Mexican and LDC's banks overseas, since as Phillip Wellons observed “[their] function [was] to act as a go-between for domestic borrowers, including their home office, and to raise money (...) in world markets for their home countries.”241

The business model of consortium banks relied, to a large extent, on interbank market deposits as a source of funds. While interbank loans normally accounted for one-fifth of the banks' assets, on the liabilities side, interbank placements ranged from 40 per cent of the banks' total liabilities, up to 100 per cent in a certain number of cases.242 Thus, by borrowing more than they lent to other banks, these banks were usually net takers of funds within the international interbank market. Mexican commercial banks also conformed to the same net borrowing pattern, and, as explained by Banamex Director José Manuel Rivero, the modus operandi consisted in “making placements with [creditor banks], for example, placing $10 million with an institution that is providing $20 million to Banamex.”243 In the same vein, Serge Bellanger, vice-president of the Institute of Foreign Bankers and Crédit Industriel et Commercial's New York branch manager, pointed out that when examining the liabilities side of foreign banks in general, the “interbank borrowings from the domestic and Eurodollar markets still remain a major component of the funding strategy.”244

Figure 1. Total combined assets and liabilities of U.S. agencies and branches of Mexican banks Breakdown by instrument, June 1982, million dollars Liabilities Assets

Source: FFIEC 002 Reports.

This business pattern can be illustrated by examining the case of the agencies and branches of Mexican banks in the U.S. As of June 1982, total combined assets and liabilities of the six bank agencies in New York and four in Los Angeles reached US$ 2.91 billion. Figure 1 provides the composition of the agencies' assets and liabilities at a consolidated level. On the liability side, the breakdown shows that borrowed money was their main fundraising instrument, followed by federal funds and deposits and credit balances. Taken together, they accounted for US$ 2.35 billion or 80.7 per cent of total liabilities. The fact that only US$ 10.9 million or 0.5 per cent of this amount was due to creditors other than banks highlights the prominent role of financial institutions as virtually the only suppliers of funds for Mexican agencies. The remaining 562 million (or 19 per cent) of agencies'

241 Wellons (1977, p. 77). 242 See Davis (1980) and Dufey and Giddy (1994). 243 FRBNY Archive, Central Records, BAC 1983: Office Memorandum, November 22, 1983. 244 "The foreign challenge to US banks," The Banker, October 1978, p. 40.

83 total liabilities, basically consisted of transfers from their head offices in Mexico and banker's acceptances. These figures give a clear sense of the large extent that their funding relied on the interbank market.

On the assets side, loans were agencies' most important claims, accounting for US$ 1.9 billion or 65.4 per cent of the total. They consisted of direct or purchase loans that were mainly granted to commercial and industrial (public or private) enterprises (73.9 per cent), and, to a lesser extent, to other financial institutions (15.4 per cent) and to foreign governments (9.6 per cent). Agencies' second largest asset was net transfers to their head offices, followed by banker's acceptances and Federal Funds sold to U.S. banks. Overall, an estimated 80-90 per cent of the U.S. agencies total assets represented claims on Mexican borrowers, and about 60 per cent represented loans to the Mexican Government or the public sector.245 This asset and liability composition makes the business model clear: borrow from commercial banks in the U.S. interbank money market and lend to final users in Mexico. As extensions of their parent banks, foreign agencies channelled U.S. money to Mexico and managed the liquid dollar assets of their international networks.

The significance of the foreign agencies of Mexican banks lay not only in their international business, but in the fact that they became the main door to international wholesale liquidity. As stressed by the BIS Study Group on interbank markets, although creditor banks seemed to have regularly lent to other banks in London or to banks of similar standing in other major financial centres or offshore centres, cross-border transactions with banks operating in remote financial centres raised more concerns.246 Data from Banco de Mexico Annual Reports show that by end-1981, borrowing from foreign banks had provided commercial banks with US$ 6.5 billion, of which as much as 70 per cent had been raised by foreign agencies while the remaining 30 per cent were loans directly granted to the head offices in Mexico.247

Through their foreign agencies, commercial banks found the international interbank markets to be new, low-cost funding opportunities. Figure 2 shows the evolution of the average domestic cost of funding compared to interbank interest rates in the U.S. and London, as well as the monthly depreciation of the peso-dollar nominal exchange rate during the 1977-1982 period. International interest rates were significantly below domestic levels throughout the entire period. This means that, given a virtually fixed exchange rate, it was cheaper for commercial banks to borrow dollars abroad than to raise pesos in Mexico. In fact, between 1977 and 1980, the cost of funding in London and New York was, on average, between 40 and 60 per cent lower than in Mexico; a difference that would eventually become even greater as the spread between domestic and international interest rates increased in subsequent years. Additionally, the fact that foreign branches or agencies were not subject to reserve requirements either in Mexico or in host countries further reduced the relative cost of foreign borrowing.248

245 FRBNY Archive, Central Records, C261 - Mexican Government 1917-1984: Office Memorandum, August 30, 1982. 246 See BIS (1983, p. 35). 247 It is worth mentioning that as foreign banks were not allowed to open branches or agencies in Mexico (Citibank being the only exception), these cross-border interbank transactions represented foreign liabilities of Mexican banks and not inter-offices business of foreign banks. For an account of the presence of foreign banks in Mexico, see del Angel (2002, pp. 139-148), and Sánchez Aguilar (1973) for the particular case of U.S. banks. 248 Since 1 April 1977, the reserve requirement for multipurpose banks was set at the uniformed rate of 37.5% for liabilities in the national currency. In the case of dollar denominated liabilities held in Mexico, reserve requirements ranged from 70 up to 100% during some years.

At a microeconomic level, the rationale behind international activities by the commercial banks relied on interest rate arbitrage operations in domestic and foreign markets. As financial historian Carlos Marichal explains, the banks' “purpose consisted in obtaining cheap funds overseas to lend domestically at higher rates, ergo recycling them locally.”249 With inflation and interest rates at double-digit levels whilst the peso-dollar nominal exchange rate held practically fixed from 1977 until early 1982, the potential for financial gains were significant. As Diaz-Alejandro (1985) observed in the case of Chile, the slow convergence (even divergence) of inflation and interest rates toward international levels, plus the fixed permanent nominal exchange rate, also yielded great incentives for private capital inflows into Mexico and its leading domestic banks were important intermediaries.

Figure 2. Domestic and international cost of borrowing for Mexican banks, 1977-82

Source: Banco de Mexico's Historical Statistics.

4.4. The risks behind Mexican foreign agencies

Mexican banks' foreign agencies, which made up the working base of their international businesses, operated under asset-liability imbalances that would prove to be very serious in the run-up to the crisis. Although no complete and systematic information exists about the banks' overseas branches or agencies in Mexican banking and financial statistics, the FFIEC 002 Report of Assets and Liabilities of U.S. Branches and Agencies of Foreign Banks provides balance sheets for those operating in the United States. U.S. agencies accounted for 10 of the 21 foreign branches and agencies of Mexican commercial banks, or an estimated of 40 per cent of their total combined assets and liabilities. Given that agencies in London and elsewhere followed a similar business model, the financial position of the U.S. agencies gives a representative sense of the general trend.

249 Marichal (2011, p. 119).

Table 1 shows the cross-border, maturity and interest rate balance sheet composition for Mexican agencies in the U.S. as of June 1982. A first worrisome complication could be found in cross-border currency mismatches. The international business model of commercial banks, as previously developed, led to a concentration of their foreign agencies' liabilities within the financial centre where they operated and of their assets abroad. Columns 2 to 4 show the general pattern. While average obligations to creditors domiciled in the U.S. accounted for 67.6 per cent of U.S. agencies' total liabilities, 73.3 per cent of their claims were due to clients domiciled outside of the United States, mostly in Mexico. Loans were made in dollars, but to borrowers operating mainly in pesos and not necessarily to exporting firms. Therefore, despite the fact that banks were not currency mismatched in their cross-border operations, their borrowers were. They were consequently exposed to currency risk and to the balance sheet effects associated with an eventual devaluation of the Mexican peso as described by Krugman (1999).

A second mismatch involved the maturity composition of the agencies' assets and liabilities. Their heavy reliance on interbank funding came with a liability structure that was almost necessarily biased toward very short-term debts, normally between overnight and six months. As of June 1982, up to 27.6 per cent of agencies' total combined liabilities consisted of borrowings (federal and borrowed funds) dues within a day and 19.1 were deposits and credit balances for 30 days ending with call date (See Table 1). Internal computations by the FRBNY staff estimated that around US$ 1.8 billion would mature between late August and the end of 1982, an amount representing almost two- thirds of their total liabilities.250 Conversely, the loans that these short-term instruments funded had been extended at much longer maturities. According to information from the FFIEC 002 reports, 73 per cent of the commercial and industrial loans, which were the major component of U.S. agencies assets, were due within the following year and the remaining 27 per cent had a maturity of over one year (See Table 1). As will be developed in further detail, the agencies' maturity imbalances would become even worse with the outbreak of the crisis.

Agencies also incurred interest rate mismatches on their borrowing and lending activities. At that time, interbank placements or credit lines were typically arranged at LIBOR plus a modest premium, which would depend on the risk associated with the borrowing bank, meaning that virtually all agencies' debts were contracted at a variable interest rate. In contrast, an important part of the agencies' loan portfolio consisted of claims arranged at predetermined or fixed interest rates. Table 1 shows that only US$ 416 million or 30 per cent of the commercial and industrial loans granted by U.S. agencies had a floating interest rate, while the remaining US$ 992 million or 70 per cent had been arranged at fixed rates. In this context, the sharp increase in international interest rates of the late seventies and early eighties would seriously damage the financial positions of these agencies. While their obligations and debt repayments increased along with the rise in interest rates, only a minor portion of the loan portfolio (which was the main source of their revenues) could adjust upward and benefit from the higher rates.

250 Mexican agencies outside the U.S. seem to fit into the same pattern. In addition to the US$ 1.8 billion in the U.S., Mexican agencies outside the U.S. had dollar liabilities of US$ 4 billion maturing between August and the end of December 1982, representing together an approximate of 75% of overseas agencies' total liabilities. FRBNY Archive, Central Records, C261 - Mexican Government 1917-1984: Office Memorandum, August 30, 1982.

Table 1. Asset and liability structure of U.S. agencies and branches of Mexican banks June 1982, millions of dollars Cross-Border Currency Maturity Interest Rate

Total Non-U.S. U.S. Addressees Loans due Borrowings TD &CB Com. & Ind. Loans Asset & Addresses due in a for 30 Liability Within 1 Over a Fixed Floating Asset Liability Asset Liability day* days** year year Rate Rate Bancomer 1,385 366 975 1,019 410 554 134 361 350 393 295 Banamex 741 170 362 572 380 198 64 136 17 261 0 Multibanco Comermex 421 129 341 292 80 154 97 199 71 130 121 Banca Serfin 261 88 206 174 56 98 57 58 69 155 0 Banco Internacional 106 26 89 80 17 22 31 50 52 53 0 Banco Somex 1 1 0 0 1 0 0 0 0 0 0

Total U.S. Agencies 2,915 779 1,972 2,136 943 1,026 382 804 558 992 416

Notes: *Federal Funds and borrowed funds of immediately available funds with one-day maturity. ** Total deposits and credit balances for 30 days (month) ending with call date. Source: FFIEC 002 Reports.

By the time of Mexico's default, foreign agencies had become important extensions of their parent banks and their financial fragility was a latent threat for them as well as the Mexican banking system as a whole. A Centro de Información y Estudios Nacionales (CIEN) report estimates that total liabilities of the foreign agencies and branches of the six Mexican international banks reached approximately US$ 7.64 billion in August 1982, an amount accounting for as much as one-fourth of their parent banks' total liabilities and 20 per cent of the Mexican banking sector's total liabilities. On the asset side, foreign agencies' credits represented over 50 per cent of the total loan portfolio of commercial banks in Mexico.251 The exposure of parent banks to their foreign agencies and the large share of these banks in the commercial banking sector made the domestic banking system vulnerable to the risks behind these international operations.252

4.5. The making of a distressed banking system

Even more striking than the volume of credit is the extent to which Mexican banks' foreign agencies impacted the risks faced by the domestic banking system. In a context of economic and financial liberalization, banks are usually confronted with novel forms of risk that can enhance market failures and risk mismanagement already present in the banking sector. Large inflows of foreign capital into a newly liberalized domestic banking sector might further aggravate these problems. In the absence of an appropriate supervisory and regulatory framework, an increase in the availability of loanable funds for domestic financial institutions is likely to raise credit, liquidity and settlement risks; eventually leading to greater systemic risk in the banking sector (McKinnon and Pill 1998).

Between 1977 and 1982, as Mexican banks expanded their international footprint and, as a consequence, their funding possibilities and lending capacities significantly increased. During this period, dollar denominated loans were the most active component of commercial banks' lending, as previously mentioned, and the private sector was the largest recipient of these funds, accounting for 50 per cent of banks' dollar claims as of early 1982.253 In fact, the increase in Mexico's private sector external debt during the late 1970s relied, to a large extent, on the foreign activities of the commercial banking sector. Given that Mexican international banks belonged to conglomerates which also controlled the companies that were borrowing most actively abroad, it is likely that these firms were the main beneficiaries of the dollar loans provided by those banks. According to Gutierrez (1992, p. 853), between 1979 and 1981, up to two-thirds of private sector foreign indebtedness was intermediated by Mexican commercial banks.

The rise of commercial banks' international financial intermediation came at the expense of higher risks to the banking sector. Table 2 shows the evolution of capital and reserves, cash and non- performing (troubled) assets relative to total assets for the commercial banking sector between 1978 and June 1982, directly prior to Mexico's default. In finance theory, the first two values, the leverage ratio and the ratio of (riskless) cash to assets, are considered determinants of the default risk of a bank and are commonly used in the literature as measures of bank risk-taking.254 They

251 CIEN-A19/E-89, Marzo de 1983, pp. 16-8. 252 In studying the Thailand financial crisis of 1997 and the maturity and currency mismatches of the banking sector with regards to the foreign sector, Allen et al. (2002, pp. 50-59) found that up to one-fourth of the commercial banks' total liabilities were foreign currency denominated, of which 60% fell due in the short term, which they state is enormous. The mismatches of the Mexican foreign agencies and their share on the domestic banking sector's total liabilities look similar, if not worse, to these figures. 253 As for the remaining, 26.4% corresponded to the Mexican public sector and 23.6 to foreign borrowers. 254 See, for instance, Calomiris and Carlson (2014).

display a clear deterioration in the health of the Mexican banking system: both leverage and cash to total asset ratios were strongly reduced during the period. As for troubled assets, it starts to rise as a proportion of total bank assets only after the February 1982 devaluation. This ratio provides a measure of the riskiness of the assets of a banking sector that have significantly increased the share of dollar loans in their lending portfolio (relative to those denominated in local currency), as discussed above.

Table 2. Riskiness indicators for commercial banks In percentages

Troubled Assets / Leverage Ratio* Cash / Total Assets Total Assets Dec- Dec- Jun- Dec- Dec- Jun- Dec- Dec- Jun-

78 81 82 78 81 82 78 81 82 International Banks 2.6 1.8 1.5 6.3 4.2 3.5 1.8 1.6 2.1 Bancomer 2.5 2.0 2.0 7.3 4.3 2.9 1.9 1.5 2.1 Banamex 3.0 2.1 1.6 5.1 4.9 4.6 1.5 1.6 1.9 Banca Serfin 2.0 1.5 1.6 10.6 3.0 2.9 2.1 1.3 1.8 Multibanco Comermex 1.9 1.4 0.8 4.9 3.7 3.3 2.2 1.8 2.8 Banca Somex 2.9 1.4 0.9 1.5 2.5 1.7 1.8 1.9 2.9 Banco Internacional 2.6 2.0 1.2 5.0 7.7 5.3 1.3 1.9 1.4

Local Banks 3.0 2.2 2.0 4.8 3.8 3.8 1.6 1.8 3.0

Total 2.7 1.9 1.6 5.9 4.1 3.5 1.8 1.7 2.3 * Computed as the ratio of equity and reserves to total assets

Source: Comisión Nacional Bancaria y de Seguros.

Indeed, risk taking seems to have been higher in the case of banks involved in international ventures. As of the early 1980s, the group of international banks showed greater levels of leverage and lower cash ratios than the banks operating solely on a domestic level. Moreover, the percentage reduction in the ratios is larger in international banks than in local banks, which would suggest that the former were much more aggressive in terms of risk taking than the latter. By participating in international capital markets, these banks received more funds and increased their liabilities without necessarily adding more capital. Figure 3 makes clear the extent to which banks leveraged on foreign funding and increased dollar lending without proportional increments in their capital base. In fact, while lending in pesos relative to capital stands at about 16, the dollar loan portfolio doubled in terms of banks' total capital between 1977 and 1982 before the devaluation of the peso.

An abundant amount of literature has analysed the factors accounting for international creditor banks' high risk-taking during the 1970s Euromarket lending boom to developing countries. While excess liquidity and poor regulation have been highlighted as major problems, scholars have also argued that creditor governments and IMF financial support to countries in payment difficulties might have deterred banks from properly assessing the risks behind these loans (Edwards, 1986; Folkerts-Landau, 1985). No conclusive evidence has been found, however, that points to moral hazard or banks' poor lending decisions, in terms of the quantity or the pricing of the bonds and syndicated loans. At a national level, although there are no records on the direct cross-border loans that Mexican domestic commercial banks granted to local borrowers, a number of elements may suggest that the presence of moral hazard likely encouraged too much dollar lending.

In the first place, the long-standing pegged exchange rate prior to the 1982 crisis could have actually introduced a source of moral hazard among private banks. As argued by Eichengreen and Hausmann (1999), low volatility in exchange rates may have led private investors to believe that authorities implicitly insured them against exchange risk, which in addition to a financial safety net, can cause a large amount of foreign capital inflows to be intermediated through the banking system. In fact, as previously discussed, the Mexican commercial banking system strongly increased their international network, foreign borrowing and lending activities starting in 1977 up until the outbreak of the crisis in 1982.

Figure 3. Evolution of the loan portfolio of commercial banks relative to capital, 1977-82

Source: Banco de Mexico's Historical Statistics

Secondly, commercial bank lending to public development banks also points to the existence of moral hazard stemming from implicit guarantees. Between 1970 and 1979, as much as 44.6 per cent of syndicated loans to Mexico, with the participation of Mexican banks, went to public financial institutions (Quijano 1987, pp. 246-247). Intermex and the Libra Bank, followed by Bancomer, were the main creditors, providing up to 84.3 per cent of the financing. In terms of their Euroloan portfolio, development banks accounted for 47, 87 and 30 per cent of the banks' claims in Mexico respectively. Indeed, the fact that Intermex was jointly owned by Banamex, along with state-owned Nacional Financiera (Nafinsa), Mexico's biggest development bank, and Banco Nacional de Comercio Exterior underscores the close ties in international lending between Mexican private banking and the public sector.255 Furthermore, board members and counsellors at commercial banks usually held

255 Nafinsa and Banco Nacional de Comercio Exterior bought into Intermex in 1978 buying 13% of the shares each. For a description of development banking in Mexico and of the international activities of Nafinsa at that time, see Ramírez (1986).

managerial positions in public development banks, as well as in the Mexican government, the Banco de Mexico and other official agencies.256

4.6. Banking difficulties in the face of crisis

On 20 August 1982, the Mexican government announced a temporary debt moratorium on principal payments that brought the country into default and launched the international debt crisis of the 1980s. Although frequently overlooked, the Mexican private sector was also confronting serious debt payment difficulties at the same time. The Alfa Industrial Group, Mexico's main economic conglomerate and the largest private international debtor suspended principal payments of its foreign debt even before the sovereign debt crisis broke.257 Similar to other major economic groups and private companies borrowing abroad, such as the Visa Group, the rise in worldwide interest rates from the previous years, paired with the peso devaluations of early 1982, increased the burden of their dollar debt and forced them to eventually go into default and debt restructuring.258

The increase in private and public sector external debt repayment problems would have serious repercussions for the domestic banking system. As previously noted, with half of their dollar loan portfolio in the hands of Mexican private companies and an additional quarter owed by the government and public agencies, Mexico's leading international banks and the commercial banking system were highly exposed to financial difficulties. Figure 3 shows the significant extent to which the banking sector was exposed to the risk of debt-servicing difficulties from both the private and public sector when the prospects of devaluation loomed. As a matter of fact, dollar claims went from 6 times the capital base of the banking sector in the early 1982, to more than 11 times after the February devaluation. The situation would only get worse during the second half of 1982, with the government moratorium, new defaults in the private sector and further devaluations of the peso.

As evident in Figure 4, the market perceived troubles in major international banks well before the Mexican government and private sector defaulted. After a period of dizzying rates of expansion, the rise of the Mexican stock market index stopped in 1979 and stagnated for the next two years, before finally busting in 1981.259 During the late 1970s and early 1980s, Bancomer's, Banamex's and Banca Serfin's share prices coincided with the general trend observed in the stock market. However, in the spring of 1980 - one year before the bust of the stock market - the share prices of these large international banks collapsed. By June 1980, in just six months, banks' stock prices had plummeted to almost half their January value. From there on, banks' share prices continued a downward trend

256 For instance, Manuel Espinosa Yglesias, President of Bancomer, was a Board Member at Banco de Mexico from 1977; Prudencio López Martínez, Alternate Member at Bancomer, was also an Alternate Member at Banco de of Mexico and Nafinsa as well as General Director of Consejo Nacional de Fomento Educativo (Conafe); Finally, Bernardo Quintaja Arrioja, President of the ICA Group, was also a Regional Counselor of Banamex and served as a member of the management board of Banco de Mexico from 1977 (CIEN-A13/E- 68/Agosto de 1982, pp. 22-23). 257 In trouble since late 1981, when 2,000 highly ranked executives were laid off and some of the company's assets were put up for sale, the Alfa group informed its international creditors on 21 April 1982 that it could no longer pay the principal on its US$ 2.3 billion foreign debt. See "Mexico's Alfa Tightens Belt," The New York Times, 21 October 1981 and "The Debt Burden on Alfa of Mexico," The New York Times, 10 May 1982. 258 See Gutierrez (1986). 259 For the role of the Mexican stock exchange in the national economy and its relation with the commercial banking system during those years see Quijano (1987, pp. 182-199).

until August 1982. The State took the stocks out of the market after the 1 September 1982 nationalization.260

Figure 4. Monthly share prices for international banks and Mexican stock exchange, 1977-82

120 1800

100 1600

0 1400

80 e

r 1200

1 1000

l 60

i p

800 c

A i

P e 40

d 600

I 400 20 200

0 0

7 8 9 0 1

8 9 0 1 2

7 7 7 8 8

7 7 8 8 8

7 7 7 7 8 8 8 8 8 8

- - - - -

------

v v v v v

y y y y y

g g g g g

b b b b b

a a a a a

o o o o o

e e e e e

u u u u u

F F F F F

A A A A A

N N N N N

M M M M M Bancomer Banamex Banca Serfin Stock Exchange

Source: Anuario Financiero y Bursátil (several issues).

At the international level, there were also early signs of Mexican banking fragility. Previous research has shown that the use of borrowed funds from other banks - in the form of interbank certificates of deposits, due bills, and rediscounts - is a forecast of bank distress and can thus be considered a forward-looking risk measurement (Calomiris and Carlson 2014, Calomiris and Mason 1997, White 1984).261 Since these borrowed funds are not low-cost interbank lines, but short-term higher-interest funding or 'hot debt', bank recourse to this source of funds suggests greater levels of risk. Figure 1 shows how much the U.S. agencies of Mexican banks relied on borrowed money as of mid-1982, a funding pattern that is confirmed by the previous FFIEC 002 reports that those same agencies filed. Overall, from June 1980 to June 1982, borrowed funds represented an average share of 48 and 50.8 per cent of Bancomer's and Banamex's total liabilities and as much as 57.4 and 60.3 per cent in Multibanco Comermex and Banca Serfin respectively, with peaks ranging from 70 to 84 per cent in some periods. It was, indeed, their most important source of funding during the period.

Mexican banks' agencies in the U.S. would encounter fundraising difficulties as the overall perception of country risk increased. More generally, as observed in Figure 5, September 1981 marked a turning point for the operations of foreign banks in the U.S. interbank market. After years of solid, rapid growth, foreign banking interbank liabilities due to U.S. banks fell in the fourth quarter

260 Curiously, in late 1981 and the beginning of 1982, when Bancomer stock prices were at 35% of their January 1980 value, under the mandate of Bancomer President Manuel Espinosa Yglesias, an offer was made to Citibank and to Bank of America to buy a majority of the shares and take control of the bank. The offer was finally declined. See Carral (2010, pp. 128-130). 261 Calomiris and Mason (1997, p. 874) observed that during the 1920s and 1930s, examiners of the Comptroller of the Currency used reliance on borrowed money as a clear indicator of banks having troubles.

of 1981, again in early 1982, and then stagnated. This meant that U.S. commercial banks, which were net providers of funding in the interbank market, not only failed to place any new interbank funds with foreign banks as a group, but also stopped renewing past credit lines and even withdrew deposits from them. However, whether the U.S. interbank credit crunch distressed the funding positon of Mexican agencies, the final blow to their money market funding activities came with the onset of the debt crisis in August 1982. Figure 5 shows that during 1982, borrowed money liabilities fell progressively, until September, when they dropped to half the value they had held as late as the previous December. Federal Funds and banks acceptances, which were the agencies' second major source of funding, would virtually disappear among agencies' liabilities by the end of 1982.

Figure 5. Interbank transactions of foreign banking offices in the U.S.

Source: U.S. Financial Account and FFIEC 002 Reports.

In this context, it did not take long for Mexican agencies, engaged in term transformation in their international businesses, to face grave liquidity problems when the debt crisis broke out. Deprived of their single most important source of immediate liquidity, interbank liabilities fell due more rapidly than mature assets became available. Lacking alternative funding sources, the agencies' financial position was seriously compromised. These agencies were not FDIC insured and they were unable to access the Federal Reserve's discount window facilities. Financial assistance from Mexico would also prove to be limited. At a time when Banco de Mexico was running out of foreign reserves and parent banks were experiencing difficulties in the reimbursement of dollar claims, home country financial institutions could not offer a definitive solution to U.S. agencies' dollar liquidity needs. After all, Mexican foreign agencies had been working as instruments of their head offices to raise dollars abroad and there was little reason to think the arrangement could work the other way around.

Agencies made up the shortfall in conventional funding by increasing recourse to time deposits from correspondent banks. In this respect, FFEIC 002 reports exhibit a change in the fundraising structure

of Mexican agencies from 1982 (see Figure 5). While in 1980 and 1981, total deposits and credit balances accounted for, on average 10 per cent of agencies' liabilities, by the end of 1982, they reached US$ 1.008 billion and 40 per cent of total liabilities. In 1983, total deposits and credit balances reached US$ 1.46 billion, an amount that was up to 40 per cent of their total liabilities. Virtually all of these funds (over 98 per cent) consisted of time deposits with six day terms - mainly in the form of open-account - with developed countries' banks in the U.S. or in foreign countries. FRBNY's internal documents and memorandums stress that Mexican agencies' representatives had been struggling to arrange credit lines with correspondent banks in the U.S. and Europe to ensure the availability of needed funds and avoid liquidity strains.262

The interbank market upheaval occurred both in the U.S. and also on a broader international level. As demonstrated by BIS (1983), the international interbank market was truly integrated, with substantial volumes of transactions between banks in the same centre, as well as cross-border transactions. Banks from developing countries also participated, whether located in major financial centres or in their home country. The policy of major banks from industrial countries placing and lending in the international interbank market was based on the creditworthiness of the borrower, which relied on a country risk analysis that looked at the nationality and location of the bank. Under this policy in such an integrated, international market, the BIS reported "it might be, for example, that the market comes to regard all banks of a certain nationality (e.g. Mexican) with some suspicion, perceiving the interbank operations with them more risky and therefore want to reduce their involvement with them."263

Indeed, an interbank run on Mexican banks did break out on Tuesday, 7 September 1982. Boughton (2001, p. 301) reports that on that Black Tuesday a panic began in the international wholesale markets and international banks refused to roll over lines of credits to Mexican banks on a massive scale. On that same day, officials of the Fed, the FRBNY and the Bank of England worked the telephones to persuade creditor banks to maintain the level of interbank credits that Mexican banks held with them. He notes that a substantial part of a recently approved BIS bridge loan was used to repay some portion of outstanding claims and that the creditor banks agreed to maintain the rest, thereby succeeding in stabilizing the market without a default. From that point on, as addressed below, Mexican banks could only access the international interbank market and meet liquidity needs because of maintenance commitments that were part of Mexico's debt rescheduling and stand-by agreements.264

4.7. Interbank funding during debt renegotiations

From the outbreak of the debt crisis in 1982 to the launching of the Brady plan in 1989, Mexico went through multiple reschedulings of its external debt. In total, there were four renegotiation rounds: each round led to a corresponding restructuring agreement between Mexico and its international

262 See FRBNY Archive, Central Records, C261 - Mexican Government 1917-1984. Although there is no evidence on how the banks attracted these monies, it is likely possible that they offered depositors higher interest rates. Banamex New York agency's officials declared that by September 1984 “Mexican banks continue[d] to pay roughly 3/4 of 1% over LIBOR on their interbank deposits.” FRBNY Archive, Box 142529, Mexico: Office Memorandum, September 18, 1984. 263 BIS Archive, Box 1/3A(3)M Vol. 1., Study Group on the International Interbank Market: Policy Issues Paper, Draft of 25.11.82. 264 Kraft (1984, pp. 25-27) provides second-hand evidence on the serious difficulties foreign agencies of Mexican banks were going through in the interbank markets after the moratorium declaration.

creditors. The principles and the strategy underlying these agreements essentially consisted in rescheduling the existing debt and extending new lending facilities which were conditioned on the agreement to an IMF adjustment program.265 A device associating new bank finance, IMF finance, and other government or multilateral finance was established by creditors to cover Mexico's financial needs. The arrangements reached between Mexico and its creditors aimed to conserve the country's much needed foreign exchange and allow Mexican banks to preserve their dollar funding base.

During debt renegotiations, Mexican banks' interbank funding was an issue of major concern. With their medium and long-term assets being restructured along with the country's other external debts, the banks' solvency position was under serious threat. Banks were forced to confront increasing difficulties in the renewal of short-term interbank credit lines that had been used to fund these loans. In fact, after the moratorium declaration, the Mexican government and the central bank stepped in to support banks in financial difficulties. Mexico's Public Credit Director and leading negotiator Angel Gurría expressed his serious concerns in a conversation with FRBNY officials about the critical financial position of the U.S. and London branches of Mexican banks. Gurría recounted that he had met with 140 bankers in Mexico City that day. He stated that he "would point out as emphatically as he could that no bank had ever been allowed to fail in Mexico,266 and that the Government and the Banco de of Mexico stood strongly behind the banks."267

Despite their willingness, Mexicans authorities lacked the financing required to assist the dollar funding needs of its banking sector.268 In such a context, as pointed out by Gurría himself, the understanding and cooperation of creditor banks was crucial. He therefore urged "not to create a problem by drawing down credit lines."269 Mexico's position was targeted for having international commercial banks keeping open funding lines and preventing interbank credit retrenchment and deposits withdrawals with Mexican banks' foreign agencies. In fact, unlike the bulk of the country's public external obligations, the Mexican government remained current on interbank foreign debt payments even after the moratorium declaration.270 By not defaulting or rescheduling this debt, they expected that the interbank market would stay open and creditor banks would renew outstanding placement and provide new credit facilities to the banks.

265 In total, there were four renegotiation rounds: 1982-83, 1983-84, 1984-85 and 1986-87. See Negrete Cárdenas (forthcoming) for a description of the debt renegotiation process during this period. 266 Up to that point, there had not been a bank failure since 1937. 267 FRBNY Archive, Central Records, C261 - Mexican Government 1917-1984: Office Memorandum, August 30, 1982. A few days after this talk, on 1 Septembre 1982, the Mexican government nationalized the commercial banks, in what scholars have suggested could have been a mechanism to rescue a banking system on the brink of collapse (del Angel 2002, p. 229; Marichal 2011, p. 122). In this vein, Gurría stated that, although perhaps done for the wrong political reasons, the takeover was a way to solve the financial difficulties of banks that would otherwise have had to declare themselves insolvent. Source: Interview held on 9 July 2013. 268 During the previously mentioned conversation, Gurría made clear to FRBNY officials that neither the Mexican government nor the Banco de Mexico could deal with the banks' dollar needs on its own because, as he expressed, "[they were] a little short of cash." FRBNY Archive, Central Records, C261 - Mexican Government 1917-1984: Office Memorandum, August 30, 1982. 269 Ibid. 270 With the 1 September 1982 nationalization, private commercial banks' foreign liabilities (as well as their assets) became the responsibility of the Mexican government. Other facilities that were excluded from the restructuring scheme and serviced when due were international organization's credits, bonds, private placements, leases, banker's acceptances and trade credits. See Gurría (1988, p. 77).

The interbank issue was also important for Mexico's official creditors. In the U.S., several interviews were arranged by FRBNY officials with Mexican bankers and government authorities to discuss the situation of the U.S. offices of the Mexican banks. The goal was to assess the real financial position of these agencies and discuss how to deal with their dollar liquidity needs.271 As for the IMF, the interbank element was not only necessary to secure the domestic banking system but also, more generally, to implement Mexico's stabilization program. In fact, Fund officials underlined that international commercial banks' roll-over operations could not be limited to medium and long-term debt but also needed to integrate "the inter-bank element related to the euro-market operations of agency banks, which attract short-term euro-market deposits to re-lend to banks in their own countries at longer maturities."272 In Jacques de Larosière's words, "it could undermine the rest of the rescheduling operation if the base of the iceberg (the large interbank element) were to dissolve."273

A compromise was eventually reached as part of the first rescheduling agreement. To insure that the country's foreign bank agencies did not experience a large scale leakage of funding, outstanding interbank loans for Mexican banks were frozen at the August 1982 pre-moratorium level. On one hand, with the implementation of the Mexican stabilization program providing the basis for an IMF money facility for Mexico, creditor banks responded to Mexico's request and agreed to "maintain current exposure to foreign branches and agencies of Mexican banks, concurring in de Larosière's assessment that it was critical that all banks continue to do so."274 On the other hand, Mexico committed to make sufficient funds available to such agencies and branches to process market interest payments on their interbank account. In the end, with the restructuring loan documentation, creditor banks committed to not letting deposits fall below US$ 5.2 billion until the end of 1986. In practice, interbank commitment agreements to keep deposits rolling over 90 days were renewed and then renewed again, whenever they were about to expire.

The US$ 5.2 billion threshold commitment on interbank outstanding debt to Mexican banks' foreign agencies was to be maintained for ten years. Arguing that Mexican banks needed the interbank placements as a long term source of funding for their loans to Mexico governed by the restructuring agreements, Mexican government officials asked for an extension of the covenant on two occasions. With the 1986-87 Financing Packages, the expiration date was extended to June 1989 and then again as part of the 1982-89 Financing Package of the Brady Plan, which set the final expiration date on 31 December 1992. The final market-oriented solution came in 1991 and consisted in exchanging the interbank deposits for a new instrument, the Floating Rate Privatization Note, which was a direct obligation of the United Mexican States and could be used to purchase shares of Mexican commercial banks under re-privatization.275

271 Several documents and internal reports were prepared by the staff of the FRBNY regarding how to proceed in the case of default by an agency. See, in particular, FRBNY Archive, Central Records, C261 - Mexican Government 1917-1984: Office Memorandum Attachment D, August 30, 1982 272 IMF Archive, OMDF Jacques de Larosière's chronological files, Box 3, File 4: Office Correspondence, December 20, 1983. 273 Ibid. 274 FBRNY Archive, Central Records, BAC 1982: Office Memorandum, November 18, 1982. 275 FRBNY Archive, Box 142529, Mexico: Telex, June 4, 1991.

4.8. Conclusion

Based on this analysis, a number of conclusions can be drawn that both shed new light on the 1982 Mexican crisis and also have wider implications for the Latin American and international debt crisis of the 1980s. It is revealing that in the years preceding the country's default, leading domestic banks became heavily involved in the international financial system and relied, to a large extent, on foreign interbank borrowing to fund their dollar loans.

A major finding of this study is that in running their international businesses, Mexican bank's foreign agencies accumulated significant maturity and interest rate imbalances. By the time of the crisis, although many of their foreign liabilities consisted of very short-term interbank deposits, the bulk of their dollar denominated assets had much longer maturities. Furthermore, while these interbank credit lines had been set at floating rates, a significant part of the loans were arranged at predetermined fixed rates. Additionally, while balance sheets did not register currency mismatches they were still indirectly exposed to the risk of an eventual currency crisis since their dollar liabilities were from foreigner creditors and their dollar claims were mainly with Mexican debtors running their businesses largely in pesos.

Important questions remain: How could Mexican banks have increased their risk position to such dangerous levels? Who was responsible for allowing that to happen? Although beyond the scope of this study, it is difficult to believe that such evident mismanagement would have gone unnoticed. In Mexico, it was surely possible for financial regulators and the country's most seasoned bankers to see what was happening. The reasons why banks engaged in foreign lending and took such risky positions in such a tenuous environment must go well beyond their individual initiative and, perhaps, be part of broader scheme that included the government, as well as the public and non- banking private sectors, in a time of great need for financing. The crucial question of the interplay between domestic banks and policymakers in the international banking setting and external indebtedness during the 1970s has not yet been addressed in the literature and deserves further investigation.

A final issue that this chapter raises is with respect to our understanding of the international debt crisis of the 1980s. The existing literature, in overlooking the involvement of domestic banks in debtor countries in the petrodollar recycling process of the 1970s, has implicitly assumed that they did not play an important part in the making of the crisis. However, there is no reason to think that my story about Mexican banks reveals a pattern that is exclusive to them. To the contrary, banking institutions from other large Latin American debtor countries, such as Brazil and Argentina, were also heavily engaged in foreign financing based on a similar business model. Therefore, considering the participation of commercial banks from borrowing countries in international capital markets, may imply that the origins of the debt crisis should be revised and reconsidered.

The fact that interbank deposits from foreign banks with the overseas agencies and branches of Mexican commercial banks had to be frozen at pre-moratorium levels for almost ten years is a clear sign of their financial weakness and critical dependence on foreign finance. Similar interbank arrangements were also undertaken during debt renegotiations and rescheduling agreements in Brazil and Argentina. This finding suggests the possibility of further work on the link between sovereign and domestic banks during the Latin American debt crisis of the 1980s.

Chapter Five

Domestic Banks and the Mexican External Debt

Renegotiating Position

Abstract

A striking feature of the 1980s international debt crisis is the sharp contrast of the burden borne by creditor countries and that borne by Mexico and other debtor economies. While a number of scholars have stressed that tough financial and adjustment conditions imposed on debtor countries were the logical market results of dealing with risky economies, it has also been argued that renegotiating conditions were not the result of competitive market forces, but instead, mainly determined on the basis of the weaker bargaining power of debtors relative to creditors. A central issue that has not been satisfactorily explained is the factors behind debtors' powerless renegotiating position in a context where the entire international banking and financial system hinged upon Mexico and other borrower countries' repayments. This chapter shows that compared to foreign creditor banks, large domestic Mexican banks were more seriously compromised, and that Mexican policymakers didn't have any realistic way to stabilize the banking system without foreign financial support. Because of their heavier reliance on international interbank funding and larger exposure to their own private and public sector debts, a default or major disruption on external debt services would have probably been less damaging for foreign banks than for Mexicans themselves. The chapter provides new insights for understanding why Mexican negotiators went to such great lengths to avoid default and failed to put up stronger resistance and succumbed to creditors' renegotiating conditions.

5.1. Introduction

After years of dizzying expansion of external borrowing from the international banking system, in August 1982 a government moratorium on principal payments was declared and the supply of foreign loans to Mexico was suspended. Mexico's default, however, did not bring about a unilateral suspension or repudiation of foreign debt, but rather an orderly rescheduling with no major interruption in the service of its external obligations. To keep current on its payments, the country required additional foreign financing, which banks and official creditors agreed to provide conditional upon an agreement between the government and the International Monetary Fund (IMF). Access to foreign capital proved to be costly for Mexico, not only in terms of the financial conditions, but also because of the consequences of the IMF austerity and adjustment programs on the domestic economy, which entailed great social costs.

The literature largely agrees that the burden of the international debt crisis of the 1980s was borne principally by debtors, and, to a much lesser extent, by their international creditors.276 In the case of Mexico, the debt crisis extended from 1982 to 1989, and its management involved four conditionality rescheduling agreements and corresponding financing plans. By 1986, Mexico had already transferred a large amount of resources to creditor countries, but the weight of both debt and service payments on GDP and exports continued to rise. The domestic economy remained in recession. By the end of the decade, Mexico's GDP per capita was still lower than at the outbreak of the crisis, and unemployment, along with other social indicators, had worsened considerably. As for Mexico's creditor banks, despite their great exposure to the country and other highly indebted economies, no major failures or collapses occurred. Indeed, interest receipts exceeded new lending, banks remained profitable, and suffered no reduction in the value of their loan portfolio until very late in the decade.

The political economy of the debt management strategy and the way the burden of the crisis was distributed among debtors and creditors have interested a great number of scholars. Following an approach initially designed by William Cline, a number of works have stressed that the high cost of credit and the tough rescheduling conditions were the logical result stemming from the proper function of international financial markets. From this perspective, banks needed to be compensated for greater risks and debtors penalized for bad behavior, in order to avoid the moral hazard problems that plague private market-based final arrangements.277 In contrast, Robert Devlin and other economists from the Economic Commission for Latin America and the Caribbean (ECLAC) have argued that renegotiating conditions was not the result of a competitive credit market dynamic, but rather a bilateral monopoly, where financial terms were determined on the basis of the relative bargaining power of debtors and creditors.278 According to this approach, banks managed to impose rescheduling with conditions that allowed them to escape losses and consequent devaluation of assets based on a strong negotiating position, and obtained rents that could have been captured by debtors if their bargaining power had been greater.

This chapter takes the later interpretation as its starting point and seeks to shed new light on the factors that determined the balance of power between debtors and creditors. In fact, the factors

276 See, for instance, Cohen (1989), Feinberg (1986), Fischer (1987) and Devlin and Ffrench-Davis (1995). 277 Cline (1984). 278 Devlin (1989).

behind debtors' weak bargaining position were absent in Devlin and ECLAC's explanation. In a context where the world's biggest banks and the whole banking and international financial system depended on repayments from Mexico and other borrower countries, one would have expected major debtor countries to use their leverage to drive bargaining outcomes in their favor and impose a greater part of the cost of the crisis management on creditors. In contrast, within certain limits, debtors appeared to accept whatever it took to reimburse their debts and to access the foreign credits they needed. Although such behavior reveals a clear, and indeed strong, dependence on international finance, the economic or political reasons for this dependence and the sectors most vulnerable to it have received little attention in the literature. Although obviously the vulnerabilities of highly exposed individual sectors of the domestic economy to the lack of foreign financing might have impaired the bargaining position of Mexican negotiators, we still have not yet identified these sectors or their financial condition.

This chapter focuses on the situation of the Mexican banking sector and its dependence on foreign finance. Until now, the fragility of Mexican commercial banking system and its links with external capital have been largely overlooked in the accounts of Mexico's debt renegotiation. In this chapter, I explore how Mexican policymakers confronted the external difficulties of the domestic banking sector in the face of the crisis, and the extent that banking stability was dependent upon access to the international capital markets. Previous chapters have shown that in the years preceding the debt crisis, Mexico's commercial banking system displayed increasing levels of risk, which had primarily concerned major domestic banks and was mainly related to their international activities. Through their foreign offices in the U.S. and London, these banks had become increasingly involved in intermediating international interbank liquidity with domestic borrowers, and in doing so, they accumulated significant maturity, interest rate and currency imbalances on their balance sheets.279 By the time of the crisis, the mismatches of banks' foreign agencies, whose liabilities were owed to the interbank market and whose assets were not liquid, had become a major potential problem for their parent banks and the entire domestic banking system.

Indeed, after the declaration of the moratorium, Mexican agencies experienced increasing difficulties rolling over foreign deposits and faced progressive drains on interbank funding. Willing to reduce their country exposure, both large and small international creditor banks retrenched deposit lines with Mexican agencies, which, in turn, put growing pressure on their liquidity position. Deprived of foreign finance, agencies turned to their home country and looked for assistance from parent banks and local financial authorities to meet their foreign obligations. But by that time, neither their head offices nor the Bank of Mexico could realistically support the dollar needs of these agencies. Instead, they assisted them by drawing on Bank of International Settlement (BIS) financial facilities and swaps from the U.S. Federal Reserve, but were unable to stop the funding drain. The final solution to the Mexican agencies' liquidity problems came as part of the first rescheduling program, where creditor banks committed to keep their interbank liabilities with them at the August 1982 pre-moratorium level.

In this chapter, I argue that Mexican banks' foreign agencies were the Achilles' heel of the domestic banking system: protecting access to interbank deposits constrained the bargaining position of the Mexican state in restructuring its debt. Had foreign creditor banks not agreed to freeze US$ 5.2

279 See Chapter Four.

100

billion of interbank deposits with Mexican agencies, the agencies would most likely fail to service their interbank commitments, in which case host countries' financial authorities would have to close them. Given the important financial ties between these agencies and the head offices, this would have had important repercussion back home running the risk of spreading the banking crisis and making the burden of the debt crisis even worse. For their part, creditor banks leveraged Mexico's need for interbank liquidity, among other vulnerabilities, to negotiate favorable conditions.

The rest of the chapter is structured as follows. The next section stresses the high cost that the debt restructuring deals entailed for Mexico in comparison to its international creditors. In Section 5.3, I explore the underlying factors of the weak bargaining power, by drawing on the sovereign debt theory and interbank debt as bargaining tool for creditors. Section 5.4 analyzes the exposure of Mexican banks to Mexico own default in relation to foreign creditor banks. Section 5.5 discusses the funding and liquidity problems encountered by Mexican foreign agencies in the wake of the crisis, while Section 5.6 demonstrates that the stability of the domestic financial system depended on continued access to foreign funding. The last section concludes.

5.2. The puzzle behind Mexican debt deals

Mexico's debt crisis broke out on August 20, 1982, when government officials approached the international financial community to announce that they could no longer meet payments on external debt. By that time, the Bank of Mexico had run out of reserves and the funds gathered from international private financial, and capital markets were insufficient to service the monthly foreign currency obligations, whose main creditors were commercial banks from developed countries.280 At the international level, Mexico's difficulties were a major concern; a failure to repay this debt would have implications far beyond national boundaries. Major world banks, especially U.S. and some European, held substantial amounts of Mexican debt on their books.

The first official measures adopted to assist Mexico were securing U.S. emergency funding and arranging a bridge loan from the BIS.281 Additionally, Mexican authorities looked for extra funding from the IMF and soon decided to negotiate the terms for a three-year Extended Fund Facility (EFF) program.282 At Mexico's request, but with the leverage of their own governments and financial authorities, private creditor commercial banks accepted the rollover of the principal payment coming due in the months following the moratorium and created a Bank Advisory Committee (BAC) responsible for restructuring outstanding loans. 283 In the end, as part of the financial and restructuring program prepared by the IMF and creditor governments, the banks committed to grant new loans to help the country meet its financial requirements.

280 See Kraft (1984) for an account of Mexican payment problems on the eve of the crisis and the measures undertaken to rescue the country. 281 See United States General Accounting Office (1997) for a detailed account of the measures taken by U.S. authorities to assist Mexico during 1982, both before and after the crisis broke out in August. 282 For a description of the early stages of the Mexican rescue, see Kraft (1984) and Boughton (2001, pp. 289- 296). 283 They were also referred to as Bank Advisory Group, Steering Committees or Working Committees, or more generally, the London Club. See Rieffel (2003), Lomax (1987) and Rhodes (1994) for more details on the nature of these committees and their role during debt renegotiations.

101

The basic strategy of international creditors in dealing with the Mexican crisis, therefore, consisted of rescheduling the amortization of principal with commercial banks and extending new money facilities conditioned on IMF agreement.284 On one hand, the Mexican government committed to adopting austerity measures and introducing a number of economic reforms to address the underlying fundamental imbalances of the domestic economy and improve balance of payment performance. On the other hand, the group of creditors, which included commercial banks, governments from industrial countries and international organizations, were to jointly provide the financing and foreign exchange the country needed to implement the adjustment program.285

Table 1. Mexico's restructuring agreements with creditor banks, 1982-1989

Rounds 1982-1983 1984-1985 1986-1987 1989 Rescheduling terms Outstanding principal (Mil. US$) 35,150 48,700 61,950 48,900 Public Sector 23,150 48,700 52,250 48,400 Private Sector 12,000 9,700

Maturity date 1982-1984 1985-1990 1985-1990

Interest rate (%) LIBOR+1 7/8 LIBOR+1 1/8 LIBOR+13/16 LIBOR+13/16 Prime+1 3/4 Prime+7/8 (first 2 years) 6.25 (Fixed)

+1 1/8 (next 5 years)

+1 1/4 (last 7 years)

Commission (%) 1

Amortization period (years) 8 14 20 30 Grace period (years) 4 5 7

New financing terms

Banks loans (Mil. US$) 5,000 3,800 7,700

Interest rate (%) LIBOR+2 1/4 LIBOR+1 1/2 LIBOR+13/16

Prime+2 1/8

Commission (%) 1 1/4 5/8

Amortization period (years) 6 10 12

Grace period (years) 3 6 4

Source: Devlin (1989) and Gurria (1988, 1995)

The first round of negotiations between Mexico and the commercial banks extended from the outbreak of the crisis up to the end of 1983. An agreement was finally reached to restructure US$ 35.1 billion of current maturities of principal due by the public and private sector between August

284 It was critical for the banks keep treatment of interest as distinct from amortization. Under U.S. banking law, if interest payments were interrupted the banks had to reclassify the loans as non-accruing or non- performing, which would have affected their financial position and the general confidence. See Sachs and Huizinga (1987) for a discussion of the U.S. banking rules and the application of supervisory criticism during the debt crisis, and Sgard (2016) for an account of the working relationship between the IMF and commercial banks in lending and debt restructuring issues with developing countries during the 1970s and 1980s. 285 See, in particular, James (1996) and Boughton (2001).

102

23, 1982 and December 31, 1984. The negotiated terms of rescheduled debt, as Mexican leading negotiator Angel Gurría remarked, were tight: tenor of eight-years with a four-year grace period and interest rates of LIBOR plus 1 7/8 points and prime plus 1 3/4 points.286 As for the US$ 5 billion granted by the banks as part of the refinancing package, in the words of Gurría himself, "the cost was even more onerous than the restructured amounts" (see Table 1).287 Moreover, a front-end fee of one per cent of the aggregate amount of outstanding debt was additionally charged, as part of restructuring contracts and of 1 1/4 points for the new lending facilities. In the end, according to Robert Devlin's estimations, the final negotiated price of credit for Mexico was about 2.8 times more expensive than what the country had been paying before the crisis in 1980-81.288

Although the 1982-1983 restructuring agreement bought time and provided some breathing room, it didn't accommodate Mexico's payment capacity with the debt service schedule after December 31, 1984. In fact, two new debt refinancing packages would be implemented in the upcoming years: the multi-year rescheduling agreement (MYRA) of 1984-1985 and the restructuring program of 1986- 1987. The financial strategy of these two arrangements was also to reschedule the debt profile and grant fresh resources. Similarly, the broad economic approach to manage the crisis continued to be based on IMF structural adjustment programs. In both cases, the negotiated terms and conditions were better than those that had been agreed to in the first rescheduling: repayment and grace periods were longer and spread over interest rates were considerably lower (see Table 1).289 However, despite the improvement in terms of deferrals of principals and the reduction in surcharges, these packages still did not provide a satisfactory long-term solution to the crisis and the country's foreign debt payment problems.

The burden of these arrangements was not only related to financial terms of the loans, but also to the implications of IMF conditionality for the domestic economy. Upon the agreements signed with the Fund, Mexican policymakers committed to reduce the fiscal deficit and the use of foreign indebtedness by the public sector as well as to limit the increase of the net assets of the Banco de Mexico. The letter of intent for the EFF program agreed to reduce the public deficit from the estimated 16.5 per cent of the Mexican GDP in 1982 to 8.5 per cent in 1983, 5.5 per cent in 1984, and 3.5 per cent in 1986. As for the public sector's use of foreign credit, which amounted US$ 6.9 billion in the first nine months of 1982, should not exceed US$ 5 billion in 1983 as a whole. The program also established a considerable slowdown in the level of growth of the Banco de Mexico's net claims on the public sector as well as of its net foreign and domestic assets.290

286 Gurría (1995, p. 124-125). Lender banks were given the possibility to choose between Libor and U.S. prime rates, which was generally more than the interbank rate that was a deposit rate. 287 Gurría (1988, p. 77). These interest rates margins (spreads plus fees) look high even by bankers standards since, as Friedman (1983, p. 24) observed, "in early 1982, it was safe to say that a country that offered to pay a spread of 3% over LIBOR would (…) be regarded as too dangerous a country risk." 288 Devlin's computations of the negotiated price of credit are based on an index made up of three of the elements of the cost of credit that were subject to renegotiations: commissions, amortization period and margin over LIBOR. See Devlin (1989, p. 192-194). 289 Devlin's computations shows that the negotiated prices of credit in the second and third restructurings were respectively 70 and 84% lower than the terms agreed to in the first rescheduling, but still higher than the levels prevailing before the crisis. 290 For an outline of the economic program and policy targets see IMF EBS/82/208, December 30, 1982. The adjustment programs of the 1986 stand-by and 1989 extended arrangements went along the same lines. See IMF EBS/86/161, July 23, 1986; and IMF EBS/89/66, April 12, 1989.

103

To conform with the IMF's credit condition Mexican policymakers engaged, therefore, in restrictive fiscal and monetary policies. As part of the stabilization program taxes were raised, government spending cut, interest rates boosted, wage increases restrained and price controls lifted. As a result, while between 1977 and 1982 government spending and the money supply increased by 17.1 and 15.1 per cent annually in real terms, the average annual growth rate for the period 1983-89 was -2.7 and 2.3 per cent respectively. In a context of a severe economic and financial crisis, these austerity measures had an additional negative impact on the Mexican domestic economy. While GDP in real terms increased at an average annual rate of 9.2 per cent between 1977 and 1981, it fell by 0.6 per cent in 1982 and grew by 1.6 per cent per year between 1983 and 1989.

A central issue in the failure of these restructuring and economic packages was that the magnitude of interest payments and the scarcity of new foreign financing resulted in significant outward transfers of resources. As is evident from Figure 1.1, after the outbreak of the crisis, Mexico went from being a net recipient to a net exporter of capital, with the amount of net foreign transfers representing about 6 per cent on average, between 1982 and 1988. The negative transfer of resources added to recessionary pressures, while the principal indicators of foreign debt burden showed no signs of improvement. To the contrary, the ratios of foreign debt to GDP and of interest and principal payments to total exports continued to grow over time, and only started to decline toward the end of the decade (see Figure 1.2). As the years passed, the Mexican GDP per capita, employment and other social and living standard indicators deteriorated, while debt problems remained unsolved.291

Figure 1. The net transfer and external indebtedness problems, 1980-1990

Figure 1.1. Nets transfers to Mexico* Figure 1.2. Mexico's external debt burden

* New lending minus total debt servicing on existing debt Source: CEPALSTAT and World Debt Tables.

291 See Krueger (1987), Dornbusch (1986), Sachs (1986) and Devlin (1987) for an analysis of the relationship between high debt services and low growth in developing countries; Thorp (1998), Bértola and Ocampo (2012) and ECLAC (1990)for an overview of the economic and social performance of Latin American economies during the debt crisis of 1980s and their connection with the debt crisis.

104

The last and definitive round of Mexican debt restructuring came with the launch of the Brady Plan in 1989. This new scheme introduced elements of debt forgiveness and uncoupled restructuring agreements with commercial banks from IMF adjustment programs. The deal aimed to restructure US$ 48.4 billion of Mexican debt that was in the hands of commercial banks, who had to exchange the old loan claims for new debt instruments created under the agreement. The options included two debt and debt-service reduction bonds and four old fashioned new-money facilities, with the two debt-relief instruments accounting for about 90 per cent of total converted debt.292 Although the amount of the actual debt relief and its importance for the success of the program are still a matter of debate, consensus exists on the crucial role the Brady plan played in reducing the burden of net transfers and allowing Mexico to overcome the crisis (see Figure 1).293

The discussion above demonstrates how costly debt deals were for Mexico and its domestic economy, but it does not speak to the implications for creditors. A remarkable feature of the international debt crisis of the 1980s was that despite the serious exposure of creditor countries' banks and the fears of a collapse of the international banking system, no bank failure occurred because of sovereign debt problems.294 The fact that Mexico and other highly indebted countries continued to service their debts, although with delay and arrears, allowed these banks to raise capital, increase reserves, and reconstitute their balance sheets. These banks did not suffer any reduction in the value of their loan portfolio or losses until very late in the decade and, indeed, remained profitable during most of the period.295 The burden of the debt crisis for commercial banks seems to have been limited to lowering stock values and increasing primary capital ratios.296 In short, as Kamlani (2008) argues, the protracted debt management process was instrumental in helping the banks bolster their loan-loss reserves and build their equity bases through nearly seven years of healthy profits.297

But what can explain the outcome of debt negotiations and the decidedly uneven distribution of the burden of the adjustments that the management of the crisis and rescheduling agreements

292 These instruments were the discount bond, which entailed a reduction of 35% paying LIBOR plus 13/16 of a point, and the par bond that kept the same original value, but paid a fixed 6.25% interest. Both bonds were to be amortized in one single payment 30 years after issue and had the principal been fully collateralized through a 30-year zero-coupon bond placed in at the Federal Reserve Bank of New York. See van Wijnbergen (1991) for a detail analysis of the implementation and results of the Brady Plan in Mexico. 293 See Gurría (1995). 294 See "The disaster that didn't happen," The Economist, September 12, 1992, and "Latin American debt: The banks' great escape," The Economist, February 11, 1989. 295 Sachs and Huizinga (1987) shows that the large fees and higher interest rates of rescheduling agreements allowed banks to enhance income and that, in some cases, earnings as percentage of assets even exceeded the levels earned before the crisis. Moreover, until 1986, major banks managed to continue to pay dividends to shareholders "as if the crisis hadn't even occurred." Ibid, p. 574. See also "New York banks show strong gains," Financial Times, January 1983; "Banks gains from fees by altering Latin Debt," The New York Times, January 10, 1983; and "Banks greedy over Third World," Financial Times, March 31, 1983. 296 Cohen (1989, p. 4) argues that creditor banks paid an indirect price, as the financial market effectively discounted their LDC papers for them and bid down the value of bank equity instead. This lowered the quotation of their share prices, which was attributed to their heavy third-world exposure, especially in Latin America. In the same vein, when asked whether commercial banks would eventually share some of the debt burden in a workshop at Harvard in 1984, New York Fed President Anthony Salomon expressed the view that "depressed stocks prices [were] a real cost for commercial banks." FRBNY Archive, Central Records, File Bank Advisory Group 1984: Presentation by A. Salomon at Harvard, May 1984. 297 Kamlani (2008, p. 198).

105

entailed? The rationale behind the creditors' strategy for dealing with developing countries' debt problems is found in the influential work of economist William R. Cline.298 He argued that debtor economies faced liquidity problems, rather than a structural-type insolvency and, therefore, should be supported with further lending until the situation was reversed. Financial discipline and adjustment was a necessary counterpart of financial assistance to address their structural economic imbalances and restore market confidence. Under this perspective, the high cost and onerous financial terms that resulted from rescheduling negotiations with Mexico, as well as with other heavily indebted countries, were a logical outcome of the greater risks and tough international credit market conditions that followed the moratoriums and defaults.299

According to this approach, the reprograming terms charged to borrower countries were determined by marketplace logic. Spread over Libor, commission levels, and other components of the cost of credit, such as amortization and grace periods, were to be determined by competitive market negotiations between lending banks and borrowing countries. Better economic performance and consequent lower risk would alleviate the high costs and burdensome financial terms of rescheduling agreements. As opposed to these arguments, Robert Devlin and ECLAC economists claim that debt rescheduling and interest payment refinancing constituted a non-competitive market transaction in which debtor and creditors sat down to administratively decide how to share the loses on a weak portfolio.300 Indeed, if anything, the market framework of these transactions was a bilateral monopoly and the high cost of credit on rescheduled debt represented monopoly rents gained by the creditor banks. Nevertheless, the outcome of debtor-creditor negotiations was uncertain and depended on the bargaining power of parties involved.

Devlin's theoretical framework is well-suited for the actual debt renegotiating process of the 1980s. On one side of the table were the banks, who although big in number, had organized themselves into Bank Advisory Committee. This collusive nature of the committee provided commercial banks with a monopoly power. This power was further increased by the support of creditor governments and international organizations, as, together, they controlled access to credit. The borrowing country with questionable creditworthiness sat at the other side of the table. Such country faced organized creditors individually, and thus negotiated from a structurally weak position. However, given the high exposure of major industrial countries' banks to Mexico and other heavily indebted economies, one would expect these countries to have some negotiating leverage to turn conditions in their favor. In fact, the claim that debtor countries submitted to difficult conditions because they found themselves in weak bargaining positions is not substantiated by Devlin and ECLAC economists.301 What do we know about the factors underlying the powerless negotiating position of Mexico and the burdensome rescheduling terms it agreed to?

298 See Cline (1984, 1985, 1987, 1995) and Bergsten et al. (1985). 299 Compensation for greater risks, premiums against moral hazard, and adjustment on banks' rationing positions were among the factors underlying the outcomes of rescheduling negotiations. See Gasser and Roberts (1982) and Cline (1984, p. 81). 300 Devlin (1989, p. 210). 301 Devlin (1983) cites the case of Nicaragua as an example where the Sandinista government succeeded in rescheduling debt without being charged fees and no spread over Libor, but he does not explains what the basis of its stronger bargain power was.

106

5.3. Excluded debt, bargaining tools, and sovereign debt theory

Although the restructure agreements mentioned above concerned most categories of Mexico's foreign debt with commercial banks, a number of obligations were deliberately pulled out of the generalized rescheduling principles by negotiators. There were, in particular, three categories of bank indebtedness excluded from the rescheduling deals: short-term trade facilities, such as letters of credits and banks' acceptances; tradable instruments, such as publicly issued bonds and floating rate certificates of deposits or notes; and the interbank deposits or placements with the foreign agencies and branches of Mexican commercial banks.302 As of 1983, these three categories of so- called excluded debts accounted for at least as much as US$ 15 billion, or 17 per cent of Mexico's external indebtedness with private creditors (see Table 2).

Table 2. Composition of Mexico external indebtedness in 1983 Millions of dollars

Total External Debt 92,831 Private Creditors 84,993 Commercial Banks 75,174 Public Sector 48,544 that was rescheduled (principal maturing 1982-84) 23,150 Private Sector 26,630 Long-Term 16,490 Publicly guaranteed 1,690 Non-guaranteed 14,800 that was rescheduled under FICORCA 12,000 Short-Term 10,140 Trade-related debt* 4,164 Interbank deposits* 5,976 Publicly -issued bonds* 4,589 Others public sector creditors 5,230 Official creditors 7,838 Multilateral agencies 4,432 Bilateral 3,406

*Excluded from the restructure agreements Source: World Bank's World Debt Tables (several issues) and Gurria (1995)

The outstanding feature of excluded debts was that, contrary to restructured debt, amortizations and interests were promptly serviced at all times during the crisis.303 Mexico's decision to leave certain facilities out of the restructuring scheme and to keep fully current with their payments can be interpreted as an attempt to avoid the implications that this would otherwise entail. Within a rescheduling process, as sovereign debt lawyer Lee C. Buchheit explains, the main objective of the

302 FRBNY Archive, Central Records, Bank Advisory Group Nov-Dec 1982: Telex, December 6, 1982. A similar approach seems to be observed in negotiations with other Latin American debtor countries. Archival evidence shows that in the cases of Argentina, Brazil and Venezuela there were also discussions about whether to include trade credits and interbank liabilities in the restructure agreements or to keep them as distinct from the rest of foreign bank debt See Alvarez and Flores (2014, pp. 130-131) for an analysis of the role that trade finance played during external debt renegotiating in the 1980s and Buchheit (1991) for interbank credit lines.

107

debtor country government is debt relief, and to that end it "will seek to restructure as many categories of its existing debt as possible, and it will forbear to do so in respect to a particular category only if it is persuaded that the attempted restructuring of that category will result in a disproportionate injury or inconvenience to the debtor country's economy or longer-term financial interests."304 Put in terms of sovereign debt theory, it is to say that the costs of defaulting on such debts must have been quite important to the Mexican government.

Indeed, the economic literature on sovereign debt provides us with an interesting framework to evaluate the Mexican debts deals in the light of excluded debts. A main underpinning of this literature is that defaults come with costs for debtors' economies, and that at least a part of these costs relates to possible retaliatory actions by creditors. In this respect, Bulow and Rogoff (1989) argue that the penalties or sanctions that lenders can impose on borrowers, as identified in the literature beginning with the pioneering contribution of Eaton and Gersovitz (1981), are relevant in determining the threat points for renegotiation. In the case of the Mexican debt renegotiating process, where the country faced a syndicate of international banks, excluded debts do not only indicate the costs the debtor is trying to avoid, but also reveal what its vulnerabilities are. Thus, for creditors, these are bargaining tools that can be used to improve terms and conditions for rescheduling.

From this perspective, the decision to not reschedule trade-related debt suggests that external trade was one of Mexican policymakers' vulnerabilities when negotiating with creditors. As a large body of literature has demonstrated, default and debt rescheduling are often associated with a contraction of bilateral transactions between debtor and creditor nations.305 The words of Secretary of Finance Jesus Silva-Herzog, in persuading President Lopez Portillo not to declare a moratorium, makes clear that this was an issue of major concern for Mexican officials: "Mexico import[ed] more than half of its corn consumption from the United States; if we had no money, as we didn't have, and if we had no credit, as we wouldn't have had, in two months the Mexican people would have run out of tortillas."306 More damaging than restrictions in consumer goods would have been a reduction of raw materials and capital goods. A substantial portion of these items were imported from creditor countries and were also heavily dependent on the availability of trade financing.307

Indeed, a main objective of Mexican policymakers during negotiations was to have creditor banks and official export agencies keep credit lines open and avoid a lack of trade finance.308 Back then, commercial banks not only provided long-term finance to the developing world, but they were also

303 See Gurría (1988, p. 76). 304 “Of creditors, preferred and otherwise,” international Financial Law Review, June 1991, p. 12. 305 See, for instance, Rose (2005). 306 Tortillas are a type of thin, unleavened flat bread, made from corn, which are an indispensable product of the consumption habits and pantry staple' in Mexico. Silva-Herzog, "1982: The President's Decision," documentary directed by Diego Delgado and Luciana Kaplan, 2008. 307 At this respect, a study by the Instituto Mexicano de Investigaciones Tecnólogicas (IMIT) estimated that the Mexican private manufacturing sector imported about US$ 1.5 billion in capital equipment in 1982 and would need to import about US$ 1 billion for replacement and spares in 1983; an amount estimated to represent at least 35% of the sector's capital requirements. FRBNY Archive, Box 111377, Mexico Trip – Feb. 21-24, 1984: Internal report, June 13, 1983. 308 Alvarez and Flores (2014) show the important role that trade finance played in the decline of defaulting countries' imports after the outbreak of the crisis and during the subsequent recovery, once defaulting countries had subscribed to IMF programs and committed to creditors' debt management strategy.

108

important intermediators in international trade and export and import financing, where official credit agencies usually acted as a guarantor for emerging economies.309 Like the rest of the international lending to Mexico, access to trade finance became severely distressed with the outbreak of the crisis in August 1982, and the government decision to keep current on these obligations was to prevent short-term trade credit from drying up. In fact, when asked about the reasons why this debt was excluded from rescheduling, Angel Gurría answered "that [it] was the bloodline of trade relationships, and otherwise the structure of trade would [have] become very difficult [since] trade-related financing was largely self-liquidating and self-renewing."310

Less clear, however, are the costs Mexican negotiators expected to suffer if they rescheduled capital market instruments. The bulk of Mexican foreign securities consisted in bonds issued by the government and larger public agencies in the Euromarkets, underwritten by creditor commercial banks and, by the time negotiations started in 1982, in the hands of individual investors or bondholders. According to Kraft (1984), long discussions were held between Mexico and international banks in the early stages of negotiations about whether to include the bonds in the rescheduling exercise or not: the reason why they were finally excluded was the negligibility of the amount.311 This amount, however, was very similar to other excluded debt facilities, but in those cases, a fundamental reason behind the exclusion existed. One reason could be that bonds, which were in the hands of a large number of individual and institutional investors, were simply more difficult to restructure than bank loans. Other possibility is that with the international banking credit market closed to Mexico after the moratorium on bank loans, the Mexican government looked to protect the bond market because it was the remaining gateway to foreign finance.312 Non- restructuring bonds would prevent the country from harming its credit standing in that market and facilitate the use of bonds as substitute for international bank loans. This policy may have actually been sound, since under the auspices of Brady Plan, troubled loans were successfully converted into securities issued in the Eurobond market.313

The third and largest category of excluded debt was interbank obligations, which consisted essentially in deposits placed by creditor banks with the foreign agencies and branches of Mexican commercial banks. From the mid-1970s, as the previous chapters show, leading Mexican banks had opened overseas offices that relied on borrowing in the Euromarkets and in the U.S. money markets to fund loans to final borrowers back home and in other Latin American countries. When the crisis hit, interbank lending also shrank, and the agencies were confronted with liquidity stress. Mexican negotiators quickly looked to secure their financial situation and, as in the case of trade finance,

309 Letters of credits issued by commercial banks, particularly to finance imports of developing countries, were the most common instrument and were usually insured by export credit agencies. 310 FRBNY Archive, Central Records, C261 Mexican Government 1917-1984: Office memorandum, August 23, 1982. 311 See Kraft (1984, pp. 43-44). 312 An influential body of research has largely discussed why the repercussions of failing to respect debt contracts, such as increases in borrowing costs or exclusion from credit markets, are a main concern for debtor governments and a fundamental reason why they may want to keep honoring their debts. See Eaton, Gersovitz, and Stiglitz (1986), Eaton and Gersovitz (1981), Grossman and Van Huyck (1988), Kletzer and Wright (2000), and Cole and Kehoe (1995). 313 Mexican bond issues were interrupted with the outbreak of the crisis. The last bond was issued by Telmex on August 4, 1982, and started again almost seven years later, on June 8, 1989, in the wake of the Brady Plan. See Negrete Cárdenas (1999, p. 410), Table B.15.

109

their position was targeted for having international creditor banks keeping open wholesale funding lines, and rolling over interbank placements as they fell due.314

The question of whether to restructure interbank deposits was a matter of thoughtful discussion between Mexican policymakers and their international creditors. In talks with the members of the Bank Advisory Committee, the Mexican government inquired about the possibility of folding the agencies liabilities into the broader debt rescheduling agreement: "the unanimous view of the bankers was '[d]on't try it.'"315 A formal restructuring of these facilities was a de facto default, which would disappoint creditor banks' expectations and was likely to shrink the interbank credit market for Mexican banks.316 According to the international banks, "no rescheduling had ever covered these interbank deposits" and any attempt to do so "would not only fail, but would also cause problems for the proposed rollover."317 With regard to this issue, Angel Gurría himself pointed out some years later that "the Mexican private banks could not be allowed to default abroad in their money-market activities."318

Moreover, although within certain limits, it appeared that Mexican negotiators went to great lengths to protect domestic private banks' access to international wholesale liquidity. The establishment of extraordinary facilities for the settlement of private external debt under the FICORCA program, for instance, or the outright nationalization of these banks could be related to the need to secure interbank deposits.319 The accumulation of arrears on interest payments of Mexican private debt seems to have been a major obstacle for the rollover of interbank placements, especially with small regional banks. In the words of Larry Miller, the Chemical Bank official responsible for handling the Mexican agencies problems for the Advisory Group, "the fact that the banks were not getting any payment to speak of on private debts was tending to make them more aggressive in trying to draw money out of the agencies."320 In that same vein, in a meeting with IMF officials, William Rhodes of Citibank stressed that "partly as result of the small banks' frustration over the private interest problem, the Mexican banks (all state owned) have lost about $500 million in deposits through the agencies." 321 In fact, the nationalization of the banking system on September 1, 1982 itself, though controversial at a domestic level, was perceived as a good sign by creditor international banks, which

314 When asked by creditor banks about their short-term deposits with the foreign agencies and branches of Mexican banks, Angel Gurría said that "the Mexican banking sector was the backbone of the country's economic progress [and that] he would hope that Mexican banks would be supported for a return to normalcy." He added that "if other banks have a perception of difficulty in collecting from Mexican banks, if they withdraw deposits, they make the perceived problem a real problem." FRBNY Archive, Central Records, C261 Mexican Government 1917-1984: Office memorandum, August 23, 1982. 315 FRBNY Archive, Box 108406, Sam Y. Cross Chronological Files August-December 1982: Notes on the Meeting of G-10 Governors and Switzerland at the BIS, September 27, 1982. 316 See Buchheit (1991). 317 FRBNY Archive, Box 108406, Sam Y. Cross Chronological Files August-December 1982: Notes on the Meeting of G-10 Governors and Switzerland at the BIS, September 27, 1982. 318 See Gurría (1988, p. 76). 319 Through the mechanism FICORCA, from Spanish Fideicomiso para la Cobertura de Riesgo Cambiario (Trust Fund for Covering Exchange Risk), the public sector took over the exchange risk but not the private sector commercial credit risk, which remained with the original creditors. For san explanation of how the mechanism worked see Gurría (1988, pp. 79-83). 320 FRBNY Archive, Box 108406, Sam Y. Cross Chronological Files August–December 1982: Office Memorandum, October 19, 1982. 321 FRBNY Archive, Central Records, Bank Advisory Group Nov-Dec 1982: Office Memorandum, November 18, 1982.

110

had feared that the Mexican government would not stand behind the debts of the private financial institutions. It could have even been instrumental for the interbank cause since, as The New York Times reported, "it [was] expected that, as a result of the nationalization, international banks [would] be willing to place funds with Mexican banks."322

5.4. Default and implications for the banking systems

Within the negotiation, a main implication of the Mexican government's strong determination to look after interbank and trade finance was to give creditor banks a lever to step up their bargaining power. But while the costs of being deprived from international funding would certainly have been substantial for Mexico, the bargaining over external debt rescheduling was not one-sided. In fact, it could be said that the Mexican government actually did have leverage, since creditor banks would not have directly benefitted from cutting the country off from the international capital markets. Moreover, in pushing the country to refuse to reschedule and declare outright default, creditors risked inflicting damage on their own banking system because their large commercial banks were significantly exposed to Mexican debt.

Indeed, from the outset of the Mexican crisis, the primary U.S. and European concerns were the risk to their major commercial banks and the vulnerability of the international banking system. When arguing for approval of credit facilities to Banco de Mexico by the U.S. Congress, President Ronald Reagan stated that Mexico's "extreme balance of payments problems [were] of a magnitude and character which (…) could [have led] to substantial disruption of international money, financial and exchange markets."323 The possibility that Mexico might stop servicing its debts and that major banking institutions would no longer be able to attract deposits was a major fear of industrial countries' policymakers.324 In fact, as developed in Section 5.2, the creditors' strategy to deal with the crisis appears to have been trying to keep the country afloat and the banks profitable. This would give them the time and resources to raise capital and increase reserves until they were in position to make loss provisions and write down the value of their portfolio, without compromising confidence in the banking sector.325

But Mexico could not easily think of leveraging foreign banks' financial exposure if their own domestic banks had also been compromised. In a context where both sides are simultaneously exposed to default, the balance of bargaining power between debtor and creditors is determined by the relative exposure of their banking systems. It is difficult to think that the Mexican government could realistically push creditors into concessions if the exposure of domestic banks was larger than that of its foreign counterparts. To the contrary, in such a case, foreign creditors would call the shots within the external debt renegotiating game. Overall, the risk of a default for both the Mexican and creditor countries' banking system depended on the size and concentration of the loan portfolio of

322 “Takeover pleases U.S. Banks,” The New York Times, September 2, 1982. 323 Letter to the speaker of the House and the President for the Senate on the provision of credit facilities to the Bank of Mexico. 324 See Volcker (1983) and Volcker and Gyohten (1992, pp. 187-227), Chapter 7. 325 In the words of Paul Volcker "when credit [was] negotiated, the creditors and the borrower normally reached agreement on some increase on interest (and/or fees) (…), banks [had] more income with which to make provisions for any expected losses. In many cases, this income [was] clearly used to supplement reserves or augment capital." FRBNY Archive, Central Records, Bank Advisory Group 1983: Letter, March 7, 1983.

111

major banks to Mexican debt, but also on the potential disruptive effects this could entail in the international interbank markets where these banks operated.

Although largely overlooked in the literature, by the time of the default, large domestic banks had a considerable amount of Mexican external debt in their balance sheets. Historical records of Banco de Mexico show that by late 1982, the loan portfolio of the foreign agencies and branches of Mexican commercial banks totaled US$ 4.68 million. Of this amount, US$ 4.27 million or 91.3 per cent were credits granted to Mexican borrowers while the balance of US$ 408.5 million or 8.7 per cent consisted of claims on foreign clients.326 The public sector, which included the central government, public companies and agencies and development banks, was the main recipient of the loans granted by these agencies. They accounted for 62.6 per cent of the total loan portfolio while the remaining 37.1 per cent consisted in lending to the private sector.

Table 3 provides a comparative perspective on the Mexican exposure of leading banks in the U.S. and Mexico.327 Major U.S. money-center banks are a convenient benchmark for the comparison because they were Mexico's main foreign creditors and far more exposed than any other U.S., European or Japanese bank.328 In 1982, the six major U.S. commercial banks, which accounted for about 45 per cent of the total exposure of the U.S. banking system to Mexico, had outstanding claims on Mexico of US$ 11.2 billion. The two largest, Citibank and Bank of America, who were also Mexico's main creditor banks, had US$ 2.7 and US$ 2.5 billion of Mexican debt in their books: these amounts represented about half their capital base.329 For Manufacturer Hanover and Chemical Bank, as Table 3 exhibits, the exposure was a little larger with claims on Mexico representing 58 and 62 per cent of the banks' capital base, respectively.

These values show that the exposure of individual large U.S. banks to Mexico was high, but the figures for their counterparts in Mexico were even more worrisome.330 Mexican bank exposure is calculated as the ratio of total lending to Mexico by overseas agencies and branches to the capital base of the parent bank. By that time, Mexico's six largest banks were involved in foreign lending, and their international loan portfolio consisted of direct or syndicated credits granted from overseas offices. These foreign agencies and branches were not independently capitalized, but rather integrated to the balance sheet of the parent bank in Mexico, like any other domestic branch.331 By late 1982, total outstanding loans to Mexico from the overseas banking offices of Banamex and Bancomer, Mexico's two largest private commercial banks, represented 4 and 4.6 times their capital base, respectively (see Table 3). These two banks accounted for about half of the domestic banking market share, and along with Serfin, Comermex, Banco Internacional and Somex, whose exposure

326 Banco de Mexico Archive, UT C3405Exp.8: Crédito otorgado por agencias y sucursales de bancos mexicanos en el exterior, 1982-1984. 327 Exposure to capital provides a visible picture of the vulnerability of a bank because it indicates the potential disruptive effects of loans loss provisions and the possibility that depositors became unwilling to leave their funds in a bank with capital and reserve levels that seemed insufficient to withstand the impact. 328 Data compiled by Huizinga (1989) for top banks in major creditor nations shows that the Mexican exposure as percentage of equity of the 5 largest U.S. banks was 49 in average during 1987-88 while for their counterparts in the UK, Japan and Germany the average ratios were 28.2, 21.8 and 18.6% respectively. See Table 7-7, p. 136. 329 This means total capital, which includes equity capital, surplus and undivided profits. 330 Normal levels of non-performing loans as percentage of capital would have been in the 1 to 2% range. 331 Mexican banks' international lending consisted of direct or syndicated credits they granted through overseas agencies and branches as shown in Chapter Four.

112

was even larger, they represented up to three quarters of the Mexican commercial banking system. These figures make clear that Mexico's domestic banks were much more seriously exposed to their home country's debt than any U.S. bank, and possibly more than any other foreign creditor bank.

Table 3. Exposure of leading commercial banks to Mexico, December 1982

Millions of dollars Percentage

Foreign Loan Total Total Loans to Portfolio Assets Capital Mexico FLM /LP LM / TA LM / TC

(FLM) (LP) (TA) (TC) US Banks Citicorp 2,725 79,224 121,482 5,495 3.4 2.2 49.6 Bank of America 2,500 72,523 119,869 5,247 3.4 2.1 47.6 Chase Manhattan 1,687 52,057 77,230 3,844 3.2 2.2 43.9 JP Morgan & Co 1,082 30,376 56,766 3,306 3.6 1.9 32.7 Manufacturers Hanover 1,730 42,222 59,195 2,945 4.1 2.9 58.7 Chemical Bank 1,500 29,740 45,011 2,413 5.0 3.3 62.2 Total Six U.S. Banks 11,224 306,142 479,553 23,250 3.7 2.3 48.3 Mexican Banks Banamex 1,135 3,178 7,767 280 35.7 14.6 404.7 Bancomer 1',200 3,167 8,006 260 37.9 15.0 462.3 Serfin 428 1,807 4,351 114 23.7 9.8 375.0 Comermex 624 1,651 3,074 52 37.8 20.3 1'202.5 Banco Internacional 266 1,475 2,004 42 18.1 13.3 641.1 Somex 621 2,399 3,520 73 25.9 17.7 855.0 Total Six Mexican Banks 4,275 13,676 28,723 820 31.3 14.9 521.1

Note: FLM for Mexican banks are the loans granted from their foreign agencies and branches to Mexican borrowers. Source: Salomon Brothers, CNBS, Call Report FFIEC 002 and Bank of England.

Although more modest than exposure to capital, Mexican banks were also in a relatively weaker situation than U.S. creditor banks, when comparing the shares of total bank loans. While foreign loans to Mexico represented between 3.2 and 5 per cent of the loan portfolio of the six largest U.S. banks, the corresponding values for Bancomer and Banamex were 26.4 and 22.2 per cent, respectively. Banca Serfin, Comermex and Banco International also held higher shares than U.S. banks (see Table 3). Additionally, Mexican banks had a substantially less diversified international portfolio. At an aggregate level, lending to Mexican borrowers represented about 6.9 per cent of the foreign total loans portfolio of all U.S. banks but as much as 76 per cent for Mexican banks'.332 These figures illustrate that Mexican financial institutions would most likely have been unable to withstand the impact debt servicing problems by their own government and private sector.333

332 Based on data from FFIEC's Country Exposure Lending Survey for December 1982 and Banco de Mexico's Historical Financial Statistics. 333 The corollary is that default on external debt was a much more serious threat for Mexico than for creditor countries. Partial or selective default was not a realistic possibility, since Mexican banks were participating with foreign banks as part of syndicated loans. This made it impossible to default on some banks and not

113

A second channel through which the Mexican crisis could potentially endanger the banking system was the international interbank market. During the 1970s, as the Euromarkets grew, the interbank market operations came to play a larger role and, indeed, formed the majority of the international activities of commercial banks from developed countries. As shown in Figure 2, in 1982 commercial banks from the BIS reporting area had foreign claims for about US$ 1.7 trillion, of which 30 per cent was due by non-banking institutions, while the remaining 70 per cent or US$ 1.2 trillion were outstanding interbank claims. Although the bulk of these interbank claims corresponded to transactions among banks from reporting countries, about US$ 401 billion or 40 per cent were owed by banks from all around the world. By that time, outstanding interbank claims of commercial banks from the BIS reporting area in Mexico reached around US$ 11.2 billion or 1.1 per cent of total interbank claims of BIS reporting countries' banks.

Figure 2. Evolution and Importance of the interbank market

Note: BIS reporting countries are Austria, Belgium, Luxembourg, Denmark, France, Germany, Ireland, Italy, Netherlands, Sweden, Switzerland, U.K., Canada, Japan, the U.S. (including offshore branches) Source: BIS.

Lack of liquidity in the interbank market as a consequence of a default by a major borrower country, such as Mexico, was considered the largest threat to the stability of the international banking system. The collapse of Herstatt Bank and the Franklin National Bank in 1974 as a result of foreign exchange problems had already shown the consequences of payment disruptions in the interbank market to central banks from industrial countries, and forced them to intervene in order to avoid a liquidity crisis and prevent more banks from failing.334 The high volume and the uncollateralized nature of the funds transferred informally by telephone (over-the-counter), as well as the cumulative structure of deposits relent several times between banks, and the low level of regulation were all factors that exposed the market to potential contagion effects and caused financial authorities to worry about its stability. In a context where outstanding interbank claims formed over

others. Besides, through the operation of cross-default clauses common in most loan agreements, a technical default call by one bank would trigger all creditor banks to declare the same. 334 See Spero (1999) for an account of the failure of Franklin National Banks and the challenges it posited for the stability of international banking.

114

two thirds of international banking in BIS reporting countries, the risks of a major disruption in the market for the entire banking system was considerable.

Indeed, central bankers from industrial countries had long discussed the dangers of a moratorium or major delays on debt servicing from a large borrower for the interbank market. As early as 1977, the Bank of England started to prepare a series of reports and papers about the importance of large debtor countries vis-à-vis the international banking system, in which they studied the possible consequences of default by Brazil and the Eastern Bloc. In particular, the file "Apocalypse Now" discussed the vulnerability and the possible collapse of U.S. and European banks as a result of a loss of confidence linked to international and domestic loan losses. This paper was discussed extensively among G-10 central bank governors at the BIS meetings of the Standing Committee on the Eurocurrency Market and the Basel Committee on Banking Supervision. They particularly focused on the implications for crisis management in the interbank wholesale dollar markets.335

The level of exposure to interbank market liquidity problems and its implications for national banking systems varied depending on the individual circumstances of the borrower bank. One important difference to be drawn is between banks that primarily fund their international lending from retail deposits and those that rely mainly on the interbank wholesale market for funding. A second distinction, partially related to the previous one, is between banks with their own dollar funding base and those with lower capacity to access foreign exchange, or, in other words, non- dollar based banks. And finally, there is the issue of the relative amount of outstanding loans to the defaulting country or to other countries perceived to be in similar positions. Banks that lacked their own dollar base, relied primarily on the interbank market for funds, and were largely exposed to the troubled country's debt, were likely to first face interbank liquidity problems and to have the greatest difficulties in overcoming possible funding strains. This was precisely the situation of Mexican banks operating in the international capital markets.

Therefore, the difference between Mexican and U.S. banks was not only related to exposure, but also to the fact that Mexican banks were non-dollar based and largely dependent on the interbank market for funding.336 By 1982, interbank money from foreign banks represented as much as a fifth of total liabilities of the six largest Mexican banks, but it were their single and most important source of foreign exchange. In contrast, highly exposed U.S. money-center banks were net lenders in the interbank markets, had large retail deposits and were naturally dollar-based. Therefore, a liquidity crisis in the international wholesale markets promised to be potentially much more problematic for Mexican banks than for their U.S. counterparties.

5.5. Mexican foreign agencies and the interbank situation

As of June 1982, 21 agencies and branches of Mexican commercial banks were operating in the main international financial centers with consolidated balance sheets of about US$ 7.7 billion. The U.S. was the main destination, with six offices in New York and four in Los Angeles, accounting for 37.7 per cent of Mexican foreign agencies' total liabilities. In the branch hierarchy, London held second place, with a 27.5 per cent share, but there were also considerable levels of activity in Caribbean offshore centers, especially the Grand Cayman Islands (see Table 4). Bancomer and Banamex, the

335 Bank of England Archive, Apocalypse Now, 3A143/1 to 3A143/5. 336 See Chapter Four.

115

largest market players, had four overseas offices each and a total US$ 4.2 billion on their balance sheets in equal shares. They were followed by Comermex and Banco Internacional (US$ 1.1 billion each), then Serfin (US$ 0.7 billion) and Banca Somex (US$ 0.4 billion).337 At a consolidated level, as much as 95 per cent of agencies' total liabilities consisted of funding facilities from banking institutions or interbank liabilities while the remaining 5 per cent were deposits from the non-bank sector.338

Table 4. Total liabilities of Mexican commercial banks' foreign agencies June 1982, millions of dollars Foreign agencies and branches Total New Los Cayman London Nassau Total Bank York Angeles Islands Banamex 580.0 159.2 887.7 0.0 420.1 2,047.0 13,528.5

Bancomer 1,004.6 376.6 523.3 260.7 0.0 2'165.2 13,252.1

Banca Serfin 33.1 228.2 310.0 0.0 201.3 772.7 5,635.7

Cormemex 204.9 215.6 400.9 280.0 25.0 1,126.4 4,724.7

Somex 1.1 0.0 0.0 479.3 0.0 480.4 5,157.7

Banco Internacional 105.8 0.0 0.0 1,015.4 0.0 1,121.2 2,854.0

Total 1,929.6 979.5 2,122.0 2,035.4 646.4 7,712.9 45,152.7

Source: FFIEC 002, Bank of England's archives and CNBS.

Up to the outbreak of the crisis in fall 1982, uncollateralized interbank operations between developed countries' banks and the foreign agencies of Mexican and Latin American banks in general were common business. Interbank deposit placements or credit lines within banks in major financial centers were considered to be risk-free, spreads were small, and the volume traded was relatively high.339 During normal times, when interbank deposits came due, Mexican agencies would call the creditor banks to roll over their debts, either by renewing the deposit directly with them, or by borrowing from some other bank and refunding the first. However, as bad times emerged and the crisis broke out, the view of the risk behind interbank transactions changed: greater caution, higher spreads, stricter bank and country limits, reductions in liquidity, especially for banks from troubled countries.340 Lending banks became more selective on which banks they dealt with. Indeed, there was a flight to quality, with international banks being offered increasing amounts of interbank deposits.341

337 They had five, two, four and three foreign offices each, respectively. 338 CIEN-A19/E-89/Marzo de 1983, "La banca antes de la nacionalización," p. 17. For a detail description of the funding structure for the case of the U.S. agencies and branches of Mexican banks see Chapter Four. 339 BIS (1983, pp. 9-15). 340 Ibid, pp. 32-35. 341 As head of Citibank's international money markets division Mike Rice stated, "People were so risk-averse that banks like Citibank are offering more deposits than ever." "From Brazil? Just a minute, Sir," Euromoney, July 1983.

116

As expected, Mexican agencies experienced a worrying deterioration of interbank funding lines in the aftermath of the government moratorium declaration. As the perception of country risk increased, creditor banks wanted to reduce their involvement in Mexican agencies and stopped rolling over interbank deposits with them to reduce exposure. U.S. regional banks, along with Japanese banks, were the first to pull back and demand payment on their interbank claims when they came due. Indeed, two waves of large withdrawals took place before the bank nationalization and the 1982 IMF–World Bank Annual Meetings in Toronto, with a run against Mexican banks in the interbank market on September 7 that was successfully controlled by the end of the day.342 During the following months, creditor banks continued to draw money out of the agencies. Although withdrawals were mostly in modest amounts, the cumulative effect of these drains created considerable liquidity pressures.343

Mexican agencies managed to prevent further erosion and leakages of interbank funding at a high cost in terms of interest. During the meeting with international commercial bankers of August 20, Mexican external negotiator Angel Gurría asked them to roll over liabilities with Mexican agencies, which, in return, had revised spreads upward and "increased yield substantially."344 While, for instance, the spread for Eurodollar certificate of deposits would have normally been 25 basis points prior to the crisis, by the autumn of 1982, the yield had risen to 300 basis points over T-bills.345 As for interbank deposits, while Mexicans used to pay a 1/8 per cent over Libor or prime, rate premiums rose up to 3/4 or 1 per cent, depending on individual bank and its creditors. Additionally, creditor banks also began to require a fee or commission of 1/8 to 1/4 per cent on their funding lines that had not been charged before.346

In terms of the maturity schedule, there was a considerable shortening with agencies being forced to rely more on placements up to three months and less on six-month liabilities as they had done before the crisis. As of September 1982, approximately 80 per cent of total aggregated liabilities of Mexican foreign agencies came due before the end of the year and the remaining 20 per cent over the following months.347 The problem was that while agencies' liabilities were highly concentrated in very short-term interbank instruments, because most assets were long term loans or illiquid claims, the adjustment capacity was limited. In this respect, the Bank Advisory Committee for Mexico estimated that Mexican foreign agencies were about US$ 6 to 6.5 billion mismatched in terms of

342 See Boughton (2001, p. 301) and Rhodes (2011, p. 185). 343 BAC estimated that Mexican agencies lost about US$ 800 million—an erosion of 10-15%—of interbank deposits between August and December 1982. FRBNY Archive, Central Records, C261 1981-1982 Brasil – Banco Central do Brasil 1981-82: Office memorandum, December 24, 1982. 344 FRBNY Archive, Central Records, C261 1917-1984 Mexican Government: Office memorandum, August 23, 198 345 "From Brazil? Just a minute, Sir," Euromoney, July 1983. 346 Premiums of 200 points were considered very expensive at the time. For instance, by the time of Herstatt crisis, Japanese banks were paying premiums of up to 2 percentage points over and above the going market rate, the Japanese authorities requested their banks "to refrain from rushing to borrow and from paying extremely high premiums." Bank of England Archive, Apocalypse Now, 3A143/4: Bank of Japan's paper, December 1980. 347 Bank of England Archive, Task Force, 13A195.2: Note, October 1982. An estimate of about US$ 750 million by September 3, US$ 1.25 billion by mid-September, and US$ 6-6.5 billion by December 31, 1982. FRBNY Archive, Central Records, C261 1917-1984 Mexican Government: Office memorandum, August 30, 1982.

117

their dollar balance sheets. In the words of advisory committee's chairman, William R. Rhodes of Citibank, this was a "real bomb."348

Table 5 exhibits the net position of Mexican foreign agencies in London as the percentage of total claims in different maturity bands and similar data for their interbank businesses.349 It shows the great degree of maturity transformation done by the Mexican agencies and how substantially the mismatch increased between August and November 1982. Lending in the three year and over band increased from 29.5 per cent of total claims to 37 per cent of total claims, while the sources of funding shortened. In August, the proportion of liabilities with a maturity in excess of three month amounted to 36 per cent, while by mid-November this had fallen to 17 per cent. The table illustrates the extent that the Mexican branches were relying on the interbank market to fund longer term lending to non-banks and how dramatically their funding sources shortened after the outbreak of the crisis. Both in terms of assets and liabilities, interbank operations in the three-month and above bands increased considerably between August and November.

Table 5. Maturity analysis of the business of Mexican agencies and branches in London, 1982 Liabilities (-) / Claims (+)

Net position Interbank business as percentage of as percentage of as percentage of total interbank total interbank total claims liabilities claims 18-Aug 17-Nov 18-Aug 17-Nov 18-Aug 17-Nov Less than 8 days -5.8 +0.6 -6.3 -1.7 -11.8 -3.6

8 days to less than 1 month +2.1 -13.5 -1.6 -24.2 -2.9 -51.6

1 month to less than 3 months -14.2 -27.8 -22.1 -33.9 -41.1 -72.2

3 months to less than 6 months -14.1 +1.1 -21.6 +0.1 -40.1 +0.3

6 months to less than 1 year +2.0 -0.1 -0.1 -1.0 -0.2 -2.0

1 year to less than 3 years +7.0 +7.1 +1.1 +2.1 +2.0 +4.6

3 years and over +23.6 +30.4 +4.5 +5.5 +10.8 +11.8

Source: Bank of England's archives.

Within this balance sheet structure, dire interbank funding began to place strong pressures on the funding and liquidity position of the Mexican foreign agencies. As a matter of fact, after the crisis broke out, agencies' officials scrambled to secure their liquidity position. First, they reacted by reducing, indeed "shrinking," the short-term portfolio where they could, while rolling over the long- term portfolio to adjust their balance sheet position. More importantly, they worked hard to confirm overdraft, advance and stand-by back-up lines with their correspondents and partner banks in the

348 FRBNY Archive, Box 108406, Sam Y. Cross Chronological Files August-December 1982: Office memorandum, August 25, 1982. 349 Interbank here is defined as positions with UK banks and with banks overseas. The data is given as a proportion of both total interbank liabilities and claims, since total interbank liabilities of these branches are significantly larger than their total interbank claims.

118

U.S. and Europe.350 Overall, although most agencies still managed to purchase overnight and term money in the Federal funds and Eurodollar markets, in the eyes of the managers, the prospects for future funding looked "awfully shaky."351

The financial position of the agencies was further complicated by a number of additional factors. First, while the cost of interbank funding considerable increased for Mexican banks, a significant portion of their assets continue to pay fixed interest rates. Second, interest rates on the foreign loans to Mexican borrowers were to be restructured in the long term, but the interbank liabilities continue to be arranged on a very short term. Finally, there was the currency mismatch between liabilities owed in dollars and claims that, although also denominated in dollars, were mostly due by Latin American debtors operating in their own national currencies. In a context where debtor countries suffered from balance of payment difficulties and their currencies were frequently devalued, the credit risk on their portfolio loan would only increase further.

Thus, the interbank situation not only entailed liquidity strains, but it also threatened to turn into a solvency problem. Mexican agencies had no liquidity cushion and very limited access to alternative non-interbank dollar funding. Since none of them had subsidiary status in the host countries, they could not operate in the retail market and thereby did not have recourse to deposits from the public for funding. In fact, agencies and branches were not technically defined as banks, which implied that they were not subject to reserve requirements nor separately capitalized from their parent banks. Indeed, they fell outside the scope of host countries' banking regulations and were primarily supervised by their home authorities. This means that Mexican agencies in the U.S. and London were not under the protection of the Federal Reserve System or the Bank of England, respectively, and had to look back home for help in case of an emergency.

5.6. Who bail them out?

The interbank difficulties encountered by Mexican foreign agencies went beyond the border and became a matter of international concern. In September 1982, when reporting the situation of the interbank market to the Central Bank Governors of the G-10, Alexandre Lamfalussy from the BIS described what he saw as a consistent and gloomy picture: increasing tiering among banks and banking systems, a progressive cutback in credits and deposit lines, eventually leading to shrinkage of interbank position and a halt in cross-border transactions between banks.352 As expected, the outbreak of the crisis in Mexico had not only created problems for Mexican banks, but it also disrupted the overall functioning and stability of the international interbank and money markets.353

350 Although agencies originally dealt trough money market brokers, since the brokers had been unable to find them money, they started to deal directly with banks with whom they have credit lines. FRBNY Archive, Central Records, C261 1917-1984 Mexican Government: Office memorandum, August 25, 1982. 351 Ibid. 352 FRBNY Archive, Box 108406, Sam Y. Cross Chronological Files August-December 1982: Notes on G-10 Governors meeting held at the BIS, September 27, 1982. 'Tiering' occurs when, instead of having uniform interest rates applicable to all participants, there is a differentiation in the spread according to the nature and the nationality of the borrowing bank. Immediately after the Herstatt crisis, the number of tiers increased and the range of rates expanded to exceed 1%, when the usual range was ¼%. "Eurobanks and the inter-bank market," Quarterly Bulletin of the Bank of England, September 1981. 353 Looking at the correlation of risk premium between developed and developing countries' banks, Saunders (1988) finds that while the correlation was very in 1981, it rose to +0.5 at the time of the debt crisis in the

119

In this respect, during that same meeting of the Eurocurrency Standing Committee at the BIS, Bank of England Governor, Gordon Richardson pointed out that the Mexican interbank market situation involved more than 1,000 banks and that it "did not affect just a few financial centers in the U.S., U.K., Switzerland, etc., but concerned everyone."354 Moreover, the "problem of branches and agencies, not only [involved] Mexican banks but also Brazilian, Argentinian and Korean, and others, whose liabilities were owed to the interbank market and whose assets were not liquid."355 Foreign agencies of commercial banks from other large debtor countries suffering from external debt payment difficulties were also coming under growing pressures in the interbank market, raising more general systemic concerns. This led financial authorities of host countries to discuss how to deal with these liquidity problems.

The appropriate lender of last resort for foreign banking offices had been heavily debated since the early 1980s. Controversy arose among G-10 Governors over whether subsidiaries and branches experiencing liquidity difficulties should be assisted by the host authority or by their head offices that, in turn, if necessary, should be supported by their own authorities. Overall, the position of the supervisory institutions was that "the host authority should be the primarily responsible party for the supervision of foreign establishments' liquidity in the domestic currency of the host country, but that for all other currencies, the parent authority should have primary responsibility."356 These basic guidelines for the supervision of banks' foreign establishments had been worked out after the Herstatt-Franklin crisis and set up in the 1975 Basle Concordat, which provided a framework for the allocation of certain responsibilities.357 The main point made on liquidity issues was that while responsibility for supervising the liquidity of foreign banking establishments rested primarily with the host authority, home authorities could not be indifferent to the moral responsibilities of the parent institutions.

But what raised many more concerns was as to whom, if anyone, would support banks from developing countries in the major international centers if they were to run into liquidity difficulties. A statement from a Bank of England official illustrates the dichotomy central bankers faced: "if [the necessary support for a foreign bank in London] was not forthcoming would [the Bank of England] be prepared to stand by and watch foreign bank become insolvent or would [it] find [it] necessary to act to protect London's reputation as an international center?"358 Unlike industrial countries, no clear safeguard existed, since reserve exchanges could only provide token debt support, and interbank claims on these countries were also limited.359 Had a bank defaulted on interbank obligations, it would likely have triggered a dangerous domino or knock-on effect, raising the prospect of a systemic collapse in London, as well as in other financial centers. The Herstatt collapse

autumn of 1982, and declined again by April 1983. This means that contagion was high around the crisis and that the problems of developing countries’ banks were affecting the industrial countries as well. 354 FRBNY Archive, Box 108406, Sam Y. Cross Chronological Files August-December 1982: Notes on G-10 Governors meeting held at the BIS, September 27, 1982. 355 Ibid. 356 Bank of England Archive, Apocalypse Now, 3A143/2: Paper draft, June 1980. 357 The Concordat was a supervisory agreement for international banks drawn up by the Basel Committee on Banking Supervision, formed by G-10 central banks in 1975. 358 Bank of England Archive, Apocalypse Now, 3A143/2: Paper draft, June 1980. 359 Kane (1983, p. 117).

120

had made very clear to central bankers that the failure of small active banks could transmit large losses to solid banks and spread to the rest of the banking system through the interbank market.360

The liquidity problems of Mexican agencies during the aftermath of the moratorium brought these policy issues into sharp focus.361 Along with securing interbank funding lines with creditor banks, the agencies and host countries' financial authorities also discussed the possibility of accessing central bank financing. In the U.S., for instance, agencies officials acted to ensure they were in a position to borrow from Federal Reserve discount facilities in the event of an emergency. None of these agencies had deposit insurance from the Federal Deposit Insurance Corporation (FDIC) or had filed the documents and met the necessary requirements to access the discount window. Even if they had, the Fed's position on this was far from clear: the use of discount facilities was to be studied carefully and, if available, would be very limited in nature.362 Borrowing from the Bank of England did not look promising either since, as stated by Governor Richardson, "they would not provide any help, since the need of Mexican agencies was for dollars and not sterling."363 Deprived from lender of last resort coverage in host countries, Mexicans agencies were pushed to handle funding problems with their own resources.

Indeed, Mexican foreign agencies were receiving shipments of currency from Mexico around that time. In August 25, 1982, in talks with FRBNY officials, Banamex New York agency vice president Clifton T. Hudgins stated that agency had received US$ 31 million from Mexico during the previous week and that they were expecting more to come.364 This represented dollars collected by the head office through foreign exchange conversions in Mexico. Moreover, after the nationalization on September 1, the Mexican central bank began to strongly support the foreign agencies of domestic commercial banks, so they could meet dollar liquidity needs. Although no systematic records exist, data compiled in a FRBNY memorandum shows that between Tuesday 7 and Wednesday 22 of September 1982, "known" funding from Banco de Mexico to Mexican foreign agencies reached US$ 311.3 million; of this amount, US$ 162.1 million or 52 per cent had been granted during the first week at the time of the interbank run. Banamex was the main recipient with US$ 118.5 million (38 per cent), followed by Multibanco Comermex and Bancomer with about US$ 62 million (20 per cent) each.365

Parent banks in Mexico had no means of providing their overseas agencies with the dollars they needed to overcome the leakage in interbank funding. As previously mentioned, foreign banks were

360 The Herstatt crisis exerted the most contractionary impact the Euromarket had hitherto experienced, with interbank position being the most severely affected. According to Moffitt (1984, p. 90), at the time of the failure, Herstatt and Franklin had outstanding interbank foreign exposure of US$ 200 million and US$ 2 billion respectively. These figures give an idea of how damaging the US$ 6-6.5 billion interbank mismatch of Mexican foreign agencies could have been at an international level. 361 Discussions about this were also pushed by the Banco Ambrosiano issue. 362 That was the message of FRBNY Official Sam Cross to Mexicans when discussing on this matter. FRBNY Archive, Central Records, C261 1917-1984 Mexican Government: Office memorandum, August 30, 1982. 363 FRBNY Archive, Central Records, C261 Mexican Government 1917-1984: Office memorandum, August 30, 1982. 364 FRBNY Archive, Central Records, C261 1917-1984 Mexican Government: Office memorandum, August 25, 1982. 365 Banca Serfin, Banco Internacional and Banco Mexicano-Somex completed the list with 9, 7 and 6% respectively. It is likely that Banco de Mexico was actually transferring more funding in unknown amounts. FRBNY Archive, Box 108406, Sam Y. Cross Chronological Files August–December 1982: Note, September 1982.

121

virtually the only suppliers - and the international interbank market the solely source - of genuine dollars for Mexican banks. At an aggregate level, dollar liabilities of the commercial banking sector reached US$ 24.5 billion by June 1982. Of this amount, US$ 10.4 billion or 42.4 per cent was due to foreign banks, while the balance consisted mostly of saving instruments with the domestic non- financial sector. These were the so-called Mexdollars (dollar-denominated deposits held in Mexican banks) and did not represent a real source of foreign exchange. They were old peso deposits that had been converted into dollar-denominated deposits. They were mostly held in the Bank of Mexico because they were subject to high reserve requirements, ranging from 70 to 100 per cent in some years. Thus, deprived from foreign finance, the only way Mexican banks could afford to send dollars to the foreign agencies was through the support of the central bank.

In a context of dwindling reserves, however, Banco the Mexico did not have any realistic way to support the foreign agencies with its own resources. Figure 3 makes explicit the limited scope of international reserves compared to the potential foreign exchange needs of the agencies. The estimated US$ 6-6.5 million mismatch on their dollar balance sheet represented about 3 to 3.5 times the volume of international reserves during the August to December 1982 period. In fact, a considerable part of the dollars being shipped to cover the liquidity of the agencies came from the BIS and U.S. facilities granted to Mexico after the moratorium. As Gordon Richardson reported to his counterparts in a BIS meeting, of the US$ 311.3 million paid by Banco de Mexico to the overseas agencies and branches of Mexican banks during September 1982, US$ 218.3 million or 70 per cent was covered by drawing on U.S. swaps. Additionally, a "substantial portion" of the US$ 1.85 billion BIS bridge-loan approved in August 1982 was parceled out to repay a part of the outstanding claims with Mexican banks during interbank panic of Tuesday, September 7.366 Indeed, "most of the BIS- U.S. swap drawings have been for the purpose of providing funds for the Mexican offshore agencies and branches."367 These figures paint a clear picture of how much Mexican financial authorities relied on foreign capital to deal with the difficulties of domestic banks' foreign agencies.368

Hence, the liquidity position of the foreign agencies was a major headache for Mexican policymakers and the main challenge for external debt negotiators. As the central bank lacked the resources to stabilize their financial position, and had great difficulties and costs in securing the rollover of interbank facilities - even after excluding them from general restructuring agreement, one alternative would have been to simply let these overseas agencies and branches fail. Had this happened, the supervisory authorities in the host country would have taken control of the agencies' business and properties and liquidate them to reimburse creditor banks.369 But in such an event, it is

366 Boughton (2001, p. 301). 367 These are Governor Richardson's words, to which he added that this was not what the facilities were designed for. However, he said there were still funds available in the first tranche of the U.S. and BIS facilities and agreed to Wallich’s proposal "to handle the use of the facilities for the benefit of the agencies day by day." FRBNY Archive, Box 108406, Sam Y. Cross Chronological Files August–December 1982: Note, September 1982. 368 Banco de Mexico could not meet the Mexican banks' need for foreign exchange to settle external obligations. When discussing the possibility for Mexican agencies to access the Fed discount windows with FRBNY officials, it was suggested that Mexicans should consider dealing with any problems with their own resources, to which Angel Gurria replied "[they] would love to happily support Mexican banks" but, he added, "[they were] a little short of cash." FRBNY Archive, Central Records, C261 Mexican Government 1917-1984: Office memorandum, August 30, 1982. 369 In 1980, the New York State Banking Department took possession of the New York branch of Banco de Intercambio Regional, declared bankrupt by order of the Argentine Central Bank, and was liquidated to

122

unlikely that the proceeds from the liquidation of the assets would have been large enough to enable head offices to recover the amounts they had assisted the agencies wit, and thereby compromise the solvency position of parent banks. In the case of Banamex, for instance, the US$ 118.5 million sent by the head office to its foreign agencies between September 7 and 22 represented about 40 per cent of its capital base.370 This helps to understand why Mexican authorities went to such great lengths to protect foreign agencies' access to international interbank liquidity.

Figure 3. Monthly reserves of Banco de Mexico, 1980 - 1982

Source: Banco de Mexico's Annual Report (1983)

In the end, Mexican negotiators and the Advisory Committee devised an interim solution to the interbank deposits problem. As part of the rescheduling agreement, creditor banks agreed to maintain interbank deposits with Mexican agencies at the August 1982 pre-moratorium levels. A covenant was introduced in the restructuring loans documentation, which provided that an event of default would be triggered if the aggregate level of interbank liabilities with the offshore agencies and branches dropped below US$ 5.2 billion. Although this did not convey any legal assurance to Mexican policymakers, it did raise the stakes for a creditor bank seeking to withdraw its deposits. For their part, the Mexican government committed to making sufficient funds available to process market interest payments on agencies' interbank accounts and to provide the dollars that domestic debtors would need to reimburse their external debts.

The freezing of interbank deposits, which was initially arranged to conclude in 1986, would be extended for two additional periods. First, with the Financing Packages of 1986/1987, the expiration date was extended to June 1989, and then again to late 1992 as part of the 1989-92 Financing reimburse foreign banks not protected on their US$ 80 million claims, which caused a run against other Argentine banks with a total loss of reserves of about US$ 2 billion. IMF Archive, ERB/83/2000 Supplement 1, November 29, 1983. 370 By late September 1982, Banamex's head office had net deposits with its U.S. agencies of US$ 103.1 million or 35% of its capital. In the case of Bancomer, the corresponding figures were US$ 169.7 million and 55%. These figures would have certainly been higher if the agencies in London and other offshore centers were taken into consideration.

123

Package of the Brady Plan. In practice, the agreement implied that creditor banks would ensure that deposits rolling over 90 days were renewed and then renewed again, whenever they were about to expire. The freezing of interbank deposits proved to be an effective solution to liquidity drain with Mexican agencies and would be maintained for almost ten years, until 1991. This reflects the extent to which Mexican policymakers could not afford to be cut off from the international financial market and the extent that they depended on foreign capital to protect the agencies and the domestic banking system.

True, non-compliance on the part of global banks on maintaining interbank deposits with Mexican agencies that triggered a default could give some leverage and strengthen the bargaining position of Mexicans officials when negotiating with international creditors. However, as discussed above, Mexican banks were in considerable worse shape than foreign creditor banks to face funding strains and deal with a potential liquidity crisis in the international wholesale markets. Arguably, it could be expected that creditors have less to lose from a default on interbank debt than the Mexicans themselves. Hence, in terms of the bargaining power the gun was most likely in the hand of foreign creditor banks rather than the other way around.

5.7. Conclusion

This chapter on the situation of Mexican banks in the face of the crisis shows the importance of considering Mexico's renegotiating positions as imbedded in the problems of its domestic banking sector. The fact that domestic banks were more exposed to Mexican debt and to liquidity pressures in the international wholesale markets are revealing findings. They signify that the impact of a default would most likely have been much worse for Mexican banks than for its foreign creditors.

Also revealing is the fact that much of the BIS-U.S. facilities granted to the Mexican government in the aftermath of the moratorium declaration were used to support the foreign exchange needs of the overseas agencies of Mexican banks. This does not only reveal the willingness and decision of Mexican financial authorities to bail agencies out and stand behind parent banks, but it also shows that they didn't have the foreign exchange to do it and depended on foreign financial assistance to support them. But not even these facilities were enough and the stabilization of the Mexican agencies situation eventually required the commitment of creditor banks not to pull over and freeze interbank deposits with Mexican foreign agencies at pre-moratorium levels. The volume of interbank deposits frozen reached the significant amount of US$ 5.2 billion, representing several times the capital base of the banks and as much as 20 per cent of the total liabilities of the domestic banking sector. This amount was even larger to the US$ 5 billion in new loans that creditor banks agreed to grant to Mexico as part of the first rescheduling agreement.

These findings suggest that there may have not been any realistic exit option to deal with Mexican banks' problems. The lack of alternatives and its decisive dependence on foreign capital to stabilize the situation of the domestic banking sector impaired the negotiating position of Mexican negotiators when facing international creditors. External financial assistance came at a high price, not only in terms of the financial conditions but also because of the tough adjustment and austerity programs Mexico accepted to carry out in order to access foreign capital. In a context in which Mexican banks were much more compromised than foreign banks, the threat of a default could not look like a credible move to scare creditor countries policymakers and force them to ease

124

renegotiating terms. The prospects of a domestic banking crisis is a forgotten but crucial element in the long list of costs the Mexican economy could expect to suffer in case of refusing orderly restructuring or repudiating its debts, such as loss of trade finance, international trade retaliation or legal harassment. Collapse of major domestic banks looks like a price Mexican policymakers could not afford to pay.

Finally, an important implication of this chapter is with respect to the weak bargaining position of debtor countries as a group. Contrary to the group of international creditors, which included international organizations, industrial countries' commercial banks and creditor governments, Latin American countries failed to consolidate a common negotiating position. A number of attempts were made, such as the Action Plan of Quito and the meeting in Cartagena in June 1984, but they all failed to form a cartel of debtors. Although this issue requires more detailed research, one of the causes accounting for the failure in coordinating debtors' collective action relies precisely on the situation of their banking sectors. In fact, large commercial banks of main Latin American countries were not only exposed to debt in their home countries but in their regional neighbor countries as well. In the case of Mexican banks, 24 per cent of their foreign claims were owed from Latin American borrowers. This shows the complexity of interest at stake in a situation in which Mexico had on the one hand a common interest with other Latin American debtors to collude against its creditors, but on the other hand sat on the creditors' side during the renegotiating rounds with their regional neighbors.

125

Final Remarks

The expansion of foreign bank lending to the developing world was a salient feature of international finance during the decade leading up to the international debt crisis of the 1980s. After the oil shock of 1973 and the onset of big trade surpluses of oil exporting countries to the Eurodollar markets, international commercial banks from advanced industrial countries relent these funds to developing countries. Although largely neglected in the literature, state-owned and private commercial banks from borrowing countries also went international during those years. They became increasingly involved in the petrodollar recycling process through syndicated and individual bank loans to government and private-sector borrowers both at home and abroad.

This thesis has focused on the international expansion of Mexican banks in connection with the sovereign debt crisis of 1982. Between 1972 and 1974, leading Mexican commercial banks took their first step into the international capital markets through the creation of London-based consortium banks in partnership with banks from developed and developing countries. To more fully engage in international business, their next step was to open their own branches and agencies in London and the U.S., the major world financial centers at that time. From their associated consortium banks and the network of foreign offices, Mexican banks took on an important role in international lending. This study shows that through their involvement with international finance, domestic banks became increasingly intertwined with external debt payment problems of the public and private sector in Mexico.

A major concern of this research is the factors that drove Mexican bank internationalization and their implications on the financial condition of the banking system. Chapter Three shows that domestic funding problems, linked to high inflation and negative real interest rates, coupled with increasing competition from foreign bank loans, were at the base of the banks' decision to go abroad. Foreign finance offered banks new funding sources for their lending and to secure their market share from foreign banks. Allowing the banks to reach the international capital markets proved helpful in providing financing for the Mexican government's state-led development strategy and the fixed-exchange rate policy of the central bank. Indeed, the Mexican government was directly involved through its partial ownership of Intermex, the biggest Mexican consortium banks, as well as Banco International and Banca Somex, two of the six commercial banks involved in international lending.

The financial position of the Mexican domestic banking sector increasingly weakened in the buildup to the debt crisis. Mexico's six largest commercial banks, Bancomer, Banamex, Serfin, Comermex, Mexicano-Somex and Banco Internacional, were among the most affected. Overall, they displayed considerable worse than average capital adequacy levels, a less liquid asset portfolio, and a more instable funding base than the rest of the domestic banks. These problems were, to a large extent, related to their growing involvement in the international financial system, and, in particular, to their increasing reliance to foreign borrowing for funding. The more dependent they were on foreign funding, the larger the exposure to currency risk and the more susceptible the banks became to international financial fluctuations. Mexico's international banks were much more aggressive in

126

financing their growth with debt instead of shareholder equity than domestic banks with no access to foreign finance, and they were thus more leveraged by the time the crises broke out.

The second concern of this thesis is the mechanism employed by the banks to intermediate foreign liquidity with domestic borrowers. As developed in Chapter Four, Mexican banks could not easily access foreign liquidity from their headquarters in Mexico, so they opened their own agencies and branches in major international financial centers to do so. These banking offices were not strictly defined as banks: they were not independently capitalized nor subject to reserve requirements, and, although they could not raise deposits from the public, they did have access to the large international wholesale markets. The business model, therefore, consisted in borrowing in the U.S. dollar and Eurocurrency interbank markets to relending these funds in Mexico. In the context of fixed exchange rates, foreign liquidity allowed the banks to raise funding that was cheaper than domestic deposits, and offer credit lines to their clients at interest rates that were competitive to those of foreign banks.

Such a business model was fragile by design and only further aggravated by risky asset liability management decisions. The use of very short-term interbank borrowing to fund longer terms loans led to the accumulation of important maturity mismatches on banks' balance sheets. Additionally, while these interbank credit lines were set at floating rates (LIBOR or U.S. prime), a significant part of the loans were arranged at predetermined fixed rates. The third problem was linked to currency risk and the cross-border nature of the activities they undertook. While banks' dollar liabilities were owed to foreign bank creditors, their dollar claims were mainly with Mexican clients, which ran their businesses largely in pesos and were not necessarily exporting firms. The increase of international interest rates in the early 1980s, the devaluation of the peso in early-1982, and the ensuing external debt payment problems of Mexican debtors aggravated the interest rate, currency, and maturity imbalances accumulated by the banks, and thus compromised their financial position.

Finally, the last concern of this study relates to the exposure of the domestic banking system to the debt crisis and how it affected the negotiating strategy of the Mexican government. Chapter Five shows the extent that the six largest Mexican commercial banks were exposed to their own country's default. On one hand, international loans with home country final borrowers extended several times over the capital base of the banks. On the other hand, in the aftermath of the moratorium, Mexican banks found it increasingly difficult to raise funds in the international wholesale markets. Because of such high exposure, lending banks began to regard Mexican banks with suspicion, perceived interbank operations more risky, and therefore reduced their involvement with them. The debt crisis, thus, not only put the solvency of the banks at stake, because of the large volumes of Mexican debt they held in their books, but more importantly, it also led to a considerably increase in funding risks.

The situation of the Mexican banking system constrained the renegotiation strategy of the Mexican government during the debt rescheduling process. This study shows that Mexican policymakers relied upon foreign capital to provide financial assistance to troubled banks and that they went to great lengths to protect their access to interbank deposits. The heavy dependence on foreign capital to secure the stability of the domestic financial system meant that Mexican negotiators were in a weak bargaining position when they entered negotiations with international creditors. Moreover, because the exposure of their own banks was much larger than U.S. and European banks, a default

127

or unsuccessful rescheduling would have probably been less costly for foreign creditors than for Mexico itself. This sheds light on why Mexico ended up carrying the burden of the debt crisis and explains why they had little option but to accept the conditions creditors demanded.

Insights for the literature on debt crises

This study of the Mexican debt crisis of the 1980s reveals the limits of the economic theory of sovereign debt as analytical framework for explaining defaults. According to this literature, given the absence of higher enforcement authorities, the reason why sovereigns repay debts is because defaults are costly. The alternative interpretation of this assumption is that, whether partial or total, excusable or unjustifiable, defaults occur whenever the gains from abrogating foreign debt commitments become higher than the costs of defaulting. This implies interpreting debt crises —the immediate consequence of a default— as the result of a cost-benefit analysis where, either because of unfortunate states of the world or other political or economic factors, sovereigns decide of stop servicing their debts. Thus, according to this approach, sovereign debt repayments depends more on unwillingness to pay and is less related to solvency per se.

Yet, the Mexican response to the 1980s debt crisis is one of willingness to pay. Since the very onset of the crisis in August 1982, when the debt moratorium was announced, the Mexican government went to great lengths to prevent an interruption of debt payments. The main testimonial and historical accounts of the crisis, such as Kraft (1984), Boughton (2001), or even William Rhodes (2011), the chief negotiator on behalf of the banks, each made clear that Mexican authorities were committed to reimbursing foreign creditors and avoiding unilateral decisions. But the strong determination to repay foreign debt, which was, as this thesis argues, indeed motivated by considerations about the potential costs of defaults, did not preclude the country from entering into a sovereign debt crisis. From an analytical perspective, this leads to the conclusion that solvency and capacity to pay must be factored into the explanation of the Mexican crisis of 1982.

The commitment of Mexico to repay its external debt is likely not a surprise for those familiar with the history of the crisis. When Mexico Finance Minister Silva Herzog and his team flew to Washington on Friday, August 13, 1982, there were simply not enough reserves in the Banco de Mexico to meet the debt obligations to foreign creditor banks which were to come due the following Monday. The country was finally able to repay the banks because of the emergency funding arranged with the U.S. Treasury Department as part of the frenzied negotiations undertaken during that weekend.371 The story that followed was one of seven years of numerous austerity programs, debt restructurings, financial assistance packages, and economic recession, in which the government was in constant fear of a unilateral default. Indeed, it appears that the Mexican government did whatever it took to be able to service public and private sector external debt. Hence, this calls for incorporating factors other than those strictly linked to cost-benefit considerations.

Taken in its broader international and historical context, the 1982 Mexican default is one among many other sovereign debt crises that broke out around the same time. The fact that debt payment problems affected countries with different growth performance, external circumstances, political regimes, domestic economic policies, and foreign borrowing reliance, points to the presence of common external factors behind the crises. The combination and interaction of external shocks,

371 See Kraft (1984, pp. 5–7) and Boughton (2001, pp. 290-294).

128

namely the rise of international interest rates and the deterioration of the terms of trade, with risky policies and macroeconomic mismanagement led indebted developing countries into situations where they were no longer able to repay. Another perspective places these crises in a Kindleberger- Minsky framework and look at the factors that drove the bank lending boom and unsustainable foreign borrowing by defaulting countries. After all, Mexico was forced to declare the moratorium on principal payments only when it was unable to continue rolling over bank debt. This, in turn, was the 'displacement' event that disrupted the international capital markets and triggered crises all over the world.

While these explanations focus on macroeconomic factors, this study shows the importance of looking at individual actors in debtor countries to understand the dynamics of the crisis. On one hand, the direct presence of public and private commercial banks in the international credit markets benefited the Mexican government and large private companies —normally part of the same business groups as the banks— with a more direct line to foreign capital. This infusion of foreign capital was an additional force that intensified the already ongoing process of overlending and external indebtedness, as Diaz-Alejandro has explained in the case of Chile.372 On the other hand, since commercial banks were in control of important debt flows in the balance of payment, they influenced the currency exchange market and the domestic monetary policy. In other words, the macroeconomic context that banks operated in and the macro variables that framed their decisions, such as inflation or the exchange rate, were endogenously determined by bank behavior. This draws banks to the center of almost any explanation of the Mexican debt crisis, whether focused on macro or microeconomic considerations.

A main implication of the growing involvement of Mexican banks in the international financial system relates to the political economy of foreign lending in the buildup to the debt crisis. The massive expansion of international lending to developing countries was not only supported by the governments of creditor countries, as has been largely addressed in the literature, but was also further encouraged by borrowing countries themselves. This thesis shows that, indeed, the international expansion of Mexican banks and their intermediating role with foreign capital was instrumental to accommodate the economic policy objectives of the Federal government and the central bank. However, the fact that through their international activities, Mexican banks incurred risks that also threatened the international banking system raises the question about the role of supervisory authorities in the countries they operated in. After all, foreign authorities granted them the permission to begin banking businesses in their countries, and it is highly unlikely that U.S. and British financial regulators were blind to Mexican banks' risky operations.

Thus, the previous discussion suggests that the Mexican foreign debt repayment suspension in August 1982 was not so much the result of a thoughtful cost-benefit analysis by the government, but rather a matter of circumstances. At both a domestic and international level, the economic situation of the country and its external indebtedness had become so vulnerable and unstable that a debt crisis was somehow inexorable. The increasing international interest rates, unfavorable terms of trade, high domestic inflation, large fiscal deficits, devaluating currency, and problems in the banking system were a ticking time bomb. The Mexican default, therefore, resulted from a combination of factors that related to the capacity and ability of the country to reimburse its debt, regardless of the

372 See also Arellano (1983a, 1983b)

129

government's position on payments. If anything, analysis by Mexican policymakers found the costs of defaulting to be greater than the potential benefits, but they still could not avoid entering into a moratorium.

Although not well-suited to provide the rationale to explain the outbreak of the crisis, the cost- benefit analysis approach is insightful for understanding the debt renegotiating process. Once in default and barred from the international credit market, access to foreign finance was no longer a market transaction, but rather a bilateral negotiation between Mexico and the international creditors. To be able to borrow new foreign loans, creditors demanded that Mexico meet certain requirements and conditions related to recessionary economic adjustment programs. However, if creditors pushed too far and Mexico repudiated, they could have harmed themselves because their home banks were exposed. Thus, the negotiation outcome depended on the bargaining power of the two parties, which, in turn, was determined by the relative costs of failing to reach an arrangement. The fact that Mexico finally carried the burden of the crisis creates two implications: first, that the costs of unilateral default outweighed than the benefits; second, Mexico would have suffered more than creditors would if the country had repudiated its foreign debt.

Mexican banks were more exposed to their own debt than foreign creditor banks, and they could not possibly take losses without compromising the confidence of depositors. Had a unilateral suspension of debt payments been declared, creditor banks would have pulled back and demanded payment on their interbank claims, which would have brought Mexican foreign agencies and parent banks under severe liquidity and solvency strains. In fact, as part of the rescheduling agreement, foreign banks committed to rollover interbank deposits when they came due. This thesis makes the point that domestic bank failures were a highly likely outcome and this was a price that Mexican policymakers wanted to avoid at all costs. Both the fragility of the domestic banking system, as well as Mexican financial authorities' dependence on foreign finance to secure the financial position of major banks allowed international creditors to force Mexico to participate in the debt management strategy, despite the devastating social and economic consequences it produced in the country.

Towards a new understanding of the crisis

A main contribution of this thesis is that the Mexican crisis of 1982 was not limited only to sovereign default, but also embraced the banking sector. As recent literature has shown, because financial institutions and governments are interconnected in a number of ways, distress in one sector tends to generate difficulties in the other and Mexico in 1982 was no exception. Moreover, the debt crisis and the banking problems overlapped with major devaluations of the peso after years of stability, drawing the currency exchange market into the financial turmoil. However, despite the actual coexistence of triple banking, currency, and debt problems, explanations for the crisis have been largely approached from sovereign indebtedness perspective and have not taken the situation of the domestic banking sector nor its connection with the currency crisis into account.

This raises a number of revisionist questions in light of the current literature on financial crises. What is the causality relationship between the debt crisis and problems in the domestic banking sector? Does the Mexican case follow the Diaz-Alejandro (1983) account of the Chilean crisis, where the massive use of Central Bank credit to bail out the banks seems to have been at the cause of the government's fiscal problems and the sovereign default in 1983? Or was it the debt crisis that

130

brought on the banking sector's troubles? Through what channels did the banking and public sector affect each other? How does the devaluation fit into the story? Was the currency crisis the cause of the banking sector's problems? Or did causality work the other way around, as, according to Kaminsky and Reinhart (1999), usually is the case? Although addressing causality issues is far beyond the scope of this study, it is possible to outline the anatomy of the Mexican financial meltdown of 1982 by considering the sequence of events within the crisis and the foregoing stylized facts.

This study's central argument is that the fundamental reason for the fragility of the Mexican banking system can be found in foreign finance. More specifically, the asset-liability mismanagement, along with the accumulation of interest rate, currency and maturity mismatches on the books was a consequence of bank involvement in international lending and foreign borrowing. However, although these imbalances exposed the banks to external shocks and shifts in the international capital markets, as well as to foreign exchange and debt payment problems in Mexico, they do not by themselves explain how distress built up in the banking sector. Moreover, the simultaneous presence of currency and debt crises and problems in the banking sector begs the question of the link between these problems. This thesis makes clear that the devaluation of the peso and the debt payment in both the public and private sector affected the liquidity and solvency position of the banks, but it says little about the flip side of these implications.

Since there were no banking failures in Mexico, it is difficult to analyze the causality links between banking and the currency crisis based only on the chronological order of events.373 The explanations that point to financial sector problems as the root of currency collapses usually stress the presence of central bank bailouts or government guarantees that ultimately explain the currency crash by excessive money creation, as in the Chilean case (Diaz-Alejandro, 1985; Velasco, 1987). However, despite isolated episodes of liquidity provision though special auctions in 1980, the annual reports of Banco de Mexico and the balance sheets of commercial banks show no sign of atypical or considerable use of the discount window or other financial facilities from the central bank. If anything, the bailout was arguably the nationalization of the banking system, but it took place six months after the currency crisis of February 1982. This does not mean, however, that the banks may have not tied the hands of the central bank in defending the currency. After all, the government blamed them for running speculative attacks against the Mexican peso and engaging in capital flights.

Whether at the origin of the currency crisis or not, the February 1982 devaluation directly impacted the balance sheets of the banking sector by increasing the weight of foreign liabilities on the balance sheets of the leading internationally oriented banks. Between 1975 and late 1981, foreign dollar interbank debt went from 0.7 per cent of the total liabilities of the Mexican commercial banks to 13 per cent and then to 20 per cent, after the fixed exchange rate regime collapsed. The banks, however, had sought to hedge the currency risk of their cross-border operations by lending in the same currency that they borrowed in abroad. Hence, to the extent that the devaluation increased

373 There is consensus in the literature that there was a banking crisis in Mexico surrounding the sovereign debt crisis of 1982, but there are discrepancies about the date, according to the different definitions used by scholars. For Kaminsky and Reinhart (1999), the banking crisis took place in September 1982, at the time of the nationalization of the Mexican banking system, and peaked in June 1984, while for Laeven and Valencia (2008) there was a systemic banking crisis in Mexico in 1981. In the Mexican banking historiography, however, there are no references to a banking crisis during those years.

131

the share of dollar claims in the same proportion as the foreign liabilities, there was no visible amplification of the currency mismatch on their balance sheets. The problem was that the dollar loan portfolio was largely concentrated in Mexican borrowers that operated in pesos and had no direct access to foreign currency. If the devaluation affected the banks, it would have been through the worsening of the balance sheets and the repayment capacity of their clients.

Indeed, as private Mexican firms were the first to suffer the consequences of the devaluation, banks bore the brunt of this repercussion. Two months after the devaluation, Grupo Industrial Alfa, Mexico's biggest private-sector conglomerate and international borrower, informed its national and foreign creditor banks that it was suspending the payments of all debts. Difficulties in reimbursing credits to the banking sector were not limited to Alfa, but rather a generalized phenomenon among private sector borrowers. Banks involved in international finance, as well as those operating at only a national level experienced an increase of loans arrears and a deterioration of asset quality beginning in February 1982. The fact that there were no differences between both groups of banks in this regard makes it difficult to establish a clear link between loans financed by borrowing in foreign currency and the financial position of the banks. To put it differently, the cross-border foreign exchange exposure of leading commercial banks and the currency crash do not look like the immediate cause, nor the most lethal source, of problems in the banking sector.

The outbreak of the debt crisis in August 1982 appears to have been even more damaging than the devaluation and repayment problems of the private sector. Like major creditor commercial banks from the industrialized world, the six largest domestic banks also had Mexican debt on their books, but their level of exposure was much larger. Individual and syndicated loans to the Mexican government and private-sector borrowers accounted for about a quarter of the loan portfolio and between four and five times their capital base. Faced with such exposure levels, the problem for the banks was not the potential net losses of troubled loans but, much more importantly, the threat of a run by depositors and, consequently, failure. A default or unsuccessful restructuring of Mexican debt that could lead to a write-down of assets would have ultimately made the banks insolvent. Because the banks at issue accounted for as much as three quarters of the general public's deposits, the debt crisis represented a major threat for the entire banking and domestic financial system.

The impact of the debt crisis on the banks was not limited to the deterioration of the assets but, more importantly, to the effects on the liability side. The increasing reliance on short-term foreign interbank liquidity as source of funding had raised the exposure of the banks to liquidity strains that could have developed in the international wholesale markets. The government moratorium declaration curtailed the supply of interbank credit lines to Mexican banks. The heavy exposure to its own government's debt raised concerns about bank solvency and made creditor banks unwilling to roll over maturing interbank deposits with them. But banks were ill-prepared to manage the funding risks and overcome foreign liquidity strains: they had no alternative genuine source of dollars and no lender of last resort coverage. Further erosion and leaks of interbank funding were prevented at a high cost in terms of interest and the commitment from international creditor banks to keep funding at the pre-moratorium levels, as part of debt rescheduling agreements with the Mexican government.

The fact that the Mexican government brought down the countries' major banks does not mean that the banks themselves were not the origin of the debt crisis. Access to international wholesale

132

liquidity not only provided the banks with a new source of loanable funds. Because it was not subject to reserve requirements, its effect in expanding the supply of credit was stronger. Thus, at least compared to domestically-oriented banks with no access to foreign liquidity, while it didn't appear that internationally-oriented banks necessarily made poor loan decisions, along with foreign international commercial banks, they may have contributed to the crisis by fueling the growth of public and private sector external debt. Moreover, Mexican banks were an endogenous and unstable force in the international capital market that simultaneously pushed the demand and supply of foreign capital upwards, thereby exacerbating the borrowing and lending boom that came to a definitive end in August 1982.

Arguably, Mexican banks would have succumbed to the same fate even in the absence of government or private sector debt payment problems. The business model behind international banking activities was highly risky: it led to a worsening in maturity and interest rate mismatches, and a poor capital structure that weakened the liquidity and solvency position of the banks. Presumably, this model's most dangerous flaw was its vulnerability to shifts in the international money and interbank markets and its exposure to funding shocks. A default by another large international borrowing country, such as Brazil or Argentina, or the occurrence of any event likely to distress the wholesale market would have triggered similar liquidity strains on Mexican banks. Despite their exposure to home borrowers, since they were highly reliant on interbank funding and had no dollar retail deposit base, Mexican banks were inherently risky. Had liquidity strains developed in the international interbank markets for reasons other than the outbreak of the debt crisis, Mexican banks still would have been among the first to face a curtailed funding supply.

A novel perspective on bank nationalization

Such weakness and exposure of the commercial banking system raise the question of the programme of bank nationalization. On September 1, 1982 the Mexican government decided by presidential decree to expropriate the possessions of all private credit institutions. Ranging from physical installations to financial assets and shares or participations in other enterprises, all the properties of the private banks, with the exception of those of Banco Nacional Obrero and Citibank, were transferred to the Mexican public sector. Nationalization encompassed the banking sector's Mexican assets and included the transfer of ownership, control and administration of the overseas branches and agencies of Mexican commercial banks and, as a consequence, of their foreign assets and liabilities.

The official justification for Lopez Portillo's nationalization measure was the need to deal with the large amounts of capital draining out of Mexico. In the eyes of the president, a State-controlled banking system would enable the government to stop the outward flight of dollars that was being operated through the banking system. The concomitant establishment of formal capital controls with the introduction of a two-tiered exchange rate was also part of the policy package that Carlos Tello, new Governor of Banco de Mexico, considered necessary to control capital flight. From a political perspective, the decision was embedded in a context of rising tension and conflicts between bankers and policy makers during the previous decade. The nationalization of the banking system

133

has indeed been present in left-wing parties' proposals since at least the 1960s and attracted growing support from them in the late 1970s.374

The traditional interpretation of the nationalization of the Mexican banking sector is that it was an opportunistic decision taken for political and ideological reasons. In the context of a severe crisis, the Mexican president accused the country's banks of encouraging, and providing mechanisms for, capital flight. In doing so, he was purportedly seeking to shift some of the blame for the country's economic problems from his administration in order to salvage the political image and power of a weakened government. More generally, the measure is regarded as resulting from the long-standing confrontation between two different factions within the Mexican government, neoliberals and structuralist economists, over the national development strategy and the role of the financial system within it.375 It has been understood in this regard as aiming to strengthen the position of the public sector and punish the private sector while providing nationalist policymakers with the control of the banking system they required to pursue economic restructuring.

The considerable focus that has been placed on the political motives for the nationalization has drawn attention away from the perilous situation of Mexican banks that were the target of the program. The recent book "La nacionalización bancaria, 25 años después, la historia contada por sus protagonistas" [The bank nationalization, 25 years later, the story told by its protagonists], edited by Amparo Espinosa Rugarcía and Enrique Cárdenas Sánchez and published by the Centro de Estudios Espinosa Yglesias in 2008, is probably the most comprehensive study available on the program and is a representative example of standard accounts of the Mexican bank nationalization. It investigates the nationalization, based on oral histories, and compiles in three volumes the testimonies of the main participants, including successive Mexican presidents and senior government officials, bankers, judges, private entrepreneurs and analysts, in order to revisit the process of nationalization from a historical perspective. As in other important studies such as the one by del Angel, Bazdresch and Suárez Dávila (2005), there is little reference to the condition of the Mexican banking system in the wake of the crisis and at the time of the nationalization.

Yet, a number of Mexican officials and prominent banking scholars have pointed out the potential importance of financial and economic factors for explaining the nationalization. Gustavo del Angel, for instance, notes that "the expropriation was a controversial political move, but perhaps a mechanism to bail out a banking system on the edge of collapse."376 Along the same lines, Carlos Marichal adduces that "possibly, the [nationalization] was unavoidable because after the devaluation [of 1982] many Mexican public and private banks have to be rescued."377 For Angel Gurría, the leading Mexican external debt negotiator, although acknowledging that the process could have been managed differently and that the decision might have been taken for the wrong reasons, contends that nationalization was a way of solving the financial difficulties of the banks that would otherwise have forced them to declare insolvency.378 Certainly the weak liquidity and solvency position of leading domestic banks constituted a veritable banking crisis, as I have shown in this dissertation, and is very much in line with these suggestions.

374 See del Angel, Bazdresch Parada, and Suárez Dávila (2005). 375 See Cordera and Tello (1971). 376 del Angel (2002, p. 229). 377 Marichal (2011, p. 124). 378 Interview held on July 9, 2013.

134

From an international perspective, the Mexican nationalization was not an exceptional event but similar to measures undertaken in other Latin American countries with troubled banking sectors. In Argentina, for instance, more than 70 institutions (accounting for 16 per cent of commercial bank assets and 35 of financial companies' assets) were liquidated or subjected to intervention between 1980 and 1982.379 In the case of Chile, its systemic banking crisis of 1981 led to a series of major interventions by the government and persuaded the central bank to inject massive amounts of capital into the banking system to avoid its collapse, thus constituting a de facto nationalization.380 In Peru, a systemic banking crisis broke out in 1983 and in 1987 President Alan Garcia nationalized the banking system also as a means of preventing capital flight and financial speculation that adversely affected the balance of payments. Prior to these Latin American cases, there was the French bank nationalization program, which appears to have played an important role as inspiration for the decisions taken by Latin American governments.381

In terms of the Mexican debt management strategy, the nationalization aligned with the broader strategy of policymakers to intervene and stand behind the debts of the private sector. In Mexico, as in other Latin American countries, the government provided considerable facilities to guarantee the servicing of private external debt, which in some cases was partially or completely taken over by the public sector. The FICORCA, for instance, was a subsidizing foreign exchange program that assisted private enterprises with both the pesos and the dollars they needed to serve their rescheduled foreign debt with international commercial banks. Private obligations to foreign suppliers or export credits also benefited from the assistance of Banco de Mexico and the Mexican government. As for the banking sector, while in disagreement with the expropriation, Carlos Abedrop Dávila, President of the Association of Mexican Bankers (ABM), proclaimed that the "only thing that was nationalize[d] [was] the high dollar indebtedness of the private banks."382

The position of the international financial community with respect to the nationalization was also more ambiguous that it might initially appear. Notwithstanding the fact that nationalization conflicted with their commitment to private ownership, international creditor banks did not object to the Mexican government's decision to assume ownership of the Mexican banking sector. To the contrary, most U.S. bankers applauded the nationalization as a much-needed step to bolster international confidence in the Mexican private banks and save them from insolvency.383 In the words of the spokesman for the Bank of America, "this [was] a positive step in that it puts the Mexican Government clearly behind its banking system."384 Whether the Mexican state would stand behind the debts of the private banks was a major concern for international creditor banks, which were refusing to rollover interbank deposit with Mexican banks or to lend them money, thereby putting them in the danger of failure.385

379 Laeven and Valencia (2008, p. 32). 380 See Diaz-Alejandro (1985) and Arellano (1983). 381 See Maxfield (1992). 382 Comercio Exterior, Vol. 32, N° 11, p. 1186. 383 “Takeover pleases U.S. Banks,” The New York Times, September 2, 1982. See also Secretaria de Hacienda y Crédito Público (1988, p. 82). 384 Ibid. 385 As for the role of international pressures in motivating these measures, Diaz-Alejandro (1984) mentions that in the case of Chile and Colombia Banks Advisory Committees urged the government to take up the debt of private firms and banks.

135

The discussion above makes clear the need for further work to revisit the Mexican bank nationalization. The existence of underlying financial problems in the banking sector at the time of the nationalization, although suggestive, clearly does not represent sufficient evidence to constitute the basis of a new interpretation of the political economy of the 1982 bailout of the Mexican banking sector by the country's government. It should be noted, moreover, that economic and political factors are unlikely mutually exclusive explanations of the nationalization but more plausibly complementary ones. What economic groups or sectors would have been the most severely affected by the failure of major commercial banks? Whose interests were protected and whose interests were sacrificed with the nationalization? These are important questions that remain unanswered and need to be addressed to furnish a convincing explanation of the nationalization of the Mexican banking system.

136

References