Interpreting the Asian Financial Crisis: Open Issues in Theory and Policy

This paper discusses the theoretical and policy debate on the Crisis in South East Asia. After contrasting competing interpretations of the crisis, the paper presents stylized facts about fundamental imbalances in the economies of the region before the currency collapse. Then, it reviews current theories of currency and financial instability, assessing them as interpretive schemes of the crisis. In light of facts and theories, a final section discusses emerging policy issues, including both structural and cyclical aspects of the recovery. he recent financial crises starting from one country (e.g. Mexico in 1994, Thailand in 1997) and spreading as a contagion to others (i.e., the “tequila” effect of the Mexican crisis on Argentina and other Latin American countries; the spread of the “Asian flu” to Hong Kong, China; Indonesia; Korea; Malaysia; Singapore; and Taipei, China from the Thai crisis; and the rapid spread of financial instability after the Russian crisis in the summer of 1998) have stimulated a vast debate on the analytics and empirics of currency crises and their policy implications, particularly for crisis prevention and management. This paper discusses the theoretical debate on the Asian crisis and its key stylized facts. After reviewing the two competing theories of the crisis in the first section, the second section analyzes the T 2 Asian Development Review fundamental imbalances in the crisis countries before and after the summer of 1997. The third section delves into an assessment of existing theories of currency and financial crises as interpretive schemes for the Asian case. In light of these theories, the fourth section discusses emerging policy issues, covering both structural and cyclical aspects of the recovery policy strategies. Alternative Perspectives on Currency and Financial Instability The Asian crisis has revived the long-standing debate between two competing interpretations of currency and financial collapses: one viewing crises as essentially caused by fundamental weaknesses and policy inconsistencies, the other as due to selffulfilling expectations and financial “panics”. Since the two interpretations of the crisis also lead to very different policy prescriptions, it is appropriate to start our analysis with a review of their empirical and theoretical foundations. The “panic” view, which may encompass such diverse phenomena as “bank runs”, “fickle investors”, and “hot money”, is one explanation of the fact that the absence of traditional fundamental weaknesses associated with currency and financial instability does not preclude a crisis from occurring. For example, as shown in Table 1, the precrisis budget deficits in Southeast Asia were low or nonexistent; public debt contained; inflation rates were in single digits; and investment, saving, and growth rates quite high by international standards (see Chang and Velasco 1998, Sachs and Radelet 1998). The “fundamental view” challenges this conclusion by claiming that in fact most crises, including the Asian crisis, are caused by fundamental weaknesses. In Asia, the strength of the macroeconomic outlook in the crisis countries was only apparent, and the financial and currency collapses could be attributed to serious institutional weaknesses, policy inconsistencies, and structural problems that were not necessarily recorded by past trends in macro variables. This interpretation calls attention to the size of the current account deficits, as well as to a number of indicators revealing the weak conditions of the financial systems in the region. These weak conditions translated into the possibility of a bailout, and thus implied a sizeable contingent fiscal imbalance (see Corsetti, Pesenti, and Roubini 1999a, b, c; Lane et al. 1999). 1 An updated account of the theoretical debate below, and a large number of contributions on the crisis are available on Roubini’s Asian Crisis Homepage at http://www.stern.nyu.edu/~nroubini/asia/AsiaHomepage.html. Interpreting the Asian Financial Crisis 3 These two views also underlie competing theories of how a crisis spread from one country to others. Fundamental and structural theory of the international transmission of a crisis stress common shocks as well as trade and financial linkages among countries, with special reference to policy spillovers. Contagion is instead referred to the spread of financial and currency instability that cannot be explained in terms of the aforementioned factors. The debate between the “fundamentals” and the “panic” view of crises is by no means specific to the Asian case. In the five years preceding the crisis in Southeast Asia, the two views have confronted each other in at least two other episodes of currency instability. The first was the demise of the target zone system based on narrow bands in the European Monetary System (EMS) between 1992 and 1993; the second was the collapse of the Mexican peso at the end of 1994. In the case of the European currency crises in 1992, the most widely used model of self-fulfilling expectations focuses on the credibility of monetary policy in the presence of an inflation-unemployment tradeoff. A policy of fixed exchange rate contributes to price stability, but rules out stabilization via monetary policies. In each period, monetary authorities decide their optimal course of action by assessing the costs and benefits of defending the exchange rate, against the cost and benefits of devaluing the currency. Key to the theory is that the short-run costs and benefits of defending the exchange rate crucially depend on prices, such as the interest rate and the wage rate, which private agents set based on their expectations of future exchange rate depreciation. This circularity is what raises the possibility of self-fulfilling expectations. If private agents expect exchange rate stability, they rationally keep nominal interest rate and wages in line with low inflation—this improves the macroeconomic outlook of the country and strengthens the ability of the government to defend the exchange rate against adverse shocks. In turn, the government does not change the exchange rate, thereby fulfilling the expectations of private agents. If, instead, private agents raise interest rates and wages expecting a devaluation, the resulting loss of competitiveness undermines the ability of the government to defend the exchange rate. The government devalues, thereby fulfilling the expectations of private agents. In other words, there is more than one equilibrium that can be rationally anticipated by economic agents—and in the presence of equilibrium multiplicity, financial and currency instability can be driven by sudden changes in agents’ coordination from one equilibrium to another, for given fundamentals of the economy. An important property of these models is that the possibility of self-fulfilling expectations arises only if the fundamentals of the economy are “weak enough” to start with. In other words, private agents will rule out the possibility of devaluation if either macroeconomic conditions or the government resolve to fight inflation are sufficiently strong. Weak fundamentals are then to be seen as a necessary condition for selffulfilling speculative attacks to occur. 4 Asian Development Review In the case of the Mexican and the Asian crises, a leading interpretive model draws an analogy between currency and financial crises and “bank runs”. The key argument in this case is that the investors’ refusal to rollover debt to a country may lead to a disruptive, early liquidation of investment projects, which implies real costs. These costs in turn compromise the country’s external solvency even if, sans the run, all projects would be economically viable. Whether “fundamentals” such as inflation, growth, unemployment, the current account, and the budget deficit are in good shape does not really matter. The “fundamental variables” that are relevant in this theory are the maturity structure and the currency denomination of the external and domestic debt. Mismatches in debt maturity and currency denomination are what expose the country’s vulnerability to a run. Of course, it may well be that imbalances and weaknesses of the economic fundamentals of a country affect the scale of the aforementioned mismatches. If the Asian events were ultimately caused by a panic, rather than by macroeconomic and structural imbalances (such as persistent current account deficits and widespread fragility of financial firms), these imbalances could have nonetheless played an essential role in the crisis. Whether espousing the financial panic thesis or the fundamental thesis, economic and policy analyses of the crisis in Southeast Asia have progressively focused on what caused the Asian countries to become fragile and vulnerable to shocks in the mid1990s. A common view attributes the crisis to premature financial liberalization, that is liberalization of financial markets prior to the institution of mechanisms of adequate supervision and prudential regulation of financial intermediaries. Such a view identifies a unifying pattern in the crisis region. Thus, it tends to blur structural and policy differences across countries, and implicitly considers the crisis since 1997 as a single episode. Hopefully, future research will shed light on country-specific features that will help fill the many gaps in the currently available explanations of the Asian events. Yet, at the time of this writing, all the crisis economies in the region are showing clear signs of recovery: output is growing again—in some cases at a very fast pace— the stock of foreign reserves is replenished, exchange rates are relatively stable, interest rates are at or below the precrisis level, and equity markets have recorded important gains. Even in Indonesia, the country with the biggest internal political problems, output seems to have bottomed out at the end of 1998. While different fundamentals may explain differences in intensity of these phenomena, the recent events in Southeast Asia confirms a key lesson from the many episodes of financial instability in the 1990s: the waves of crisis and recovery tend to be correlated across countries with different fundamentals (see Buiter et al. 1998a, ch. 4). In what follows, we review key macroeconomic and structural indicators of countries in the crisis region, assessing the evidence in light of the theoretical debate on the roots of financial and currency instability. Interpreting the Asian Financial Crisis 5 Fundamental Imbalances at the Root of the Crisis At the onset of the crisis in 1997, the South East Asian countries had two decades of rapid and sustained growth with high saving and investment rates, high rates of human capital accumulation, and a pronounced work ethic. Associated with the “Asian miracle” was a disciplined macroeconomic policy management, keeping both budget deficit and inflation in control. To a large extent, these performances of the East Asian economy were the result of growth-oriented policy strategies based on various forms of centralized coordination of production activities and resource allocation, protection of domestic industries, and implicit or explicit government guarantees on private investment projects. Such a policy strategy allowed firms to undertake highly risky projects by relying heavily on bank credit. For instance, the average debt-to-equity ratios estimated by Claessens et al. (1998) for 1996 were as high as 355 percent in Korea and 236 percent in Thailand. Lower but still high ratios were to be found in Indonesia (188 percent), Hong Kong, China (155 percent), and Philippines (128 percent). The corresponding figure for Japan is 221 percent. Even less balanced is the picture provided by the World Bank. For instance, the estimated debt-to-equity ratio in Korea is 620 percent (World Bank 1998, 55); see the discussion in Appendix I. The prevailing model of “relationship banking” provided the financial resources required by the high-growth objectives—but somehow downplayed the role of prices in project selection, and gave incentives to develop strong informal ties between creditors and debtors. The Asian miracle has occurred despite—or, as some would say, because of— significant distortions of the market mechanism in the financial sector. In the presence of extensive controls and limits to foreign borrowing, however, these distortions had not translated into high domestic vulnerability to external shocks. This key feature changed with the process of domestic and international liberalization in the 1990s. As international capital markets became progressively more accessible and domestic markets were deregulated, supervision of the financial system was inadequate—the bestknown example being provided by the strong unregulated growth of financial companies in Thailand. At the same time, the prevailing policy model kept alive a strong corporate bias in favor of debt financing. The World Bank (1998) points out that in some countries, leverage rose sharply prior 1997 in the crisis countries (see Appendix 1). Between 1991 and 1996, it doubled in Thailand and Malaysia, and grew by one third in Korea (World Bank 1998, 55). First, deregulation and liberalization did very little or nothing to break the strong ties between banking, 2 The literature has pointed out a long list of structural distortions in the precrisis Asian financial and banking sectors. These include lax supervision and weak regulation, low capital adequacy ratios, lack of incentive-compatible deposit insurance schemes, insufficient expertise in the regulatory institutions, distorted incentives for project selection and monitoring, and outright corrupt lending practices. See the analysis in Corsetti, Pesenti, and Roubini (1999a). 6 Asian Development Review nonfinancial firms, and the government, reinforcing expectations of government interventions and bailouts in the event of a crisis. Second, the risk premium on borrowing in dollars was kept low by exchange rate policies de facto pursuing stability vis-à-vis the US currency. On the creditors’ side, international investors were keen on supplying funds to the region, with apparent neglect of proper risk assessment. Two factors are often mentioned as underlying causes. The first is the role of low interest rates in the G7 area in motivating an international portfolio reshuffle in favor of high-yield assets in emerging economies. As the interest differentials may have not been large enough to compensate for risk, the second factor explaining the large capital inflow into Asia is excessive optimism about the future prospects of countries in the region. The two “views of the crisis” discussed in the previous section offer quite a different interpretation. The “panic view” stresses that there is no irrationality or myopia in the optimistic beliefs prevailing before the crisis, which simply reflected rational expectations consistent with a “good equilibrium”. The “fundamental” view explains investors’ optimism in terms of their beliefs about the willingness of domestic and international institutions to guarantee international lending. Although specific characteristics vary across countries, a similar pattern of increasing vulnerability to external shocks characterizes all economies of the region prior to the crisis. First, short-term borrowing to finance long-term projects became increasingly important, especially in Korea, Malaysia, and Thailand. Second, while borrowing short-term in foreign currency (dollar and yen), domestic banks loaned to domestic firms in both local currency and foreign currency, but neither domestic banks nor domestic firms hedged their positions. This created a sizeable currencydenomination mismatch in the balance sheet of domestic intermediaries as well as of corporate firms. Third, the availability of credit fueled investment in increasingly risky assets. In some countries, the credit boom was mirrored by real estate and property booms. In other countries, financial resources were directed toward investment in narrowly specialized industries subject to large terms of trade fluctuations. Poor and risky investment in turn deteriorated the quality of the portfolio owned by domestic financial intermediaries, enhancing the likelihood of panics and crises. In the mid-1990s, several factors contributed to deteriorate the economic outlook of the region, including Japan’s prolonged recession, adverse terms of trade fluctuations, the rapid appreciation of the dollar since 1995 as well as the increasing presence of People’s Republic of China (PRC) in the markets for exports from the region (see Corsetti, Pesenti, and Roubini 1999a). As a result, export growth slowed down considerably in 1996, and many Asian countries experienced financial difficulties well before the outburst of the crisis. These difficulties were reflected by falling prices in most 3 It should be observed that, in countries pursuing a unilateral peg against a basket of currencies, the weight of the dollar was formally about 30 percent. What made the effective weight of the dollar much higher was the fact that all the regional trade partners of these countries were also keeping their currencies in line with the dollar. Interpreting the Asian Financial Crisis 7 stock markets of the region and, in some cases, they developed into a widespread bankruptcy crisis (see Appendix II and III). In 1996, for instance, 20 of the largest 30 Korean conglomerates displayed a rate of return below the cost of invested capital, and in the first months of 1997, seven of the 30 largest conglomerates could be considered effectively bankrupt. At macroeconomic levels, declining exports and sustained investment rates implied that current account deficits either widened (as was the case for Korea) or persisted at high levels. While financial difficulties started to emerge in the region, domestic intermediation kept growing. According to anecdotal evidence, some governments took on an active role in reassuring international investors about their willingness to back domestic financial firms. Capital inflows did not slow down, but increasingly took the form of short-term interbank loans—that could be readily withdrawn and could count on formal guarantees in the interbank markets. Tables 2a-2c report various measures of the size of short-term debt in relation to international reserves. When the Thai crisis erupted in June 1997, most countries in the region were not in a condition to pursue effective and convincing policies to contain a financial and currency crisis. Financial sector weaknesses constrained the use of interest rates to discourage speculation, but, at the same time, a large external debt denominated in foreign currency implied extremely high costs of currency devaluation. As new information about the state of fundamentals worsened the investors’ perception of the region, the rapid reversal of capital flows caused sharp depreciation and asset deflation. The dimension of the capital reversal was unprecedented. Indonesia, Korea, Malaysia, Philippines, and Thailand received a net inflow of private capital as high as $60 billion in 1995 and $63 billion in 1996, but suffered a net outflow of private capital as high as $22 billion in 1997. Between 1996 and 1997, the net difference in these flows is a negative $85 billion, close to 10 percent of their combined GDP (see IMF 1999b). Most of the capital flow reversal of course happened in the second half of the year, when international banks stopped lending and started to call in their loans. In particular, commercial banks alone, which were lending $58 billion in 1996, withdrew US$ 29 billion in 1997. Private capital outflows continued in 1998 and 1999, at the rate of $29 and $18 billion, respectively. Net private outflows are expected to fall to $8 billion in 2000. As the inflow of private capital has virtually ground to a halt during the crisis, the role of official creditors and international reserves has correspondingly increased. Net official 4 An important case study is the collapse of the large Thai Bank firm, Finance One. According to the press, in the months preceding the crisis, Bank of Thailand would repeatedly confirm to foreign investors its willingness to “back Finance One all the way”. 5 The response to the crisis presented governments with a policy dilemma that has not been sufficiently understood. Fighting exchange rate depreciation required letting interest rates grow as high as necessary to discourage speculation. High interest rates, however, could exacerbate the financial imbalances in the country, leading to domestic bankruptcies and increasing the risk of a systemic collapse. 8 Asian Development Review flows from the IMF and other official creditors have increased from a negative $4.6 billion in 1996 to a positive $ 31 billion in 1997, and $20.2 billion in 1998. Official flows are estimated to be again negative for $4 billion in 1999. By the same token, the stock of international reserves, which had increased by $14 billion in 1995 and $ 19 billion in 1996, dropped by $31 billion in 1997. Corresponding to these large outflows of funds between 1997 and 1998 are equally stunning corrections of the current account of the order of 15 percent of GDP for Korea and Thailand, 6 percent of GDP for Indonesia, and 4 percent of GDP for Malaysia and the Philippines. While the dramatic drop in relative prices driven by exchange rate depreciation has played some role in the correction, the bulk of it is due to a dramatic drop in domestic demand, dragging down domestic output. Novel Elements and Déjà vu in the Asian Crisis An analysis of the evidence on the fundamental imbalances at the onset of the Asian crisis suggests the presence of many elements specific to capital market crisis due to the lack of regulation and supervision during a process of financial liberalization. East Asian countries built an increasing stock of short-term debt denominated in foreign currency. International lenders were keen on supplying funds at low costs, especially but not exclusively short-term interbank loans. Underregulated domestic financial intermediaries channeled funds from abroad toward the acquisition of highly risky assets and/or toward the financing of low-profitability and dubious investment projects. The problem induced by weak regulation that may lie at the root of capital market crises is by no means an exclusive feature of the Asian crisis. Writing about the financial crisis experienced by Chile during the process of deregulation and liberalization in the early 1980s, Díaz-Alejandro (1985, 374, 379) observes: Whether or not deposits are explicitly insured, the public expects governments to intervene to save most depositors from losses when financial intermediaries run into trouble. Warnings that intervention will not be forthcoming appear to be simply not believable. [Agents will therefore expect a bailout regardless of] laissez faire commitments which a misguided minister of finance or central bank president may occasionally utter in a moment of dogmatic exaltation. This quote stresses that expectations of a future bailout need not be based on an explicit promise by the government, or on existing institutions (such as deposit insurance). Investors understand that the balance sheet of financial firms has a public good nature, namely, it is the foundation of financial and monetary stability. They therefore rationally expect that, in the event of a crisis, government will not let its domestic fiInterpreting the Asian Financial Crisis 9 nancial system go bankrupt. It should be stressed that the same argument applies equally well to nonfinancial firms whose collapse has significant externalities, economic or political, externalities that are not necessarily restricted to its effect on the financial health of creditor banks and financial institutions. In the light of these considerations, the claim that the Asian crisis is different because the East Asian countries did not display the “usual symptoms” preceding a crisis (high inflation, high budget deficits, high unemployment) needs to be qualified. Imbalances and difficulties in the financial sector are a contingent public liability, which is real even if it is not reflected in official data on budget deficits until a crisis occurs. For instance, regarding the fiscal dimension of the Chilean crisis, Díaz-Alejandro (1985, 372) writes: The massive use of central bank credit to ‘bail out’ private agents raises doubts about the validity of pre-1982 analyses of the fiscal position and debt of the Chilean public sector. The recorded public-sector budget deficit was nonexistent or minuscule for several years through 1981, and moderate during 1982. The declining importance of ostensible public debt in the national balance sheet was celebrated by some observers; ex-post it turned out that the public sector, including the central bank, had been accumulating an explosive amount of contingent liabilities to both foreign and domestic agents who held deposits in, or made loans to, the rickety domestic financial sector. This hidden public debt could be turned into cash as the financial system threatened to collapse. At the onset of the crisis, the exact magnitude of the public contingent liability may be subject to considerable uncertainty. The required fiscal adjustment to guarantee public sector solvency raises both distribution issues (who is going to bear the costs of the adjustment?) and macroeconomic policy issues (how can a collapse of domestic demand be avoided?). It thus induces policy and political uncertainty. Note that when the size of the public contingent liability is high enough relative to the ability of 6 Sachs and Radelet (1998) and Krugman (1999a), among others, have expressed skepticism on the importance of moral hazard in the Asian crisis, pointing out that the “overborrowing and overinvestment” syndrome involved sectors and institutions that could not count on direct government guarantees. For instance, Krugman (1999a) claims that, in the presence of moral hazard, sectors benefiting from government guarantees (banking and large firms) should crowd out sectors without guarantees (property?), but this is not the case in Asia. More crucially, Sachs and Radelet point out that many firms borrowed directly from foreign intermediaries. If moral hazard is associated with the implicit guarantees to the banking sector, it cannot explain overborrowing at the firm level. Independently of one’s belief about the role of moral hazard in the crisis, the above objections are not convincing. First, if one admits that financial intermediaries are the only firms benefiting from public guarantees, competition and excessive risk taking in lending implies that doubtful projects are financed in all sectors. Moreover, if banks are guaranteed, they may be willing to offer financial assistance to firms in case of need. The possibility of counting on emergency lending at home provides an incentive to nonfinancial firms to increase the scale and scope of projects, and reduce the incentive to cover open positions in foreign currency. Second, it is not clear why the implicit bail-out guarantees cannot extend to nonfinancial firms. 10 Asian Development Review a country to raise taxes/cut spending, a financial crisis cannot but generate expectations of future inflation, via the need for seignorage revenue. As the financial crisis exposes fiscal imbalances and induces expectations of inflation, the economy suffers because of macroeconomic distress. It is sometimes argued that, if the Asian crisis was caused by a fundamental imbalance, the adjustment in the level of exchange rates and relative price should have paved the way to a swift and rapid recovery: no marked fall in economic activity should have been observed. In light of a fundamental interpretation of the crisis centered on financial distortion, this view is not correct. Before the crisis, implicit guarantees on investment projects led the private sector to undertake projects that are not profitable. In the tradables sector, the scale and type of technology adopted is not optimal. In the nontraded sector, the profitability of investment suffers from changes in the real exchange rate accompanying the devaluation, changes that do not depend on the presence of nominal rigidities. Whether or not a financial panic occurs, the adjustment to the existing fundamental imbalance may take more than a correction in relative prices, as the economy faces the cumulative bill from distorted investment decisions in the past. In the “fundamental interpretation” of the crisis, weak financial supervision and regulation plays a crucial role in explaining why Asian banks and financial intermediaries followed strategies that exposed them to a high risk of a liquidity crisis, and international investors supplied funds without proper risk assessment. Some observers have questioned an interpretation of the crisis focused on “moral hazard” in lending, pointing to an alternative theory. Key to this theory is the existence of imperfections in capital markets such that investment by firms is constrained by the net worth of their owners, that is, by the amount of collateral that is available to them. Suppose that the net worth of firms’ owners is affected by movements in the exchange rate. This would be the case if these agents have borrowed in foreign currency, and home currency prices are not very flexible. Then, the direct effect on the economy of a negative shock leading to depreciation is compounded by the indirect effect of the exchange rate movement on the ability of firms to finance their investment. Multiple equilibria are 7 In this interpretation, speculation in the foreign exchange market and a financial collapse are strictly interwoven. This is because, by depleting international reserves and forcing a devaluation, speculation translates into a sudden increase the stock of net public liabilities that worsens international investors’ confidence in the ability of the government to back its private sector. On the other hand, a run on reserves takes place when fundamentals are weak enough (that is, when the stock of private debt to be backed by the government in case of crisis is sufficiently high) to induce expectations of monetary expansion and depreciation. A crisis thus takes the form of a “twin-run” on the currency and on the liabilities of the financial and corporate sector (the international creditors withdraw their loans triggering a financial crisis). Note that the mechanism highlighted by this scheme hinges upon a link between market confidence and a low stock of reserves relative to the explicit and implicit liability of the public sector. See Corsetti, Pesenti and Roubini (1999b) for a formal model. Interpreting the Asian Financial Crisis 11 possible in such a model. Even mild shocks can have a quite dramatic economic impact. A piece of evidence that is consistent with an interpretation of the crisis centered on fluctuations in the value of collateral available to domestic borrowers is the fact that, in some of the Asian countries, the real estate sector had been booming before the crisis (see Appendix 2). A cycle of boom and bust in real estate may have played a role in generating financial stability. To the extent that one is willing to see such a cycle as the emergence and burst of a market bubble, the explanation of the crisis needs not include weak financial market regulation. Note that, in practice, it is very hard to disentangle the two views. Was excessive borrowing driven exclusively by overvaluation of the real estate used as collateral? To what extent did the property market bubble mirror a lack of prudential supervision? In addition, real estate and land price did not seem to play a crucial role in the crisis in a number of countries, including Korea. Ongoing Challenges: Bank Restructuring and Corporate Reform The crisis has deeply affected corporate and banking sectors. This is true for the countries that have suffered a collapse of their currencies, as well as for countries that have successfully defended their exchange rate either by sheer determination, such as Hong Kong, China, or by shielding their markets domestic through capital controls. The data on the banking and corporate crisis are striking. As regards the financial system, data on nonperforming loans (NPL) as of the end of March 1999 provides a rather gloomy picture. For Indonesia and Thailand, the estimated ratio of NPL to total loans reaches 55 percent and 52 percent, respectively. This ratio is 25 percent for Malaysia and 16 percent for Korea. Correspondingly, the net loss due to the declining quality of assets in the balanced sheet of banks is estimated to be at least 20 percent of GDP, with a peak of 52 percent for Thailand (see IMF 1999b). By the same token, the result of simulation exercises by the World Bank on the effect of the devaluation, interest, and credit shocks accompanying the crisis show that, on average, corporate firms lost about half of their equity value, with losses exceeding equity value for one firm out of three. At the same time, leverage increased by 50 percent and external debt by 30 percent (World Bank 1998). By early 1998, as a result of high interest rates and an increasing amount of NPLs, a large share of the banking systems in Indonesia, Korea, and Thailand were de facto bankrupt. In these countries, many banks stopped making new loans, going as far as denying trade credit and working capital, causing problems to many corporations that would have been solvent under normal credit conditions. In support of this view, 8 See for instance the analysis in Furnam and Stiglitz (1998); Krugman (1999a); and Aghion, Bacchetta, and Banerjee (1999). 9 For Japan, available estimates of troubled loans for the entire private banking system in 1998 reached 73.1 tril-