Policy, Research, and Exbtrnal Affairs | t(5 aS 2 WORKING PAPERS Debt and Internatlonal Finanoe International Economics Department The World Bank February 1991 WPS 592 Official Credits to Developing Countries Implicit Transfers to the Banks Asli Demirguq-Kunt and Harry Huizinga The stock maiket expects virtually all additional resources provided to debtor countries to be used for debt service to commercial banks. The stock market capitalization of banks increased about $6 billion at the time of the 1983 U.S. proposal to increase its quota to the IMF by $8.5 billion, and by a low estimate of $22.4 billion at the time details of the Brady Plan were recorded. The Policy, Rearch. and Extena Affaiz Complex distnbute PRE Working Papen todisseminatethe frndinpgf wo k inprogea and to rncounp the exdiange of idea anong Bank staff and ae other, intrested in devloprnent issue. These papa carry the namer of the authos rflect only their views, and should be used and cited accordingly. The findings, interpmuation, and conclusions are the author' own. They ahould not be arnibuted to the World Bank, its Board of Dircts, iu management, or any of iu manber countries. Policy, Resrch, and Extnal Affairs Debt and Intwonathnl Finance WPS 592 l This paper - a product of the Debt and Intemational Finan',e Division, International Economics Department - is part of a larger effort in PRE to understand commercial bank lending behavior. Copies are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Sheilah King-Watson, room S8-025, extension 33730 (29 pages). Two types of event have affected returns of estimate of $22.4 billion at the time details of the banks that are heavily exposed to third world Brady Plan were recorded. debt in the 1980s: actions by the debtor coun- tries (such as declarations of moratorium) and The estimate of the magnitude of these official actions (such as changes in regulations effects is infornative, but the emphasis should and in the provision of official monies to the be on the direction of these effects, as they are debtor countries). robust to overestimation problems. The effect of the first type of event has been Clearly official resources provided to debtor extensively investigated. There are fewer studies countries do devolve to creditor banks. But the analyzing the effect of official actions on bank debtor countries should at least gain insofar as stock retums. DemirgUc,-Kunt and Huizinga the reduction of a debt overhang eliminates investigate to what extent official money tvail- investment distortions. able to debtor countries has devolved to the banks, as reflected in stock market prices. The results here stem from the fact that some of the monies provided by the multilaterals are They find that the stock market expects specifically earmarked for debt service or are in virtually all additional resources provided to the form of general balance-of-payments support debtor countries to be used for debt service to that the developing countries can use for private commercial banks. The stock market capitaliza- debt service. Official creditor resources that arc tion of banks increased about $6 billion at the specifically provided to finance developrnient time of the 1983 U.S. proposal to increase its projects are less likely to be allocated to bank quota to the IMF by $8.5 billion, and by a low debt service. The PRE Working Paper Series disseminates the findings of work under way in the Bank's Policy, Research, and External AffairsComplex. An objective oftheseries is to getthese findings out quickly, even if presentations are less than fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent official Bank policy. Produced by the PRE Dissemination Center Official Credits to Developing Countries: Implicit Transfers to the Banks* by Asli Demirguc-Kunt and Harry Huizinga Table of Contents I. Introduction 2 II. Methodology 5 III. Events and Findings 9 a. IMF loans in 1982 ard 1983 9 b. Increase in U.S. quota to IMF in 1983 11 c. The Brady Plan 13 d. Recent World Bank and IMF quota increases 15 IV. Conclusion 17 References 19 Tables 21 2 I. Introduction Two types of events have affected returns of banks that are heavily exposed to third world debt during the 1980s: actions by the debtor countries, such as declaration of moratoriums, and official actions such as changes in regulations and in the provision of official monies to the debtor countries. The effect of first type of events has been extensively investigated. Among these studies Schroder and Vankudre (1986), Cornell and Shapiro (1986), Bruner and Simms (1987), and Smirlock and Kaufold (1987) study the effect of Mexico's 1982 default; Sachs and Huizinga (1987) and Musumeci and Sinkey (1987) study the effect of Brazil's 1987 debt moratorium, and Ozler (1990) investigates the effect of 1978-1983 international loan reschedulings on bank stock values. There are fewer studies analyzing the effect of official actions on bank stock returns. Change in regulations, for instance, is analyzed by Eyssell, Fraser and Rangan (1989) who investigate the effect of amendments in regulations governing international banking operations. The effect of official monies, more specifically the effect of indirect provisions made available as increases in resources of international financial institutions, has been studied by Cornell, Landsman and Shapiro (1988) and Billingsley and Lamy (1988). They show that the 1983 increase in the U.S. quota to the IMF by $8.5 billion materially affected bank stock returns. However, official monies provided directly as loans to debtor nations are also important. For instance, earlier in 1982 and in 1983 the IMF provided a series of large balance of payments loans to Argentina, Brazil, Chile, and Mexico that 3 similarly can be expected to have affected bank stock returns. More recently, as part of the Brady Plan, the IMF and World Bank have made around $24 billion available for developing country debt reduction. This paper investigates to what extent these official monies made available to debtor countries have devolved to the banks, as reflected in stock market prices. For several episodes, we calculate the increase in expected repayment by debtor nations reflected in stock market pri_es. Stock market prices are found to have responded strongly to the announcements of large IMF loans to Latin American debtors in late 1982 and to a lesser extent in early 1983. In the period from 1984 to 1988, we find no clear effects on bank stock prices of announcements of large commitments by the IMF, World Bank or national governments. Apparently, the stock market went through a learning process early in the debt crisis after which a pattern of large official balance of payments loans from the multilateral institutions that were partly used for private debt service was clearly established. Of the $8.5 billion U.S. quota increase of the IMF, we estimate that about $6 billion indirectly accrued to private banks worldwide. The recent World Bank quota increase of $74.8 billion however, did not clearly affect bank stock return at its passage as the increase had been fully anticipated. The IMF quota increase of around $60 billion announced in May 1990 negatively affected bank stock returns. This is due to the fact that the market expected a greater increase whereas the United States was able to prevent the quota subscriptions from increasing by more than 50 percent. In the case of the Brady Plan, the paper abstracts from the details of the menu by which debt reduction actually takes place that can be important to the banks as shown by Demirguc-Kunt and Diwan (1990). While the initial reaction to the debt reduction plan wps unclear, during the period of March 16 to March 20 when the extensive IMF and World Bank involvement in debt reduction was secured, bank returns showed a significantly positive reaction. Interestingly, heavily exposed banks seem to have benefited less per dollar of LDC debt than the lowly exposed banks, although for both types of banks repayment prospects should have been affected equally. Heavily exposed banks may have benefited less, as their contingent claim on the FDIC was reduced while repayment prospects improved. Huizinga and Ozler (1990) have shown that the relationship between LDC exposure and bank valuation is nonlinear due to., federal deposit insurance. Important news concerning the repayment prospects of LDC debt of course affects the value of the banks' contingent claim on the Federal Deposit Insurance Corporation (FDIC). Judging from the lowly exposed banks, the $24 billion made available for debt reduction appears to have increased the present value of debt payment by $22.4 billion. This result confirms Bulow and Rogoff (1988) who have shown that the Bolivian debt buyback of (1988) mainly benefited the banks. Comparing the experience of the lowly and highly exposed banks, we estimate that the U.S. banks contingent claim on the FDIC has been reduced by approximately $9 billion. The remainder of this paper is as follows. Section 2 describes the empirical methodology and the data. Section 3 discusses the events and presents the main results. Section 4 concludes. 5 I1. Methodol,&X The main aim is co infer from stock prices the transfer to the commercial banl:s implicit in the provision of official monies to the debtor nations. To start, let us consider che following bank valuation equation: (1) MVi - LDCi + NLDCi + NBi - LI1 where MVi is the bank value for bank i, LDCi is the present value of the expected LDC debt repayment, NLDCi is the market value of the bank's non-LDC assets, LI1 is bank iiabilities, and NB1 is the value of the bank's off balance sheet items, and in particulir its contingent claim on the FDIC. MV1 is measured as the stock price times the number of shares outstanding. Now let the official creditor make availaile resources Li to country j. Lj can be a direct loan to country j or an indirect transfer of resources to a multilateral agency to be channeled to country j at a later point. For bank i, this affects expected repayment LDCij and claim NB1. From (1) we can dlerive: A Pi E. dL. (2) _+ I J P. 1 P. MV. E . . .1. 13f dSLDC dNB. E E.. where 6 1 ij and e _ 1 Pj p is the stock price, and E dL; dL E j denote expsu ij denotes exposure of bank i to countr-t' j. 6 The coefficient 6 in (2) measures the proportion of the loans that is expected to be transfered to the canks. All banks are expected to be repaid in proportion to their exposure, i.e., __Cij__ E i The coefficient e S LDCJ S E.. 1. ij ~I. 2.J measures the indirect iripact of official transfers to debtor nations on banks' claims on the FIDIC. For a lowly exposed bank. ei is close to zero while for highly exposed banks ei may be substantially negative. As for each bank S + ei will be estimated jointly, values of ei can be inferred by comparing 6 + ei for highly and lowly exposed banks. Our sample of banks consists of roughly 21 exposed and 9 non-exposed U.S. banks, depending on the particular event. A list of banks is given in Table 1. As shown, exposed banks are grouped as highly and lowly exposed banks based on their exposures. Data on daily bank and market returns, for the period January 1, 1983 to December 31, 1988, are obtain-' from the tapes constructed by Center for Research in Security Prices (CRSP) at the University of Chicago. The market return is dividend inclusive return on the S&P 500 index. Individual bank exposure data on individual countries is obtained from The Country Exposure Lending Surveys. Following Smirlock and Kaufold (1987) and Eyssell, Fraser and Rangan (1989), the following set of n linear equations is estimated first. 7 (3) Rit- 1 + iRMt + YdlDt + elt' R2t a2 2RMt + 'd2 Dt 2t' Rnt a nn nmt +YdnD t nt where Rit is the return on the stock of bank i on day t. Rmt is the market return, Dt is a duimmy equal to 1 during the event period of three days inc'.uding the day before and after the event, and zero otherwise.) The system is estimated for the complete year, with daily returns for each year in which events took place. For the set of exposed banks only, the following alternative system is estimated: (3) Rit 1i 1 Rmt 7 Yel Dt l et' R2t- a2 +2Rmt + Ye2 D + e2t' Rnt- an + nRmt + 7enDt + ent where Ei is Eij given L:. Now -6 + e mv i f Eij ~~enn The systems are estimated using seemingly unrelated regressions (SUR) 1 Using the same dummy variable for multiple dates is to capture leakages and lags of information, and is common in event studies. See for instance, Eyssell, Fraser, and Rangan (1989), and Grammatikor and Saunders (1990). 8 technique, which allows for contemporaneously correlated disturbances.2 Th4s technique is most appropriate for estimation of a system of equations which have nonzero correlation across their residual terms due to implicit relationships. In the above systems implicit relationships exist since all banks are members of the same industry. The hypotheses to be estimated are: H1: The event parameters are zero for a group of banks. H2: The event parameters are equal to each other for a group of banks. The groups of banks we consider are the set cf exposed banks, the set of non- exposed banks, and all banks together. Hypotheses are tested separately for all three groups of banks. For system (3), we expect the event parameters 7di as a group to be different from zero for the exposed banks and for all banks togA.ther while the parameters should be zero (and equal to each other) for the non-exposed banks. If the event parameters for the non-exposed banks are different from zero, this indiLates investors can not correc:- ..j distinguish between exposed and nonexposed banks, which is a form of contagion. Also, if the event parameters for the exposed banks are equal to each other, this points at contagion as it indicates stock market investors can not distinguish between heavily and lowly 2 See Zellner (1962) for a discussion of the technique. Using SUR in estimation of system (3) is not necessary since SUR estimator collapses to OLS estimator when all the independent variables are the same. However, using this technique leads to efficiency gains in estimation of system (4). 9 exposed banks For system (4), we expect again the event parameters -y* to be significantly different from zero for the group of heavily exposed banks. The event parameters should be equal to each other if ef-O for all banks, i.e. the FDIC insurance does not affect bank valuation. If the hypothesis of equal event parameters is rejected, this could po'.nt to (i) contagion or (ii) a significant relationship between the ei's and the Ei's. A negative correlation between the ei's and Ei's suggests the importance of the FDIC claim in bank valuation. III. Events and Findings a. IMF loans in 1982 and 1983. The announcement dates of the events that are examined are reported in Table 2. The first five dates represent news concerning large IMF loans to Latin debtor countries in late 1982 and early 1983. The first of these, in October 1982, was a $2 billion loan to Argentina. The loan came at a time that Argentina had $1.7 billion in arrears on $40 billion of debt, and just 2 months after Mexico declared it was unable to service its debt in August 1982. Upon hearing the news, a banker said, "This is much the best news we have had in one of the bleakest years I can remember.3 3 WSJ, October 29, 1982. 10 Stubsequently, the IMF reacLed agreements on large loans to Brazil, Mexico, and Chile in December 1982 and the first two months of 1983, The loan to Brazil was tentatively agreed in December 1982, and formally approved in February 1983. These large loans, unlike some smaller loans ,rom the multiLateral lending agencies, are not earmarked to finance specific projects, and thus the funds are generally available for debt service. Regulations tbat required banks to disclose their LDC exposures in the 10K and lOQ reports were not announced till October 1982. Thus during this period bank stock investors had very incomplete information about individual bank exposure, and we cannot estimate system (4). Estimation of system (3) for the loan to Argentina is reported in Table 3. Fourteen of the eighteen exposed banks are shown to have a positive return during the three day event period. The hypothesis that event parameters are zero is rejected for the exposed banks, and for all banks together, but only at 10 percent level for the non-exposed banks. The hypothesis that the event parameters are equal is rejected for all the three groups of banks. The means for the event parameters (7 the heavily and lowly exposed banks are 0.88 and 0.52 percents respectively, indicating that some information about individual bank exposure was known to investors. Bank investor response to the agreement between Brezil and the IMF in December 1982 was much less favorable. During the event period, 16 of the 18 banks experienced negative excess returns. The hypotheses of zero event parameters is rejected at the 5 percent level for all banks, but only at the 11 10 percent level for the exposed banks. Apparently, stock market investors had anticipated a s ghtly more favorable loan. Also, for the announcement of the large IMF loan to Mexico, we find that the hypothesis of zero event parameters can not be rejected. The final loan of this sequence to Chile was approved on January 10, 1983. According to a WSJ article of January 4, 1983, there was considerable doubt whether this loan would be approved. The results of Table 3 show that at the time of announcement 3 exposed banks had significantly positive excess returns. Howeve:, the hypothesis of zerc event parameters is not rejected for all banks, and only at 10 percent level for exposed banks. The results suggest that the stock market, after the initial large IMF commitment to Argentina, anticipated that large commitments to other indebted countries would follow, which explains the absence of strong stock market effects at announcements of later commitments. For later announcements of large IMF and World Bank loans, such as the IMF commitment of $1.8 billion of loans to Argentina in January 1987, and the concommittant commitment by the World Bank of $2 billion to the same country, we similarly find ins.gnificant stock market effects. These results are not reported. b. Increase in U.S. guota to IMF in 1983. In 1983, the U.S. passed legislation to increase the U.S. quota to the IMF by $8.5 billion. If as suggested above, IMF resources to some extent are used to enable debtor nations to service their commercial banks costs, then an increase in U.S. funding to the IMF should positively affect shareholder wealth. Cornell, Landsman and Shapiro (1986) found that at the time of the passage of the bill 12 to increase the U.S. qucea in the Senate on June 8, 1983, bank stocks were negatively affected. Billingsley and Lamy (1988) show, however, that bank stocks were positively affected when the bill was introduced in the Senate on March 7, 1983, and that cumulate excess returns were positively related to the ratio of bank exposure to LDCs to bank assets plus loan-loss reserves. This formulation does not allow one to measure the increase in shareholder wealth. The results of estimating systems (3) and (4) are in Table 4, showing six of eighteen exposed banks experience positive excess returns at least at the 10 percent significance level. Estimates of the market model parameters a; and pi are the same as those reported in Table 3. The non-adjusted event parameters are significantly different from zero for all banks, and the hypothesis that they are the same is rejected. However, the event parameters for the non-exposed banks for themselves are also different from zero, and in fact 3 non-exposed banks have significantly positive individual returns. This points to contagion, where investors can not distinguish between exposed and non-exposed banks. However, this does not point to market inefficiency if stock holders did not yet have information about individual bank exposures. Bank annual reports for the year 1982, which were published around March and April of 1983 were the first to contain obligatory information on individual bank exposure. The exposure-adjusted event parameters are jointly different from zero, and the hypothesis that they are equal to each other can not be rejected. This is strong evidence that stock investors indeed are aware of bank exposures. 13 The means of the exposure adjusted event parameters are equal to 0.142 and 0.398 for the sets of highly and lowly exposed banks. This difference can be attributed to contagion which causes investors to bid up stock of lowly exposed banks too much relative to the stock of highly exposed banks, or it may reflect the role of deposit insurance. Heavily exposed banks have a relatively large claim on the deposit insurance agency. Thus as the repayment prospects of LDC debt improve, heavily exposed banks stand to see their claim on the FDIC go down in value more than lowly exposed banks. Hence, one expects the stock of heavily exposed banks to rise proportionally less, even if markets are fully rational. The estimated values of the exposure-adjusted event parameters in Table 4 can be used to estimate the increase in shareholder wealth during the estimation period. The estimated mean value of the exposure-adjusted event parameter is 0.242. This means that, as there is a three day event window and as the quota was to be increased by $8.5 billion, that stockholders wealth was expected to increase by $6.2 billion. As U.S. banks hold roughly 24 percent of LDC debt, at the time, this means that U.S. bank stock rose by $1.5 billion while foreign bank stockwealth rose by $4.7 billion. Thus, the U.S. quota increase to the IMF appears to have significantly increased the value of non- U.S. banks.' c. The Brady Plan. In March 1989, details of the Brady Plan were 4 The statements about the non-US banks are accurate to the extent they are affected similarly by these events. Since we use only US bank data, extending these results to other banks worldwide is at best an approximation. 14 announced. The IMF and the World Bank were to provide developing countries with funds for debt reduction. Debt Reduction could take the form of debt buybacks or the exchange of debt for exit bonds which were partly guaranteed by the multilateral institutions. As documented by Madura, Tucker and Zarruk (1990), and Unal and Demirguc-Kunt (1990), earlier announcements in December 1989 and January 1990 already suggested that official policy would be reformulated towards debt reduction, but the generous support from the multilaterals, as announced in March 1990, appears to have been largely unexpected. Although initially (March 10) there was no significant market reaction, when support from multilaterals was secured (March 17), bank stocks reacted positively. The results of Table 5 indicate that 11 exposed banks experienced significantly positive excess returns. Interestingly, the hypothesis that the non-adjusted event parameters are equal can be rejected. Indeed, the exposed banks appear to have very similar parameter estimates of around 0.02. Consequently, the hypothesis that exposure adjusted event parameters are equal is rejected. The lowly exposed banks appear to have benefited disproportionately to their exposure. This relationship is confirmed by a correlation coefficient between the estimated parameter and the exposure adjusted event parameter of -0.54 which is significant at the 3 percent level. Again, the different experience of the highly and lowly exposed banks can be attributed to either contagion or to changes in the value of FDIC claims that differ systematically with exposure across banks. If there is no contagion, then the mean event parameter estimate of 2.158 for the lowly exposed banks, and the $24 billion amount used for Lj gives us the estimate 15 that expected bank repayment went up by $155 billion. As contagion is ruled out to arrive at this estimate, it must be an upper limit. The $155 billion roughly corresponds to 20 cents on the dollar of the amount of $622 billion of commercial bank debt to developing countries outstanding at the end of 1989, which corresponds to the initial debt reduction aim of the Brady Plan.5 The mean event parameter of 0.38 for the highly exposed banks yield a lowest estimate of $27.4 billion in increased repayment to commercial banks worldwide. The estimate is a lower limit as it ignores possible changes in the banks' claim on the FDIC. As U.S. commercial banks held around 14.5 percent of commercial bank debt to developing countries as of the first quarter of 1989, the low and high estimates of expected additional repayment to U.S. banks range from $3.97 to $22.4 billion.6 The highly exposed banks own roughly half of U.S. LDC debt. Thus the estimate of the reduction in expected FDIC payments to U.S. banks as a result of the Brady Initiative is $9.2 billion. Again, this is a high estimate, as it rules out contagion. d. Recent World Bank and IMF Ouota Increases. In the last three years, both the World Bank and the IMF had recent announcements of considerable quota subscription increases. On February 19, 1988, the World Bank obtained a $74.8 4 Ouarterly Review, June 1990, Table 5. Of course it is more than 20 percent of the debt of the Brady countries that were identified for debt reduction. 6 The U.S. commercial bank lending to developing countries as a percentage of total commercial bank lending is obtained from Tables 5 and 7A, Ouarterly Review, September 1989. 16 billion general capital increase to be subscribed by member countries before September 30, 1993. On May 7, 1990, the IMF similarly obtained an increase in its resources by 50 percent, from around $120 billion to roughly $180 billion. These increases, unlike the U.S. increase of its IMF quota in 1983, were the result of lengthy reviews within the multilaterals and of negotiations between principal member countries. Thus, bank stock response at the time of the fiscal agreements is only relative to previous market expectations. Table 6 shows the results of estimating (3) and (4) for a 3-day event period surrounding the announcement of the World Bank capital increase. One highly exposed bank experienced a significantly positive excess return, and one lowly exposed bank experiences a significantly negative excess return. The hypotheses that the event parameters for the exposed banks are zero or equal are both rejected. Moreover, no clear pattern is evident in the estimated event parameters. Evidently, the actual acceptance of the World Bank capital increase was not major unexpected news. The IMF quota increase of 50 percent was passed officially on Monday, May 7, 1990. However, the day before the G-7 already released a communique endorsing the 50 percent increase. According to a later WSJ article this accord represented a victory of the U.S. which aimed to limit the increase in IMF capital.7 France and the IMF itself had sought a 100 percent increase. Thus the passage of the accord can be expected to be negative news to the banks. This is confirmed by Table 7, which shows that excess returns on Friday, May 4 were negatively correlated to the ratio of total bank exposure 7 Wednesday, May 29, 1990. 17 to market capitalization. Apparently, news of the limited increase in IMF resources leaked to the market on Friday. Using the estimate of -0.013 as an approximate figure for all banks, and taking the latest figure on commercial bank claims on developing countries for the fourth quarter of 1989, one can compute that bank industry market capitalization was reduced by around $8.2 billion worldwide on account of the limited increase in IMF resources.8 IV. Conclusion This paper has investigated the impact on the wealth of bank share holders of the transfer of official resources to the debtor countries. The main aim has been to derive actual estimates of increases in shareholder wealth following important news concerning future transfers from the multilaterals to the debtor nations. The main result, consistant with Bulow and Rogoff (1988), is that the stock market expects virtually all additional resources provided to debtor countries to be used for debt service to commercial banks. Bank stock market capitalization increased around $6 billion at the time of the 1983 U.S. proposal to increase its quota to the IMF by $8.5 billion, and by a low estimate of $22.4 billion at the time details of the Brady Plan were recorded. While the estimated magnitude of these effects are informative, the emphasis should be on the direction of these effects as they are robust to overestimation problems. Clearly, official resources provided to debtor 8 Ouarterly Review, June 1990, Table 5. 18 countries do devolve to,creditor banks. However, the debtor countries should at least gain in so far as the reduction of a debt overhang eliminates investment distortions. Our results sterA from the fact that some of the 'monies provided by the multilaterals are specifically earmarked for debt service or are in the form of general balance-of-payments support that the developing countries can use for private debt service. Official creditor resources that are specifically provided to finance development projects are less likely to be allocated to bank debt service. 19 References Billingsley, R, and R. Lamy. "The Regulation of International Lending, IMF Support, the Debt Crisis and Bank Stockholder Wealth," Journal of Banking and Finance, 12, 1988, 255-274. Bruner, R. and J. Simms. "The International Debt Crisis and Banking Security Returns in 1982," Journal of Money Credit and Banking, 2, 1987, 46- 55. Bulow, J. and K. Rogoff. "The Buyback Boondoggle," Brookings PaDers onEconomic Activity, 2 , 1988, 675-698. Cornell, B., W. Landsman, and A. Shapiro. "The Impact on Bank Stock Prices of Regulatory Responses to the International Debt Crisis," Studies in Banking and Finance, 3, 1986, 161-178. Cornell, B. and A. Shapiro. "The Reaction of Stock Prices to the International Debt Crisis," Journal of Banking and Finance, 3, 1986, 55-73. Demirgu9-Kunt, A. and I. Diwan. "The Market-Based Menu Approach: An Analysis of Commercial Bank Choice Behavior," PRE Working Paper, World Bank, 1990. Eysell T., D. Fraser and N. Rangan. "Debt-Equity Swaps, Regulation K and Bank Stock Returns," Journal of Banking and Finance, 13, 1989, 853-868. Huizinga, H. and S. Ozler. "Secondary Market Prices for Developing Country Debt," mimeo, Stanford University, 1990. Madura, J., A. Tucker and E. Zarruk. "Reaction of Bank Share Prices to the Thirld-World Debt Reduction Plan," mimen, Florida Atlantic University, 1990. Musumeci, J. and J. Sinkey. "The International Debt Crisis, Investor Contagion, and Bank Security Returns in 1987: The Brazilian Experience," forthcoming, Journal of Money. Banking and Credit. Ozler, S. "On the Relation Between Reschedulings and Bank Value," American Economic Review, 12, 1989, 1117-1131. Sachs J. and H. Huizinga. "U.S. Commercial Banks and the Developing Country Debt Crisis," Brookings Papers on Economic Activity, 2, 1987, 555- 606. 20 Schoder, S. and P. Vankudre. "The Market for Bank Stocks and Banks' Disclosure of Cross-Border Exposure: The 1982 Mexican Debt Crisis," Studies in Bankin; and Finance, 3, 1986, 179-202. Smirlock, M. and H. Kaufold. "Banh Foreign Lending, Mandatory Disclosure Rules, and Reaction of Bank Stock Prices to the Mexican Debt Crisis," Journal of Business, 60, 1987, 347-364. Unal, H. and A. Demirguc-Kunt. "Menu Approach to International Debt Reschedulings and Bank Stock Returns in the United States and J&apan," mimeo, The World Bank, 1990. Zellner, A. "An Efficient Method of Estimating Seemingly Unrelated Regressions and Tests for Aggregation Bias," Journal of the American Statistical Association, 1962, 348-368. 21 Table 1. List of Exposed and Non-exposed Banks. Highly Exoosed Banks EXP/MV EXP/BV (14) Manufacturers Hanover 418.8 212.7 (18) Chase Manhattan 239.7, 145.4 (17) Chemical Banking 215.6 142.0 (15) BankAmerica Corp. 180.9 173.1. (16) Continental Bank Corp. 154.7 136.0 (19) Citicorp 101.7 101.5 (21) Bankers Trust NY 100.3 81.4 (22) First Chicago 87.7 83.1 (24) J.P. Morgan Co. 59.3 67.7 (20) First Pennsylvania Corp. 56.2 i06.6 (23) Bank of New York 56.2 55.7 Lowly Exoosed Banks EXP/MV EXP/BV (26) Southeast Banking Corp. 30.0 33.2 (25) Republic NY Corp. 28.8 33.3 (29) Northern Trust Corp. 25.2 26.6 (27) Bank of Boston Corp. 18.2 15.9 (28) Manufactures National 15.3 17.9 (30) Security Pacific 12.8 14.7 (31) Wells Fargo & Co. 12.1 17.7 (10) NBD Bancorp 8.9 10.5 (12) Midatlantic Corp. 3.5 3.7 (13) NCNB Corp. 0.2 0.3 Non-Exposed Banks EXP/MV EXP/BV (1) Dominion Bankshares 0.0 0.0 (2) First Alabama Bankshares Inc. 0.0 0.0 (3) Crestar Financial Corp. 0.0 0.0 (4) Baybanks Inc. 0.0 0.0 (5) U.S. Trust Corp. 0.0 0.0 (6) State Street Boston Corp. 0.0 0.0 (7) Citizens and Southern 0.0 0.0 (8) Barnett Banks inc. 0.0 0.0 '9) First Virginia Banks, Inc. 0.0 0.0 Notes: EXP/MV and EXP/BV are LDC exposure as percentages of market and book. value of bank's capital respectively. LDC exposure is taken as exposure to Argentina, Brazil, Mexico, and Venezuela. All data are as of December 30, 1988. Numbers in parantheses correspond to bank numbers in Tables 3-6. 22 Table 2. Dates and Events. Date Event October 28, 1982 IMF tentatively arranges $2 billion in assist.ance for Argentina. December 15, 1982 IMF reaches tentative agreement to provide about $4.9 billion in loans to Brazil. December 23, 1982 IMF formally approves $3.96 billion loan to Mexico. January 10, 1983 IMF approves $882.5 million in loans to Chile. February 28, 1983 IMF approves $5.5 billion in loans to Brazil March 7, 1983 Proposed to increase U.S. quota to IMF by $8.5 billion introduced in U.S. Senate. June 8, 1983 Proposal to increase U.S. quota to IMF passes in Senate. February 16, 1988 World Bank Executive Directors agreed on a $74.8 billion general capital increase. December 15-20, 1988 World Bank proposes commercial banks with heavy exposure reduce debt. Mexico seeks debt restructuring. March 10, 1990 Details of the Debt Reduction Plan were announced. Banks would be asked to forgive some of their debt. The percentage of debt to be forgiven was uncertain although rumors centered around 30 percent. March 17, 1990 Additional details of the Debt Reduction announced. The Treasury proposed reduction of bank debt by convertion into bonds whose principal and interest would be guaranteed by the World Bank and the IMF. 23 Table 2 (continued). 2UEvent May 6, 1990 G-7 endorse 50% increase in IMF funds. May 7, 1990 IMF policy-making committee increase institutions resources 50% from about $120 billion to about $180 billion. 24 Table 3. LArge IMF Commitments to Latin Debtors in 1982 and 1983. BANK MARKET MODEL PARAMETERS 821028 821215 821223 Cl pi 7di -foi 'di 14 0.0002 0.91 0.0040 0.001 15 -0.°005 1.14 0.0080 -0.009 16 0.0003 0.81 0.0020 -0.005 17 0.0002 0.79 0.0060 -0.002 18 -0.0006 1.02 0.0040 0.001 19 0.0006 1.45 -0.0090 -0.006 20 0.0010 0.56 0.0850* -0.002 21 0.0004 1.00 -0.0080 .0.024* 22 -0.0001 1.03 0.0060 0.0210 23 0.0005 0.55 -0.0004 -0.003 24 0.0007 0.89 0.0001 -0.008 25 -0.0001 0.56 0.0010 -0.009 26 0.0009 0.19 0.0010 -0.012 27 0.0005 0.48 0.0280* -0.007 28 0.0010 0.87 0.0020 -0.026* 29 0.0010 0.22 -0.0010 -0.000 30 -0,0002 0.71 0.0020 .0.020* 31 0.0001 0.80 0.0040 -0.014 HYP1* 1.59* 1.47# 0.91 HYP2* 1.66* 1.25 0.93 1 -0.0003 0.27 0.0070 -0.0150* 2 -0.0001 0.19 0.0001 0.0006 3 0.0008 0.22 0.0010 0.0010 4 0.0006 0.24 0.0080 0.0040 5 0.0010 0.20 -0.0030 -0.0020 6 0.0020 0.34 -0.0030 0.0060 7 0.0010 0.08 -0.0050 0.0007 8 0.0003 0.45 0.0280* -0.0150 9 0.0010 0.40 0.0110 0.0070 HYPln 1.72# 1.36 0.61 HYP2n* 1.93* 1.53 0.69 HYP14 1.68* 1.48* 1.01 HYP2a 1.73* 1.48* 1.05 25 Table 3 (continued). BANK MARKET MODEL PARAMETERS 830110 830228 ai A,i -di Yeid 'Ydi 'Yi 14 -0.0009 0.99 0.0180* 0.1100* 15 -0.0003 1.05 0.0180* 0.2700* 16 -0.0008 1.02 -0.0080 -0.1300 17 -0.0003 0.94 0.0006 0.0008 18 -0.0008 1.11 -0.0010 0.0080 19 -0.0004 1.50 0.0100 0.1600 20 0.0010 1.08 -0.0110 -0.0900 21 0.0004 1.17 0.0008 0.0100 22 0.0005 1.48 -0.0110 -0.1800 23 0.0010 0.48 -0.0040 -0.1000 24 -0.0005 0.82 0.0050 0.1200 25 -0.0003 0.49 -0.0010 -0.0200 27 0.0001 0.76 0.0200* 0.7200* 30 0.0010 0.74 -0.0060 -0.2000 31 0.0010 0.87 -0.0070 -0.1100 10 0.0010 0.42 -0.0100* -1.2100* 12 0.0020 0.14 0.0090 0.6500 13 0.0010 0.48 -0.0003 -0 0500 HYP1l 1.48# 1.58* 0.91 0.91 HYP2. 1.54# 1.60* 0.92 0.72 1 0.0010 0.41 -0.0030 2 0.0010 0.31 -0.0010 3 0.0010 0.10 -0.0080' 4 0.0010 0.04 -0.0030 5 0.0005 0.18 -0.0009 6 -0.0030 0.56 -0.0030 7 0.0020 0.23 0.0010 8 0.0010 0.34 0.0030 9 0.0010 0.56 0.0040 HYP1n. 0.41 0.58 HYP2ne 0.38 0.59 HYP1, 1.15 0.93 HYP2, 1.19 0.96 Notes: * and # indicate significance at 5 and 10 percent levels respectively. Significance levels for market parameters are not reported. Hypothesis 1 tests whether all coefficients are equal to zero and hypothesis 2 tests whether they are all equal. Subscripts e, no, and a refer to tests for the groups of exposed,non-exposed, and all banks respectively. F values are reported. 26 Table 4. U.S. Quota Increase to the IMF in 1983. BANK 830307 'Ydi ħi 14 0.0040 0.040 15 -0.0140# -0.2300 16 0.0220* 0.386* 17 0.0190* 0.187* 18 0.0190* 0.186* 19 0.0100 0.178 20 0.0110 0.130 21 0.0050 0.105 22 0.0170# 0.291# 23 0.0009 0,024 24 0.0100# 0.271# 25 0.0220* 0.621* 27 0.0100 0.431 30 0.0110 0.404 31 0.0130 0.229 10 0.0090 0.696 12 -0.0010 -0.143 13 0.0040 0.553 HYP1* 1.73* 1.73* HYP24 1.44# 1.37 1 0.018* 2 0.011# 3 0.001 4 0.023* 5 0,005 6 0.018* 7 0.003 8 -0.003 9 0.015 HYP1n. 3.18* HYP2n, 1.93* HYP18 2.17* HYP2a 1.61* Notes as for Table 3. 27 Table S. Brady Plan Announcement of March 1989. BANK MARKET MODEL PARAMETERS 890310 890317 ai Oi 'Ydi 'Yei ddi e 14 -0.00040 1.16 0.0120 0.062 0.022* 0.122* 15 -0.00005 1.56 -0.0030 -0.073 0.022* 0.282* 16 -0.00080 0.92 -0.0003 -0.013 0.004 0.106 17 -0.00150 1.39 0.0003 -0,007 0.027* 0.268* 18 -0.00050 1.28 0.0130# 0.)31# 0.015* 0.144# 19 -0.00100 2.12 0.0100 0.283 0.034* 0.776* 20 -0.00050 0.51 -0.0040 -0.193 0.003 0.017 21 -0.00080 1.41 0.0100 0.243 0.023* 0.525* 22 -0.00030 1.17 0.0040 0.080 0.022* 0.585* 23 -0.00030 0.96 -0.0027 -0.270 0.004 -0.025 24 -0.00050 1.23 0.0040 0.279 0.028* 1.330* 25 0.00030 0.40 0.0050 0.114 0.0006 0.590 27 -0.00100 1.05 0.0030 0.378 0.0130 1.860* 30 -0.00070 1.19 0.0070 1.490 0.0170* 3.330* 31 -0.00009 0.85 0.0070 1.500 0.0130* 2.840* 10 0.00010 1.09 -0.0030 -1.090 0.0060 2.170 HYP1* 0.55 0.71 2.35* 2.96* HYP2, 0.55 0.75 1.50# 3.15* 1 0.00020 0.65 0.0030 -0.010 2 0.00030 0.56 -0.0040 -0.006 3 0.00070 0.42 -0.0060 -0.012* 4 -0.00100 0.57 -0.0010 0.003 5 -0.00005 0.16 -0.0020 0.001 6 0.00090 0.70 -0.0003 -0.002 8 -0.00080 1.15 -0.0006 0.005 9 0.00070 0.46 -0.0040 -0.006 HYPlne 0.24 1.00 HYP2n, 0.21 1.10 HYPla 0.42 2.18* HYP2. 0.44 1.94* 29 Table 7. The IMF Quota Subscription Increase of May 1990. Constant EXP/MV 0.016 -0.013 R-0.22 N-24 (2.48) (-2.51) R-0.19 Note: The dependent variable is the bank stock return on'May 4, 1990. EXP/MV is total LDC bank exposure to Argentina, Brazil, Mexico, and Venezuela divided by bank market value. 28 Table 6. World Bank Capital Increase of February 1988. BANK MARKET MODEL PARAMETERS 980219 0i tzYdi 'Yi 14 0.0003 1.03 0.0210* 0.031* 15 0.0030 1.06 -0.0130 -0.021 16 -0.0060 0.72 0.0030 0.007 17 0.0010 1.19 -0.0040 -0.010 18 0.0050 1.10 0.0010 0.002 19 0.0006 1.48 -0.0070 -0.040 21 0.0001 1.00 -0.0006 -0.004 22 0.0010 1.10 -0.0030 -0.020 23 0.0010 0.58 0.0040 0.020 24 -0.0006 1.19 0.0090 0.140 25 -0.0002 0.32 0.0090 0.270 27 -0.0003 1.19 -0.0030 -0.130 30 0.0015 0.80 -0.0170* -0.770* HYP1l 1.90* 1.90* HYP2* 1.98* 1.83* 1 -0.00040 0.55 -0.0040 2 0.00040 0.42 -0.0140* 3 0.00030 0.40 -0.0060 4 0.00050 0.36 -0.0080 5 0.00003 0.16 0.0050 6 0.00060 0.76 -0.0080 8 0.00030 0.91 0.0007 9 -0.00020 u.31 -0.0005 HYPlne 0.84 HYP2n* 0.75 HYPl4 1.47# HYP2, 1.54* Notes as for Table 3. PRE Working Paper Series Contact U2 Aulhor afor gape WPS568 Higher Wages for Relief Work Can Martin Ravallion January 1991 C. Spooner Make Many of the Poor Worse Off: Gaurav Datt 30464 Recent Evidence from Maharashtra's Shubham Chaudhuri "Employment Guarantee Scheme' WPS569 Domestic Purchase Requirements for Wendy E. Takacs January 1991 S. Fallon Import License Allocations in Mali 37947 WPS570 Debt Concentration and Secondary Raquel Fernandez January 1991 S. King-Watson Market Prices Sule Ozler 31047 WPS571 Credit's Effect on Productivity in Gershon Feder January 1991 C. Spooner Chinese Agriculture: A Lawrence J. Lau 30464 Microeconomic Model of Justin Y. Lin Disequilibrium Xiaopeng Luo WPS572 Capital Positions of Japanese Edward J. Kane January 1991 S. King-Watson Banks Haluk Unal 33730 Asli Demirguc-Kunt WPS573 Malaysian Labor Markets Under Dipak Mazumdar January 1991 M. Schreier Structural Adjustment 36432 WPS574 Public Policies and Saving in Vittorio Corbo January 1991 S. Jonnakuty Developing Countries Klaus Schmidt-Hebbel 39074 WPS575 Household Saving in Developing Klaus Schmidt-Hebbel January 1991 E. Khine Countries: First Cross-Country Steven B. Webb 39361 Evidence Giancarlo Corsetti WPS576 Lessons from Tax Reform: Wayne Thirsk January 1991 A. Bhalla An Overview 37699 WPS577 Africa's Rising Inflation: Causes, Ajay Chhibber February 1991 B. Dhomun Consequences, and Cures 39413 WPS578 The Bank's Use of Technical Beatrice Buyck January 1991 E. Madrona Assistance for Institutional 37496 Development WPS579 Chile's Experience with Stabilization Vittorio Corbo January 1991 E. Khine Revisited Andr6s Solimano 39361 WPS58O Do Natural Resource-Based Alexander J. Yeats January 1991 J. Jacobson Industrialization Strategies Convey 33710 Important (Unrecognized) Price Benefits for Commodity-Exporting Developing Countries? PRE Working Pagor Series Contact lug Aid=hDAr for pe WPS581 How Successful is World Bank Patrick Conway January 1991 S. Fallon Lending for Structural Adjustment? 37947 WPS582 Adjustment Programs and Bank Vittorio Corbo January 1991 L. Oropesa Support: Rationale and Main Stanley Fischer 39075 Results WPS583 World Bank Lending for Education Marlaine E. Lockheed January 1991 C. Cristobal Research, 1982-89 Alastair G. Rodd 33640 WPS584 Whither Hungary and the European Alfred Tovias January 1991 G. Ilogon Communities? Sam Laird 33732 WPS585 Financial Innovation and Money Patricio Arrau January 1991 S. King-Watson Demand: Theory and Empirical Jose de Gregorio 31047 Implementation WPS586 The Challenging Arithmetic of Poverty Martin Ravallion February 1991 WDR Office in Bangladesh 31393 WPS587 Quantifying the Magnitude and Martin Ravallion February 1991 WDR Office Severity of Absolute Poverty in the Gaurav Datt 31393 Developing World in the Mid-1980s Dominique van de Walle Elaine Chan WPS588 Obstacles to Developing Small and Brian Levy February 1991 E. Madrona Medium-Sized Enterprises: An 37496 Empirical Assessment WPS589 To Prescribe or Not to Prescribe: Jeffrey S. Hammer February 1991 0. Nadora On the Regulation of Pharmaceuticals S6-065 in Developing Countries WPS590 The Domestic Financial Market and Premachandra Athukorala February 1991 M. Kienzle the Trade Liberalization Outcome: Sarath Rajapatirana 30733 The Evidence from Sri Lanka WPS591 Global Indicators of Nutritional Risk Rae Galloway February 1991 0. Nadora 31091 WPS592 Offiial Credits to Developing Asli Demirg0c-Kunt February 1991 S. King-Watson Countries: Implicit Transfers to the Harry Huizinga 33730 Banks