Printer Friendly

The effect of the Brazil suspension announcement on returns to large U.S. bank stocks.

The Effect of the Brazil Suspension Announcement on Returns to Large U.S. Bank Stocks


"In February 1987, Brazil stunned the world by unilaterally suspending interest payments on its commercial-bank debt for an indefinite period" [5, p.617]. This quote from a respected international economics text accurately reflects the general reaction, as seen in the popular and financial press, to the Brazilian announcement of February 20, 1987. However, an interesting question arises as to the degree to which financial markets were "stunned" as a result of Brazil's action. An analysis of the performance of stocks of large United States banks during the period just prior to and following Brazil's action sheds some light on the effect which such disturbances have on securities markets.

Because large U.S. banks are major holders of Brazilian debt, as well as debt of other Latin American countries which could be influenced by Brazil's actions, it would be reasonable to assume that bank stocks would be adversely affected by the announcement. If markets are operating efficiently, such effects should occur only in bank stocks exposed to Latin American debt and should be quickly discounted into stock values. If, on the other hand, markets were "stunned" the effects might spread to non-exposed banks, and stock prices may take longer to adjust fully as market participants try to evaluate the new information.

The following study is an empirical analysis of the returns on large U.S. bank stocks during the thirty-six day period surrounding Brazil's announcement of unilateral suspension of interest payments. The results may be useful as an example of how the stock market reacts to an international disturbance.

Study Results

A sample of 52 banks stocks was analyzed, including nine that were identified as being heavily exposed to Latin American debt. All returns were adjusted for movements in the Standard and Poor's 500 index. The returns for the entire sample did not fall just prior to, on, or just after February 20, 1987. In fact, there were slight gains immediately following the announcement. By contrast, when only the nine most heavily exposed banks were analyzed returns fell immediately following the announcement and continued to fall for 17 of the next 22 days.

There are two conclusions of the analysis. First, market participants discriminated in their reaction. Losses in value of the most exposed stocks were offset by gains in the less exposed stocks, perhaps reflecting a flight to safety. There was no general panic affecting bank stocks as a group.

Second, the most exposed banks adjusted as expected, but not quickly. Twenty-one days is an exceptionally long adjustment period in the stock market where most new information such as earnings and dividend announcements is fully discounted within hours. The Brazilian announcement may have been only the first piece of an information flow which actually occurred over a period of several weeks as negotiations and further developments regarding the crisis unfolded.

Debt Crisis Background

A detailed discussion of the Latin American debt crisis is beyond the scope of this study; however, a brief background of the context of the Brazilian suspension is relevant [1,9]. The first overt manifestation of the debt crisis is generally perceived to be the announcement on August 19, 1982 by the government of Mexico, that a moratorium was being placed on repayment of debt principal to foreign creditors. That announcement was followed by a series of debt restructuring agreements between Mexico and its international creditors.

Problems with other South American countries followed at a rapid pace. By early 1984, all except Colombia and Paraguay had undergone debt rescheduling. Brazil underwent a rescheduling every year from 1982 through 1985. After a remarkable economic resurgence in 1985, the Brazilian economy suffered a severe reversal in 1986 with the failure of the "Cruzada Plan" to halt inflation while maintaining a trade surplus. By February 1987, with foreign reserves all but depleted, President Sarney announced the debt payment moratorium [1, pp. 121-122].

The continuing debt problems of Latin America generally and Brazil specifically provided a flow of information to security markets as to the risk exposure of the lending banks. However, the profits resulting from fees and higher margins on the rescheduled debt may have masked the downside potential and left the market unprepared for the Brazilian action.

The impact of the 1982 Mexican moratorium announcement on the returns of U.S. banks has been the subject of several recent studies [4,7,8]. Bruner and Simms, using the cumulative average residual event study methodology found a negative effect for the Mexican announcement on the returns to U.S. bank stocks [3]. Additionally, they found that at first the degree of exposure to Mexican loans was positively related to returns, but by the sixth day after the announcement, the effect had turned significantly negative. Bruner and Simms suggest an initial contagion effect which was soon replaced by rational pricing. The Bruner and Simms results support the hypothesis that financial markets were in fact "stunned" by the Mexico announcement. A comparison between the effects of the 1982 announcement and the subsequent 1987 Brazilian announcement on bank stock returns yields useful insights about the learning function in financial markets.

The Data

The sample used for this study was drawn from the 60 largest U.S. banking firms. Firms which underwent significant mergers or for which complete data were not available were eliminated. The remaining sample consists of 52 firms. Daily closing stock price data were collected for these 52 banks over a period beginning 61 trading days prior to the February 20, 1987 announcement and ending 25 trading days after the announcement. Adjustments were made for dividends and stock splits. These prices were converted to price relatives so that 86 days of return data were available.

The first 50 days of data were used to determine intercept and slope coefficients in the market model as daily return relatives for each of the 52 companies were regressed against the S&P 500 Index daily return relatives. The remaining 36 days of return relatives - ten days before the announcement the announcement date, and twenty-five days after the announcement - were examined for unusual performance by the bank stocks. The 52 banks used in this study are listed in Table 1. The data were collected from the Dow Jones News/Retrieval Service. [Tabular Data Omitted]

The Methodology

The purpose of this paper is to determine whether bank stocks exhibited abnormal performance immediately before or after Brazil announced its intention to suspend interest payments on its foreign debt. In effect, the paper is a test of market efficiency assuming the capital asset pricing model adequately describes the investor tradeoff between risk and return. If the market had properly discounted this information prior to the announcement, then one would expect no abnormal performance. If investors failed to anticipate the suspension of interest payments, the bank stocks would likely under-perform the market on a risk-adjusted basis. A third outcome would occur with bank stocks outperforming the market if the Brazilian announcement were less serious than investors had anticipated. Given the realistic possibility of any of these three outcomes, any abnormal performance should occur on (or possibly, before) the announcement date if the markets are semi-strong efficient. There should be few abnormal returns following the announcement.

Abnormal performance is determined by taking the difference between the actual return on a stock on a given day and its expected return. The expected return is computed using the market model. One form of the market model suggests that: [R.sub.jT]=[a.sub.j]+[b.sub.j][R.sub.mT]+[e.sub.jT] where [R.sub.jT] is the return relative for stock j in period T, [R.sub.mT] is the return relative for the market portfolio in period T, and [e.sub.jT] is a randomly distributed error term. In this paper, the proxy for the market portfolio is the S&P 500 Index. (The results of the analysis were unchanged when the New York Stock Exchange Composite Index was used.) The [a.sub.j] and [b.sub.j] values of the estimating equation were found by regressing actual stock return relatives against market return relatives for each bank stock for the 50 day period beginning 60 days prior to the Brazilian announcement.

The market model was then used to measure abnormal stock returns by rewriting equation (1) as: [U.sub.jT] = [R.sub.jT] - ([a.sub.j] + [b.sub.j] [R.sub.mT]) where [U.sub.jT] is the abnormal return for stock j in period T. These abnormal returns were calculated for each of the ten trading days before the announcement date, the announcement date itself, and the twenty-five trading days immediately following the announcement date.

The cumulative average residual (CAR) was employed to analyze these residuals [6]. Using the CAR, the average residual for all banks is calculated each day and these mean figures are summed over succeeding days:

[CAR.sub.T]= the cumulative average residual on day T (starting on day - 10 and running to day T, where T =[is less than or equal to]) 25.

These same calculations were made for a subset of the bank sample which consisted only of those banks whose Latin American loan exposure exceeded 100 percent of equity. There were nine such banks in the sample. The high exposure banks are shown in Table 1 with Latin American loan exposure as a percent of equity recorded after the bank's name. The logic behind examining this subset of banks separately is based on the assumption that the stocks of banks with the greatest (proportionate) loan exposure would be most affected by the Brazilian announcement. A comparison of these nine banks with the broader market of 52 banks will indicate whether the contagion effect found by Bruner and Simms in the 1982 case also existed in the 1987 event.

Result for All the Banks

Using the cumulative average residual technique, the results for the entire sample indicate positive excess returns occurred in the period immediately before and after the Brazilian announcement suspending interest payments on foreign loans. The excess returns were positive on five of the ten days preceding the announcement, with the magnitude of the positive residuals exceeding the negative residuals so that by the day preceding the announcement (T=-1) the CAR had a value of +.0253. On the announcement date the residual was +.0036. Over the 25 day period following the announcement, the residuals were positive on 14 of the 25 days. The CAR advanced from .0289 on day zero to .0387 on day 25.

The CAR for the sample of banks for the period T = -10 to T = +25 is plotted in Figure 1. In the ten day period including observations two days before the announcement through seven days after the announcement, the residuals were positive for nine days, with the CAR advancing from .0059 to .0512. From day T =+7 through T=+25, the CAR trended downward so that much of the positive abnormal performance had dissipated by the end of day T=+25.

The t-values are calculated by dividing the CAR over a given time interval by the product of the standard deviation of the residuals and the square root of the time interval: [Mathematical Expressions Omitted] where (3) [Mathematical Expressions Omitted] and [Mathematical Expressions Omitted] where N is the number of firms in the sample. The results of the t-tests are shown in Table 2.

Table : Table 2 Cumulative Average Residual t-Statistics
Returns All Exposure
Period Banks Banks
-10 to -1 2.957* 0.603
 0 1.395 3.535*
0 to + 1 0.247 7.219*
0 to + 2 1.208 5.559*
0 to + 3 1.182 6.791*
0 to + 4 1.092 6.719*
0 to + 5 1.978 4.833*
0 to + 6 3.311* 3.850*
0 to + 7 3.549* 3.815*
+7 to + 25 1.142 3.673*

* Significant at the 1% level.

Results for the Highly Exposed Banks

For the nine banks with a Latin American debt exposure of more than 100 percent of equity, the results were quite different. Negative excess returns were found immediately before and after the Brazilian announcement. The excess returns were positive on six of the ten days preceding the announcement, while the CAR was virtually unchanged from day minus 10 to day minus 1,+.0075 to +.0082. On the announcement day, the residual was -.0152. The CAR fell for 15 of the 22 days following the announcement, rising slightly in the three days thereafter. It fell from +0.0082 on day minus one to -0.1206 on day 22. These results are shown in Figure 2. The t-tests results are given in Table 2.

Further Inferences

One of the significant implications of this study is the noticeable absence of any evidence of a contagion effect. The large bank group as a whole was not negatively affected by the announcement, while the large exposure group was negatively affected. By contrast, Bruner and Simms found that there was initially a major contagion effect following the 1982 Mexican default announcement. Market participants apparently have been made aware of the risks of Latin American loans and have apparently educated themselves as to which banks are significantly exposed. Thus, the negative impacts were observed on a selective basis. By contrast in 1982, Latin American risk exposure was not properly evaluated. The market ignorance as to degree of exposure may have created the contagion found by Bruner and Simms. The market thus displayed an effective learning function that led to rational discrimination in the Brazilian case.

The results of this study have interesting implications regarding the efficiency of the market for bank stocks. One would expect that in a semi-strong efficient market any abnormal return should occur on the announcement day, or in the days preceding the announcement if there are information leakages. If the abnormal returns continue for a period of time after the announcement date, there is evidence of market inefficiency.

Looking at the total sample of banks, large abnormal returns occurred on the two days prior to the announcement and continued until seven days after. During the seven day period following the announcement, the CAR was 2.53 percent. This result offers some evidence that the new information was not impounded in the stock price by the end of the announcement day and suggests inefficiency in the market pricing mechanism.

For the sample of banks with substantial exposure, the residuals were negative for six of the seven days preceding the announcement, again indicating some information leakages. The abnormal returns were negative on 17 of the 22 days following the announcement, and the CAR during the period was -11.36 percent. This result again suggests some inefficiency in the pricing of bank stocks. An alternative explanation might be that the large, negative residuals reflect a continuing flow of detrimental information regarding the Brazilian action.

A third result of the study is that it presents some evidence in support of a market substitution effect. Why did the larger market of bank stocks react positively to the Brazilian announcement despite the major negative reaction of the exposed bank subsample? A possible scenario is that as a result of the announcement, portfolio managers may have moved out of exposed bank stocks. In order to continue to pursue their previous portfolio strategy, they may have substituted the stocks of less exposed banks. The sample of bank stocks used in the study was unweighted. However, a review of the exposed bank group shows that they are among the largest banks in the total sample.


An announcement of a suspension of interest payments on foreign bank debt by the Government of Brazil was made on February 20, 1987. This study analyzed the impact of the announcement on the returns on large U.S. bank stocks. The results showed that the news of suspension had a positive effect on the sample of 52 bank stocks examined. However, the effect on a subsample consisting of the nine most exposed banks was highly negative.

The study shows the absence of the contagion effect found in a similar event in 1982. The results cast doubt on the semi-strong form of the efficient market hypothesis for the large bank stock market. They also suggest evidence of a substitution effect among bank stocks as a result of the announcement.


1. Branford, S. and B. Kucinski. The Debt Squads. London: Zed

Book, 1988. 2. Brown, S. and J. Warner, "Measuring Security Price

Performance." Journal of Financial Economics, Vol. 8, September

1980, pp. 205-58. 3. Bruner, R. and J. Simms, Jr. "The Internatioanl Debt Crisis and

Bank Security Returns in 1982." Journal of Money Credit and

Banking, Vol. 19, February 1987, pp. 46-55. 4. Cornell, B. and A. C. Shapiro. "The Reaction of Bank Stock

Prices to the International Debit Crisis." Journal of Banking

Finance, Vol. 10, March 1986, pp. 55-73. 5. Krugman, P. and M. Obstfeld. International Economics Theory

and Policy. Glenview, Illinois: Scott Foresman and Co.,

1988. 6. Pettway, R. and J. Trifts. "Do Banks Overbid When Acquiring

Failed Banks?" Financial Management, Vol. 14, Summer

1985, pp. 5-15. 7. Schoder, S. and P. Vankudre. "The Market for Bank Stocks

and Banks' Disclosure of Cross-Border Exposure: The 1982

Mexican Debt Crisis." Working Paper, Wharton School of Finance,

University of Pennsylvania, 1986. 8. Smirlock, M. and H. Kaufold. "Foreign Lending, Disclosure,

and the Mexican Debt Crisis." In Bertinetti, R., ed. Proceedings:

A Conference on Bank Structure and Competition.

Federal Reserve Bank of Chicago, 1985. 9. Wesson, R. ed. Coping with Latin American Debt. New York:

Praeger, 1987.

PHOTO : Figure 1 CAR for all U.S. Banks

PHOTO : Figure 2 CAR for Banks Heavily Exposed to Latin American Debt
COPYRIGHT 1991 St. John's University, College of Business Administration
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1991 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Rivard, Richard J.; Meyer, Richard L.; Hovik, Conrad E.
Publication:Review of Business
Date:Mar 22, 1991
Previous Article:New product forecasting horizons and accuracy.
Next Article:The application of expert systems to securities analysis.

Related Articles
The effects of announcements of bank lending agreements on the market values of U.S. banks.
Staff Studies.
Does long-term performance of mergers match market expectations? Evidence from the US banking industry.
The impact of investment opportunities and free cash flow on financial liberalization: a cross-firm analysis of emerging economies.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters