The effects of announcements of bank lending agreements on the market values of U.S. banks.

In seeking to provide theoretical perspectives on the uniqueness of banks, a number of authors, including Campbell and Kracaw |5~, Diamond |9~, and Fama |11~, have focused on the role of banks in the information production process for corporations. Banks, the argument goes, are especially adept at evaluating risky lending situations, and their decisions to grant loans signal the borrowers' financial strength to other agents. These other agents, by taking into consideration the signals associated with bank loans, can lower their own information generation costs by engaging in less costly evaluations and thus can provide their services at lower costs. Fama |11~ argues that corporate and other borrowers, by signaling their creditworthiness through bank borrowing, can lower their information production costs.

Empirical tests on the uniqueness of bank loans provide evidence in support of the information production hypothesis. James |19~ and Lummer and McConnell |21~ show that announcements of bank loan agreements result in statistically significant, positive abnormal returns for samples of nonfinancial borrowing firms. These as well as other studies indicate that bank loan agreements serve as signals and provide information to other market participants. Surprisingly, very few, if any, studies have been conducted to examine the parallel wealth effects for the banks allocating the loans. It is possible that the lack of research on this topic may be due to the lending process for banks. Since banks are involved in the information production process through their lending agreements, their shareholders should experience wealth effects at the time banks announce acquisitions of information-producing capabilities. Subsequently, no wealth effects should be observable for banks when the loan agreements themselves are announced, since the potential benefits from the loans should have been recognized at the time of the announcements of the acquisitions of information-producing capabilities. However, wealth effects should be observable if banks use their lending agreements to convey some information not related to their information-producing capabilities to market participants. It is feasible that these other signals may be related to efforts by banks to diversify, manage their risks, and deal with regulatory factors.

While domestic lending agreements by major U.S. bank-holding companies (USBHCs) may signal some effort by these banks to diversify, manage risks, and deal with regulatory factors, these signals may not be as strong as similar signals associated with foreign lending agreements by USBHCs. That is, USBHCs may seek to signal their diversification and risk management efforts, and possible arbitrage of regulatory barriers to lending, by originating loans overseas. The two major propositions implied here are tested in this paper. Samples of domestic and foreign lending agreements are examined to see if their announcements have any wealth effects. The a priori hypothesis is that wealth effects for the foreign case should be more pronounced than for the domestic case due to the signaling effects.

The results of this paper support the a priori hypothesis. No wealth effects are identified for domestic bank lending agreements. However, significant, negative abnormal returns are associated with announcements of foreign lending agreements by USBHCs. Cross-sectional regressions corroborate the presence of identifiable determining factors.

The paper is organized as follows. The conceptual framework is presented in Section I. The data and methodology are contained in Section II. The results are discussed in Section III, and Section IV concludes the paper.

I. Conceptual Framework

Announcements of domestic lending agreements by banks may not convey additional information to market participants. Announcements of foreign lending agreements may signal both the potential benefits and costs of foreign lending.

A. Domestic Lending Agreements

With announcements of their domestic lending agreements, USBHCs provide information about their corporate clients to other capital market participants. However, these same announcements may not provide these participants any additional information regarding the banks themselves. Market participants, who are aware of a bank's investment in information-producing capabilities, may have already accounted for the announcement of a series of lending agreements that would exhaust or fully utilize the bank's acquired capability to produce information.(1) Thus, the abnormal returns for bank stocks should not be significantly different from zero when domestic loan agreements are announced.

B. Benefits of Foreign Lending Agreements

Through their foreign lending agreements, USBHCs may be seeking to signal to market participants their efforts to gain benefits by exploiting market imperfections, responding to regulatory factors, and managing risk through diversification.

1. Exploiting Market Imperfections

Banks, in providing loans to corporations, acquire considerable information about them. This information base may allow banks to finance the exports of their clients through loans to foreign entities. Empirical work by Jain |17~ supports the notion that foreign lending by U.S. banks can be explained by the export behavior of their domestic clients. In a more speculative vein, Darity and Horn |7, p. 33~ argue that financing customers' exports tends to increase deposits at banks and enhance their ability to originate new loans. Banks who announce foreign loan agreements may be signaling the capture of certain market imperfections and, thus, may expect to see positive abnormal returns associated with these announcements.

2. Regulatory Factors

During the 1960s and 1970s, the U.S. government imposed certain banking restrictions to improve the U.S. balance of payments. The Interest Equalization Tax (IET) was imposed in 1963 and made it more expensive for foreign borrowers to borrow from U.S. banks. The Voluntary Foreign Credit Restraint Program (VFCRP) was instituted in 1965 and made it more difficult for U.S. banks to originate foreign loans. Both of these restrictions on foreign lending were lifted in 1974.

The 1982 debt crisis, the subsequent IMF regulations, and the International Lending Supervisory Act (ILSA) mandated provisions to increase reserves and maintain minimum capital requirements might have made bank managers more prudent and risk-averse. Evidence presented by Billingsley and Lamy |4~, though not conclusive, seems to suggest that the variance of stock returns may have decreased over the legislative timeframe preceding ILSA without a significant decrease in the mean stock return. The decrease may have been due to a decrease in the unsystematic part of the risk. The Cornell, Landsman, and Shapiro |6~ study provided inconclusive evidence of the effects of ILSA on banks' market values. On the other hand, Bennett and Zimmerman |3~ note that exposure to troubled LDCs had risen from 26.1% of banks' international loan portfolio in 1982 to 29.4% in 1986. Different levels of abnormal returns should be observed under different regulatory regimes.(2)

3. Risk Diversification

International lending might provide risk reduction benefits for U.S. banks in addition to those that could be achieved from domestic loan portfolio diversification alone (Goodman |13~). Low levels of correlation between international assets returns imply that gains are associated with international diversification. There is ample evidence to indicate that national equity markets usually exhibit heterogeneous performance in any given period (see, for example, Grubel |15~). If fundamental economic activities fuel equity markets, then transnational involvement in these national economic activities should lower the variabilities in operating incomes for U.S. banks in a manner similar to the reduction in portfolio risk due to global diversification.

A variety of other factors also indicate that global loan portfolio diversification by U.S. banks is desirable. First, Gray and Gray |14~ posit that diversification benefits exist because credit market cycles are not perfectly correlated among countries. Second, U.S. banks may be able to lower their cost of capital because they may be able to lower their risk through global diversification. Finally, foreign lending by country and by type of customer may reduce the risk of catastrophic losses to a bank, since loan defaults in one country may not be perfectly correlated with similar defaults in another country (Eiteman and Stonehill |10~). Davis |8~ argues that the positive factors of foreign lending such as financial strength, resilience of most international borrowers, risk diversification, and growing expertise of international lenders outweigh the negative factors such as structural risks and uncertainties of overseas lending. Rugman and Kamath |25~ provide evidence suggesting that international diversification by banks results in risk reduction.

C. Costs of Foreign Lending Agreements

It may be the intent of bank managers to signal potentially positive news through the announcement of foreign loans. However, these announcements may also be signals, albeit unintentional, of managerial behavior and increased riskiness due to country risk.

1. Managerial Behavior

Darity and Horn |7~ argue that bank managers have been very aggressive in international lending and have provided incentives to borrowers that are out of step with the risk characteristics possessed by borrowers. These incentives might include a reduction in the spread, increase in the size of the loan to be offered, or extended maturities on loans. Kindleberger |20~ states that whenever there is a decline in interest rates for any reason, bank managers, in an attempt to maintain profitability, start allocating loans to borrowers who are willing to pay the highest interest rate, irrespective of the riskiness of the borrower. The "hubris hypothesis" is another reason why the value of a bank may fall as a result of international expansion. Roll |20~ hypothesizes that the corporate managerial market is not efficient with respect to evaluating the benefits of expansion projects. Managers of firms believe that they are capable of discovering "bargains"; however, they end up paying too much for the investments.

2. Country Risk

Country risk is another drawback for banks allocating foreign loans. Country risk involves the possibility of losses due to country-specific economic, political, and social events. Haner and Ewing |16~ identify three types of country risks resulting from international operations: First, operating risk may increase due to differences in business practices at home and abroad, and from possible discriminatory practices by the host country. Second, sociopolitical risks may arise due to national policy changes caused by changes in governments. Third, financial risk may increase due to constraints on converting funds held in the local currency to other currencies. U.S. banks lending abroad expose themselves to these country risks and face the possibility of inconvertibility and inability to transfer foreign funds into the numeraire currency.

Monti |23~ states that ever-increasing competition among lenders in international markets has resulted in the reduction of profits to the extent that they are insufficient even to cover operating costs. Aliber |2~ argues that the riskiness of foreign loans is also affected by the focus of bankers on maintaining market share regardless of costs. TABULAR DATA OMITTED In effect, banks may not be adequately compensated when they originate overseas loans.

The above discussion suggests that no abnormal returns should be observed for banks for announcements of domestic lending agreements. With regard to foreign loans, if investors believe that foreign loan originations have positive expected net present values, then positive abnormal returns should be associated with these loan announcements. On the contrary, if shareholders evaluate these loans to be unsound and capable of creating levels of risk not commensurate with the potential returns, then negative abnormal returns should be associated with the loan announcements.

II. Data and Methodology

A. Data

The sample analyzed in this study contains USBHCs engaged in both domestic and foreign lending activities over the 22-year period from 1968 through 1989.(3) The sample is limited to only those USBHCs whose stock was trading on either the New York Stock Exchange (NYSE) or the American Stock Exchange (AMEX) on or after July 2, 1962. The selection criterion was imposed due to the data restrictions of the Center for Research in Security Prices (CRSP) NYSE/AMEX Daily Return Tape, which has data on securities starting July 2, 1962. Twenty-five USBHCs that met the data requirements were identified. The Wall Street Journal Index (WSJI) for the ten-year period from 1980 to 1989 was searched to identify announcements of domestic lending agreements by these USBHCs. One hundred forty-five announcements of lending by the 25 USBHCs to domestic corporations were identified for this time period.

The WSJI was also searched over a longer time period, from 1962 to 1989, to similarly identify announcements of foreign lending agreements by the sample USBHCs. No announcements were identified for the 1962 to 1967 period. Two hundred twenty-five announcements by 16 USBHCs were identified for the 22-year period from 1968 to 1989. Exhibit 1 provides a distribution of the 225 announcements by year and bank. The peak activity year was 1981, with numerous announcements annually from 1975 to 1984. Post-1984 announcement activity was relatively slow in comparison with the activity in the previous decade. The largest number of announcements, 48, was made by Wells Fargo, while four banks made only one announcement of foreign lending agreements. The average number of announcements per bank over the 22-year study period was 14.

B. Methodology for Wealth Effects

As is the case with the studies by James |19~, and Lummer and McConnell |21~, standard event study methodology is used to identify the wealth effects associated with the announcements of bank lending agreements. The return generating process is specified using ordinary least squares market model, whose parameter estimates for the banks are obtained by using the estimation period from t = -170 to t = -21 relative to the announcement date t = 0.(4) Daily average abnormal returns (ARs), and cumulative average abnormal returns (CARs) for multiple-day intervals in the event period, are computed. Test statistics for ARs and CARs are based on average standardized abnormal returns and average standardized cumulative abnormal returns.(5) The returns for the Center for Research on Security Prices (CRSP) value-weighted index are used in this study.(6) A binomial sign-test (B-value) is also used that examines whether the proportion of positive (negative) abnormal returns in the test period is significantly different from the proportion of positive (negative) abnormal returns expected under the null hypothesis (0.50).

C. Cross-Sectional Regressions

Cross-sectional regressions are run with standardized CARs for the foreign sample as the dependent variable to explain variations in cumulative abnormal returns. Market imperfections are modeled by the size of the foreign loan, SIZE. Regulatory factors are modeled with two variables. The effects of the interest equalization tax and the Voluntary Credit Restraint Program are captured through the use of a dummy variable IET = 1 if the announcements were made prior to December 31, 1974, and 0 otherwise. The effects of ILSA are also modeled with the dummy variable ILSA = 1 for post-ILSA (1984-1989) announcements, and 0 otherwise.

One measure of risk used in the regressions is the percentage change in the variance of banks' stock returns before and after announcements of foreign lending agreements. The variance of stock returns VAR(|R.sub.j~) can be partitioned into two components (Aharony, Jones, Swary |1~):

|Mathematical Expression Omitted~ (1)

where the first term on the right-hand side is the systematic component of risk and the second term is the unsystematic or firm-specific component. |Delta~TRSK = (VAR(|R.sub.j,post~) - VAR(|R.sub.j,pre~))/VAR(|R.sub.j,pre~) = |Delta~VAR(|R.sub.j~), where pre and post are with reference to the announcement date for the loan allocation, is used as a regressor to identify the diversification effects associated with foreign lending.(7) Another regression is also estimated that uses |Delta~BETA and |Delta~UNSYS instead of |Delta~TRSK so that the components of any changes in the banks' risks can be identified.(8) |Delta~BETA = (Post-Announcement |Beta~ - Pre-announcement |Beta~)/Pre-announcement |Beta~, and |Delta~UNSYS = (VAR(|E.sub.j,post~) - VAR(|E.sub.j,pre~))/VAR(|E.sub.j,pre~).(9)

Although depositors are insured, bank stockholders may view foreign lending as risky. Thus, based on the Country Exposure Lending Survey (a publication of the interagency Federal Financial Institutions Examination Council), the host countries for the foreign loans are classified into developed nonrisky, and developing categories. DVP = 1 if the host country is a developed nonrisky country, DVP = 0 otherwise.

The following general regression model is estimated to identify determinants of wealth effects associated with announcements of foreign lending agreements:

CAR = |a.sub.0~ + |a.sub.1~|Delta~TRSK + |a.sub.2~IET + |a.sub.3~DVP + |a.sub.4~ILSA + |a.sub.5~SIZE (2)

where

CAR = Standardized cumulative average residuals for the three-day interval (-2, 0).

|Delta~TRSK = |Delta~VAR(|R.sub.j~), where VAR(|R.sub.j~) is defined in Equation (1).

IET = 1 if announcement occurred in the 1968-1974 period, 0 otherwise.

DVP = 1 if host country is developed nonrisky, 0 otherwise.

ILSA = 1 if announcement occurred in the 1984-1989 period, 0 otherwise.

SIZE = Size of loan.

In a second regression, based on Equation (1), |Delta~BETA and |Delta~UNSYS are substituted for |Delta~TRSK. The intercept |a.sub.0~ is suppressed in regressions so that the null hypothesis that the dummy variable coefficient is zero can be tested. The regression coefficient estimators are White's |26~ heteroskedasticity consistent estimators.

III. Empirical Results

A. Announcements of Domestic Lending Agreements

Exhibit 2 reports the results for announcements of domestic loans. The AR for the event date as well as the CARs for event period windows varying in length from two to 30 days are positive but not significantly different from zero. Similarly, the percentage positive residuals for the reported windows are not significantly different from what one would expect under the null hypothesis. These results substantiate the conjecture that bank stock prices should not react to news regarding domestic lending agreements.

B. Announcements of Foreign Lending Agreements

The AR for the event day and the CARs for event period windows of various lengths are reported in Exhibit 3. The AR for the event day is -0.14, which is insignificant. However, the CARs for the (-1, 0), the (-2, 0) and several other windows are negative and significant. The results indicate that shareholders of bank stocks suffer negative abnormal returns of about one-half percent over the days surrounding the announcement of foreign lending agreements by USBHCs.

C. Comparison of Domestic and Foreign Announcements

Exhibit 4 presents the mean differences of CARs for foreign and domestic lending agreements. While the mean difference for the ARs for the event date is not significantly different from zero, the mean differences for CARs for all event windows, with the exception of the (-15, 15) window, are significantly negative. It is possible that announcements of foreign lending agreements may be reflective of factors such as increased riskiness of the banks due to country risk, regulatory factors, diversification efforts, and managerial factors.

Exhibit 2. Cumulative Average Residuals for Domestic Loan
Announcements(a)

               CARs(b)                Perpos(d)
Window         (%)       Z-Value(c)   (%)         B-Value(e)

(-15 to +15)   0.50      0.83          53          0.66
(-5 to +5)     0.19      1.13          53          0.78
(-3 to +3)     0.19      0.98          53          0.72
(-1 to +1)     0.22      0.88          54          0.92

(-2,0)         0.07      0.59          55          1.10

(-1,0)         0.15      0.77          56          1.51

(0)            0.01      0.22          53          0.75

Total Number of Announcements from 1979-1989 = 145

Notes:

a The sample consists of 145 announcements of domestic loan
allocations for 25 U.S. banks collected from the Wall Street
Journal Index for the period 1980-1989.

b CARS represent cumulative average residuals.

c The null hypothesis is that standardized cumulative average
residuals equal zero.

d Represents percentage positive cumulative average residuals.

e B-value tests whether the percentage of positive (negative)
cumulative average residuals in the test period are
significantly different from the proportion expected under the
null hypothesis (50%). B-value is estimated as (S - PM) /
|square root of~P(1 - P)M |is similar to~ N(0,1), where M is
the number of announcements in the sample, P is the proportion
of positive (negative) cumulative average residuals under the
null hypothesis and S is the number of positive (negative)
cumulative average residuals observed.

D. Cross-Sectional Regressions

Exhibit 5 reports the results for the cross-sectional regressions specified in Equation (2). Regression (1A) shows that |Delta~TRSK (= |Delta~VAR(Rj)) is highly significant in explaining variations in the dependent variable CAR (-2, 0). The negative coefficient for |Delta~TRSK means that decreases in the bank's total risk are associated with relatively higher levels of CARs. Calculations for pre- and post-announcement total risk, based on Equation (1), indicate that, on average, VAR(|R.sub.j~) declines by 1.6% subsequent to announcements of foreign lending agreements by banks. It appears that investors value positively the risk reduction benefits associated with banks' foreign lending agreements.

If announcements of foreign lending agreements result in a change in the risk of the banks' portfolios to the degree that they result in declines in the market values of such banks, then a significant relationship between abnormal returns and the systematic risk component of these banks should be observed. On the other hand, if foreign lending agreements are expected to possess default risks that are different than the ones for the banks' existing loan portfolios, thereby affecting the bankruptcy probability for these banks, then a significant relationship between abnormal returns and the unsystematic risk components of these banks should be observed.(10)

Exhibit 3. Cumulative Average Residuals for Foreign Loan
Announcements(a)

               CARs(b)                 Perpos(d)
Window         (%)       Z-Value(c)       (%)       B-Value(e)

(-15 to +15)   -0.06       -0.04          49        -0.20
(-5 to +5)     -0.64       -2.25(**)      39        -3.40(***)
(-3 to +3)     -0.65       -2.46(**)      41        -2.81(***)

(-1 to +1)     -0.37       -2.07(**)      42        -2.33(**)

(-2,0)         -0.44       -2.43(**)      41        -2.60(***)
(-1,0)         -0.26       -1.77(*)       41        -2.82(***)

(0)            -0.14       -1.47          43        -2.10(**)

Notes:

a The sample consists of 225 announcements of foreign loan
allocations by U.S. banks collected from the Wall Street
Journal Index for the period 1968-1989.

b CARS represent cumulative average residuals.

c The null hypothesis is that standardized cumulative average
residuals equal zero.

d Represents percentage positive cumulative average residuals.

e B-value tests whether the percentage of positive (negative)
cumulative average residuals in the test period are
significantly different from the proportion expected under the
null hypothesis (50%). B-value is estimated as (S - PM) /
|square root of~P(1 - P)M |is similar to~ N(0, 1), where M is
the number of announcements in the sample, P is the proportion
of positive (negative) cumulative average residuals under the
null hypothesis and S is the number of positive (negative)
cumulative average residuals observed.

* Significant at |Alpha~ = 0.10.
** Significant at |Alpha~ = 0.05.
*** Significant at |Alpha~ = 0.01.

The components of |Delta~TRSK, namely |Delta~BETA and |Delta~UNSYS replace |Delta~TRSK in Regression (1B). The coefficient for |Delta~BETA is negative but not significant. However, the coefficient for |Delta~UNSYS is significantly negative. This decomposition of |Delta~TRSK indicates that decreases in banks' total risks subsequent to announcements of foreign lending agreements are primarily due to decreases in the banks' unsystematic risks. Comparisons of estimations of pre- and post-announcement unsystematic risk indicate that for the average foreign lending agreement announcement, banks experience a 4.3% decline in unsystematic risk. The results appear to imply that investors, in general, associate a lower probability of bankruptcy risk with announcements of foreign lending agreements.

Exhibit 4. Mean Difference (Foreign Loans - Domestic Loans) of
Cumulative Average Residuals Between U.S. Banks Making
Announcements of Foreign Loans and Domestic Loans

Window            Mean Difference (%)      Z-Value(a)

(-15 to +15)            -0.56              -0.66
(-5 to +5)              -0.83              -2.30(***)
(-3 to +3)              -0.84              -2.33(***)
(-1 to +1)              -0.59              -1.99(**)

(-2,0)                  -0.51              -2.00(**)
(-1,0)                  -0.41              -1.72(*)

(0)                     -0.15              -1.10

Notes:

a Z-value is estimated as (|SCAR.sub.1~ - |SCAR.sub.2~) /
|square root of~1/|M.sub.1~ + 1/|M.sub.2~, where |M.sub.1~ and
|M.sub.2~ represent the number of announcements in the foreign
and domestic samples, respectively.

* Significant at |Alpha~ = 0.10.
** Significant at |Alpha~ = 0.05.
*** Significant at |Alpha~ = 0.01.

Regression (2) in Exhibit 5 reports the results of regressing both |Delta~TRSK and IET against CARs. IET is representative of regulatory factors in the U.S., which made foreign borrowing more expensive in the U.S. in the 1963 to 1974 period. A priori, it could be argued that due to regulatory factors during this time period, USBHCs might have felt pressure to offer highly attractive interest rates to foreign borrowers to maintain or enhance their multinational capabilities. If this were the case, then one would expect to see nonpositive CARs associated with announcements of foreign lending agreements during the period in question. Regression (2) provides very weak support for this proposition. The coefficient for lET has the right sign, but has a significance level of only 11.5%. The two variables combined explain 7.41% of the variation, with the regression being significant at the 0.01% level.

DVP is added to |Delta~TRSK and IET in Regression (3). The regression is highly significant with 7.37% of the variation being explained by the three variables. |Delta~TRSK is significant, as in the case for the previous regression. IET becomes significant at the ten percent level in this regression. Its negative coefficient provides some support for the conjecture advanced previously. DVP has a positive coefficient that is not significantly different from zero.

The last regression in Exhibit 5 includes ILSA as a dummy variable. The coefficients for the three variables have the same signs and significance as in Regression (3). TABULAR DATA OMITTED ILSA has a negative but insignificant coefficient.(11) The intent of the ILSA was to improve foreign lending practices by U.S. banks. However, the absence of a positive significant coefficient appears to indicate that the ILSA may not have had the desired effect of positively influencing foreign lending by U.S. banks.

A final variable, namely SIZE, was also proposed in the methodology section as representative of market imperfections. The size of the loan was available for 177 announcements. Separate regressions were run for this particular subsample with size of loans used in one set of regressions and size of loans/total assets in a second set of regressions. However, the coefficients for SIZE had values close to zero, with t-values ranging from 0.18 to 0.38. If the results from this subsample can be generalized, it can be stated that announcements of foreign lending agreements by USBHCs do not appear to signal capture of any gains associated with exploiting market imperfections.

IV. Summary and Conclusions

The purpose of this study was to examine the market reaction to announcements of lending agreements, especially foreign ones, by USBHCs. The results provide support for the notion that banks, as vehicles for information generation about their corporate clients, do not provide any information about themselves to market participants when they announce domestic lending agreements. However, the evidence in this paper suggests that banks do in fact reveal information about themselves to investors when they announce their foreign lending agreements.

The results in this paper show that shareholders of USBHCs experience significant negative abnormal returns at the time of the announcement of foreign lending agreements. The results suggest that the risks and costs associated with foreign loans outweigh any diversification and synergistic benefits that may be derived from them. Lower interest rate spreads during the 1965 to 1974 period may have been justified to remain competitive under the regulatory factors prevailing during the time. However, average negative abnormal returns for the whole period from 1968 to 1989 are puzzling. It may not be feasible, and is beyond the scope of this paper, to try to identify the reasons that may account for the propensity of managers of USBHCs to engage in what appears to be unsound lending practices. It is possible that these managers unwittingly invested in "bad" loans because their salaries, bonuses, and commissions were tied to the total volume of loans allocated, rather than to the ultimate profits from these loans. It is also feasible that these loans were allocated with the best of intentions but showed the effects associated with poor training of lending officers, charging low spreads (Folkerts-Landau |12~), following the herd behavior (Jain and Gupta |18~), and aggressive marketing of loans.

A number of cross-sectional regressions were run to explain variations in abnormal returns associated with announcements of foreign lending agreements. The coefficient for the change in pre- and post-announcement total risk for the banks was significantly negative. On the average, a bank announcing a foreign lending agreement could expect to see a 1.65% decline in its total risk. Decomposition of this total risk into its systematic and unsystematic components indicated that the source of the decline in total risk was due to a decline in unsystematic risk. Unsystematic risk, on the average, declined by 4.3% subsequent to announcements of foreign lending agreements. The results suggest that foreign lending does provide diversification benefits, especially in the form of lower probability of bankruptcy for the lending banks.

The results for the period characterized by U.S. government efforts to dampen foreign lending by U.S. banks seem to suggest that banks tried to continue to maintain or enhance their global endeavors despite certain restrictive measures. It appears that banks tried to remain competitive by providing costly loans to foreign borrowers. In essence, the evidence indicates that the primary beneficiaries of restrictive foreign lending regulations in the U.S. were the foreign borrowers. In a similar vein, the lack of a significant difference in abnormal returns before and after the passage of the ILSA suggests that it has not been an effective regulatory tool in modifying the foreign lending practices of USBHCs.

1 A bank's effort to increase its information-producing capabilities may lead it to open new branches, make capital investments in equipment, acquire other banks, and so on. The exploration of the market reaction to these efforts by banks is beyond the scope of this paper.

2 There are other regulations pertaining to foreign banking. It is not the intent here to identify all of them. Rather, the focus is on identifying a few major ones.

3 The sample includes foreign loans made to foreign governments as well as to companies within the foreign countries. However, the nature of the reported news is such that it was not possible to separate the two.

4 Other pre-event estimation periods of t = -150 to t = -30 and t = -120 to t = -30 were also used. The results were not different from the ones reported here. Similar results were obtained for tests based on post-event estimation periods.

5 See Mikkelson and Partch |22~ for examples of the tests.

6 All of the tests in the study were replicated by using the CRSP equal-weighted index. The results were found not to be significantly different.

7 The direction of change in risk, rather than the magnitude, was also tested. The result was found to be not significantly different from the one reported here, although it was less powerful.

8 Theoretically, the change in unsystematic risk should not affect the value of the bank for fully diversified investors. However, in reality it may affect the value of the bank because higher unsystematic risk means higher total risk and higher resultant bankruptcy probability and expected bankruptcy costs (Aharony, Jones, and Swary |1~).

9 Pre- and post-estimation periods of t = -170 to t = -21, and t = 21 to t = 170, respectively, were used. Other estimation periods of t = -150 to t = -30 and t = 30 to t = 150, and t = -120 to t = -30 and t = 30 to t = 120 were also used. The results remained the same.

10 Aharony, Jones, and Swary |1~ find a significant positive relationship between unsystematic risk and the probability of bankruptcy.

11 On October 26, 1982, the SEC mandated that banks must reveal the size of their exposure to problem foreign loans if they exceed one percent of their loan portfolio. On November 17 and 18, 1983, the Senate and the House, respectively, passed the ILSA bill and on November 30, 1983, President Reagan signed the bill. All announcements of foreign lending agreements for the period starting October 1, 1982, and ending December 31, 1983, were excluded from the sample and a new regression was estimated. The results were similar to those in Regression (4).

References

1. J. Aharony, C.P. Jones, and I. Swary, "An Analysis of Risk and Return Characteristics of Corporate Bankruptcy Using Capital Market Data," Journal of Finance (September 1980), pp. 1001-1016.

2. R.Z. Aliber, "International Banking: A Survey," Journal of Money, Credit, and Banking (November 1984), pp. 661-684.

3. B. Bennett and G. Zimmerman, "U.S. Banks' Exposure to Developing Countries: An Examination of Recent Trends," Federal Reserve Bank of San Francisco Economic Review (Spring 1988), pp. 14-29.

4. R.S. Billingsley and R. Lamy, "The Regulation of International Lending, IMF Support, The Debt Crisis, and Bank Stockholder Wealth," Journal of Banking and Finance (June 1988), pp. 255-274.

5. T. Campbell and W. Kracaw, "Information Production, Market Signaling, and the Theory of Intermediation," Journal of Finance (September 1980), pp. 863-882.

6. B. Cornell, W. Landsman, and A. Shapiro, "The Impact on Bank Stock Prices of Regulatory Responses to the International Debt Crises," Studies in Banking and Finance 3, 1986, pp. 161-178.

7. W.A. Darity, Jr., and B. Horn, The Loan Pushers: The Role of Commercial Banks in the International Debt Crises, Cambridge, MA, Ballinger Publishers, 1988.

8. S.I. Davis, "How Risky is International Lending?," Harvard Business Review (January/February 1977), pp. 135-143.

9. D. Diamond, "Financial Intermediation and Delegated Monitoring," Review of Economic Studies (July 1984), pp. 393-414.

10. D.K. Eiteman and A. Stonehill, Multinational Business Finance, Boston, MA, Addison-Wesley Publishing Company Inc., 1986.

11. E.F. Fama, "What's Different About Banks?," Journal of Monetary Economics (January 1985), pp. 29-39.

12. D. Folkerts-Landau, "The Changing Role of International Bank Lending in Development Finance," International Monetary Fund Staff Papers (June 1985), pp. 323-327.

13. L.S. Goodman, "Can Risks in LDC Lending be Diversified?," Business Economics (March 1982), pp. 12-19.

14. J.M. Gray and H.P. Gray, "The Multinational Bank: A Financial MNC?," Journal of Banking and Finance (March 1981), pp. 33-63.

15. H. Grubel, "Internationally Diversified Portfolios: Welfare Gains and Capital Flows," American Economic Review (December 1968), pp. 1299-1314.

16. F.T. Haner and J. Ewing, Country Risk Assessment: Theory and Worldwide Practice, New York, Praeger Publishers, 1985.

17. A.K. Jain, "International Lending Patterns of U.S. Commercial Banks," Journal of International Business Studies (Fall 1986), pp. 73-88.

18. A.K. Jain and S. Gupta, "Some Evidence on 'Herding' Behavior of U.S. Banks," Journal of Money, Credit, and Banking (February 1987), pp. 78-87.

19. C. James, "Some Evidence on the Uniqueness of Bank Loans," Journal of Financial Economics (December 1987), pp. 217-235.

20. C.P. Kindleberger, Manias, Panics, and Crashes: A History of Financial Crises, New York, Basic Books, 1978.

21. S.L. Lummer and J.J. McConnell, "Further Evidence on the Bank Lending Process and the Capital-Market Response to Bank Loan Agreements," Journal of Financial Economics (November 1989), pp. 99-122.

22. W.H. Mikkelson and M.M. Partch, "Valuation Effects of Security Offerings and the Issuance Process," Journal of Financial Economics (January/February 1986), pp. 31-60.

23. A. Monti, "Recent Trends in International Banking," Journal of Banking and Finance (September 1982), pp. 389-399.

24. R. Roll, "The Hubris Hypothesis of Corporate Takeover," Journal of Business (April 1986), pp. 197-216.

25. A.M. Rugman and S. Kamath, "International Diversification and Multinational Banking," Recent Developments in International Banking and Finance 1, 1987, pp. 35-59.

26. H. White, "A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for Heteroskedasticity," Econometrica (May 1980), pp. 817-838.

The authors thank Marcia Millon Cornett, Dave Davidson, Yoonbai Kim, Mark Flood, Nanda Rangan, Stuart Rosenstein, Tom Schwarz, Andy Szakmary, three anonymous reviewers of this journal, and the journal editor James S. Ang for helpful comments on earlier drafts of the manuscript, and Pauletta Masterson for her assistance in preparing the manuscript.

Amjad Waheed is an Assistant Professor of Finance at East Tennessee State University, Johnson City, Tennessee. Ike Mathur is a Professor of Finance at Southern Illinois University at Carbondale, Carbondale, Illinois.