Department of Economics Working Paper Series

The Performance of Domestic and Foreign Banks: The Case of

Korea and the Asian Financial Crisis
Yongil Jeon
Central Michigan University

Stephen M. Miller
University of Nevada and University of Connecticut

Working Paper 2002-28

May 2002

341 Mansfield Road, Unit 1063

Storrs, CT 06269–1063
Phone: (860) 486–3022
Fax: (860) 486–4463
http://www.econ.uconn.edu/
 Abstract
This paper considers the performance of banks, domestic and foreign, in Korea
prior to, during, and immediately after the Asian financial crisis, examining how
the profitability of those banks differed and identifying factors that explain why
those differences existed. The performance of Korean banks deteriorated dramat-
ically in 1998 with most banks recovering somewhat in 1999. Foreign banks did
not experience the same negative effect on their returns on assets and equity as a
rule. Several standard findings emerge. For example, equity to assets correlates
positively with domestic, but not foreign, bank performance, as measured by the
returns on assets and equity, even when the government recapitalized institutions
that were performing quite badly. Also, foreign-currency deposits significantly
and negatively correlate with domestic Korean bank performance, although only
in the post-crisis period for regional banks. In sum, the domestic Korean banks
suffered more severely from the Asian financial crisis than foreign banks.
Journal of Economic Literature Classification: E44, G21, O16, O53
Keywords: Asian financial crisis, Korean commercial banks, profitability
I. Introduction:

The world’s economy experienced several financial crises in the last decade – for example, the

1994 Mexican peso collapse and the resulting turmoil in Latin America, and the 1997-98 Asian

crisis.1 The devaluation of the Thai baht initiated the Asian financial crisis on July 2, 1997 that

quickly spread to a number of Asian countries. The effects of the crisis on countries differed with

some countries experiencing severe economic contractions while others saw modest effects on

their domestic economies.2

Our paper focuses on the performance of nationwide (national) and regional Korean

banks as well as foreign banks (bank offices) prior to, during, and immediately after the Asian

financial crisis.3 We examine how the profitability of these banks differed and identify factors

that explain why these differences existed. Our paper also adds significant value in two areas.

First, we assemble probably the best panel data set on Korean domestic – nationwide and

regional – and foreign banks during the mid and late 1990s. Second, we capitalize on that panel

data structure and employ the fixed-effect regression technique.

Various analysts suggest that foreign bank lending played a unique role in the Asian

financial crisis vis-à-vis other similar events (Cho and Hong 2001, Kaminsky and Reinhart 2001,

and Tornell 2001). Domestic banks supplied major quantities of credit to domestic firms.

Domestic banks also came to rely more heavily on foreign bank lending. When the crisis reared

1
Ito and Krueger (2001) recently edited a series of papers that examine the causes and consequences of the Asian
financial crisis with comparisons to similar prior crises.
2
Corsetti, Pesenti, and Roubini (2001), for example, argue that Hong Kong, Singapore, and Taiwan did not suffer
as much as other countries because of trade and current account surpluses, significant holdings of foreign exchange
reserves, and the relative absence of “crony capitalism.” Other discussions of financial crises include Kaminsky and
Reinhart (2001), Tornell (2001), and Hahm and Mishkin (2000)
3
Jeon and Miller (2002) provide an analysis of Korean nationwide banks with a slightly longer time frame and a
2
its ugly head, the supply of foreign lending evaporated quickly, confronting the domestic banks

with a liquidity crisis. Moreover, some commentators indict the initial International Monetary

Fund (IMF) rescue programs as worsening the liquidity crisis by requiring tighter credit (Radelet

and Sachs 1998).

Noland (2000) differentiates the Korean crisis from other Southeast Asian crises, since

the Korean investment boom occurred in the manufacturing sector, especially the chaebols,

rather than in real estate and since investment growth was funded in large part by short-run

capital inflows. That is, short-term capital controls were liberalized while the long-term controls

were not.4 In short, the financial crisis caused some important corporate borrowers to default on

their loans to banks. That negative shock was reinforced and compounded by the loss of foreign

lending to domestic banks. Impending bank failures necessitated the intervention by the central

bank to assist in finding merger partners (possibly foreign) or to take over operations of the

failed banks itself.

Our analysis supports several conclusions. The performance of Korean banks

deteriorated dramatically in 1998. Most banks recovered somewhat in 1999. Foreign banks did

not experience the same negative effect on their returns on assets and equity as a rule. Several

standard findings emerge. First, equity to assets correlates positively with domestic, but not

foreign, bank performance, even when the government recapitalized a number of institutions that

were performing quite badly. Second, provisions for loan losses generally correlates negatively

richer set of balance sheet and income statement data.

4
Noland also argues that the initial IMF program exacerbated problems by confusing the Asian financial crisis with
the earlier Latin American crisis. The Asian crisis differed according to Noland, because the corporate expansion
was loan, not equity, based. Thus, the financial crisis raised interest rates, triggering a liquidity crisis. The IMF’s
prescription to tighten credit worsened the liquidity crisis.
3
with bank performance. Finally, decreases in foreign-currency deposits significantly and

negatively correlate with domestic Korean bank performance as measured by the returns on

assets and equity.

II. Foreign Banks: A Blessing or A Curse in Domestic Financial Markets

The Asian financial crisis underscores the importance of strong, stable financial markets for the

maintenance of economic development, since weak, unstable financial markets can push an

economy to its knees. In this regard, some analysts argue that foreign bank participation in

domestic financial markets strengthens the domestic economy. Other analysts argue that the

financial service industry possesses public good characteristics and that the unfettered private

interests (markets), especially interests with foreign connections, should not control credit

allocation decisions. That view, at a minimum, implies that foreign banks should not operate in

the domestic economy. An even more stringent view argues that state ownership and state-

mandated credit allocation needs to send credit to those sectors most crucial for economic

development.

Korea transversed the spectrum from a system with large elements of state ownership and

state-directed credit flows to a more open and competitive financial markets with a significant

presence of foreign banks. And a large privatization of state-owned banks has occurred since

Korea decided to travel down the path of deregulation.

Proponents of foreign bank entry make several arguments (e.g., Claessens, Demirgüç-

Kunt, and Huizinga 2001, Demirgüç-Kunt, Levin, and Min 1998, Goldberg, Dages, and Kinney

2000, and Levin 1996). First, foreign banks provide the channel through which capital inflows

finance domestic activities. To the extent that such funding adds to, rather than substitutes for,

4
domestic funding, it will stimulate the domestic economy, at least one that experiences a

shortage of available funding from domestic sources. Second, the increased competition among

banks – domestic and foreign –- will improve the performance of banks and provide financial

services at a lower average cost. Finally, the experience of foreign banks in their home country

may lead to better regulation and supervision in the foreign markets where they do business.

Opponents of foreign bank entry also make several arguments (e.g., Claessens,

Demirgüç-Kunt, and Huizinga 2001, and Demirgüç-Kunt, Levin, and Min 1998). First, unlike

the optimists, the pessimists see the capital flow channel as a path for capital flight when the

“going gets tough”. That is, the pessimists place a much higher weight on the negative

consequences of capital outflows during bad times than on the positive consequences of capital

inflows during good times. Second, foreign banks may have a competitive advantage that allows

them to “cherry” pick among the available domestic funding options, choosing the more-

profitable, low-risk options and leaving the less-profitable, higher-risk options for domestic

institutions. Third, foreign banks from developed countries may introduce complexities not seen

by domestic regulators and supervisors, worsening, rather than improving, the regulatory and

supervisory process.

The movement in recent decades toward more-open financial markets and the increased

activity of foreign banks in domestic financial markets suggests that the proponents have

currently won the day. The Asian financial crisis raises the issue of the role, if any, of foreign

banks in creating or continuing the crisis.

Claessens, Demirgüç-Kunt, and Huizinga (2001) note a conundrum in understanding

foreign bank operations. To wit, foreign banks that operate in developing countries generally

5
achieve higher profitability than domestic banks;5 the opposite occurs in developed countries.

They first articulate the popular explanation: banks from developed countries follow domestic

customers into foreign markets, even when the profitability is lower.6 They then offer several

rationalizations for the differences between the profitability of foreign banks in developed and

developing countries. First, low net-interest margins in developed countries may reflect

participation in wholesale, rather than retail, markets with smaller net-interest margins. Second,

the technical advantages for foreign banks in developed countries may be too small to cover

informational disadvantages. Of course, those two explanations of low net-interest margins may

reverse themselves in developing countries. That is, foreign banks may enter retail markets more

fully in developing countries and/or they may possess higher levels of technical efficiency that

overcomes any informational disadvantages.7

Claessens, Demirgüç-Kunt, and Huizinga (2001) examine foreign bank operations in 80

countries. They conclude that foreign banks experience lower net-interest margins, overhead

expenses, and profitabilities than domestic banks in developed countries.8 The opposite ranking

emerges in developing countries. They also consider how the operation of foreign banks may

affect the performance of domestic banks as well as how the performance of domestic banks may

5
The Korean experience matches that observation over the 1994 to 1999 period. See Table 1 below.
6
Nolle and Seth (1996), however, examining the experience of U.S. banks operating in 6 developed countries
(Canada, France, Germany, Japan, the Netherlands, and the U.K.), conclude that “ … the ‘follow the customer’
hypothesis may have more limited applicability … than previously supposed.” (p. 2). DeYoung and Nolle (1996)
argue that lower profitability of foreign banks in the U.S. reflects a preference for growth over profits.
7
On this point, see our discussion of Berger, DeYoung, Genay, and Udell (2001) at the end of this section.
8
Terrell (1986) reports similar findings.
6
attract foreign banks. They conclude that the expansion of the foreign bank presence associates

with a lower profitability and a higher provisioning for bad loans by domestic banks.9

Berger, DeYoung, Genay, and Udell (2001) explore this conundrum from a slightly

different point of view. They consider two alternative hypotheses that can explain differences in

foreign and domestic bank performance – the home-field- and global-advantage hypotheses. The

home-field-advantage hypothesis argues that domestic banks generally outperform foreign banks

because of informational and cost advantages. The global-advantage hypothesis argues that

banks from some countries possess sufficient efficiency gains, allowing them to overcome any

home-field advantages accruing to domestic banks. They examine the X-efficiency of domestic

and foreign banks within the borders of a given country. They calculate X-efficiency in five

developed (OECD) countries – France, Germany, Spain, the U.K., and the U.S. Moreover, the

foreign banks also come from developed (OECD) countries, except for Korea. They conclude

that domestic banks exhibit higher cost- and profit-efficiencies than foreign banks, supporting

the home-field-advantage hypothesis. Decomposing the foreign banks by country of origin

provides some support for the global-advantage hypothesis. That is, foreign banks from the U.S.

generally exceed the cost- and profit-efficiencies of domestic banks.10

The empirical observation that foreign banks perform better than domestic banks in

developing countries implies that the technical savvy of banks from developed countries

9
Amel and Liang (1997) describe similar results.
10
Demirgüç-Kunt, Levin, and Min (1998) employ the same data as Claessens, Demirgüç-Kunt, and Huizinga
(2001) and draw three conclusions concerning the effects of foreign bank operations on the domestic economy.
Greater participation by foreign banks tends to (1) reduce the probability of a banking crisis, (2) improve the
efficiency of domestic banks, and (3) boost indirectly economic growth by improving domestic bank efficiency.
Further, the effects of foreign bank operations relate to the number of foreign banks and not the size of their
operations.
7
generally overcomes the home-field-advantage in developing countries, especially when the

domestic economy has relatively unsophisticated financial markets and institutions. Of course,

that conjecture only makes sense in those countries where financial markets allow the entry of

foreign banks.

III. Data Sources and Descriptive Analysis

Our balance sheet and income statement data for nationwide and regional Korean banks and

foreign banks come from Bank Management Statistics, published by the Financial Supervisory

Services (1999, 2000). Sixteen nationwide and 10 regional banks and 59 foreign banks enter our

database for at least one year in the sample from 1994 through 1999.11 Several bank entrances,

mergers, acquisitions, and conversions occurred over the sample period, as the Asian financial

crisis threw a roadblock across the path of deregulation and privatization of the financial sector

begun by the Korean government and the Bank of Korea in the early 1980s.12

During the 1960s and 1970s, major components of the Korean financial system were

nationalized. Lending was targeted toward favored sectors (and firms), such as exports and

heavy industries (Bank of Korea 1994). Moreover, regional banks, which could operate only in

their own provinces (and a branch in Seoul), entered the scene in 1967 to encourage regionally

based development.13 Plans to deregulate the financial system and place Korean commercial

banks in the private sector began in the early 1980s.

11
See the appendix for the banks and country of origin in the sample.
12
Bank of Korea (1994) and Gilbert and Wilson (1998) provide information on the Korean financial system.
13
Over 1980 to 1994, Gilbert and Wilson (1998) calculate that nationwide banks experienced significant, large
productivity improvement while regional banks experienced mixed results.
8
 Deregulation in the early 1980s expanded the power of commercial banks, who could

now, for example, offer credit cards, issue negotiable certificates of deposit, provide automated

teller machines, and so on (Gilbert and Wilson 1998). Simultaneously, foreign exchange controls

and restrictions on foreign ownership of Korean assets eased. The government’s hand still

wielded, nonetheless, a potent force, controlling interest rates on certain types of loans and

deposits. Further, the government’s informal credit policy continued to favor selected sectors.

Gilbert and Wilson (1998) argue that the Korean commercial banking system

experienced a crisis in the mid-1980s with significant levels of bad loans. No Korean bank failed

at this time, however, as charge-off rates for bad loans were allocated slowly enough to maintain

individual bank viability. No such luck (skill) graced the Korean commercial banking industry

during the Asian financial crisis.

While the Asian financial crisis produced the dramatic domestic economic crisis in

Korea, more fundamental causes also added to its severity.14 The corporate sector overextended

itself with too much investment and borrowing. Commercial banks overused short-term foreign

lending as a source of funds. Finally, the lack of transparency of balance sheets, income

statements, and management practices all led to a crisis of confidence in Korean institutions. In

sum, the Korean economy was an “accident waiting to happen.”

Korea First and Seoul became insolvent during the Asian financial crisis. They were

judged as “too-big-to-fail” institutions. Thus, the government nationalized and recapitalized

them in January 1998. The Bank of Korea sought private (foreign) buyers for both banks after

14
The next few paragraphs rely on information from Bank of Korea (1998).
9
recapitalization. After protracted negotiations, Newbridge Capital acquired Korea First in

December 1999. Seoul still sits on the auction block.

Having determined that Korea First and Seoul were too-big-to-fail, the Monetary Board

of the Bank of Korea in February 1998 identified 12 of the remaining 24 Korean banks as falling

below the Bank of International Settlements (BIS) capital adequacy requirement of 8 percent.

After examining the financial conditions of those twelve banks, the Financial Supervisory

Commission ordered the closure of 5 Korean banks – 3 nationwide and 2 regional – since they

were judged to have little chance of recovering. Those banks were closed through purchase and

assumptions (P&As) where the acquiring banks assumed the liabilities and purchased only the

“sound” assets.

Chase Manhattan entered the Korean economy as the first foreign bank in 1967.15 The

participation of foreign banks grew at a good pace through out the 1970s and 1980s, but

stabilized in the 1990s and then fell somewhat after the Asian financial crisis. Foreign banks

came to Korea during the 1970s and 1980s, partly because they received more favorable

treatment in certain areas than domestic banks. In the mid-1980s, regulatory change began eating

away at the preferential treatment of foreign banks. But along with the elimination of preferential

treatment in some areas, other regulatory changes reduced barriers and restrictions on foreign

bank activities. Thus, the playing field was basically leveled between foreign and domestic

banks.

Demirgüç-Kunt, Levin, and Min (1998) argue that the foreign bank presence in Korea

provided positive, competitive pressure on domestic banks through the end of their sample in

15
The discussion in this paragraph relies on Demirgüç-Kunt, Levin, and Min (1998).
10
1996, just prior to the Asian financial crisis. They attribute the better foreign bank performance

to more economical labor use and better underwriting of loans. That is, foreign banks possessed

much higher ratios of profit to employee, loans to employee, and expenses to employee and

much lower ratio of non-performing loans to loans than domestic banks (Demirgüç-Kunt, Levin,

and Min, 1998, Table 10, p. 102).

From 1994 through 1999, 59 foreign banks operated in Korea – some for the full sample

period, others for only parts. The 59 banks include 14 each from the U.S. and Japan, 6 from

France, 4 each from Canada and Singapore, 3 from the U.K., and 2 each from Australia, China,

the Netherlands, and Switzerland.16

Our database includes information on the asset and liability holdings and income and

expense information of Korean and foreign banks. Before performing more rigorous analysis of

the database, we first provide an overview discussion of a number of key variables related to

bank performance over the 1994 to 1999 period.

Table 1 lists the average returns on assets and equity for foreign, nationwide, and

regional banks. The foreign banks consistently experience higher average returns on assets and

equity, consistent with half of the conundrum noted by Claessens, Demirgüç-Kunt, and Huizinga

(2001). Moreover, foreign banks average returns remain positive through the Asian financial

crisis, unlike Korean banks – nationwide and regional. Regional banks first lead nationwide

banks in average returns on assets and equity while returns remain positive. Once returns flip

negative, regional banks sustain a larger hit in 1998 and become more negative than nationwide

16
The following countries each had one bank represented – Germany, Hong Kong, India, Jordan, and Pakistan.
11
banks. Regional banks recover more quickly than nationwide banks and exhibit higher returns,

although still negative, than nationwide banks in 1999.

Table 2 enumerates the number of foreign, nationwide, and regional banks over the

sample period. Foreign banks outnumber Korean banks – nationwide plus regional – by about 2

to 1. Bank closures in 1998 and 1999, however, push this ratio to nearly 2.5 to 1. Japanese and

U.S. banks surpass by a large margin the presence of banks from other countries. In 1994, 14

Japanese banks operated in Korea. As just noted, this number falls to 7 by 1999 with all but one

of the exits occurring after the Asian financial crisis. In 1994, 11 U.S. banks operated in Korea.

After 3 entrances and 3 exits, 11 U.S. banks remain in business in 1999. France and Singapore

each held steady with 6 and 4 banks, respectively, in all sample years.

Superiority in numbers does not translate into dominance in the markets, since Korean

banks uniformly exceed, on average, the foreign banks in asset size. Table 3 reports the average

assets held by foreign, nationwide, and regional banks. Nationwide banks possess about 4 times

the assets of regional banks and about 35 times the assets of foreign banks. Those ratios rise to

about 6 and 45 by 1999. Foreign banks achieved the highest average size in assets in 1998. The

top 15 foreign banks in 1998 in asset size include 2 U.S. banks, 5 Japanese banks, 4 French

banks, and one bank each from Germany, Hong Kong, the Netherlands, and Switzerland.

Table 3 also suggests that Korean banks did respond to the shocks from the Asian

financial crisis to the extent that they could. In general, loans decreased as a percent of assets in

both 1998 and 1999 as banks tried to reduce the income risk that they faced. Moreover, deposits

increased as a percent of assets in 1998 and 1999 as banks tried to reduce the expense risk that

they faced.

12
 Researchers suggest that foreign lending to domestic banks played an important role in

the Asian financial crisis (Radelet and Sachs 1998, Cho and Hong 2001, Kaminsky and Reinhart

2001, and Tornell 2001). That is, the Asian crisis precipitated a loss of foreign-source liabilities,

exerting strong pressure on those banks with an illiquid asset base. If accurate, then we expect to

see retrenchment in bank portfolios – declining assets and/or deposits. We do not observe such

movements by in large for the Korean or foreign banks.

Total assets climbed continually from 1994 through 1999 for Korean nationwide and

regional banks, and fell only slightly in 1999 for foreign banks (see Table 3). Deposits also rose

steadily over the entire 1994 to 1999 period (see Table 4). In short, the consolidated balance

sheet information does not provide much ammunition for the hypothesis that the withdrawal of

foreign-source liabilities played a significant role in the Korean economic woes after the Asian

financial crisis. If foreign-source liabilities were withdrawn from Korean nationwide banks, then

such loses were more than replaced from domestic sources.

To offer a related insight to that last observation, Tables 4 and 5 also report information

on foreign-currency deposits and foreign-currency loans, respectively. Both items, measured in

Won, decrease after the Asian financial crisis in absolute terms (not shown) and as a percent of

loans and discounts and deposits, respectively. Unfortunately, we do not know whether the lost

loans and deposits were domestic or foreign residents, since that data are not available.

The holding of foreign-currency loans and foreign-currency deposits do expose banks to

foreign exchange risk as long as foreign-currency loans exceed, or fall short of, foreign-currency

deposits. For example, if foreign-currency loans exceed foreign-currency deposits, then a

weakening Won increases the Won value of foreign-currency loans more than deposits, adding

13
to the equity base. Of course, a strengthening Won squeezes the equity base.17 We note that

foreign-currency loans and deposits rise and fall together with those loans exceeding deposits in

every year except 1997 for only nationwide banks (not shown).

Finally, Tables 4 and 5 also offer some evidence on the effects, if any, of foreign-

currency deposits and foreign-currency loans on bank performance. Nationwide banks rely more

heavily on foreign-currency deposits and foreign-currency loans than regional banks.18 Both

nationwide and regional banks did reduce their dependency on foreign-currency deposits and

foreign-currency loans in 1999. Regional bank performance exceeded nationwide bank

performance over 1994 to 1996, deteriorated more significantly during the 1997 and 1998

period, and then recovered more briskly in 1999. In addition, foreign banks by 1999 relied more

heavily on both foreign-currency deposits and foreign-currency loans than Korean nationwide or

regional banks. Also, foreign bank performance dominated Korean nationwide and regional

banks throughout the sample period. Thus, it is improbable that the holding of foreign currency

deposits or foreign currency loans per se systematically contributed to poor bank performance

across all bank types. If that were the case, then Korean regional bank performance should not

have suffered so significantly and foreign bank performance should have suffered more severely.

IV. Explaining Bank Performance in Korea

Our data sample includes all nationwide and regional banks and all foreign banks in operation in

any year from 1994 to 1999. Since some banks entered and/or exited over the sample period, we

have an unbalanced panel data set of 441 observations – a panel of 510 observations with 69

17
The database does not provide the breakdown of foreign-currency loans and deposits into individual currencies.
18
Negotiable certificates of deposit include both won- and foreign currency-denominated accounts. Total loans and
discounts do not equal the sum of won- and foreign-currency-denominated loans, since total loans and discounts
14
missing values. As noted above, the data include balance sheet and income statement data on

these banks. In addition, we collected some macroeconomic information that change over time,

but do not differ between banks at a point in time.

Our econometric analysis considers possible correlations between the balance sheet and

income statement data as well as the macroeconomic data and our measures of bank performance

returns on assets and equity. We produce two sets of regressions for the foreign, nationwide, and

regional banks. In each set, the first regression considers three different types of individual bank

explanatory variables: (1) portfolio distribution variables – won loans to assets, foreign currency

loans to assets, won deposits to assets, negotiable certificates of deposit to assets, foreign

currency deposits to assets, and equity to assets; (2) a risk variable – provision for loan losses to

loans; and (3) a scale variable – total assets. The second regression broadens the analysis to

include macroeconomic variables -- the unemployment rate, the rate of growth of real gross

domestic product, the rate of depreciation of the Won, the fiscal budget surplus as a fraction of

nominal gross domestic product, and the rate of inflation in the consumer price index.

We anticipate that interest-earning assets – Won loans to assets and foreign currency

loans to assets --- will garner positive effects on bank profitability while interest-earning

liabilities – Won deposits to assets, negotiable certificates of deposits to assets, and foreign

currency deposits to assets – will garner negative effects. We expect that equity to assets will

possess a positive effect, which reflects complementary factors. Higher equity to assets implies

lower interest earning liabilities to assets. So net income, which includes any dividend payments,

should increase as equity to assets rises, other things constant. Higher equity to assets also may

include domestic import usance bills and advances to customers.

15
imply a well-run bank operation with fewer problems. And a well-run bank may imply higher

net income. We foresee a negative effect of our risk variable, provision for loan losses to loans,

on bank performance. Since we predict that poor macroeconomic performance will lead to poor

bank performance, then negative effects on bank performance should associate with higher

unemployment, a faster depreciation in the exchange rate, and a higher inflation rate. Further, a

higher growth rate of real gross domestic product should associate with better bank performance.

Finally, we do not have prior expectations about the effects of bank size of a higher fiscal surplus

as a fraction of nominal gross domestic product on bank performance.

The standard method in empirical bank studies estimates regression equations with

ordinary least squares (OLS), which assumes that the omitted variables are independent of the

regressors and are independently, identically distributed. Such estimation, however, can create

problems of interpretation if bank-specific characteristics, such as bank management, that affect

performance are not considered. If those omitted bank-specific variables (both observed and

unobserved) correlate with the explanatory variables, then OLS produces biased and inconsistent

coefficient estimates (see Hsiao, 1986). Using panel data, however, the fixed-effect model

produces unbiased and consistent estimates of the coefficients.19

The fixed-effect model assumes that differences across banks reflect parametric shifts in

the regression equation. Such an interpretation becomes more appropriate when the problem at

hand uses the whole population, rather than a sample from it. Since our sample considers all

19
Another method of excluding unobserved country-specific variables estimates the first-differenced regression (see
Hsiao 1986, and Westbrook and Tybout 1993). Also see Woolridge (2000, p. 447) for the choice between using
fixed effects and first differences. Another popular method is the random-effects model, which assumes that
individual specific constant terms are randomly distributed across cross-sectional units. See Greene (2000, Ch. 14)
for details.
16
domestic and foreign banks over a particular time period (i.e., 16 nationwide, 10 regional, and 59

foreign banks), we adopt the fixed-effect model for our analysis.

Before reporting the results of our regression analysis, some background discussion on

the sequence of events in our research will provide useful information. Jeon and Miller (2002)

performed similar analysis on a richer data set for nationwide banks from 1991 to 1999.20 They

found that including the 1999 data made the regression models less precise. Moreover, they

discovered that Seoul appeared as an outlier.

Such large changes in results from adding data from 1999, with hindsight, seems a

probable outcome, since the post-Asian-financial-crisis data are much noisier. The government’s

hand in nationalization of several institutions and in recapitalizing many others likely altered

normal relationships. A quick look at Table 1 suggests that amongst the nationwide banks, Seoul

appears to have followed a different path in 1999. Seoul’s returns on assets and equity

deteriorate further in 1999 when other banks experience some relief from the difficulties in 1998.

Moreover, while the government finally did find a foreign purchaser for Korea First, Seoul

remains at the altar awaiting a proper suitor. Jeon and Miller (2002) deleted the 1999 Seoul

observation from their data set to improve the performance of the regression estimates. Thus, we

also delete the 1999 Seoul observation, converting our data set to an unbalanced panel of 510

with 70 missing values. And lastly, before examining our econometric results, we allow for

differentiation between the effects of independent variables before and after the Asian financial

crisis. To do so, we construct an Asian-financial-crisis dummy variable (equal zero for 1994 to

20
The data provided on foreign banks required that we limit the variables considered in our econometric analysis.
That is, the portfolio distribution and income and expense variables provide much less information in our current
data set, although we are considering a much larger number of banks.
17
1996; one otherwise) and interact it with each independent variable. We drop all interaction

terms for which the coefficients do not significantly differ from zero at the 20-percent level and

re-estimate.

Table 6 and 7 report the regression results for the returns on assets and equity,

respectively, excluding the 1999 Seoul observation and omitting the interaction terms with

insignificant coefficients. Several observations deserve mention. First, F-tests determine whether

foreign and domestic banks and then nationwide and regional banks experience different

relationships. The foreign and domestic bank regressions differ significantly from each other at

the 1-percent level in every case – returns on assets and equity both with and without

macroeconomic controls. The nationwide and regional bank regressions also differ significantly

from each other at the 1-percent level, except for the return on asset regressions without

macroeconomic controls.21 In sum, we report separate regression results for nationwide, regional,

and foreign banks. Within each bank subgroup, however, we employ panel data estimation, using

the fixed-effect technique.

Second, the adjusted R2 for the regional-bank regressions suggest a reasonably good fit.

The fit for the nationwide-bank regressions falls somewhat compared to the regional-bank

regressions. The fit for the foreign-bank regressions leaves much unexplained variability.22

Third, higher capital adequacy (equity to assets) associates positively with both the rates

of return (return on assets and equity) for domestic (nationwide and regional) banks, but not for

21
Here, the F-test is not significant even at the 20-percent level.
22
The home-country variables for the foreign banks may explain movements in foreign bank performance. To
pursue such issues goes beyond the intent of the current paper. Jeon, Natke, and Miller (2002) address those issues.
18
foreign banks.23 While that effect occurs for both the pre- and post-crisis periods for the return on

assets specification, it only appears in general in the post-crisis period for the return on equity

specification. What does that difference imply? When equity to assets increases, that may reflect

higher equity, lower assets, of both. Thus, the return on assets more easily rises with an increase

in equity to assets, than can return on equity. Nonetheless, we still find that positive correlation

between return on equity and equity to assets during the post-crisis period, implying that the size

of equity to assets provides a strong signal. Moreover, the positive effect of equity to assets on

the returns on assets or equity for domestic banks emerges even as the government recapitalized

several domestic banks and arranged for the merger of many others. That is, banks with

significant financial problems receive an injection of new equity that should raise the equity to

asset ratio. Such recapitalizations presumably impart a negative correlation. Thus, the highly

significant and positive association between capital adequacy and rates of return must offset this

government-induced negative association.

Fourth, provisions for loan losses to total assets has a strong negative correlation with the

returns on assets and equity across all three types of banks except for the regional banks in the

rate of return on equity equations. The provision for loan losses crudely signals the riskiness of

banks. Thus, higher loan loss provisions signal higher risk and associates negatively with bank

returns.

Fifth, the ratio of foreign-currency deposits to total assets displays a significantly

negative effect on the returns on assets and equity for both nationwide and regional banks. For

23
The fact that capital adequacy does not significantly affect foreign bank performance probably reflects the fact
that the equity positions of foreign banks comprise a small part of the larger equity positions of the international
banks to which they belong.
19
nationwide banks, that effect occurs in both the pre- and post-crisis periods; for regional banks, it

occurs only in the post-crisis period. The foreign banks did not experience that significant effect.

Jeon and Miller (2002) do not find that conclusion in their study of 16 nationwide banks. They

considered a longer time period – 1991 to 1999. Differences in sample length may explain the

difference in findings as to foreign currency deposits. That is, the Asian financial crisis began in

mid-1997. As such, the sample from 1994 to 1999 has a larger percentage of sample years

coming from after the crisis. Less influence is potentially wielded by the post-Asian financial

crisis years in the 1991 to 1999 period. In additions, Jeon and Miller (2002) use end-of-year

(last-quarter) data. Here, we employ quarterly averages. Thus, the well-known “window-

dressing” effect may explain the findings in Jeon and Miller (2002).24

Sixth, the ratio of won-denominated loans to assets possesses a positive correlation with

the rates of return on assets and equity for foreign banks, but not for domestic national banks.

Moreover, that effect seems isolated to the post-crisis period. Nationwide banks possess a

positive correlation between foreign currency loans and the rates of return on assets and equity.

Regional banks possess a significant positive correlation only for the return on equity.

Finally, the macroeconomic variables exert the most effect on the foreign and regional

banks and the least effect on the nationwide banks. A rising unemployment rate and an increase

in surplus to GDP ratio each associate with a lower rate of return on assets and equity for

regional banks, but not for nationwide or foreign banks. The inflation rate and exchange rate

depreciation negatively affect the rates of return on assets and equity for foreign banks, but

generally do not affect the rates of return on assets and equity for domestic banks.

24
“Window dressing” refers to the accounting “tricks” used at the close of the year to make an institution’s
20
 In sum, the empirical results accord with our prior expectations with a few exceptions.

All significant financial and macroeconomic variables possess the anticipated correlation with

bank performance for Korean banks. A few more unexpected significant correlations emerge in

the foreign bank regressions. Moreover, unexpected negative effects of foreign currency loans to

assets, equity to assets, and inflation on bank performance appears in the foreign bank

regressions as well as an unexpected positive effect of the unemployment rate.

V. Conclusion

The Asian financial crisis is but one of several recent crises that hit the world’s economies.

Analysts suggest that it differs from prior crises in the importance of foreign lending. That is,

recent capital flows into many Asian countries in response to the Asian miracle quickly exited

once the crisis emerged. The sharp loss of lending quickly plunged financial institutions and

corporations into a liquidity crisis. In addition, some analysts (e.g., Radelet and Sachs 1998, and

Noland (2000) cite the initial IMF rescue programs that required credit tightening as contributing

to the severity of the problems. The problems in Korea mirrored many of the problems

confronted by other countries. The financial institutions needed recapitalization and

restructuring. Furthermore, a number of important chaebols faced imminent default on their

obligations.

If the exodus of capital precipitated a liquidity crisis in Korean banking, the data should

reveal dramatic retrenchment in assets and deposits. We find no such evidence. Thus, although

the data do not allow us to disentangle foreign-owned assets and liabilities from their domestic

counterparts, any loss of foreign-owned liabilities that may have occurred was more than offset

performance appear better than it actually is.

21
by an expansion in domestically owned liabilities, since total assets continued to increase before,

during, and after the Asian financial crisis.

This also paper considers the performance of banks in Korea – nationwide, regional, and

foreign -- before, during, and immediately after the Asian financial crisis that is dated in July

1997. The two largest banks – Korea First and Seoul – were judged as “too-big-to-fail.” The

government sought foreign buyers for those banks. It took nearly two years to reach agreement

with Newbridge Capital to acquire Korea First. As of this writing, Seoul was still on the auction

block. Government assistance was given to a number of banks to facilitate an acquisition by

other Korean banks.

The performance of Korean banks took a big hit in 1998. Most banks recovered

somewhat in 1999 with the notable exception of the further deterioration of Seoul. The foreign

banks did not experience the same negative effect on their returns on assets and equity as a rule.

Several other factors also possess a strong correlation with bank performance as

measured by the returns on assets and equity. Equity to assets correlates positively with

domestic, but not foreign, bank performance, even when the government recapitalized a number

of institutions that performed quite badly. Decreases in foreign-currency deposits associate

significantly and negatively with domestic Korean bank performance. Provisions for loan losses

generally correlates negatively with bank performance as one expects.

22
 In sum, the Korean economy and financial sector has so far weathered a huge financial

storm. But the oceans are not yet safe; the storm continues. Much progress has occurred in

restructuring the financial sector. Less progress has occurred in the restructuring of the chaebols.

This story has not yet seen its last chapter.

References:

Amel, D. F., J. N. Liang, 1997. “Determinants of Entry and Profits in Local Banking Markets.”
Review of Industrial Organization 12, 59-78.

Bank of Korea, 1994. Financial Liberalization and Internationalization in Korea. The Bank of
Korea, Seoul.

Bank of Korea, 1998. Bank Restructuring in Korea. The Bank of Korea, Seoul.

Berger, A. N., R. DeYoung, H. Genay, and G. F. Udell, 2001. “Globalization of Financial

Institutions: Evidence from Cross-Border Banking Performance.” In R. E. Litan and A.
Santomero (Eds.), Brookings-Wharton Papers of Financial Services, 3.

Cho, D., and K. Hong, 2001. “Currency Crisis of Korea: Internal Weakness or External
Interdependence?” In Regional and Global Capital Flows: Macroeconomic Causes and
Consequences, eds. T. Ito and A. O Krueger, East Asia Seminar on Economics, Volume
10.

Claessens, S., A. Demirgüç-Kunt, and H. Huizinga, 2001. “How Does Foreign Entry Affect the
Domestic Banking Market?” Journal of Banking and Finance 25, 891-911.

Corsetti, G., P. Pesenti, and N. Roubini, 2001. “Fundamental Determinants of the Asian
Financial Crisis: The Role of Financial Fragility and External Imbalances.” In Regional
and Global Capital Flows: Macroeconomic Causes and Consequences, eds. T. Ito and A.
O Krueger, East Asia Seminar on Economics, Volume 10.

Demirgüç-Kunt, A., R. Levin, and H.-G. Min, 1998. “Opening to Foreign Banks: Issues of
Stability, Efficiency, and Growth.” In The Implications of Globalization of World
Financial Markets, Conference Proceedings, The Bank of Korea, Seoul.

DeYoung, R., and D. E. Nolle, 1996. “Foreign-Owned banks in the United States: Earning
Market Share or Buying It.” Journal of Money, Credit and Banking 28, 622-636.

23
Financial Supervisory Services, 1999, Bank Management Statistics, Seoul.

Financial Supervisory Services, 2000, Bank Management Statistics, Seoul.

Gilbert, R. A., and P. W. Wilson, 1998. “Effects of Deregulation on the Productivity of Korean
Banks.” Journal of Economics and Business 50, 133-155.

Greene, W. H., 2000. Econometric Analysis, Fourth edition. Prentice-Hall, Inc., Upper Saddle
River, NJ.

Goldberg, L., B. G. Dages, and D. Kinney, 2000. “Foreign and Domestic Bank Participation in
Emerging Markets: Lessons from Mexico and Argentina.” Economic Policy Review,
Federal Reserve Bank of New York, Volume 6, 17-36.

Hahm, J. H., and F. S. Mishkin, 2000. “Causes of the Korean Financial Crisis: Lessons for
Policy.” National Bureau of Economic Research Working Paper #7483.

Hsiao, C., 1986. Analysis of Panel Data. Cambridge, UK: Cambridge University Press.

Ito, T., and A. O. Krueger, 2001. Regional and Global Capital Flows: Macroeconomic Causes
and Consequences, East Asia Seminar on Economics, Volume 10.

Jeon, Y., and S. M. Miller, 2002. “The Effect of the Asian Financial Crisis on the Performance
of Korean Nationwide Banks.” Working Paper, University of Nevada, Las Vegas.

Jeon, Y., P. A. Natke, and S. M. Miller, 2002. “Were foreign bank operations a stabilizing
influence in Korea?” Working Paper, Central Michigan University.

Kaminsky, G. L., and C. M. Reinhart, 2001. “Bank Lending and Contagion: Evidence from the
Asian Crisis.” In Regional and Global Capital Flows: Macroeconomic Causes and
Consequences, eds. T. Ito and A. O. Krueger, East Asia Seminar on Economics, Volume
10.

Levin, R., 1996. “Foreign Banks, Financial Development, and Economic Growth.” In C. E.
Barfield (Ed.) International Financial Markets, American Enterprise Institute Press,
Washington, DC.

Noland, M., 2000. Avoiding the Apocalypse: The Future of the Two Koreas. Institute for
International Economics, Washington, DC.

Nolle, D. E., and R. Seth, 1996. “Do Banks Follow Their Customers Abroad?” Research Paper
#9620, Federal Reserve Bank of New York, NY.

24
Radelet, S., and J. Sachs, 1998. “The Onset of the East Asian Financial Crisis.” National Bureau
of Economic Research, Working Paper #6680.

Terrell, H. S., 1986. “The Role of Foreign Banks in Domestic Banking Markets.” In H. Cheng
(Ed.) Financial Policy and Reform in Pacific-Rim Countries, Lexington Books,
Lexington, MA.

Tornell, A, 2001. “Lending Booms and Currency Crises: Empirical Link.” In Regional and
Global Capital Flows: Macroeconomic Causes and Consequences, eds. T. Ito and A. O
Krueger, East Asia Seminar on Economics, Volume 10.

Westbrook, M. D., and J. R. Tybout, 1993. “Estimating Returns to Scale with Large, Imperfect
Panels: An Application to Chilean Manufacturing Industries,” The World Bank Economic
Review, January, 85-112.

Woolridge, J. M., 2000. Introductory Econometrics, South-Western College Publishing,

Cincinnati, OH.

25
Table 1: Average Rates of Return on Assets and Equity (percentage)
1994 1995 1996 1997 1998 1999
ROA
Foreign Banks 1.23 1.39 1.56 3.74 2.43 0.92
Nationwide Banks 0.62 0.27 0.31 -1.03 -3.60 -1.89
Regional Banks 0.81 0.55 0.40 -2.13 -7.53 -0.09
ROE
Foreign Banks 6.67 9.25 9.74 23.02 18.97 7.96
Nationwide Banks 5.45 3.01 3.66 -15.04 -55.80 -65.57
Regional Banks 5.74 4.88 4.40 -20.51 -175.45 -1.61

Note: See Appendix for bank names and country of origin.

Table 2: Number of Foreign, Nationwide, and Regional Banks

1994 1995 1996 1997 1998 1999
Number of Banks
Foreign Banks 52 52 49 53 51 46
Nationwide Banks 14 15 15 16 13 11
Regional Banks 10 10 10 10 8 6

Note: See Table 1.

Table 3: Average Total Assets, Deposits, and Loans and Discounts

1994 1995 1996 1997 1998 1999
Assets (millions of Won)
Foreign Banks 2847 3659 4637 6142 7967 7693
Nationwide Banks 103361 129683 154244 204071 275763 350231
Regional Banks 24793 31054 37840 44331 47603 60727
Deposits (% of assets)
Foreign Banks 11.8 9.2 7.9 8.2 12.3 16.3
Nationwide Banks 57.7 60.2 59.5 58.0 60.7 63.6
Regional Banks 69.2 67.2 64.8 61.7 61.3 71.4
Loans and Discounts (% of assets)
Foreign Banks 34.4 28.2 27.2 25.8 23.9 22.0
Nationwide Banks 57.3 56.5 55.7 57.9 50.7 48.6
Regional Banks 56.0 54.3 52.3 53.0 49.7 45.3

Note: See Table 1.

26
Table 4: Average Deposits and Their Distribution
1994 1995 1996 1997 1998 1999
Deposits (million of Won)
Foreign Banks 337 338 368 504 979 1255
Nationwide Banks 59656 78065 91811 118318 167481 222852
Regional Banks 17165 20861 24523 27358 29191 43382
Won Deposits (% of deposits)
Foreign Banks 29.9 26.1 29.9 34.6 36.8 44.9
Nationwide Banks 60.2 64.7 65.9 68.3 68.0 81.3
Regional Banks 80.1 76.3 72.9 73.2 75.0 90.2
Negotiable Certificates of Deposit (% of deposits)
Foreign Banks 60.6 66.5 62.6 53.6 36.7 33.3
Nationwide Banks 13.5 13.8 12.1 8.3 5.0 3.9
Regional Banks 15.1 18.1 21.1 20.2 16.1 6.3
Foreign Currency Deposits (% of deposits)
Foreign Banks 10.1 7.5 7.4 11.8 26.6 21.8
Nationwide Banks 26.3 21.5 21.9 23.4 27.0 14.9
Regional Banks 4.9 5.6 5.9 6.6 8.9 3.5

Note: See Table 1.

Table 5: Average Loans and Discounts And Their Distribution

1994 1995 1996 1997 1998 1999
Loans and Discounts (million of Won)
Foreign Banks 979 1031 1262 1585 1906 1694
Nationwide Banks 59226 73208 86872 118059 139836 170226
Regional Banks 13890 16871 19801 23490 23668 27496
Won Loans (% of loans and discounts)
Foreign Banks 68.0 66.6 57.1 47.4 45.2 55.0
Nationwide Banks 67.8 69.6 66.9 65.6 70.1 77.9
Regional Banks 95.5 94.6 93.7 92.8 88.2 94.3
Foreign Currency Loans (% of loans and discounts)
Foreign Banks 27.1 27.1 36.5 47.0 51.0 39.6
Nationwide Banks 29.5 28.3 31.0 32.3 27.5 19.8
Regional Banks 3.0 3.9 4.6 5.3 7.0 4.3

Note: See Table 1.

27
Table 6: Panel Regressions of Return on Assets: 1994-1999
Variable ROA
Foreign Nationwide Regional
WL/A -0.0150 0.0523 -0.0441 -0.0615 0.0755 0.0855
(-0.44) (1.37) (-0.59) (-0.86) (0.60) (1.18)
D*WL/A 0.1177* 0.0837**
(2.89) (2.09)
FCL/A -0.0048 -0.0553‡ 0.1625** 0.1512‡ -0.0362 -0.0449
(-0.17) (-1.86) (2.38) (1.93) (-0.06) (-0.14)
E/A -0.0012 -0.0098 0.3463* 0.1580 0.3191* 0.2593*
(-0.06) (-0.55) (2.96) (1.15) (3.50) (3.67)
D*E/A

PLL/A -0.6030* -0.5149* -0.0193 0.2496 -1.2800* -1.4269*

(-3.38) (-2.88) (-0.0528) (0.58) (-4.68) (-3.72)
D*PLL/A -0.6059 -0.8661** 0.7555‡
(-1.55) (-2.06) (1.97)
A 0.0021 -0.0055 -0.0003 0.0000 -0.0024 -0.0024
(0.49) (-1.27) (-1.24) (0.05) (-0.65) (-0.86)
WD/A -0.0850 0.0416 0.0403 0.0333 -0.1178 -0.1303**
(-0.67) (0.33) (0.72) (0.56) (-1.44) (-2.32)
D*WD/A

CD/A -0.0629 -0.0411 0.0715 0.0293 -0.2511**

-0.2281*
(-1.14) (-0.78) (0.88) (0.37) (-2.07)
(-3.54)
FCD/A 0.1012 0.1155 -0.4246* -0.3412* 0.0981
0.1621
(1.00) (1.19) (-6.40) (-4.63) (0.22)
(0.64)
D*FCD/A -1.8769*
-0.5344
(-3.76)
(-1.58)
UNEM 0.9283 -0.7440 -1.2016‡
(1.49) (-1.28) (-1.86)
DGDP 0.2683 00815 0.4802‡
(0.95) (0.31) (1.99)
DEXCH -0.1987* 0.0612 -0.0473
(-2.96) (0.99) (-0.77)
SUR 0.9607 -0.8987 -1.9619‡
(0.89) (-0.94) (-1.93)
INF -0.7359** 0.6604‡ 0.0538
(-2.08) (1.96) (0.15)
Adjusted R2 0.0711 0.1861 0.6554 0.6981 0.7398 0.9307
SEE 0.0276 0.0249 0.0125 0.0117 0.0171 0.0088
Note: The dependent variable is the return on assets (ROA) as a fraction. Independent
28
 individual bank variables include won-denominated loans to assets (WL/A), foreign
currency loans to assets (FCL/A), equity to assets (E/A), provisions for loan losses to
assets (PLL/A), assets (A, in billions of won), won-denominated deposits to assets
(WD/A), certificates of deposit to assets (CD/A), and foreign currency deposits to assets
(FCD/A). Independent macroeconomic variables include the unemployment rate
(UNEM) as a fraction, the rate of growth of real GDP (DGDP) as a fraction, the rate of
change in the Won per US dollar exchange rate (DEXCH) as a fraction, the government
budget surplus to nominal gross domestic product (SUR), and the rate of inflation in the
consumer price index (INF) as a fraction. Finally, D is the Asian financial crisis dummy
variable (equals zero from 1994 to 1996; one otherwise). Summary statistics include the
adjusted R2 and the standard error of the regression (SEE). Numbers in parentheses under
coefficient estimates are t-statistics.

* means significantly different from zero at the 1-percent level

** means significantly different from zero at the 5-percent level
‡ means significantly different from zero at the 10-percent level
‡‡ means significantly different from zero at the 20-percent level

29
Table 7: Panel Regressions of Return on Equity: 1994-1999
Variable ROE
Foreign Nationwide Regional
WL/A -0.1988 0.3264** 0.9660 -1.0809 6.4047‡ 7.9386**
(-1.07) (2.08) (0.61) (-0.88) (1.72) (2.28)
D*WL/A 0.4087‡ -3.0081‡
(1.85) (-1.74)
FCL/A -0.0208 -0.2997** 3.2046* 2.2202 -15.2288 -5.9433
(-0.14) (-1.96) (2.72) (1.65) (-0.93) (-0.39)
E/A -0.0838 -0.1501 3.0559 3.6925 -1.3906 5.7026
(-0.89) (-1.63) (1.23) (1.57) (-0.30) (1.15)
D*E/A 6.5000‡ 15.6920* 10.6105**
(1.79) (2.88) (2.12)
PLL/A -9.1705* -8.2430* -8.3901* 3.6082 -67.6646* -60.7605*
(-5.25) (-5.41) (-2.95) (0.49) (-3.17) (-2.77)
D*PLL/A 4.7305** 5.2592* -12.7275‡ 82.2019* 80.1891*
(2.30) (2.67) (-1.76) (3.53) (3.57)
A 0.0748* 0.0376‡ -0.0042 0.0009 0.1099 -0.0986
(3.40) (1.69) (-0.88) (0.14) (1.03) (-0.74)
WD/A 0.5799 -0.4385 0.8379 0.3661 -2.4612 5.8787
(0.58) (-0.69) (0.88) (0.36) (-0.98) (1.31)
D*WD/A -2.3267‡ -12.1790**
(-1.80) (-2.27)
CD/A -0.1370 -0.0245 0.5417 -0.3733 -7.5209** -7.3239**
(-0.48) (-0.09) (0.40) (-0.27) (-2.11) (-2.42)
FCD/A -0.2239 -0.2036 -6.2501* -4.4065* 12.0958 5.8084
(-0.52) (-0.41) (-5.70) (-3.47) (0.87) (0.48)
D*FCD/A -58.4154* -38.8472**
(-3.63) (-2.37)
UNEM 7.2613** -12.6696 -103.3952**
(2.31) (-1.26) (-2.22)
DGDP 1.2614 0.1060 -1.5374
(0.87) (0.02) (-0.13)
DEXCH -1.0163* 1.0892 -0.0433
(-2.94) (1.03) (-0.01)
SUR 8.1054 -12.4824 -148.7384**
(1.48) (-0.76) (-2.15)
INF -4.0207** 7.5334 -17.9635
(-2.21) (1.30) (-1.04)
Adjusted R2 0.1896 0.2971 0.5903 0.6199 0.7277 0.8133
SEE 0.1371 0.1277 0.2093 0.2016 0.4994 0.4136
Note: See Table 6. The dependent variable is the return on equity (ROE) as a fraction.
30
Appendix:

The banks included in this study were in operation in Korea at some point during 1994 to

1999 and are subdivided into nationwide and regional Korean banks and foreign banks. Table A

lists the banks, their country of origin, and the number of years in the sample.

Table A: The Banks and Country of Origin

Foreign Banks
Country Bank Name Sample Period
1 United States CitiBank 1994-1999
2 United States Bank of America 1994-1999
3 United States First National Chicago 1994-1999
4 United States American Express Bank 1994-1999
5 United States Chemical 1994-1999
6 United States Chase Manhattan 1994-1995
7 United States Banker Trust 1994-1998
8 United States Bank Boston 1994-1999
9 United States The Union Bank of California 1994-1999
10 United States Bank of New York 1994-1999
11 United States Bank of Hawaii 1994-1999
12 United States Nations Bank 1997-1998
13 United States First Union National 1997-1999
14 United States Morgan Guaranty Trust Co 1998-1999
15 Japan Bank of Tokyo 1994-1995
16 Japan Bank of Tokyo Mitsubi 1994-1999
17 Japan Fuji Bank 1994-1999
18 Japan DaiIchiKankyo 1994-1999
19 Japan Sumitomo Bank 1994-1999
20 Japan Sanwa Bank 1994-1999
21 Japan ToOhKai 1994-1998
22 Japan SaKura 1994-1998
23 Japan Asahi Bank 1994-1999
24 Japan Taiwa 1994-1998
25 Japan Mitsubisu Trust 1994-1998
26 Japan Yamakuchi 1994-1999
27 Japan Japan Long-Term Trust 1994-1997
28 Japan Yasda Trust 1994-1997
29 France Credit Agricole Indosuez 1994-1999
30 France Banque Nationale de Paris 1994-1999
31 France Banque Paribas 1994-1999
32 France Credit Lyonnais 1994-1999
33 France YoBafe 1994-1999
34 France Societe Generale 1994-1999
31
 Foreign Banks (continued)
Country Bank Name Sample Period
35 United Kingdom Standard Chartered Bank 1994-1999
36 United Kingdom Barclays 1994-1995
37 United Kingdom National Westminster 1994-1997
38 Canada Bank of Nova Scotia 1994-1999
39 Canada Montreal 1994-1996
40 Canada Canada Royal 1994-1999
41 Canada National Bank of Canada 1994-1999
42 Singapore Singapore 1994-1999
43 Singapore Development Bank of Singapore 1994-1999
44 Singapore United Overseas Bank 1994-1999
45 Singapore Overseas Union Bank 1994-1999
46 Australia Australia and New Zealand Banking Group 1994-1999
47 Australia National Australia Bank Limited 1994-1999
48 Netherlands ABN-AMRO Bank 1994-1999
49 Netherlands ING Bank NV 1994-1999
50 Germany Deutsche Bank 1994-1999
51 India Indian Overseas Bank 1994-1999
52 Hong Kong Hong Kong-Shanghai 1994-1999
53 Pakistan National Bank of Pakistan 1994-1999
54 Jordan Arab Bank PLC 1994-1999
55 China Bank of China 1994-1999
56 China Industrial and Commercial Bank of China 1997-1999
57 Philippines Metropolitan Bank and Trust Company 1997-1999
58 Switzerland Credit Suisse First Boston 1997-1999
59 Switzerland UBS 1999-1999

Korean Nationwide Banks

Country Bank Name Sample Period
1 Korea Cho Hung 1994-1999
2 Korea Sang-Up 1994-1999
3 Korea Korea First 1994-1999
4 Korea Hanil 1994-1998
5 Korea Seoul 1994-1999
6 Korea Korea Exchange 1994-1999
7 Korea Kook-Min 1995-1999
8 Korea Housing &Commercial 1997-1999
9 Korea Shinhan 1994-1999
10 Korea Han Mi (KORAM) 1994-1999
11 Korea Dong-Wha 1994-1997
12 Korea Dong-Nam 1994-1997
13 Korea Dae-Dong 1994-1997
14 Korea Hana 1994-1999
15 Korea Boram 1994-1998
16 Korea Pyong-Wha (peace bank of Korea) 1994-1999

32
 Korean Regional Banks
Country Bank Name Sample Period
1 Korea Daegu 1994-1999
2 Korea Pusan 1994-1999
3 Korea Chung-Cheong 1994-1997
4 Korea Kwang-Ju 1994-1999
5 Korea Che-Ju 1994-1999
6 Korea Kyung-Gi 1994-1997
7 Korea Jeon-Buk 1994-1999
8 Korea Kong Won 1994-1998
9 Korea Kyong-Nam 1994-1999
10 Korea Chung Buk 1994-1998

33
Table 6o: Panel Regressions of Return on Assets: 1994-1999
Variable ROA

Foreign Nationwide Regional

WL/A 0.0476‡ 0.0999* -0.0333 -0.0432 0.1190 0.1024‡‡
(1.79) (3.24) (-0.44) (-0.59) (0.81) (1.37)
FCL/A -0.0055 -0.0564‡ 0.1523** 0.1640** -0.1379 -0.0780
(-0.19) (-1.88) (2.21) (2.04) (-0.21) (-0.25)
E/A -0.0016 -0.0100 0.3313* 0.1021 0.3349* 0.2758*
(-0.09) (-0.55) (2.81) (0.74) (3.15) (3.79)
PLL/A -0.6103* -0.5080* -0.5549* -0.5576* -1.0599* -0.6999*
(-3.37) (-2.82) (-3.28) (-3.17) (-3.39) (-4.21)
A 0.0025 -0.0058‡‡ -0.0003 0.0000 -0.0001 -0.0011
(0.58) (-1.33) (-1.14) (0.02) (-0.02) (-0.37)
WD/A -0.1145 0.0390 0.0273 0.0146 -0.0317 -0.0931‡‡
(-0.89) (0.31) (0.49) (0.24) (-0.35) (-1.68)
CD/A -0.0680 -0.0429 0.0512 0.0142 -0.1979‡‡ -0.2160*
(-1.21) (-0.81) (0.63) (0.17) (-1.41) (-3.23)
FCD/A 0.0904 0.1181 -0.4394* -0.3740* -0.8593** -0.1294
(0.88) (1.21) (-6.62) (-5.05) (-2.07) (-0.61)
UNEM 0.7045 -0.3122 -1.7386*
(1.14) (-0.56) (-2.81)
DGDP 0.2482 0.2118 0.5260**
(0.87) (0.80) (2.10)
DEXCH -0.2031* 0.0339 -0.0542
(-3.00) (0.55) (-0.86)
SUR 0.7299 -0.6131 -2.6178**
(0.68) (-0.63) (-2.62)
INF -0.7893** 0.7429** 0.2034
(-2.22) (2.16) (0.56)
Adjusted 0.0422 0.1743 0.6472 0.6800 0.6449 0.9245
R2
SEE 0.0270 0.0251 0.0127 0.0121 0.0199 0.0092
Note: The dependent variable is the return on assets (ROA) as a fraction. Independent
individual bank variables include won-denominated loans to assets (WL/A), foreign
currency loans to assets (FCL/A), equity to assets (E/A), provisions for loan losses to
assets (PLL/A), assets (A, in billions of won), won-denominated deposits to assets
(WD/A), certificates of deposit to assets (CD/A), and foreign currency deposits to assets
(FCD/A). Independent macroeconomic variables include the unemployment rate
(UNEM) as a fraction, the rate of growth of real GDP (DGDP) as a fraction, the rate of
change in the Won per US dollar exchange rate (DEXCH) as a fraction, the government
budget surplus to nominal gross domestic product (SUR), and the rate of inflation in the
consumer price index (INF) as a fraction. Summary statistics include the adjusted R2 and
34
 the standard error of the regression (SEE). Numbers in parentheses under coefficient
estimates are t-statistics.

* means significantly different from zero at the 1-percent level

** means significantly different from zero at the 5-percent level
‡ means significantly different from zero at the 10-percent level
‡‡ means significantly different from zero at the 20-percent level

35
Table 7o: Panel Regressions of Return on Equity: 1994-1999
Variable ROE

Foreign Nationwide Regional

WL/A 0.0582 0.3397** -0.5238 -0.8117 5.8239 6.7288‡‡
(0.43) (2.14) (-0.41) (-0.65) (1.24) (1.59)
FCL/A -0.0308 -0.3131** 2.6497** 2.4088‡ -15.4080 -14.8777
(-0.20) (-2.02) (2.30) (1.76) (-0.73) (-0.83)
E/A -0.0795 -0.1375‡‡ 5.9484* 2.8609 11.6522* 14.8876*
(-0.84) (-1.47) (3.01) (1.22) (3.40) (3.63)
PLL/A -5.7875* -5.2829* -9.0958* -8.2521* 8.5090 15.5917‡‡
(-6.21) (-5.68) (-3.21) (-2.75) (0.84) (1.66)
A 0.0753* 0.0332‡‡ -0.0035 0.0007 0.1571 0.0010
(3.41) (1.47) (-0.75) (0.11) (1.17) (0.01)
WD/A -1.0614‡‡ -0.4058 0.4744 0.0910 -1.8413 -1.1207
(-1.59) (-0.63) (0.50) (0.09) (-0.62) (-0.36)
CD/A -0.1675 -0.0330 0.2185 -0.5954 -4.9077 -5.8685‡‡
(-0.58) (-0.12) (0.16) (-0.43) (-1.08) (-1.56)
FCD/A -0.2346 -0.2042 -6.2189* -4.8895* -31.3187** -14.5412
(-0.44) (-0.41) (-5.60) (-3.87) (-2.34) (-1.22)
UNEM 6.9832** -6.3185 -88.0117**
(2.20) (-0.66) (-2.52)
DGDP 0.9494 2.0211 9.7697
(0.64) (0.45) (0.69)
DEXCH -0.9457* 0.6880 -3.1143
(-2.71) (0.65) (-0.87)
SUR 8.2428‡‡ -8.2849 -128.4488**
(1.48) (-0.50) (-2.28)
INF -4.5418** 8.7454‡‡ -4.1863
(-2.47) (1.49) (-0.20)
Adjusted 0.1680 0.2785 0.5797 0.6052 0.5504 0.7057
R2
SEE 0.1389 0.1293 0.2120 0.2055 0.6417 0.5192
Note: See Table 6. The dependent variable is the return on equity (ROE) as a fraction.

36