soft related lending: a tale of two korean banks *

Soft Related Lending: A Tale of Two Korean Banks *

John P. Bonin

and

Masami Imai

Department of Economics

Wesleyan University

Middletown CT. 06459 USA

ABSTRACT

In this paper, we investigate the effect of the impending sale of two Korean banks to foreign financial institution after the financial crisis of 1997 on the abnormal stock price returns of related borrowers. Having detailed information that links publicly traded companies to a main bank in Korea, we employ event studies using stock market data from the end of 1997 through the beginning of 2000. We find that events signaling the sale of the banks to a foreign financial institution yield an average decrease of about 2% in the stock price of related borrowers over a three-day window surrounding the event. In addition, we find that events indicating a weakening in the Korean government’s resolve to sell these two banks generate an increase in the stock price of related borrowers of about the same magnitude. Moreover, we find evidence that these events have larger impacts on firms that are both less liquid and more reliant on short-term bank loans and that the stock prices of the more unprofitable firms react more strongly to some of the events. Taken together, these results indicate that these two Korean banks engaged in soft lending practices that provided rents to related borrowers and, perhaps, kept afloat zombie insolvent firms. We conclude that credibility is the most important component to a government’s commitment to prompt credible action to stem a banking crisis and that this is best established by sale of the bank to a foreign owner in situations in which relational lending has become common practice.

JEL Classifications: G21, O53 * The authors are grateful to the Economics Department of Wesleyan University and the Luce Foundation for providing financial support and to Maiko Kondo and Kiyoki Nishio for excellent research assistance.

2

1. Introduction: Relationship Lending

Stable, long-term, relationships between banks and their clients are often claimed to

be beneficial because they allow banks to gather private information that permits more

efficient debt contracts. In economies with less developed financial sectors, bank-

centered systems are considered to be preferable to market arrangements that require

considerable supporting infrastructure1 However, the resulting related lending practices

that develop in these economies may lead to inefficient outcomes that enlarge the total

costs of resolution in the event of a banking crisis. This negative aspect of related lending

is discussed by Laeven (2001) who finds that Russian firms and banks engaged in insider

lending on the basis of loan volume. Similarly, Cull, Matesova and Shirley (2002) claim

that the Czech Republic’s initial failure to cede majority control of its banks to private

owners may explain the poor performance of the banks’ large preferred customers. In

addition, La Porta, Lopez-de-Silanes, and Zamarripa (2003) use detailed data on interest

rates and performance of loans by Mexican banks to show that interest rates on related

loans are low and insensitive to risk relative to unrelated loans and that related loans tend

to become nonperforming loans. Finally, the East Asian financial crisis of 1997 exposed

the dark side of soft related lending.

Rajan and Zingales (1998) argue that the Asian financial is attributable in part to the

prevalence of related lending and the reluctance of banks to cut off funds to weak

borrowers. Jeon and Miller (2004) identify Korea as the country that best illustrates this

problem because its macroeconomic indicators did not indicate severe problems. Rather,

the authors point to an investment boom fueled by short-term foreign capital inflows to

1Rajan and Zingales (1998) maintain that the prompt and unbiased enforcement of contracts by the courts is a pre-condition for the viability of a market-based system.

3

distinguish the Korean situation from the real-estate booms that characterized other Asian

countries. The combination of an economy dominated by large conglomerates, i.e.,

chaebols, and permissive regulatory practices toward banks proved to be lethal. By

November 1997, the central bank was unable to defend the domestic currency against the

attack that spread quickly throughout the region after the devaluation of the Thai baht in

July 1997. The Korean government sought an emergency standby credit package from

the IMF in December 1997. This action set in motion a chain of events that had

significant implications for related lending practices in two relatively large Korean banks.

Jeon and Miller (2004) attribute the flight of short-term foreign capital, especially

bank lending, in Korea to the bankruptcy of several chaebols, i.e., Kia, Hanbo, Haitai,

Sammi, and Daewoo, and the practice of syndicated loans to other chaebols. At the

beginning of December 1997, the Korean government nationalized two banks that had

extensive relationships with financially weak chaebols. Korea First Bank and Seoul

Bank, with market shares of around 5% each, were considered to be too big to fail but too

weak to be viable without recapitalization. An article in the Financial Times on January

30, 1997 reports that “Korea First, with shareholders' equity of Won 1,800bn (£1.3bn),

lent almost Won 1,100bn to Hanbo.” A subsequent article in the Financial Times on

February 7, 1997 reports: “The former Seoul Bank chief, Mr Song Hong-kyun, was

arrested in December and accused of taking $ 244,100 (£150,000) in kickbacks after

extending preferential loans to four companies. The court released the name of only one

company involved, the International Valve Company.” Moreover, despite ranking eighth

and tenth of eleven nationwide commercial banks in terms of assets, Korea First Bank

and Seoul Bank ranked third and fifth, respectively, among these same eleven banks in

4

number of companies identifying them as their main bank. Toward the end of December

1997, the Korean government and the IMF agreed on a letter of intent that singles out

these two banks for restructuring in preparation for sales to foreign financial institutions.

Such an ownership change would be expected to put an end to soft related lending

practices in these banks.

In this paper, we examine the impact of related lending on the value of the stock

prices of the client companies of these two Korean banks. In a similar study of related

lending for 29 insolvent banks in three Asian countries, one of which is Korea, Djankov,

Jindra, and Klapper (2001) find that a bank relationship adds value to a company’s stock

but that this value depends on the expected continuance of this relationship. These

authors note that in two of the three countries multiple bank relationships are common

but in Korea a company is associated with a single main bank. Hence, we are able to

identify unambiguously the relationships between companies and banks. In addition, this

relationship is quite stable over time in Korea during the sample period. We begin with

the nationalization of the two banks on December 9, 1997 and identify eight other

pertinent events through December 23, 1999 when Korea First Bank is sold to Newbridge

Capital Group. These nine events are chosen for the information that they provide about

the likelihood of continued related lending practice in these banks. Our first objective is

to evaluate the effect of these events on the abnormal stock returns of the Korean

companies having one of these two banks as their main bank.

Bae, Kang, and Lim (2002) investigate the extent to which company value is related

to the financial health of its main bank in Korea using data from 1997 and 1998. These

authors point out that Korea is a prime example of a bank-centered system because

5

companies obtain most of their external financing from banks. Bae et al. identify 113

adverse events and attribute to these an average decline in bank stock prices of 2.5% and

an average decline in company stock prices of 1.3%. Our focus in this paper is different

in that we examine the impact of news that provides information about likely changes in

bank lending practices on the stock prices of the companies. Choi and Hasan (2005)

investigate the effect of ownership and governance change on bank performance

measures in Korea from 1998 to 2002. Their results show the importance of the extent of

foreign ownership and the presence of a foreign director on the board in generating

increased return and decreased risk in banks’ portfolios. Majority stakes of both Korea

First Bank and Seoul Bank were slated to be sold to foreign financial institutions after

restructuring in an agreement between the Korean government and the IMF. Our intent is

to analyze the effect of the likely sale of a bank to foreign owner on the value of

companies relying on this bank for external financing.

The maintained hypothesis is that a foreign owner will not continue lending practices

that are detrimental to the financial performance of the bank, in particular, soft related

lending. Hence, we expect to find a negative response in the stock prices of companies

attached to the bank to news that makes a sale to a foreign financial institution more

likely and a positive response to news that suggests such a sale is delayed or even in

jeopardy. Over the approximately two-year sample period, the fates of the two banks

diverge considerably. In April 1998, Morgan Stanley was selected to advise the

restructuring and privatization of both banks. At the end of the year, a memorandum of

understanding was signed by the Korean government and Newbridge Capital concerning

the purchase of Korea First Bank. In February of the next year, a similar document was

6

signed with HSBC for the purchase of Seoul Bank. However, only the first agreement

came to fruition when the deal was consummated at the end of 1999. Negotiations over

the sales of Seoul Bank to HSBC broke down; in November 2002, a domestic Korean

bank, Hana Bank, took over Seoul Bank. We intend to exploit the differences in the

experiences of these two banks from the end of 1997 to the end of 1999 in investigate the

effects of foreign control on related lending in Korea.

In their three-country study using pooled data from 1997 to 1999, Djankov et al.

(2001) conclude that a foreign sale of a bank results in a one percent decline in the stock

value of companies related to the bank in the short term but an eight percent increase in

the long term. However, these authors incorrectly list Seoul Bank as having been sold to

a foreign owner during this period. Nonetheless, their main result that corporate

valuation depends on investors’ confidence in the continuity of management in insolvent

banks provides support for our maintained hypothesis. In addition, we expect the sale of

a bank to a foreign owner to affect the stock prices of weak companies more than strong

companies. Peek and Rosengren (2003) show that Japanese banks expand related loans

to financially and economically weak borrowers and that this tendency to prop up weak

firms is the strongest among the weakest banks. Djankov et al. show that smaller more

leveraged firms and those with less expected market growth are impacted more severely.

To investigate the effect of different company characteristics on the changes in the value

of related lending, we interact measures of liquidity, profitability, and reliance on short-

term bank loans with the events as explanatory variables in multiple regressions.

To preview our empirical findings, these related borrowers experienced negative

abnormal returns when the government of Korea nationalized these two banks and agreed

7

with the IMF to sell them to foreign financial institutions. Moreover, these firms enjoyed

positive abnormal returns whenever the Korean government had difficulty in finding

interested buyers or in agreeing to sell these banks. Finally, when Newbridge Capital

acquired Korea First Bank, firms that depended heavily on the bank for financing

experienced negative abnormal returns. Furthermore, these bank ownership effects tend

to be stronger for unprofitable firms with less liquid assets and more reliance on short-

term bank loans. Hence, our empirical evidence suggests that sales of insolvent banks to

foreign financial institutions have negative effects on unprofitable, perhaps zombie, firms

because they guarantee the end of soft related lending.

Our paper is organized as follows. The next section provides a detailed

description of the events that we identify as pertinent to our study of the impact of foreign

ownership on soft related lending and specifies their expected effects on the stock prices

of the related borrowers. Section 3 discusses the data and describes the event study

methodology that we employ. In section 4, we present and discuss the results for the

baseline model. Section 5 provides robustness checks in which we control for industry

effects and also examine the sensitivity of the results to the selection of the group of

borrowers identified as related to the two banks. Section 6 concludes with a summary of

the results and some policy implications.

2. The Events and Their Expected Effects

Soft related lending was highlighted as an important contributing factor to the

financial crisis in Korea by the IMF. In return for standby credit support at the end of

1997, the Korean government agreed to change its bank regulatory policy from

8

forbearance to prompt corrective action. Two banks, Korea First Bank and Seoul Bank,

were singled out as the major offenders in continuing to provide loans to insolvent related

borrowers. The Korean government agreed to change the governance of these two banks

and restructure them in preparation for sales to foreign financial institutions. Over the

next two years, a series of events occurred that resulted in one of the banks, Korea First

Bank, being sold to a foreign owner but the other, Seoul Bank, left at the altar. We intend

to use the variations in these events to examine the effect of soft related lending on the

value of the client companies of these two banks. The relevant events are discussed

below and their expected impact is specified. We obtained the announcement dates from

a comprehensive search of the Lexus-Nexus database, which includes the Financial

Times, AFX News-Asia, and the Korea Times. Table 1 provides a brief chronology of the

relevant events.

The first event to be highlighted is the partial nationalization of Korea First Bank

and Seoul Bank on December 9, 1997 when the government of Korea announced its

purchase of 59% of the shares of these two banks.2 In return for this massive capital

injection, the banks were required to undertake stringent restructuring. On the same day,

bank officials announced that they would lay off 1500 workers over the next 2 years. We

consider this event to indicate a critical change in government policy toward these two

banks from forbearance to more stringent prompt corrective action.3 Hence, we expect

this event to have a negative effect on the stock prices of related borrowers.

The second important event is the signing of the letter of intent with the IMF on

December 26, 1997. In negotiations with the Korean government, the IMF insisted that

2 We use Korean dates for all events.

9

the banking sector be restructured and that foreign investors be allowed to take majority

stakes in domestic Korean banks. In the letter of intent, the Korean government agreed

that, among other things, the bank of Korea would provide no short-term liquidity to

financial institutions, the government would assume complete control of insolvent banks

and remove the existing management, and the government would appoint outside experts

to oversee the restructuring and privatization of Korea First Bank and Seoul Bank.4 We

expect this event to have a negative effect on the market value of related borrowers.

The third event is the actual appointment of Morgan Stanley as lead manager of

the restructuring and privatization efforts for both banks on April 22, 1998. Although this

move is likely to have been anticipated because of the agreement with the IMF, we check

to see if it has any independent effect on related borrowers because action speaks louder

than words in financial markets.

The fourth important event is the delay in the privatization of Korea First Bank

and Seoul Bank on November 5, 1998.5 In the letter of intent with the IMF, the Korean

government promised that these two banks would be sold by November 15, 1998. Given

the difficulty it had in obtaining bids from foreign financial institutions, the Korean

government decided to postpone the sales until the end of January. Postponement of

privatization based on a lack of foreign interest by the government could be perceived as

indication that past soft lending practices would continue to some extent. Hence, this

event may provide hope to related borrowers that lending practices may not change as

drastically as they would have with imminent foreign ownership of the two banks.

3 Before this event, the Korean government purchased nonperforming loans through the Korean Asset Management Company (KAMCO) without imposing any stringent restructuring requirements. 4 See the IMF website (http://www.imf.org/external/np/loi/122497.htm) for further details. 5 See “Korean sell-offs postponed” Financial Times. November 5, 1998

10

The fifth and sixth events are the signings of memoranda of understanding for the

sales of Korea First Bank and Seoul Bank to Newbridge and HSBC on December 28,

1998 and February 22, 1999, respectively. Unlike the postponement news, these events

should have a negative impact on the stock prices of related borrowers because they

signal a renewed commitment to pursuing sales of both banks to foreign owners.

Although the memoranda of understanding were signed, negotiations for the sales

of the two banks stalled. In fact, the government and the foreign financial institutions

failed twice to reach agreement by the deadlines specified in the memoranda. In the case

of Korea First Bank, the government and Newbridge failed to reach any agreement by

April 30, 1999, which was the first deadline, or by May 12, 1999, which was the second

deadline. The corresponding missed deadlines for Seoul bank were May 31, 1999 and

June 28, 1999. These events indicate continuing difficulty with the finalization of

agreements to sell the banks to foreign owners and, as such, may give hope again to

related borrowers that soft lending practices may continue in the immediate future.

Eventually, on July 1, 1999, Newbridge Capital Group agreed to acquire Korea

First Bank. The transaction was consummated on December 23, 1999. Both of these

events signal the end of soft related lending to the clients of this bank. In the case of

Seoul Bank, HSBC dropped its bid for acquiring the bank and the government failed to

sell the bank to a foreign financial institution as it had agreed to do. The different

experiences of these two banks during this period should be reflected in different

responses of the stock prices of related borrowers. Companies using Korea First Bank as

their main bank would expect to be dealt with on commercial terms only while

11

companies affiliated with Seoul Bank may perceive some continued forbearance on the

part of the government owner.

3. The Data and the Methodology

To estimate the abnormal returns of related companies associated with the news

concerning the likelihood of foreign takeovers of Korea First Bank and Seoul Bank, we

run a standard market-model regression. Our methodology follows that used widely in

event studies; MacKinlay (1997) provides an overview of the literature. Our sample

consists of daily stock prices from November 1, 1997 to February 29, 2000 for publicly

traded firms taken from the University of Rhode Island’s Pacific Basin Capital Market

Research Center (PACAP) data base. PACAP also provides balance sheet information

that we use to construct our measures of firm characteristics.

To establish the main bank links, we use the annual publication Korean Company

Information (Kankoku Kaisha Joho), which identifies the most important bank for each

Korean firm. Initially, we use the 2000 edition and merge PACAP data with this

information. Table A2 of the Appendix identifies the number of firms related to each of

nationwide Korean banks from the 1998 and the 2000 editions. As the diagonal of the

table indicates, the relationship is relatively stable over time. Our sample for the baseline

model consists of 106 firms for which we have information about stock market returns

and that identify one of the two relevant banks as their main bank.

We regress the daily changes in stock market prices of firms on the change in a

market index given by Korean Stock Price Index (KOSPI) and dummy variables

12

associated with a three-day event window consisting of the event date plus one day

before and one day after its occurrence. The regression equation takes the following form:

it

l

kkitkmtiiit DRR εγβα +++= ∑

=1

, (1)

where itR is the change in the stock price of firm i on day t, iα is the intercept coefficient

for firm i, mtR is the change in the market index, KOSPI, for day t, iβ is the market risk

coefficient for firm i, ktD is a binary variable that equals 1 if day t is within the three-day

event window k, and kγ captures average abnormal returns associated with the event

window k. Equation (1) is estimated as a system of equations for the individual firms to

allow for contemporaneous correlation of the error terms across firms.6

Second, we investigate the differential impact on the value of firms relying on

Korea First Bank and Seoul Bank depending on their financial characteristics. For this

exercise, the regression equation is:

it

l

kkitik

l

kkitkmtiiit DXDRR εθγβα ++++= ∑∑

== 11

,

where iX is a financial characteristic of firm i that we hypothesize is correlated with

related lending. Hence, the iX variables are interacted with the dummy variables

identifying events in the regression. We consider several such variables, namely, the

ratio for gross profit to assets, the ratio of cash and bank deposits to assets, and the ratio

of short-term bank loans to assets.

6 We assume that the error terms are independent but heteroskedastic within each equation and uncorrelated with the market index and event dummies. In addition, the non-contemporaneous correlations of error terms across firms are assumed to be zero. Given this assumption, we use the cluster option of STATA’s regress command to obtain correct standard errors.

13

The ratio of gross profit to assets captures the profitability of firms. Profitable

firms are expected to remain relatively unaffected by these events because, unlike

unprofitable firms, they can be expected to continue to receive loans on a commercial

basis. However, to the extent that these profitable firms were receiving subsidized soft

related loans, they would also lose value if the bank reverts to commercial conditions.

Moreover, profitable firms may be able to obtain credit from various sources at

reasonable rates regardless of their main bank’s ownership structures. The ratio of cash

and bank deposit to assets measures the liquidity of the firms. Even if the lending is

restricted upon the foreign takeover of the bank, those firms with sufficient cash and bank

deposits will not need to rely so extensively on bank loans because of their liquid assets.

Alternatively, firms with low liquidity are more beholden to their main banks for

financing.

The ratio of short-term bank loans to assets measures the extent to which firms

actually rely on bank loans as opposed to other sources of funds. Firms having a large

percentage of short-term bank loans in their portfolios are more likely to depend on their

main bank for funds. 7 We expect the reliance on short term bank loans to magnify the

effect of a change in ownership of the company’s main bank on its abnormal returns due

to the uncertainty about continuing related lending. Table 2 provides descriptive

statistics for the variables; it indicates considerable variation in terms of liquidity,

profitability, and reliance on short-term bank loans among firms whose main bank is

either Korea First Bank or Seoul Bank. However, the three financial measures of firm

characteristics to be used interactively as explanatory variables in the regression are

7 Ideally, we would like to include in the regression a variable that measures directly the reliance of the firm on the main bank but we are unable to find such data.

14

highly correlated as expected.8 Hence, multicollinearity may affect the statistical

significance of the tests for independent effects of the change in ownership structure of

the main bank on the stock price returns of related borrowers.

4. The Empirical Results

The baseline model incorporates the nine events along with firm-specific

intercepts and risk coefficients based on the relationship between stock prices and the

Korean Stock Price Index as independent variables. The dependent variable is estimated

abnormal stock price returns. The results are reported in column 1 of Table 3 with the

coefficients for the intercepts and the risk coefficients omitted to keep the table relatively

uncluttered. The first event to signal a change in the ownership for Korea First Bank and

Seoul Bank is the announcement of the nationalization that renders the government the

majority owner, with a 59% stake, of both banks due to its capital injection. The

estimated abnormal return of affiliated firms is a statistically significant negative 2.2%.

As predicted, investors perceived this event as bad news for firms that have close

borrowing relationship with these two banks.9

Similarly, the announcement of a finalized agreement with the IMF that commits

the Korean government to restructuring these two banks and selling them to foreign

financial institutions generates a statistically significant negative 2.4% abnormal return

for related borrowers. In both cases, the anticipated change in ownership has the expected

8 The simple correlation coefficients range from 0.64 to 0.79. 9 Djankov et al. (1999) find that nationalization leads to a 3% short-term increase in abnormal returns of related borrowers for their sample of insolvent banks in three Asian countries. However, they interpret this event differently because they consider nationalized banks to be those that will continue lending relationships. For Korea First Bank and Seoul Bank, we consider nationalization to be the first step to

15

negative effect on the stock prices of companies using these banks as a main bank. The

letter of intent that the Korean government signed with IMF includes specific steps to be

taken to re-privatize Korea First Bank and Seoul Bank. In particular, the first requirement

is to appoint an outside lead manager for the restructuring and preparation for

privatization. When the Korean government actually took this step on April 22, 1998,

negative abnormal returns are indicated for related borrowers but the coefficient is not

statistically significant in Table 3.

On November 5, 1998 when the Korean government announced that the

anticipated sales of both Korea First Bank and Seoul Bank were postponed, related

borrowers of both banks experienced statistically significant positive abnormal returns of

2.5%. This evidence confirms our hypothesis that any news indicating the likely

continuation of related lending will have a positive impact on the value of companies

using these banks as their main bank. In addition, the average gain for related borrowers

of these banks is equal to the loss in value they experienced when the letter of intent with

the IMF was finalized.

After the announcement of a postponement in privatization of the banks,

memoranda of understanding were signed with two foreign financial institutions,

Newbridge Capital Group for Korea First Bank and HSBC for Seoul Bank. The

coefficient for this event is negative, as expected, but it is not statistically significant.10

However, when the first deadlines stipulated in the memoranda passed without any

formal agreement with the foreign owner for either bank, related borrowers earned

transferring ownership to a foreign financial institution. Djankov et al. recognize this expectation by classifying both of these Korean banks as foreign owned in their study. 10 This coefficient and the following two coefficients reflect the impact of the event on the related borrowers for each bank in different time periods due to the different dates of the event for each bank.

16

statistically significant positive abnormal returns of 1.6%. A second deadline was also

missed but the coefficient for this event is not statistically significant perhaps because the

expectation of a delay had already been captured by the market reaction to the first

deadline passing without any action.

Although the announcement of the sale of Korea First Bank to Newbridge Capital

did not have a statistically significant effect on related borrowers, the actual privatization

event on December 23, 1999 resulted in statistically significant negative abnormal returns

of 2.4% for related borrowers.11 Taken together, these results indicate that considerable

uncertainty surrounded the sale of both banks after the Korean government announced a

postponement in privatization plans at the beginning of November 1998. Not until more

than a year later when Korea First Bank was actually purchased by Newbridge Capital

did the market find this bank’s change in ownership credible. At that time, firms using

Korea First Bank as their main banks lost value on average equal to the value they lost

during the two first events in December 1997 when the banks were nationalized and the

letter of intent with the IMF was signed.

Turning to the issue of whether the relational benefit depends on the type of the

borrower, we examine the impact of each announcement interacted with the profitability,

liquidity, and dependence on bank loans of the related firms. We expect to find firms that

depend more heavily on related lending and unprofitable firms, which are sometimes

characterized as zombie firms, to be affected more by any perceived change in sift

lending practices. The second column of Table 3 displays the results of this regression for

the baseline assumptions. For the events representing the signing of the letter of intent

11 Only companies using Korea First Bank as their main bank are included in the estimation of this coefficient leading to relatively large standard errors.

17

with the IMF and the postponement of privatization plans, five of the six interactive

coefficients have the expected sign and are statistically significant. Stronger, more

profitable firms lose less value, as the positive coefficient indicates, when the event

signals an end to related borrowing and gain less, as the negative coefficient indicates,

when the continuance of related lending seems more likely. More liquid firms lose

considerably less in value as a result of the agreement with the IMF; their response to the

postponement event is positive but not statistically significant. In addition, both the

decrease in value accompanying negative news and the increase in value associated with

positive news is magnified for related borrowers that are more heavily dependent on their

main bank as a source of funds.

Several of the other interactive coefficients are statistically significant and all of

these have the expected signs. The interactive coefficients regarding nationalization and

the appointment of Morgan Stanley as the outside advisor indicate strong positive effects

for liquid firms and strong negative effects for firms relying heavily on bank loans.

However, both coefficients representing profitability are statistically insignificant; the

one associated with nationalization has an unexpected sign. The market uncertainty

generated by the postponement of privatization is reflected in mostly statistically

insignificant interactive coefficients. Curiously, the interactive coefficients associated

with the actual privatization of Korea First Bank are also statistically insignificant and of

unexpected sign. The large standard errors on these coefficients reflect the smaller

number of relevant firms because only related borrowers of this bank are included.

The results from the baseline model are broadly consistent with the hypothesis

that unprofitable and financially fragile borrowers were the main beneficiaries of related

18

loans. Moreover, the coefficients from the first column of Table 3 indicate that the stock

prices of related borrowers decrease by about 2% in response to news indicating the end

to soft related lending and increase by about the same percent when news signals strongly

the postponement of ownership change and the possible continuation of past lending

practices. This magnitude is roughly comparable to the average decline in the stock

prices of companies affiliated with a main bank experiencing financial difficulty found by

Bae et al. (2002) in the pre-financial crisis period in Korea. In summary, the evidence

indicates that Korea First Bank and Seoul Bank were engaged in soft related lending and

that selling these banks to a foreign owner is perceived as putting an end to such non-

commercial behavior by investors.

5. Robustness Checks

To investigate the strength of the evidence regarding related lending and foreign

ownership of the main bank, we conduct several robustness checks. First, we have

assumed that the relationship between the stock price of the firms and the Korean stock

price index (KOSPI) remains the same throughout the sample period in the baseline

regression model. However, over this more than two-year period, this relationship may

change. To allow for such differences, we divide our sample into four sub-samples,

namely, November 1, 1997 to May 17, 1998; May 18, 1998 to November 24, 1998;

November 25, 1998 to July 14, 1999; and July 15, 1998 to February 29, 2000. Any bias

caused by considering the entire period in the regression is likely to be less severe when

these sub-samples are used. Table A3 in the Appendix reports the regression coefficients.

Comparing Table A3 and Table 3, we find similar results. The major differences are the

19

statistically significant coefficients for the appointment of Morgan Stanley as the outside

advisor and the memorandum of understanding in the first column of Table A3. In the

regression that includes firm characteristics, several coefficients lose their statistical

significance when the four sub-samples are used. However, in general, the results are not

sensitive to the possible structural changes in the relationship between individual firms’

stock prices and KOSPI.

A second robustness check recognizes the possibility that related borrowers of

Korea First Bank and Seoul Bank may be concentrated in a few industries. If this is the

case, the coefficients on the event dummies may be affected by industry-specific shocks.

Table A1 in the Appendix displays the distribution of firms over different industries for

all of the listed firms in our data base and for the firms whose main bank is either Korea

First Bank or Seoul Bank. The evidence in Table A1 does not indicate that the

distribution of related borrowers of Korea First Bank and Seoul Bank is skewed by

industry any differently from the distribution of all listed firms. Nevertheless, as a

robustness check, we estimate the following two equations that include an industry stock

index as well as KOSPI:

it

l

kkitkjtimtiiit DRRR εγββα ++++= ∑

=121

it

l

kkitik

l

kkitkjtimtiiit DXDRRR εθγββα +++++= ∑∑

== 1121

In these equations, the industry index, jtR , is included to account for industry-specific

shocks so that firms in different industries may have different normal returns.

Table 4 presents the estimation results; these are qualitatively similar to the

baseline results with a few notable changes. First, the coefficients for the announcement

20

of the memorandum of understanding are statistically significant in both regressions in

Table 4 indicating that this event yielded negative abnormal returns for related firms.

Second, the coefficient associated with the eventual acquisition of Korea First Bank by

Newbridge Capital in the first column of Table 4 is no longer statistically significant,

although its sign remains negative. Third, the coefficient indicating that liquid firms did

not lose as much value as illiquid firms in response to both the appointment of Morgan

Stanley and the announcement of the memorandum of understanding is no longer

statistically significant in the first instance and has a statistically significant unexpected

sign in the second case. In summary, controlling for industry-specific shocks corroborates

the results in the baseline model for the most part, strengthening the evidence for an

impact from the signing of the memoranda of understanding to privatize both banks on

related borrowers but weakening the evidence for the effects of the final acquisition of

Korea First Bank and for the importance of the liquidity of related borrowers.

Our third robustness check concerns the selection of the sample of related

borrowers. Clearly, the choice of firms for inclusion in the regression from among the

entire sample of 751 listed firms is critical to the results. In the baseline regression, we

use the 2000 edition of Korea Company Information to identify the firms whose main

bank is either Korea First Bank or Seoul Bank. Implicit in this choice is the notion that

related borrowers stay with their main banks despite the financial difficulty that these two

banks were experiencing during the sample period. To check the stability of the main

bank relationship over time, we consulted the 1998 edition of Korea Company

Information and linked the 566 listed firms for which we have information to a main

bank in both years. Table A2 in the Appendix presents information about the pattern of

21

change in main bank for these Korean firms. The numbers along the diagonal are the

number of firms that stay with the same main bank over this two-year period. Around

70% of related borrowers of Korea First Bank and over 80% of related borrowers of for

Seoul Bank remain with their main bank during this period.

However, since some firms obviously did change their main bank during the

sample period, we re-run our regression analysis using two different samples of related

borrowers for the two banks. First, we take the firms that identify Korea First Bank or

Seoul Bank as their main bank in the 1998 edition of Korea Company Information; these

results are reported in Table A4 in the Appendix. Second, we take the firms that identify

Korea First Bank or Seoul Bank as their main bank in both 1998 and 2000; these results

are reported in Table A5 of the Appendix. Comparisons of the coefficients in these two

tables with those in Table 3 indicate that the selection of the group of related borrowers

does not have much impact on the results. The first columns of the three tables are

remarkably similar indicating that the results of the events studies are quite robust.

When the firm characteristics are included in the regressions in the second

columns of the tables, a few differences are found. Most notably, taking the 1998 group

by itself or intersected with the 2000 group generates negative coefficients for

nationalization interacted with profitability and letter of intent interacted with

profitability. In addition, by using the intersection of the 2000 group and the 1998 group,

we find a positive and statistically significant coefficient for memoranda of

understanding interacted with reliance of short-term bank loans. The fragility of some of

the coefficients reflects the possibility that some borrowers may switch their main banks

22

during the sample period in response to the events in the privatization process.

Nonetheless, most of the coefficients are robust to the choice of sample firms.

In summary, our robustness checks indicate that the results presented in Tables 3

and 4 capture well the effects of the events surrounding the sale of these two Korean

banks to foreign owners on related borrowers. The events earlier in the sample period

have statistically significant coefficients of the expected signs while the later events tend

to have statistically insignificant coefficients. Regarding firm characteristics, we find

some evidence that more profitable and more liquid firms lose less value from negative

news and gain less from positive news taking the perspective of a borrower hoping for

the continuation soft related lending practices in its main bank. In addition, we find that

borrowers relying heavily on their main bank for financing are affected more adversely

by some of the events.

6. Conclusion

In this paper, we investigate the effects of changes in the ownership of two

insolvent Korean banks, namely Korea First Bank and Seoul Bank, on the stock prices of

their related borrowers after the 1997 financial crisis until the beginning of 2000. We

find that events signaling strongly a change of management and a sale of a bank to a

foreign financial institution yield an average decrease of about 2% in the stock price of

related borrowers over a three-day window surrounding the event. In addition, we find

that events indicating clearly a weakening in the Korean government’s resolve to re-

privatize these two banks generate an increase in the stock price of related borrowers of

about the same magnitude. These results are consistent with the hypothesis and findings

23

in the literature that related lending provides rent to borrowers in a main bank financial

system. Moreover, we find evidence to support our hypotheses that these events have

larger impacts on firms that are both less liquid and more reliant on short- term bank

loans. In a few instances, the stock prices of the more unprofitable firms react more

strongly to the events. Taken together, these results indicate that the previous owners and

managers of these two Korean banks engaged in soft related lending practices and,

perhaps, kept afloat zombie insolvent firms.

The two banks are selected because they were identified by the IMF for sale to

foreign financial institutions in a letter of intent signed with the Korean government

establishing conditions for IMF financial support during the crisis. In addition, we have

information for 2000 that allows us to link publicly traded firms in Korea with a main

bank so that we can investigate related lending at the microeconomic level. Our empirical

results are derived first in a baseline regression that measures abnormal returns based on

a market index and includes dummy variables for each event. To determine the impact of

firm characteristics, we include interactive terms matching events to three financial

measures for the related borrowers. To establish the robustness of the results, we include

industry controls to capture any bias due to a non-representative group of firms across

industries. In addition, we divide the sample period into four sub-samples and re-run the

baseline regressions to take account of possible structural breaks in the relationship

between the stock prices of firms and the market index. Finally, we use two additional

specifications for determining the group of related borrowers for these two banks using

information about the links in 1998 and combining this with the links in 2000. In general,

our statistical results are robust to these changes.

24

An important lesson to be drawn from this study is that credible prompt corrective

action can change bank behavior, end soft related lending, and reduce the cost of bank

rehabilitation. Abnormal stock price returns of related borrowers decrease upon the

announcement of an event signaling the privatization of these two Korean banks and

increase when news that the process is delayed is made public. As the privatization of

Korea First Bank proceeds to eventual purchase by Newbridge Capital at the end of

December 1999, the events lose statistical significance. This situation may occur for two

reasons, one of which is statistical. Since the privatization of Seoul Bank becomes

derailed during 1999, a smaller group of related borrowers is associated with the latter

events because only those firms having Korea First Bank as their main bank are relevant.

Hence, from the statistical perspective, standard errors increase. In addition, we presume

that adjustments take place throughout the period so that, by the time the actual

privatization occurs, the stock prices of the related firms incorporate the correct

expectation that soft lending will no longer be practiced by bank management because

the foreign owner will make loans on a commercial basis only. Hence, we conclude that

credibility is the most important component to a government’s commitment to prompt

credible action so as to stem a banking crisis. Credibility is best established by sale of the

bank to a foreign owner in situations in which relational lending has become common

practice for a domestic bank.

25

References

Bae, Kee-Hong; Kang, Jun-Koo; Lim, Chan-Woo. The Value of Durable Bank Relationships: Evidence from Korean Banking Shocks. Journal of Financial Economics. Vol. 64 (2). p 181-214. May 2002.

Choe, Heungsik; Lee, Bong-Soo. Korean Bank Governance Reform after the Asian

Financial Crisis. Pacific-Basin Finance Journal. Vol. 11 (4). p 483-508. September 2003.

Choi, Sungho and Hasan, Iftekhar. Ownership, Governance, and Bank Performance:

Korean Experience. unpublished working paper, Rensselaer Polytechnic Institute Cull, Robert; Matesova, Jana; Shirley, Mary. Ownership and the Temptation to Loot:

Evidence from Privatized Firms in the Czech Republic. Journal of Comparative Economics. Vol. 30 (1). p 1-24. March 2002.

Djankov, Simeon, Jindra, Jan, and Klapper, Leora. Corporate Valuation and the Resolution of

Bank Insolvency in East Asia. unpublished working paper, World Bank, 2001. Jeon, Yongil; Miller, Stephen M. The Effect of the Asian Financial Crisis on the

Performance of Korean Nationwide Banks. Applied Financial Economics. Vol. 14 (5). p 351-60. March 2004.

La Porta, Rafael; Lopez-de-Silanes, Florencio; Zamarripa, Guillermo. Related Lending.

The Quarterly Journal of Economics. Vol. 118 (1). p 231-68. February 2003. Laeven, Luc. Insider Lending and Bank Ownership: The Case of Russia. Journal of

Comparative Economics. Vol. 29 (2). p 207-29. June 2001. MacKinlay, A Craig. Event Studies in Economics and Finance. Journal of Economic

Literature. Vol. 35 (1). p 13-39. March 1997. Peek, Joe; Rosengren, Eric S. Unnatural Selection: Perverse Incentives and the

Misallocation of Credit in Japan. National Bureau of Economic Research, Inc, NBER Working Papers: 9643. 2003.

Rajan, Raghuram and Zingales, Luigi. Which Capitalism? Lessons from East Asian

Crisis. Journal of Applied Corporate Finance, 11(3), p 40-48. Fall 1998.

26

Table 1: Relevant Events Source: Lexus-Nexus database. All dates are Korean dates.

Date Event December 9, 1997

Korean government became the majority owner of Korea First Bank and Seoul Bank and promised stringent restructuring of the banks with 1,500 workers to be laid off.

December 26, 1997 Korean government and the IMF agreed on the letter of intent, which aims at prompt restructuring and eventual sales of Korea First Bank and Seoul Bank to foreign banks.

April 22, 1998 Korean government appointed Morgan Stanley as a lead manager for restructuring and privatization.

November 5, 1998 Korean government postponed the sales of Korea First Bank and Seoul Bank due to the difficulty in obtaining foreign bids.

December 28, 1998 Korean government signed the memorandum of understanding with Newbridge Capital for the sale of Korea First Bank.

February 22, 1999 Korean government signed the memorandum of understanding with HSBC for the sale of Seoul.

April 30, 1999 (Korea First Bank) May 31, 1999 (Seoul Bank)

Korean government failed to reach any agreement with foreign institutions by the first deadline.

May 12, 1999 (Korea First Bank) June 28, 1999 (Seoul Bank)

Korean government failed to reach any agreement with foreign institutions by the second deadline.

July 1, 1999 Korean government agreed with Newbridge on the sale of Korea First Bank

December 23, 1999 Newbridge acquired 51% of Korea First Bank’s share.

27

Table 2: Descriptive Statistics Variable Mean Std.

Dev. Min Max

Daily Return (% Change of Stock Price)

.08

5.96

-53.57

69.20

Liquidity (Ratio of Cash plus Bank Deposits to Assets)

.14

.16

.01

.68

Profitability (Ratio of Gross Profit to Assets)

.21

.17

-.01

.75

Reliance of Short-Term Bank Loans (Ratio of Short-Term Bank Loans to Assets)

.25

.17

0

.76

Data Source: PACAP

28

Table 3: Estimated Abnormal Returns: Baseline Model The multiple regression models are estimated with ordinary least squares. The standard

errors are adjusted for heteroskedasticity and contemporaneous correlation across firms

using STATA’s cluster option. The dependent variable is the daily percentage change in

stock price of firms whose main bank is reported to be either Korea First Bank or Seoul

Bank. Firm-specific intercepts and risk coefficients on the Korea Stock Price Index are

included but not reported to keep the table relatively uncluttered.

Abnormal

Returns Abnormal Returns Interacted with Firm Characteristics

Nationalization (12/9/97)

-2.203*

-1.261

(1.182) (1.804)

Letter of Intent with IMF (12/26/97) -2.448*** -5.335***

(0.791) (1.787)

Morgan Stanley Appointed (4/22/98) -1.087 -0.629

(0.939) (0.868)

Sales Postponed (11/5/98) 2.537*** 1.855***

(0.728) (0.372)

Memorandum of Understanding (12/28/98 for KFB and 2/22/98 for Seoul Bank)

-0.567 (0.446)

-1.260 (1.082)

1st Deadline (4/30/1999 for KFB and 5/31/99 for Seoul Bank)

1.626* (0.913)

0.802 (0.667)

2nd Deadline (5/12/1999 for KFB and 6/28/99 for Seoul Bank)

-0.247 (0.603)

0.674 (0.637)

Agreement to Sell KFB (7/1/99) 0.223 0.133

(0.624) (0.342)

Aquisition of KFB (12/23/99) -2.335** -1.852

(0.955) (1.990)

Nationalization*Profit -2.426

(1.539)

Nationalization*Liquidity 8.924***

(3.007)

Nationalization*STLoan -4.738***

(1.597)

29

Letter of Intent*Profit

1.855***

(0.558)

Letter of Intent*Liquidity 9.418**

(4.504)

Letter of Intent*STLoan -9.780***

(2.254)

Morgan Stanley*Profit 0.845

(1.611)

Morgan Stanley*Liquidity 2.262***

(0.660)

Morgan Stanley*STloan -4.449*

(2.396)

Sales Postponed*Profit -5.430***

(0.468)

Sales Postponed*Liquidity 1.817

(1.480)

Sales Postponed*STLoan 10.888***

(2.130)

Memorandum*Profit 0.890

(2.441)

Memorandum*Liquidity -0.997

(0.625)

Memorandum*STLoan 1.821

(1.468)

1st Deadline*Profit -1.083

(1.837)

1st Deadline*Liquidity -1.690

(2.223)

1st Deadline*STLoan 6.024***

(1.625)

2nd Deadline*Profit -3.956***

(1.264)

2nd Deadline*Liquidity 0.231

(2.054)

2nd Deadline*STLoan 2.726

(1.900)

30

Agreement*Profit 0.599

(1.995)

Agreement*Liquidity 2.759

(5.514)

Agreement*STLoan -2.132

(4.395)

Aquisition*Profit -5.188

(3.581)

Aquisition*Liquidity -1.827

(6.477)

Aquisition*STLoan 7.227

(7.480)

Constant 0.172 0.181

(0.224) (0.225)

Observations

64063 64063

R-squared

0.13 0.13

Notes

i. Robust standard errors are in parentheses. ii. The symbols *, **, and *** represent significance at the 10%, 5%, and 1%

levels, respectively.

31

Table 4: Estimated Abnormal Returns with Control for Industry Index

The multiple regression models are estimated with ordinary least squares. The standard






Abnormal



-2.000***

-1.404

(0.576) (1.320)

Letter of Intent with IMF (12/26/97) -1.579*** -2.595***

(0.579) (0.817)


(0.460) (0.722)

Sales Postponed (11/5/98) 2.299*** 1.866***

(0.679) (0.544)


-0.825** (0.373)

-1.706* (1.005)


1.388** (0.646)

0.615 (0.582)


-0.054 (0.443)

0.876 (0.588)

Agreement to Sell KFB (7/1/99) -0.123 -0.437***

(0.191) (0.169)

Aquisition of KFB (12/23/99) -1.402 -0.956

(0.879) (1.832)


(1.030)

Nationalization*Liquidity 7.550**

(3.047)


(0.753)

32


0.693***

(0.248)


(4.183)


(2.109)


(1.504)

Morgan Stanley*Liquidity 1.098

(1.172)

Morgan Stanley*STloan -3.068**

(1.324)


(1.088)


(1.702)


(2.439)


(2.326)

Memorandum*Liquidity -2.136**

(0.897)

Memorandum*STLoan -0.091

(1.124)


(1.660)


(1.905)

1st Deadline*STLoan 5.444***

(1.326)

2nd Deadline*Profit -3.768***

(1.376)

2nd Deadline*Liquidity -0.433

(2.066)


(2.041)

33

Agreement*Profit

2.114

(2.477)


(6.501)


(5.682)


(3.903)


(7.598)


(7.673)


(0.225) (0.225)

Observations

62211 62211

R-squared

0.20 0.20

Notes



34

Table A1: Industry Composition of Related Firms Industry Other Listed

Firms Firms with KFB and Seoul

Bank Total Listed

Firms Chemicals

87

12

99

Communication 1 1 2 Construction 44 14 58 Distribution 52 11 63 Electrical & Elect. 96 16 112 Electricity & Gas 10 0 10 Fishing 3 1 4 Foods & Beverag. 47 5 52 Iron & Metal 36 9 45 Machinery 41 4 45 Medical 9 0 9 Medical Supplies 31 7 38 Mining 3 0 3 Non-metallic 26 2 28 Other Manufactur 22 2 24 Paper & Wood 28 6 34 Textile & Wearing

58 10 68

Transport & Stora.

17 0 17

Transport Equip.

34 6 40

Total 645 106 751 Data Source: PACAP

35

Table A2: Main Bank Relationships: Number of Firms for Each Bank

Main Bank:2000

Main Bank:1998 CHOHUNG HANA HANVIT KOOKMIN KORAM KEB KFB PEACE SEOUL SHINHAN Other Totals: 1998

CHOHUNG 76 2 5 0 2 2 1 0 1 2 5 96

HANA 0 3 0 0 0 0 0 0 0 0 0 3HANVIT 1 2 172 0 1 3 7 0 3 2 4 195

KOOKMIN 0 0 2 5 1 0 0 0 1 1 0 10

KORAM 2 0 1 0 16 0 0 0 0 0 0 19

KEB 4 0 3 0 0 62 1 0 0 1 3 74KFB 1 0 12 0 4 3 54 0 0 2 7 83

PEACE 0 0 1 0 0 0 0 0 0 0 0 1

SEOUL 1 0 5 1 0 1 2 0 45 2 2 59

SHINHAN 1 0 1 1 0 1 2 0 0 21 1 28

Other 3 1 6 0 1 1 2 1 2 3 20

Totals: 2000 89 8 208 7 25 73 69 1 52 34 22 566

Source: Korea Company Information 1998 and 2000

36

Table A3: Estimated Abnormal Returns Allowing for Potential Structural Breaks The multiple regression models are estimated with ordinary least squares. The standard






Abnormal



-1.798***

-1.047

(0.315) (0.904)

Letter of Intent with IMF (12/26/97) -2.467*** -4.819

(0.303) (4.777)

Morgan Stanley Appointed (4/22/98) -0.821*** -0.521

(0.302) (0.883)

Sales Postponed (11/5/98) 2.384*** 1.821**

(0.315) (0.921)


-0.727** (0.334)

-1.379 (0.983)


1.372*** (0.335)

0.509 (0.991)


-0.512 (0.336)

0.312 (0.997)

Agreement to Sell KFB (7/1/99) -0.000 0.027

(0.457) (1.726)

Aquisition of KFB (12/23/99) -1.799*** -1.608

(0.464) (1.743)


(2.273)


(3.015)

Nationalization*STLoan -4.980**

(2.428)

37


1.530

(2.244)

Letter of Intent*Liquidity 9.455***

(2.963)


(2.359)


(2.232)


(2.969)

Morgan Stanley*STloan -4.106*

(2.407)

Sales Postponed*Profit -5.255**

(2.330)


(3.062)


(2.454)


(2.513)


(3.206)


(2.555)


(2.674)


(3.227)

1st Deadline*STLoan 5.455**

(2.576)

2nd Deadline*Profit -3.440

(2.678)


(3.244)


(2.608)

38

Agreement*Profit

0.226

(5.528)


(5.300)


(3.532)


(5.567)


(5.334)

Aquisition*STLoan 7.864**

(3.626)

Notes


levels, respectively. iii. R-squares are 0.25, 0.13, 0.12, and 0.06 for four separate sample periods.

39

Table A4: Estimated Abnormal Returns with Main Bank as in 1998

The multiple regression models are estimated with ordinary least squares. The standard






Abnormal



-2.417**

-1.586

(1.177) (1.748)

Letter of Intent with IMF (12/26/97) -2.187*** -1.180

(0.738) (0.810)

Morgan Stanley Appointed (4/22/98) -1.426 -1.496*

(0.973) (0.800)

Sales Postponed (11/5/98) 2.285*** 2.350***

(0.609) (0.607)


-0.295 (0.377)

-1.168 (0.780)


1.713* (1.025)

1.731 (1.061)


-0.018 (0.493)

0.134 (0.959)

Agreement to Sell KFB (7/1/99) 0.669 1.464***

(0.635) (0.356)

Aquisition of KFB (12/23/99) -1.835** -4.286*

(0.878) (2.221)


(1.727)


(1.386)

Nationalization*STLoan -2.518

(1.601)

40


-1.122***

(0.341)


(4.406)


(2.455)

Morgan Stanley*Profit 3.557***

(1.039)


(1.927)

Morgan Stanley*STloan -6.998***

(0.990)


(0.599)


(1.226)


(2.009)


(3.452)


(1.488)


(3.007)


(1.868)


(1.724)

1st Deadline*STLoan 2.386

(1.464)

2nd Deadline*Profit -1.375

(2.181)

2nd Deadline*Liquidity -0.228

(2.184)

2nd Deadline*STLoan 1.872**

(0.895)

41

Agreement*Profit

-1.207

(2.694)


(2.564)


(2.977)

Aquisition*Profit 1.520

(4.474)


(6.499)


(6.187)


(0.224) (0.225)

Observations

76785 76785

R-squared

0.13 0.13

Notes



42

Table A5: Estimated Abnormal Returns: Main Bank Intersection 1998 and 2000 The multiple regression models are estimated with ordinary least squares. The standard






Abnormal



-2.226*

-0.100

(1.136) (1.803)

Letter of Intent with IMF (12/26/97) -2.472*** -1.413**

(0.775) (0.611)


(0.920) (0.622)

Sales Postponed (11/5/98) 2.285*** 1.495***

(0.699) (0.463)


-0.313 (0.377)

-1.190 (1.001)


1.505* (0.844)

1.031* (0.546)


-0.196 (0.526)

0.552 (0.759)

Agreement to Sell KFB (7/1/99) 0.531 0.810

(0.650) (0.767)

Aquisition of KFB (12/23/99) -2.068* -1.773

(1.061) (3.322)

Nationalization*Profit -5.933***

(1.495)


(3.016)


(1.319)

43


-0.018

(0.447)


(4.419)


(2.361)


(1.176)

Morgan Stanley*Liquidity 2.065***

(0.788)

Morgan Stanley*STloan -5.807***

(2.041)


(0.809)


(1.406)


(1.926)

Memorandum*Profit -0.463

(2.538)

Memorandum*Liquidity 0.008

(0.520)

Memorandum*STLoan 4.298**

(1.830)


(1.241)


(2.183)

1st Deadline*STLoan 6.017**

(2.519)

2nd Deadline*Profit -3.336**

(1.473)


(2.056)


(1.574)

44

Agreement*Profit

-0.995

(5.321)


(4.667)


(6.322)


(5.927)


(8.707)


(9.198)


(0.224) (0.225)

Observations

53513 53513

R-squared

0.13 0.13

Notes



soft related lending: a tale of two korean banks *

Documents