Corporate Governance and Firm Profitability:

Evidence from Korea before the economic crisis

Sung Wook Joh*

December 2001

Forthcoming in Journal of Financial Economics

Abstract
This study examines how ownership structure and conflicts of interest among
shareholders under a poor corporate governance system affected firm performance
before the crisis. Using 5,829 Korean firms subject to outside auditing during 1993-
1997, the paper finds that firms with low ownership concentration show low firm
profitability, controlling for firm and industry characteristics. Controlling shareholders
expropriated firm resources even when their ownership concentration was small. Firms
with a high disparity between control rights and ownership rights showed low
profitability. When a business group transferred resources from a subsidiary to another,
they were often wasted, suggesting that “tunneling” occurred. In addition, the negative
effects of control-ownership disparity and internal capital market inefficiency were
stronger in publicly traded firms than in privately held ones.

JEL classification code: G3

Key words: corporate governance, ownership, profitability, shareholder expropriation,
business group

*I am grateful to the participants at the 12th Annual NBER Seminar on the East Asian Economics and the
World congress meeting of the econometric society for comments on the earlier version of this paper. I
am especially grateful to Simon Johnson and an anonymous referee for their valuable suggestions. Their
helpful comments have greatly improved the paper. However, all remaining errors are mine.
Correspondence address: Sung Wook Joh, KDI, Tongdaemun-gu, Cheongryangri, Seoul, 130-012 Korea.
Tel) +82-2-958-4132 Fax) +82-2-964-5479 Email) joh@[Link]
1. Introduction

Many countries that suffered during the recent economic crises in Asia and other

emerging markets had weak legal environments and poor governance systems. This

observation has triggered much discussion on the importance of corporate governance.

For example, Johnson, Boone, Breach, and Friedman (2000) show that countries with

weak legal protections suffered greater exchange rate depreciation and severer stock

market declines during the crisis. Using firm level data, Mitton (2002) shows that

corporate governance measures, such as high disclosure quality and concentrate

ownership, affected stock market valuation during the crisis. Lemmon and Lins (2002)

show that, during the crisis, firms showed low performance when their controlling

managers had more control rights than ownership rights.

Most prior research has focused on the effects of corporate governance structure during

the crisis. In contrast, this study examines its effects before the crisis. Aghion,

Bacchetta, and Banerjee (2000) and Krugman (1999) argue that the financial distress of

firms helped cause the crisis. If poor corporate governance helped lower firm value and

financial survivability before the crisis, it arguably increased the economy's aggregate

vulnerability.

This study examines whether a firm's ownership structure affected its performance

before the crisis. Using ownership and financial data of firms in a country that

experienced the crisis, I examine whether owners with more control rights than

1
ownership rights expropriated firm resources before the crisis. When controlling

shareholders' control rights exceed their ownership rights they have an incentive to

expropriate firm resources, as their private benefits exceed their costs. Furthermore,

expropriation is more likely when the disparity between control and ownership is large

and when their position is secure. Firms experiencing greater expropriation of resources

likely show lower performance.

I also investigate whether these effects are stronger in business groups. Controlling

shareholders in business groups can maintain their control with the help of indirect

pyramidal ownership (La Porta, Lopes-de-Silanes, and Shleifer (hereafter referred to as

LLS), 1999; Claessens, Djankov, and Lang, 2000). These controlling shareholders

therefore have greater incentives and means to expropriate firm resources than their

counterparts in independent firms. In addition, firms affiliated with business groups can

suffer more, as their controlling shareholders have more tools to divert firm resources

through the transfer of assets from one subsidiary to another. Lamont (1997), Shin and

Stulz (1998), Scharfstein and Stein (2000), and Scharfstein (1998) argue that

multidivisional firms overinvest capital in weak divisions and underinvest it in stronger

ones. This study examines whether business groups have lower firm profitability than

independent firms do. This study also identifies a mechanism through which

controlling shareholders can waste firm resources --investment in affiliated firms.

I used 5,829 Korean firms subject to outside auditing during the pre-crisis period of

1993-1997; I have selected the Korean case for the following reasons. Many have

2
argued that Korea’s poor corporate governance system helped cause the 1997 crisis.

Moreover, the Korean economy can be characterized by the prevalence of business

groups (chaebols) that consist of legally independent, horizontally and vertically

distributed firms. According to Joh (2001a), more than 60% of Korean firms subject to

outside auditing belong to business groups (including small ones). Chung and Yang

(1992) report that the largest 30 chaebols produced 35.4% of total output and 16.3% of

GDP in 1989, respectively. The OECD (1998) reports that the top 30 chaebols

accounted for 40.2% of the value added in the manufacturing sector in 1995.

A cross-sectional, time-series country analysis can be more advantageous than cross-

country analyses. A country analysis avoids endogeneity problems. Cross-country

analyses can also underestimate the importance of country-specific laws. This study

examines the performance of firms operating with the same legal institutions, corporate

governance environments, macro- and developmental economic stages, accounting

standards, etc. Firms in the same country face the same legal protection and institutional

constraints, or lack thereof; so a country analysis can avoid the endogeneity problems

between ownership structure and institutional environments. In particular, many studies

have shown that differences in legal institutions explain much of the cross-country

ownership differences (La Porta, Lopes-de-Silanes, Shleifer, and Vishny (hereafter

referred to as LLSV) 1997, 2000, 2002). For example, in Korea until 1994, by article

200 of the security exchange law only existing incumbent controlling shareholders were

allowed to hold more than 10% of shares. In addition, by law foreign ownership was

restricted until the end of 1997; foreign individual investors were forbidden to hold

3
more than 7% of shares, and foreign ownership as a group could not exceed 26% of

total shares. Together, these laws protected the incumbent controlling shareholders from

outside investors.

Many cross-country studies likewise do not account for important, country-specific laws

that affect measures of ownership and control rights. Consider the Korean law requiring

“shadow voting” and a mandatory tender offer, for example. Korean financial

institutions were legally forbidden from affecting outcomes in corporate voting

decisions (“shadow voting”). Also, anyone acquiring 25% of a firm's shares had to

tender an offer on at least 50% of its shares. This mandatory tender offer protected

incumbent controlling shareholders from plurality ownership takeovers. Moreover, both

hostile and foreign takeovers were prohibited until 1998. As these laws reduced the

threshold shareholding needed to control a firm, they helped Korean controlling

shareholders maintain their control despite their small ownership stakes.

This study measures firm performance through its profitability. For this analysis of

Korean data, accounting profitability is likely a better performance measure than stock

market-based measures for at least three reasons. First, researchers have shown some

market inefficiencies even in the most developed countries (De Bondt and Thaler, 1985,

1987; Lo and MacKinlay, 1988; Conrad and Kaul, 1998). Developing countries also

show stock market inefficiency.1 Thus, stock prices in Korea are not likely to reflect all

available information. Second, Mossman, Bell, Swart, and Turtle (1998) show that a

1
For some evidence of inefficiency in emerging markets, see Butler and Malaikah (1992), and Kim and

4
firm's accounting profitability is more directly related to its financial survivability than

its stock market value is. Many studies use accounting measures to predict bankruptcy

(Altman, 1968; Takahashi, Kurokawa and Watase, 1984) or financial distress (Hoshi,

Kashyap and Sharfstein; 1991). Third, accounting measures allow us to evaluate the

performance of privately held firms as well as that of publicly traded firms.

The results show that a firm’s profitability is lower when the controlling family’s

ownership is lower, controlling for firm, industry, and macro-economic effects.

Likewise, firm profitability was low for firms when the difference between control

rights and cash flow rights was high. The paper also provides some evidence of

nonlinearity of ownership effects on firm profitability. Independent firms outperformed

firms affiliated with large business groups. In addition, resources transferred from firms

in business groups to affiliated firms lowered firm profitability. Moreover, negative

effects of control-ownership disparity and internal capital market inefficiency were

stronger in publicly traded firms than in those privately held.

These results contrast with Lemmon and Lins' (2002) study showing no significant

ownership effects on the changes in Tobin's Q before the crisis in East Asian countries.

The results likely differ because this study uses a better detector of financial distress

(accounting profitability), accounts for more country-specific factors, controls for more

firm and industry characteristics, examines longitudinal data, and uses continuous

control and ownership rights variables instead of a dummy variable. (See the data

Singal (1997).

5
section for more details). In addition, lower performance of large chaebols contradicts

Khanna and Palepu’s (2000) recent argument that firms affiliated with large diversified

business groups perform better than independent firms in emerging markets. This result

suggests that the advantages of business groups disappear while their disadvantages

may increase as the economy develops further.

The paper is organized as follows. I briefly discuss Korean corporate sector problems

before the crisis. Then I discuss ownership structure and the determinants of firm

profitability. Next, I describe the data used in the study. Finally, I present and discuss

the results.

2. The crisis and corporate sector problems

High debt-equity ratios and low profitability in Korean firms persisted for many years,

unobstructed by a weak corporate governance system. Together, these factors helped

increase the vulnerability of the corporate sector before the 1997 economic crisis.

Before the Korean crisis, the corporate sector showed very high debt-equity ratios and

low profitability. In 1997, the average debt-equity ratio of Korean firms far exceeded

that of other countries (Korea, 396%; U.S., 154%; Japan, 193%; and Taiwan, 86%).2 As

Joh (2001b) shows, the average debt-equity ratio of Korean firms has been very high for

a long time. It did not sharply increase in recent years.

2
For Taiwanese firms, the figure is based on 1996 data. Source: Bank of Korea’s Financial Statement

6
With high debt-equity ratios, Korean firms were expected to yield high profitability on

their equity. Yet, the average rate of return on equity was often lower than the

prevailing interest rates for loans (see Figure 1). The return on capital fell short of its

opportunity cost for almost 10 years before the crisis. While the average profitability

was much lower in 1996 and in 1997 compared to the opportunity cost and compared

with previous years, there does not appear to be a sudden drop when other factors are

controlled for. Moreover, variation across firms for 1996 and for 1997 does not seem to

be different in a systematic way than in earlier years (see Section 6.2). When firm-

specific factors, industry-specific factors, and macro economic conditions are controlled

for, profitability has been declining over time (see Section 6.3). Krueger and Yoo (2001)

show that corporate performance has deteriorated over time. They also showed that the

rate of return on assets (ROA) of the Korean manufacturing sector has been lower than

that of other countries, such as Japan, Germany, the United States, and Taiwan. Such

chronic low profitability suggests that, on average, capital was wasted on unprofitable

projects.

<insert Figure 1 around here>

Both the external and internal components of Korea's corporate governance system

failed to provide sufficient monitoring and discipline to end this waste (Joh, 2001b).

Krueger and Yoo (2001) do not examine whether and why there has been a large

Analysis for 1997.

7
variation across firms. Joh (2001b) argues that unlike small firms, large firms faced few,

if any, exit threats. Through repeated rescues of weak large firms, governments

implicitly guaranteed large firms. Korean laws protected incumbent controlling

shareholders. By prohibiting both hostile and foreign M&As. Also, anyone acquiring

25% of a firm's shares had to tender an offer on at least 50% of its shares, this

mandatory tender offer virtually prevented plurality ownership takeovers. Furthermore,

opaque accounting and management prevented banks and investors from receiving

accurate firm information. Internal governance systems did not properly monitor firm

management either. For example, controlling shareholders selected most of the directors

on the board (including outside directors) (see Jun and Gong, 1995; Seoul National

University Management Research Institute, 1985). Consequently, outside directors

rarely opposed agenda items and often did not attend meetings regarding major

transactions with controlling shareholders.3

This combination of high debt-equity ratios and low profitability was not sustainable.

Six of the thirty largest business groups (chaebols) went bankrupt before the currency

crisis,4 triggering a cascade of nonperforming loans. Starting with the default by Hanbo

(ranked #14 in 1995) in January 1997, a series of large chaebols’ defaults raised

suspicion regarding the conglomerates’ survival and the fundamental soundness of the

corporate sector. Due to their size and importance, the failure of these chaebols

3
Despite the introduction of new rules regarding directors after the economic crisis, Joh (2001b) shows
that outside directors routinely approved agenda items (over 99%). Outside directors rarely attended
meetings regarding transactions with controlling shareholders. The attendance rate was lower than 37%.
For more information on outside directors, see Joh (2001b).
4
In 1997, six of the 30 largest conglomerates went bankrupt: Hanbo, Sammi, Jinro, KIA, Haitai, and New
Core. Their default dates were January 23, March 19, April 21, July 15, November 1, and November 4,

8
devastated the economy. Their failures led to a series of subcontractor bankruptcies and

sharply increased the number of nonperforming loans. By the end of 1997, 6.7% of all

loans were nonperforming loans, totaling 64.7 trillion won (over $45.6 billion)

according to the end-of-year exchange rate, 1,415 won per dollar. By June 1998, over

10% of all loans were nonperforming loans (Financial Supervisory Commission,

Various Press Releases.). These nonperforming loans severely weakened many banks

and eventually provoked the liquidity crisis.

3. Ownership structure of Korean firms

The largest shareholder, usually the founder, typically controls a Korean firm (Seou

National University Management Research Institute, 1985; Lim, 1989). In over 80% of

large firms, the largest and controlling shareholder or family members are among the

top executives (Claessens, Djankov, and Lang, 2000). The other 20% are likely to be

state-controlled enterprises and financial institutions. Even when a hired professional

CEO manages the firm, his decision-making power and scope are often quite limited

Seoul National University Management Research Institute, 1985).

In large firms, Korean controlling shareholders and their families often own little equity,

forming a controlling minority structure as described by Bebchuck, Kraakman, and

Triantis (1999). As Table 1 shows, the largest shareholder and family of a publicly

traded firm owns less than 31.7% of shares on average (less than 13% of asset-weighted

respectively.

9
shares). In contrast, controlling shareholders of privately held (private) firms own

49.5% of shares on average.

<Insert Table 1 around here>

Despite their low ownership, controlling shareholders in publicly traded firms maintain

control for at least two reasons. First, ownership among individual shareholders is

dispersed. Large firms in Korea often have more dispersed ownership than those in most

other East Asian countries (LLS, 1999; Claessens, Djankov, and Lang., 2000). In 1997,

the aggregate individual ownership was about 40% of shares. More than 95% of all

shareholders were small individual shareholders holding less than 1% of total shares.

More than 80% of all shareholders owned less than 1,000 shares. Most individual

shareholders owned less than 500 shares. Aggregate individual ownership has been

large but declining over time, from 60% in the 1980s to less than 40% in 1997 (Korea

Stock Exchange, 1999). As most shareholder rights required at least 5% ownership,

these small shareholders could not easily oppose the controlling shareholder.

Second, institutional shareholders did not monitor firm activity, despite owning over

40% of the shares.5 Banks, virtually controlled by the government, held around 10% of

listed firm shares. Other financial institutions owned more than 10%. These included

insurance companies, security firms, and investment trust companies. The “shadow
6
voting” rule forbade all financial institutions from voting on firm decisions.

5
In 1997, the government owned 6.6%, and foreigners owned 9.1%. (Korea Stock Exchange; 1999).
6
Depositors and investors ultimately own shares owned by financial institutions. Votes by financial

10
Nonfinancial corporations held more than 20% of shares. Most nonfinancial corporation

ownership consisted of cross-holding or interlocking ownership of affiliated firms. Thus,

nonfinancial institutional investors often protected the incumbent controlling

shareholder from potential outside threats.

Using interlocking ownership, controlling shareholders of business groups (chaebols)

maintained control despite owning even less than their counterparts in independent

firms. Table 1 shows that the average controlling shareholder ownership stake was

29.2% for firms affiliated with business groups. The average ownership concentration

of controlling shareholders of the 70 largest chaebols was 17.1%. Within the largest

chaebols, larger firms showed lower ownership. For example, the asset-weighted

average ownership for the 70 chaebol-affiliated firms was 9.9%. Direct ownership

understates the true controlling shareholders’ ownership since it does not take into

account their stakes in other affiliated firms that hold shares of the firm. However, as the

overall controlling shareholders’ ownership in a group is small, the difference is also

small.

The controlling shareholder controlled chaebol-affiliated firms through extensive

interlocking institutional ownership, including many cross-holdings. Direct interlocking

ownership, in which firm A owns firm B and firm B owns firm A, is illegal, so chaebol-

affiliated firms used complex, pyramidal, or circular patterns of institutional

interlocking ownership. Several de facto holding firms (holding companies were not

institutions are allocated in the same proportion as the nonfinancial institution votes. This rule was

11
allowed until 1998) owned a large portion of affiliated firms’ stock. The indirect

ownership and institutional ownership that the controlling shareholders in business

groups essentially controlled was 32.5%. In particular, for the 70 largest chaebol-

affiliated firms, controlling shareholders essentially controlled over 43.5%.

In short, shareholders exercised control far beyond their ownership stake. They

exploited dispersed ownership and inadequate monitoring by institutional shareholders.

Moreover, controlling shareholders in chaebols maintained control with even less

ownership by exploiting affiliated firms’ interlocking ownership.

4. Determinants of firm profitability

In firms with high disparities between control and ownership rights, conflicts of interest

among shareholders can affect performance. A firm's organizational structure can

exacerbate these conflicts. Other factors that affect firm profitability include firm

attributes such as financial structure, size, market share, and business strategy. Industry

and macro-economic attributes are also included. For a brief summary of how size,

industry attributes, and firm attributes (such as market share and business strategy)

affect firm performance, see Martin (1993).

4.1 Control-ownership disparity

In a firm with a high control-ownership disparity, a controlling shareholder exercises

abolished in 1998.

12
control but owns only a small fraction of the firm’s cash flow. Bebchuck, Kraakman,

and Triantis (1999) call it a controlling minority structure. LLSV (2002) find that this

ownership structure is widespread around the world. In this firm, the controlling

shareholder's benefit from appropriating firm resources exceeds its cost. As a result,

these controlling shareholders have an incentive to pursue their private benefits at the

expense of other shareholders. Jensen and Meckling (1976) argue that this tendency

increases when the controlling shareholder owns less. Morck, Shleifer, and Vishny

(1988) argue that such effects do not have a monotonic relationship. In general, as the

control-ownership disparity increases, controlling shareholders appropriate more firm

resources (Shleifer and Vishny, 1997). Thus, conflicts of interest among shareholders

can lower firm performance. Using data from 27 countries, LLSV (2002) show that

firms with high ownership concentration show high Tobin’s Q. Lemmon and Lins

(2002) show that firms with greater separation of control and ownership rights had

severer firm devaluation (as measured in Tobin's Q and stock market returns) during the

crisis. Mitton (2002) shows that firms with high disclosure quality and ownership

concentration showed better stock market performance during the Asian economic crisis.

4.2 Firm organization

Firms affiliated with business groups are prevalent in emerging markets because they

have advantages over independent firms through intragroup trading and internal capital

markets (Leff, 1978; Hubbard and Palia, 1999; Khanna and Palepu 2000). 1 Leff (1978)

argues that business groups in less developed countries appropriate quasi-rents accrued

from control of scarce inputs. Through diversification, business groups can reduce risk

13
and uncertainty in firm operations, thereby lowering default and bankruptcy risks.

Business groups can reduce Williamsonian transaction costs through intra-group trading

(Chang and Choi, 1988), thereby overcoming imperfections in the labor and product

markets in less developed economies. Moreover, a business group can exploit its large

size to borrow money at a lower cost. It can then operate an internal capital market for

its subsidiaries, acting as the headquarters of a multidivision firm.

Yet business groups' advantages can decrease as the economy develops. In a competitive

product market, intragroup transactions are less attractive. Without competition, the

seller with a captive buyer has less incentive to lower costs and improve quality.

Likewise, a developed external capital market erodes the advantage of an internal

market. Lamont (1996), Scharfstein (1998), Shin and Stulz (1998), and Scharfstein and

Stein (2000) argue that internal capital markets in diversified conglomerates are often

inefficient. They claime that these firms overinvest capital in weak divisions and

underinvest it in stronger ones. Thus, internal capital markets tend to lower the value of

multidivisional, diversified firms. Several studies support these arguments. Wernerfelt

and Montgomery (1988), Lichtenberg (1992), and Lang and Stulz (1994) show that the

firm value and productivity of specialized firms often exceeds those of diversified

conglomerates.

Moreover, a controlling shareholder with uneven ownership of subsidiaries can exploit

the internal capital market and intragroup trading for private gain. As business groups

provide many opportunities for intragroup transactions, controlling shareholders can

14
expropriate minority shareholders through engaging in “tunneling” (Johnson, La Porta,

Lopes-de-Silanies, and Shleifer, 2000). Controlling shareholders of business groups can

move away resources for their private benefits, such as self-dealing, and divert

resources from one subsidiary in which they own less to firms in which they own more,

resulting in inefficient investment.

In sum, a business group's intragroup trading and internal capital market can provide

advantages in a less developed economy, but its disadvantages include inefficient capital

investment and expropriation by a controlling shareholder with uneven ownership.

Furthermore, a developed economy with a competitive product market and an external

capital market reduces the advantages of intragroup trading and an internal capital

market.

4.3 Financial structure

Past literature has shown mixed effects of debt on firm profitability. Debt affects

profitability positively in Hurdle (1974), but negatively in Hall and Weiss (1967) and in

Gale (1972). Debt may yield a disciplinary effect when free cash flow exists (Jensen,

1986; Stulz, 1990). A rise in debt increases default risk, firms can reduce wasteful

investment and increase firm performance to secure their survival. On the other hand,

debt can increase conflicts of interest over risk and return between creditors and equity

holders. Facing large debts, equity holders with limited liability may encourage the

firm to undertake overly risky projects (Stiglitz and Weiss, 1981).

15
5. Data

This study uses financial and ownership data from the National Information and Credit

Evaluation's (NICE) database.7 Each firm gives its financial statement to the Korea

Securities Supervisory Board. Upon receiving the financial data from the board, NICE

checks the integrity of the data. Financial statements are checked more carefully than

ownership information. Ownership information requires special care and attention as

relationships of large individual owners are sometimes misreported.8 As all firms in the

same country are subject to the same accounting standards, the potential problems

associated with poor accounting practices will likely be smaller than those in cross-

country studies. To further reduce the likelihood of these problems, this study only uses

firms subject to outside auditing. About 24% of the firms in this data set were publicly

traded. In 1997, there were 1,135 publicly traded firms; 776 were listed on the Korea

Stock Exchange, and the rest were registered with the Korea Securities Dealers

Automated Quotation (KOSDAQ). All the firms used in the analyses had at least 6

billion won in assets in 1997. The data includes 5,829 firms in the standard four-digit

Korean industrial classifications between 1993 and 1997. Financial institutions and

state-controlled firms are not included.

5.1 Variables

7
Financial statements of Korean firms are available from two major credit evaluating firms, NICE
(National Information Credit Evaluation) and KIS (Korea Information Service). There are two major
sources for ownership information in Korea; one is the Fair Trade Commission data recorded from 1987,
but this includes ownership information for chaebol-affiliated firms only, and releases the information at
the aggregate level for each chaebol. The other source is information compiled by NICE and KIS.
8
For example, a person declares no relationship one year and later declares a family relationship with the

16
This study measures firm performance through profitability. As discussed earlier,

accounting profitability is likely a better performance measure than stock market-based

measures for the following reasons. First, stock prices are less likely to reflect all

available information when the stock market shows inefficiency. Second, a firm's

accounting profitability is more directly related to its financial survivability than is its

stock market value. Third, accounting measures allow us to evaluate the performance of

privately held firms as well as that of publicly traded firms.

Several accounting profitability measures are used: ordinary income divided by assets,

net income divided by assets, and those income variables divided by sales. Ordinary

income is operating income (sales minus the cost of sales, selling expenses, and

administrative expenses) minus interest payments plus dividends and gains on securities,

etc. Net income is ordinary income minus tax and extraordinary items, etc. These

profitability measures help address concerns regarding variations in taxes and

accounting conventions. In these analyses, the results are similar for all profit measures.

Although not reported here, the results also hold for operating income as a profitability

measure. Together, these results suggest that accounting distortions, if any, are not

significant.

To examine governance effects, Lemmon and Lins (2002) define cash flow rights

leverage as the sum of ownership rights and indirect ownership over ownership rights.

In their actual analysis, however, they omit information by using a cash flow rights

controlling shareholder of the firm. Therefore, intertemporal consistency of ownership should be checked.

17
leverage dummy instead of a continuous variable. Furthermore, a leverage ratio can

distort the degree of the control-ownership disparity when ownership is very low or

very high. I therefore, I use the difference between control rights and ownership rights

instead of their ratio. This difference is the same as LLSV’s (2002) “wedge”.

I also use ownership concentration, because neither the leverage ratio nor the difference

variable encompasses all cases that cause control-ownership disparities and shareholder

conflicts. They only include cases in which the cash flow rights leverage is greater than

unity, i.e., those in which indirect management ownership occurs. Yet, control-

ownership disparities can also occur when cash flow rights leverage is unity. For

example, the “shadow voting” requirement restricts independent financial institutions'

voting rights, so financial institution ownership can facilitate controlling shareholder

control. Even without institutional ownership, a controlling shareholder with large

ownership can hold more control rights than cash flow rights. As discussed earlier, a

controlling shareholder does not face any credible external threat in a weak corporate

governance system. Thus, the controlling shareholder can exercise full control over the

firm's resources, and the controlling family’s ownership concentration can serve as a

better control-ownership disparity measure for the Korean firm data.

The ownership data provided by NICE includes the names and shareholdings of the

largest individual, family members, and institutional shareholders. I compute the

controlling family’s direct ownership stake of each firm. The family ownership for all

sample firms used in the study is nearly 46%. As shown in Table 1, ownership varies

18
depending on the type of firm. Ownership concentration in publicly traded firms is

lower than in privately held firms, and that of firms affiliated with business group is

lower than that in independent firms. At the same time, ownership changes over time

within a firm, as Table 2 shows.

Business groups are prevalent in Korea. Of the firms used in this study, over 27%

belong to large or medium business groups. The main purpose of examining the

affiliation status with large business groups is not to explore whether a change in the

ranking affects firm profitability, but to evaluate how large business groups perform

compared to other firms. While the ranking of business groups varies each year, almost

all groups remain in the largest 30 or largest 70 over the time of the study. I select the 70

largest business groups as follows. First, the selection of the 30 largest chaebols follows

the KFTC classification in 1995 based on the size of their total assets. Using a

classification based on the size of other years does not change the results. I then identify

40 additional chaebols. These chaebols have bank loan restrictions and an equity

investment ceiling. Using debt size rather that asset size to select chaebols results in

nearly the same choices. Chaebol equity investment was briefly deregulated between

1997 and 1998. In 1998, regulation was reintroduced, and will go into effect as of 2002.

Roughly 12% of all observations belong to the top 70 chaebols. In the analysis, "Large

70 chaebol dummy" has a value of 1 when the firm is affiliated with one of the 70

largest chaebols, and 0 otherwise. I also use the top 30 chaebols instead of the top 70, to

examine the rigorousness of the results. I include the names of the top 30 groups in the

Appendix.

19
On average, 13% of a Korean firm's assets are not used for production. Instead, firms

invest in financial securities, long-term deposits, loans, etc. Firms invest, on average,

4% of their assets in affiliated firms’ financial securities. To test the relative efficiency

of resource allocation among subsidiaries, financial investments in affiliated and

unaffiliated firms are distinguished. "Investment-in-affiliated-firms" and "Investment-

in-unaffiliated-firms" are the ratios of these investments over the firm's total assets.

Large firms may benefit from economies of scale, but their advantage may decrease

beyond a certain threshold, so the log value of assets (and sales) is used for firm size.

To control for capital structure, the analyses include an equity ratio, the firm’s equity

over its total assets. Other firm attributes are measured through market share, the

advertisement over sales ratio, the advertisement over sales ratio, and the export over

sales ratio.

<Insert Table 2 around here>

6. Empirical tests and results

This section examines how corporate governance affects firm profitability. In addition

to corporate governance, firm profitability can depend on industry attributes and yearly

effects, as well as firm attributes such as size, financial structure, market share, and

business strategy. Controlling for other factors, I investigate the effects of ownership

20
concentration, control-ownership disparity, and firm organization on firm performance.

The corporate governance systems of specific types of firms may differ. For example,

private firms could have weaker corporate governance systems, since they receive less

outside scrutiny. As private firms constitute 76% of the total observations, the effects of

private firms could dominate the results. Similarly, corporate governance systems of

firms affiliated with business groups could differ from those of independent firms. As

discussed earlier, controlling shareholders of business group-affiliated firms face few

threats to their control, and business groups may pursue joint maximization of the

group-affiliated firms. I therefore perform the same tests on three sets of data: all

sample firms, only publicly traded firms, and firms affiliated with business groups.

As the study uses a panel data set, I use the industry and time dummies. Schmalensee

(1985) shows that industry effects strongly influence accounting profits. Include 281

industry dummies for 4-digit industries to control for unobserved industry-fixed effects.

To control for unobserved macroeconomic effects, I use year dummies. Applying

within-unit estimation while treating each 4-digit industry as unit, the industry mean of

each variable is subtracted from the original equation specification. Each coefficient

therefore measures the effect of each explanatory variable's deviation from the industry

mean. While coefficients on industry dummies are not calculated, coefficients on time

dummies are available upon request.

In addition to using industry dummies, I also conduct tests using firm dummies to

21
control for unobserved firm-specific factors. With firm-fixed effects, we cannot

examine the effects of business organization, because a firm’s affiliation with top 70

chaebols (or top 30) does not change during the study period. After establishing that

important results do not change regardless of whether firm dummies or industry

dummies are used, I focus on the analysis with industry dummies. The tests show that

the results still hold for the top 30 chaebols as well. (I have included the results for a

few specifications when the 30 chaebols are used; other results are available upon

request.)

6.1. Impact of controlling shareholders’ ownership on firm profitability

Table 3 reports the basic results on how ownership concentration affects firm

profitability. The dependent variables are ordinary income to assets and net income to

assets. Beginning with ownership as the only explanatory variable, I add other

explanatory variable one at a time. Panel A summarizes the results when firm-fixed

effects are controlled, and Panel B reports the results when industry-fixed effects are

controlled. Each column shows how the results change as an additional explanatory

variable is added to the specification.

When firm and year dummies are included, ownership has a positive and significant

effect on firm profitability, as shown in Panel A. In both performance measures, firm

profitability was largely explained by changes in ownership. Over 70 % of the variance

of ordinary income over assets of each firm over time can be explained by the changes

22
of the ownership within the firm. Similarly, the R2 of the specification when net income

to assets is used also exceeds 63%. Including other explanatory variables such as firm

size and financial structure does not change how ownership affects firm profitability.

When firm dummies are included to control for firm-fixed effects however, the effects

of affiliation with business groups cannot be examined. Therefore, after establishing the

result that effects of ownership concentration hold even with firm dummies included, I

use industry dummies in the rest of the analysis.

In Panel B, industry dummies are included in the specification instead of firm dummies.

Still, the coefficient on ownership concentration is positive and significant. Since we

use the within-unit estimation, the positive and significant coefficient on ownership

concentration shows that when a firm’s ownership increases compared to the industry

mean, its relative performance compared to the industry mean also increases. The effect

of affiliation with a large business group is negative and strong. Including other

controlling variables does not change the results.

Both panels show that ownership concentration yields a positive and significant effect

when all samples are used. These results suggest that controlling shareholders

expropriate firm resources and lower firm profitability when their ownership is low.

<Insert Table 3 around here>

6.2. Yearly regression results

23
This section examines whether the basic regression results hold for individual years.

Table 4 reports the results of the regression on how ownership concentration and

business group affiliation affect firm performance each year, controlling for industry

fixed effects. In all years except 1994, ownership concentration yields positive and

significant effects. The ownership effect is not statistically significant in 1994. It should

be noted that the GNP growth rate in 1994 was one of the highest in the 1990s, and the

average return on equity in the corporate sector in 1994 exceeded the opportunity cost

of capital as shown in Figure 1. The positive effect of ownership is highest in 1997,

when the economic crisis hit the Korean economy. The results suggest that the

ownership effects are related to the macroeconomic conditions.

These yearly estimation results imply that variation across firms for 1996 and for 1997

do not seem to be different in some systematic way than in earlier years, although

average profitability in 1996 and in 1997 was much lower, as shown in Figure 1. The

sudden drop in Figure 1 could have been caused in part by a large magnitude of

ownership effects and the effect of chaebol affiliation. The magnitude of the effects of

ownership concentration is greater in 1996 than in earlier years, and greatest in 1997.

The results also shows a greater negative impact of being affiliated with large chaebols

in 1996.

<Insert Table 4 around here>

6.3. Determinants of firm profitability

24
This section examines whether ownership and firm organization can affect firm

performance, controlling for all other factors that affect firm profitability. These factors

include firm attributes such as a firm's market share and its business strategy.

Table 5 summarizes the results on determinants of firm profitability. The dependent

variables are ordinary income to assets and net income to assets. Panel A uses

ownership concentration as an explanatory variable. Panel B uses the difference

between control rights and ownership rights (control-ownership) as an explanatory

variable. Other explanatory variables include various firm characteristics. Each panel

shows the results when the same tests are conducted to (a) all sample firms, (b) only

publicly traded firms, and (c) firms affiliated with business groups. Measuring firm

profitability through ordinary income ratio over assets or net income over assets yields

similar results in the three sets of data.

The results in Panel A show that the earlier results still hold. Adding more firm

attributes does not change the effects of ownership concentration and firm organization.

Firms with high ownership concentration outperform those with low ownership

concentration. For a 1% increase in ownership, all else being equal, ordinary income

ratio over assets and net income ratio over assets rose by 0.018 and 0.015 respectively,

when all observations are used. As the average ordinary income ratio over assets was

1.2372% and net income ratio was 0.2478%, the effect was substantial. In publicly

traded firms and in firms affiliated with business groups, high ownership concentration

by controlling shareholders also increases firm profitability. In addition, independent

25
firms outperformed firms affiliated with large chaebols. In all three sets of data, the

negative impact of chaebol affiliation is large. For example, when all sample

observations are used, a hypothetical independent firm with mean values from the data

would have a positive net income to assets rate, 0.2478%, but a comparable firm in one

of the 70 largest chaebols would show a loss, -3.0061% at the mean ownership

concentration.

In Panel B, firms with a high control-ownership disparity show low profitability. While

the three sets of data show all negative effects, the impacts of the difference between

control rights and ownership rights on firm profitability are stronger and significant in

regressions using publicly traded firms only and using firms affiliated with business

groups. In publicly traded firms, a 1% increase in the difference between control rights

and ownership rights lowers ordinary income ratio over assets by 0.020 and net income

ratio over assets by 0.014, all else being equal. Similarly, in firms affiliated with

business groups, the ordinary income to assets ratio is lowered by 0.0086. As before,

chaebol-affiliated firms show lower profitability in all three sets of data.

In both Panel A and Panel B, other controlling variables also show significant results. A

rise in the equity ratio increases profitability. This result has at least two possible

explanations. First, severe conflicts can occur between equity holders and creditors

when the equity ratio is low. Firms with a very low equity ratio might have taken non-

profitable projects on average. Second, the debt disciplinary effect could have been

small. The debt disciplinary effect requires free cash flows, and Korean firms did not

26
have much free cash flow, as observed earlier.

Table 5 shows that the coefficients on the time dummies decline as time passes after the

peak of 1994 when 1997 is used as a base year. Although other tables do not show the

coefficients on time dummies, the overall directions remain the same.

<Insert Table 5 around here>

The positive effect of ownership concentration in Panel A and the negative effect of the

control-ownership disparity in Panel B imply that conflicts among shareholders lower

firm profitability. Contrary to the lack of significant effects on the changes in Tobin's Q

studied in Lemmon and Lins (2002) or stock returns in Mitton (2002), these results

suggest that firm performance suffered due to minority shareholder expropriation before

the crisis. These results suggest that the effects of shareholder conflicts under a poor

corporate governance system occur regardless of firm attributes.

6.4. Robustness tests

The results above are based on profitability measured in terms of income to assets ratios,

and the tests use the top 70 chaebol classification. Table 6 reports the main results when

we test the same specification with different profitability measures and with more

narrowly defined large business groups. In Panel A, I measure firm profitability with

the income to sales ratios. At the same time, the size of each firm is controlled by the

log value of sales rather than assets. Using accounting income divided by sales does not

27
change the main results. This suggests that the results are robust regardless of the choice

of profitability measures.

Panel B shows the results when regression equations include, instead of the top 70

chaebols, the top 30 chaebols identified by the government in 1995 based on the size.

The main results remain the same as before. Furthermore, compared to the results when

the large 70 chaebols are used, the negative effects of the difference between control

rights and ownership rights are stronger and significant. These results show that large

business groups affect firm performance negatively, contradicting a recent finding by

Khanna and Palepu (2000).

<Insert Table 6 around here>

6.5. Time-varying effects of ownership concentration

This section examines how the macroeconomic condition affects the impacts of

controlling shareholders’ ownership on firm profitability. A controlling shareholder

would have more incentive to divert firm resources when the shareholder expects that

the firm will suffer from financial distress. Since firms are more likely to face financial

distress when the economy is in recession, controlling shareholders expropriate firm

resources more when the economy is growing slowly.

The results summarized in Table 7 show that expropriation depends on the

macroeconomic situation. Panel A reports the results when an interaction term between

28
ownership and GNP growth rate of each year is included, and Panel B reports the results

when the restriction on the same slope of ownership concentration over time is removed.

Each year, the effect of controlling shareholders’ ownership concentration can be

different.

In Panel A, when all sample firms are used in the analysis, the negative coefficient on

the interaction term between ownership and GNP growth rate shows that the effects of

ownership concentration are small when the economy is growing fast and become large

when the economy is growing slowly. Panel B shows that the coefficient on 1997

ownership concentration is the highest, that on 1993 is the second highest, and those on

1994 and 1995 are small. During the period in this study, the GNP growth rate was

5.5% in 1993, 8.3% in 1994, 8.9% in 1995, 6.8% in 1996, and 5.3% in 1997. The results

in Panel B therefore are consistent with the result in Panel A that the coefficient on the

interaction term between ownership concentration and GNP growth rate is negative.

They are also consistent with earlier results in Table 7 when regression is conducted

each year. These results also suggest that the controlling shareholders’ expropriation

problem is exacerbated during the downturn of the economy.

<Insert Table 7 around here>

6.6. Nonlinearity of ownership effects

The above analyses are based on the restriction that ownership concentration yields a

linear effect on firm profitability. Using large U.S. firms, Morck, Shleifer, and Vishny

29
(1988) show that firm value (measured in Tobin’s Q and profit rate) first increases, then

decreases and increases again as the management ownership increases. This section

investigates whether ownership effects on profitability vary depending on the magnitude

of ownership. Ownership's nonlinear effects are estimated in two specifications. One

specification uses piecewise linear splines from 0-5%, 5-25%, and 25-100%, and the

other tries to fit a cubic ownership concentration curve to the data. As shown in Morck,

Shleifer, and Vishny (1988), the piecewise linear spline specification is calculated in the

following manner. The "ownership concentration 0-5%" variable is the value of the

controlling shareholders’ ownership if it is less than 5%, and 5% otherwise. Likewise,

the "5-25%" variable is the controlling shareholders’ ownership minus 5% if it is greater

than 5% but less than 25%. It is 20% if it is greater than 25%, and 0 otherwise. Finally,

ownership concentration above 25% is the controlling shareholders’ ownership minus

25% if it is greater than 25%, and 0 otherwise.

The results are summarized in Table 8. Panel A tests for ownership's nonlinear effects

using piecewise linear splines from 0-5%, 5-25% and 25-100%. In all three sets of data,

the results show that profitability declines when ownership concentration is less than

5%, increases sharply between 5% to 25% of ownership, and increases gradually when

ownership exceeds 25%. The results in Panel B show a cubic relation between

ownership and firm profitability. In all three data sets, the cubic relations show that

profitability generally increases as ownership increases. Profitability declines when

ownership is extremely low or extremely high. The regression estimates a local

minimum at 5% ownership and a maximum at 95% ownership.

30
 <Insert Table 8 around here>

In both Panel A and Panel B, an increase in ownership at a lower level appears to be

associated with a decrease in firm profitability. The entrenchment effect comes earlier

than in the results in Morck, Shleifer, and Vishny. (1988). The difference may come

from a poor corporate governance system in Korea. As discussed earlier, incumbent

controlling shareholders are protected from outside takeovers due to government

regulations, “shadow voting,” and mandatory tender offers.

6.7. Expropriation in publicly traded firms

As discussed above, on average, 13% of a Korean firm's assets are not used for

production but for financial securities, long-term deposits, loans, etc. Moreover, on

average, 4% of their assets are invested in affiliated firms’ financial securities. This

section investigates how financial investments in affiliated and unaffiliated firms affect

firm performance. It is worthy noting that there could be endogeneity issues because

firms make investment decisions, which are used as an explanatory variable in the

regressions. Future research therefore needs to find a good instrument variable.

Panel A of Table 9 reports the effects of financial investment in affiliated and

unaffiliated firms on firm profitability. Panel B reports the results without the

restriction that governance problems affect public and private firms in the same way. In

both panels, investment in affiliated firms and investment in unaffiliated firms are

31
measured through the ratios of these investments over the firm's total assets.

Panel A shows that financial investment in unaffiliated firms increases profitability,

other things being equal. In contrast, financial investment in affiliated firms lowers

profitability. Evaluated at the sample's mean value, investment in unaffiliated firms

raised net income ratio by 0.056 at the margin. On the other hand, investment in

affiliated firms lowered it by 0.026. This result suggests inefficiency of resource

allocation among subsidiaries. Panel B summarizes the results when the magnitudes of

these problems can vary for each type of firm. A firm’s listing status affects overall firm

profitability in three ways. The direct effect is measured through the coefficient of the

“Public firm” dummy (1 for a public firm, 0 for a private firm). Indirect effects are

measured through interaction terms with investment: "Public firm dummy * Investment

in affiliated firms" and "Public firm dummy * Investment in unaffiliated firms." When

all sample firms are analyzed, increasing controlling shareholder ownership raises firm

profitability more for public firms than for private firms. A 1% increase in ownership

would increase net income ratio by 0.0537 for public firms and by 0.0103 for private

firms. In addition, financial investment in affiliated firms also shows a larger negative

impact on profitability in publicly traded firms.

The negative effect of investment in affiliated firms implies that the internal capital

markets in business groups were run inefficiently and lowered firm profitability. The

results suggest that firm resources were wasted when they were transferred from a

publicly traded subsidiary (whose controlling shareholders have low ownership) to

32
another subsidiary. These results also suggest that conflicts of interests among

shareholders are generally more serious in publicly traded firms than in privately held

firms. Considering the legal independence of subsidiaries, these results are consistent

with controlling shareholders’ expropriation of firm resources through “tunneling”, as

discussed in Johnson, La Porta, Lopes-de-Silanies, and Shleifer (2000).

<Insert Table 9 around here>

6.8 Summary and discussion of results

The empirical analysis shows that the poor corporate governance system in Korea had

contributed to poor profitability even before the crisis. Firms with lower controlling

family ownership or higher differences between control rights and ownership rights

showed lower performance. There also exists some evidence of nonlinearity of

ownership effects on firm profitability. Morck, Shleifer, and Vishny (1988) show that as

controlling ownership increases in U.S. firms, firm performance increases at first, then

decreases and then increases again. McConnell and Servaes (1990) show a curvilinear

relationship between Tobin’s Q and insiders’ ownership in U.S. firms. At low levels of

ownership, the relation is positive. The Korean case shows that performance goes down

first, then increases and decreases again. Even when the ownership concentration is

small (below 5%), Korean firms already suffer low profitability. Such an inverse

relationship between firm performance and low level of ownership may be related to the

generally poor corporate governance system in Korea. Controlling shareholders could

exercise their influence even with a low level of ownership concentration. Without

33
much threat from outside, entrenched controlling shareholders and mangers might have

engaged in value-destroying behavior with a small ownership stake.

In addition, Korean firms affiliated with business groups in the mid-1990s showed

lower profitability than independent firms did. This contradicts Khanna and Palepu's

(2000) results in their study of Indian conglomerates and Chang and Choi's (1988)

results in their study of Korean firms. Differences in developmental stages of the sample

firms in part explain the differences in the results. While this study uses Korean firms

from 1993 to 1997, Khanna and Palepu (2000) examine financial performance of Indian

firms in 1993, and Chang and Choi (1988) analyze Korean firms between 1975 and

1984. The samples of both studies are drawn from the early development stage of each

country. In emerging markets or less developed economies, business groups are better

able to use limited resources, through internal capital markets and intragroup trading, to

overcome market imperfections. As the economy develops, the potential benefits of

overcoming these market imperfections decreases while the cost of agency problems

and conflicts of interest between controlling family shareholders and minority

shareholders can increase.

Financial investment in affiliated firms also lowered profitability. Moreover, the

negative effects of control-ownership disparity and internal capital market inefficiency

were stronger in publicly traded firms than in privately held ones. These results hold for

all three sets of data; all sample observations, public firms only, and firms affiliated with

business groups. Together, these results suggest that conflicts of interest between the

34
controlling shareholder and other shareholders reduce firm profitability.

7. Conclusion

Many argue that poor corporate governance was a major cause of the recent economic

crises in emerging markets. While previous studies have shown that poor corporate

governance lowered firm performance during the crisis (e.g., Mitton, 2002, Lemmon

and Lins, 2002), they have not shown its effects in the years before the crisis. My work

addresses this issue by showing how the corporate governance structure affected firm

profitability before the crisis. Using detailed information on Korean firms during 1993-

1997, a microanalysis of the determinants of firm profitability provides evidence that

firms with higher control-ownership disparity showed lower profitability, all else being

equal. The paper also shows some evidence of nonlinearity of ownership effects on firm

profitability. In addition, firms whose controlling shareholder had more ownership

outperformed firms whose controlling shareholder had less ownership. Independent

firms outperformed firms affiliated with large business-groups. Moreover, moving

resources to affiliated firms further lowered profitability. Such conflicts of interests

among shareholders were more serious in publicly traded firms (whose controlling

shareholders' ownership were generally small) than privately held firms. Together, these

results suggest that controlling shareholders, especially at large chaebols and publicly

traded firms, exploited the internal capital market for private gain at the expense of

other shareholders.

35
These results suggest that Korea's weak corporate governance system offered few

obstacles against controlling shareholder's expropriation of minority shareholders. Firm

performance had been deteriorating over time even before the crisis occurred. Weak

corporate governance systems allowed poorly managed firms to stay in the market and

resulted in inefficiency of resource allocation despite low firm profitability for many

years. Chronic low firm profitability over time is an important issue since it implies

that nonperforming loans will increase and weaken the financial sector. Consequently, it

would be helpful to examine the overall profitability of the corporate sector in

evaluating the soundness of the financial sector and predicting crises. Yet, we still need

more studies to examine whether and how poor firm profitability would have increased

the possibility of crisis. If further studies support these results, policies that improve a

country's corporate governance system can support its aggregate economic growth and

stability.

36
Appendix
Table 10 List of Top 30 business groups (chaebols)

This table shows the 30 largest chaebols identified each year by the Korea Fair Trade
Commission based on the size of the total assets of firms that belong to the same
business group.

(as of the end of the year)

Ranking 1993 1994 1995 1996 1997
1 Hyundai Hyundai Hyundai Hyundai Hyundai
2 Samsung Daewoo Samsung Samsung Samsung
3 Daewoo Samsung LG LG Daewoo
4 LG LG Daewoo Daewoo LG
5 SK SK SK SK SK
6 Hanjin Hanjin Ssangyong Ssangyong Hanjin
7 Ssangyong Ssangyong Hanjin Hanjin Ssangyong
8 Kia Kia Kia Kia Hanwha
9 Hanwha Hanwha Hanwha Hanwha Kumho
10 Lotte Lotte Lotte Lotte Dongah
11 Kumho Kumho Kumho Kumho Lotte
12 Daelim Daelim Doosan Halla Halla
13 Doosan Doosan Daelim Dongah Daelim
14 Dongah Dongah Hanbo Doosan Doosan
15 Hanil Hyosung Dongah Daelim Hansol
16 Hyosung Hanil Halla Hansol Hyosung
17 Dongkuk Halla Hyosung Hyosung Kohab
18 Sammi Dongkuk Dongkuk Dongkuk Kolon
19 Halla Sammi Jinro Jinro Dongkuk
20 Hanyang Tongyang Kolon Kolon Dongbu
21 Tongyang Kolon Tongyang Kohab Anam
22 Kolon Jinro Hansol Dongbu Jinro
23 Jinro Kohap Dongbu Tongyang Tongyang
24 Dongbu Woosung Kohab Haitai Haitai
25 Kohap Dongbu Haitai New core Shinho
26 Kukdong Haitai Sammi Anam Daesang
27 Woosung Kukdong Hanil Hanil New Core
28 Haitai Hanbo Kukdong Keopyong Keopyong
29 Byuksan Daesang New Core Daesang Kangwon
30 Daesang Byuksan Byucksan Shinho Saehan
Note: When a group changes its names, the latest name has been used.

37
References

Aghion, P., Bacchetta, P., Banerjee, A., 2000. A simple model of monetary and currency crises.
European Economic Review 44, 728-738.

Altman, E., 1968. Financial ratios, discriminant analysis and the prediction of corporate
bankruptcy. Journal of Finance 4, 589-609.

Bank of Korea, Various Issues. Financial Statement Analysis. Bank of Korea, Seoul.

Bebchuck, L., Kraakman, R., Triantis, G., 1999. Stock Pyramids, Cross-Ownership, and Dual
Class Equity: The Creation and Agency Costs of Separating Control from Cash Flow
Rights. Discussion Paper #249 Harvard Law School.

Butler, K., Malaikah, S., 1992. Efficiency and Inefficiency in thinly traded stock markets:
Kuwait and Saudi Arabia. Journal of Banking and Finance 16, 197-210.

Chang, S., Choi, U., 1988. Strategy, Structure and Performance of Korean Business Groups: A
Transactions Cost Approach. Journal of Industrial Economics 37, 141-58.

Chung, B., Yang, Y., 1992. Economic analysis of the Korean Chaebol Sector. KDI, Seoul

Claessens, S., Djankov, S., Lang, L., 2000. The Separation of Ownership and Control in East
Asian Corporations. Journal of Financial Economics 58, 81-112.

Conrad, J., Kaul, G., 1998. An Anatomy of Trading Strategy. Review of Financial Studies 11,
489-519.

De Bondt, Werner.F.M, Richard H. T., 1985. Does the Stock Market Overreact. Journal of
Finance 40, 793-805.

De Bondt, Werner.F.M, Richard H. T., 1987. Further evidence on investor overreaction and
stock market seasonality. Journal of Finance 42, 557-81.

Financial Supervisory Commission, Various Issues. Press Release on Current Non-Performing

Loans.

Gale, B., 1972. Market Share and Rate of Return. Review of Economics and Statistic 54, 412-23.

Hall, M., Weiss, L., 1967. Firm Size and Profitability. Review of Economics and Statistics 49,
319-331.

38
Hoshi, T, Kashyap, A., Scharfstein, D., 1991. Corporate Structure, Liquidity and Investment:
Evidence from Japanese Industrial Groups. Quarterly Journal of Economics 106, 33-
60.

Hubbard, R.G., Palia, D., 1999. A Re-examination of the conglomerate Merger Wave in the
1960s: An Internal Capital Market View. Journal of Finance 54, 1131-52

Hurdle, G., 1974. Leverage, Risk, Market Structure and Profitability. Review of Economics and
Statistics 56, 478-85.

Jensen, M., 1986. Agency Costs of Free Cash Flow, Corporate Finance and Takeovers.
American Economic Review 76, 323-29.

Jensen, M., Meckling, W., 1976. Theory of the Firm: Managerial Behavior, Agency Costs, and
Ownership Structure. Journal of Financial Economics 3, 305-60.

Joh, S., 2001a. The Korean Corporate Sector: Crisis and Reform. In: Kwon, Y., Shepherd, W.
(Eds.), Korea’s Economic Prospects: from financial crisis to prosperity. Edward Elgar,
Cheltenham, UK. pp. 116-132.

Joh, S., 2001b. Korean corporate governance and firm performance. Unpublished working paper
presented at the 12th NBER seminar on East Asian Economics.

Johnson, S., Boone, P., Breach, A., Friedman, E., 2000. Corporate Governance in the Asian
Financial Crisis. Journal of Financial Economics 58, 141-186

Johnson, S., La Porta, R., Lopes-de-Silanies, F., Shleifer, A., 2000. Tunneling. American
Economic Review Papers and Proceedings 90, 22-27.

Jun, I., Gong, B., 1995. Corporate governance in Korea. Korea Economic Research Institute,
Seoul.

Kim, E., Singal, V., 1997. Stock Market openings: experience of emerging economies.
Unpublished working paper. University of Michigan.

Khanna, T., Palepu, K., 2000. Is Group Affiliation Profitable in Emerging Markets? An
Analysis of Diversified Indian Business Groups. Journal of Finance 55, 867-891.

Korea Stock Exchange, 1999. Ownership structure of listed firms. Stock 368

39
Krueger, A., Yoo, J., 2001. Falling profitability, higher borrowing costs, and chaebol finances
during the Korean crisis. Unpublished working paper. presented at the conference on
the Korean Crisis and Recovery, Seoul.

Krugman, P., 1999. Balance sheets, the transfer problem, and financial crises. Unpublished
working paper. MIT.

Lamont, O., 1997. Cash flow and Investment: Evidence from Internal Capital Markets. Journal
of Finance 52, 83-109.

Lang, L., Stulz, R., 1998. Tobin’s q, Corporate Diversification, and Firm Performance. Journal
of Political Economy 102, 1248-80.

La Porta, R., Lopez-De-Silanes, F., Shleifer, A., 1999. Corporate Ownership Around the World.
Journal of Finance 54, 471-517.

La Porta, R., Lopes-de-Silanies. F., Shleifer, A., Vishny, R., 1997. Legal Determinants of
External Finance. Journal of Finance 52, 1131-50.

La Porta, R., Lopes-de-Silanies. F., Shleifer, A., Vishny, R., 2000. Investor Protection and
Corporate Governance. Journal of Financial Economics 58, 3-28.

La Porta, R., Lopes-de-Silanies. F., Shleifer, A., Vishny, R., 2002. Investor protection and
corporate valuation. forthcoming Journal of Finance

Leff, N., 1978. Industrial Organization and Entrepreneurship in the Developing Countries.
Economic Development and Cultural Change 26, 661-75.

Lichtenberg, F., 1992. Industrial De-diversification and its Consequences for Productivity.
Journal of Economic Behavior and Organization 18, 427-438.

Lemmon, M., Lins, K., 2002. Ownership structure, Corporate governance, and firm value:
evidence from the East Asian Financial Crisis. unpublished working paper. Utah
University.

Lim, W., 1989. Ownership structure and Protection of controlling rights of Publicly traded
Korean firms. Korea Listed Companies Association, Seoul.

Lo, A., MacKinlay, C., 1988. Stock market prices do not follow random walks: Evidence from a
simple specification test. Review of Financial Studies 1, 41-66.

40
Martin, S., 1993. Advanced Industrial Economics. Blackwell.

McConnell, J., Servaes, H., 1990. Additional Evidence on equity ownership and corporate value.
Journal of Financial Economics 27, 595-612.

Mitton, T., 2002. A Cross-Firm Analysis of the Impact of Corporate Governance on the East
Asian Financial Crisis. forthcoming Journal of Financial Economics.

Morck, R., Shleifer, A., Vishny, R., 1988. Management Ownership and Market Valuation: An
Empirical Analysis. Journal of Financial Economics Vol. 20, 293-315.

Mossman, C., Bell, G., Swartz, M., Turtle, H., 1998. An Empirical comparison of bankruptcy
models. Financial Review 33, 35-54.

OECD, 1998. OECD Economic Surveys 1998 Korea. OECD, Paris..

Scharfstein, D., 1998. The Dark Side of Internal Capital Markets II: Evidence from Diversified
Conglomerates. NBER working paper #6352.

Scharfstein, D., Stein, J., 2000. The Dark Side of Internal Capital Markets: Divisional Rent
Seeking and Inefficient Investment. Journal of Finance 55, 2537-2564.

Schmalensee, R., 1985. Do Markets Differ Much?. American Economic Review 75, 341-351.

Seoul National University Management Research Institute, 1985. Issues and Tasks of Korean
Firms. Seoul National University Press, Seoul.

Shephard, W., 1972. The Elements of Market Structure. Review of Economics and Statistics 54,
25-37.

Shin, H., Stulz, R., 1998. Are Internal Capital Markets Efficient?. Quarterly Journal of
Economics 113, 531-552.

Shleifer, A., Vishny, R., 1997. A survey of Corporate Governance. Journal of Finance 52, 737-
783.

Stiglitz, J., Weiss, A., 1981. Credit Rationing in Markets with Imperfect Information. American
Economic Review 71, 393-410.

Stulz, R., 1990. Managerial Discretion and Optimal Financing Policies. Journal of Financial
Economics 26, 3-27.

41
Takahashi, K., Kurokawa, Y., Watase, K., 1984. Corporate Bankruptcy Prediction in Japan.
Journal of Banking and Finance 8, 229-247.

Wernerfelt, B., Montgomery, C.A., 1988. Tobin’s q and the Importance of Focus in Firm
Performance. American Economic Review 78, 246-250.

42
Fig. 1. Profitability of Korean firms from 1967 to 1997

Average ordinary income on equity and average borrowing interest rate of firms covered
by the Bank of Korea’s annual survey. The survey covers all large firms and uses a
stratified random sampling for small firms.

43
Table 1 Controlling shareholders’ ownership stakes
Using the ownership information in the NICE data set, this table shows the largest
individual shareholder's and family members’ ownership and differences of control
rights and ownership rights (control-ownership) of 5,829 firms during 1993-1997. All
firms were subject to outside auditing and had assets exceeding 6 billion won in 1997.
Of the total sample of 19,497 observations, 24% are publicly traded and 76% are
privately held. Independent firms account for 73%, and group affiliated firms account
for 27% of total observations. Top 70 large business group affiliated firms (chaebols)
account for 12%. Panel A reports the simple mean and size-adjusted mean per firm over
all years when using firm asset as weight. Panel B reports the average per firm each year.

Panel A:
Difference between control rights
Ownership concentration
and ownership rights

Num. Mean Standard Weighted Mean Standard Weighted

Type of firms
of obs. (%) deviation Mean (%) (%) deviation Mean (%)

All firms
19,497 45.19 34.61 19.65 20.97 33.37 22.25
Publicly traded firms
4,702 31.67 28.10 13.05 15.22 25.54 17.89
Privately held firms
14,796 49.49 35.37 38.02 22.80 35.50 34.36
Independent firms
14,184 51.19 33.35 41.63 19.23 30.74 16.67
Group affiliated firms
5,314 29.20 32.79 12.55 32.45 37.17 23.22
Large chaebol affiliated
2,385 17.12 26.97 9.91 43.52 38.49 24.16
firms

Panel B:
Difference between control rights
Ownership
and ownership rights
Type of firms 1993 1994 1995 1996 1997 1993 1994 1995 1996 1997
All firms 43.36 43.65 46.01 46.72 46.32 18.69 20.87 20.86 21.55 23.47
Publicly traded firms 33.48 31.74 32.78 31.31 26.57 13.69 15.02 15.28 15.95 17.33
Privately held firms 49.58 48.02 50.07 50.86 51.10 20.69 23.02 22.58 23.06 24.96
Independent firms 47.33 49.87 51.63 52.34 52.45 14.38 16.35 16.84 17.26 18.73
Group affiliated firms 29.17 29.22 31.31 30.82 21.75 28.52 31.38 31.40 33.69 42.45
Large 70 chaebol 17.23 17.32 18.64 17.26 15.37 38.09 42.12 43.87 38.31 49.63
affiliated firms

44
Table 2 Summary statistics
The summary statistics show the time-series average of the cross-sectional statistics for 5,829 firms during 1993-1997 (total 19,497 observations). All firms
were subject to outside auditing and had assets exceeding 6 billion won in 1997. Ownership stake is the sum of the largest individual shareholder's and
family's ownership. The “Public firm” dummy has a value of 1 when the firm is listed in either the Korean Stock Exchange or the KOSDAQ (Korea
Securities Dealers Automated Quotation).
All year 1993 1994 1995 1996 1997
Variables
Mean Std Mean Mean Mean Mean Mean
Net income to assets (%) 0.2478 10.6112 1.0096 1.5540 0.4427 -0.5027 -1.7290
Ordinary income to assets (%) 1.2372 11.0496 2.0361 2.4466 1.4168 0.4058 -0.5094
Ownership stake (%) 45.1952 34.6140 43.3562 43.6535 46.0069 46.7204 46.3158
Difference between control rights
20.9696 33.3652 18.6912 20.8716 20.8688 21.5564 23.4692
And ownership rights(%)
Log (Asset) 3.1782 1.1770 3.1413 3.2148 3.1705 3.1644 3.2103
Equity ratio (%) 19.4856 32.8244 21.1821 20.7130 19.1823 18.1209 18.0073
Market share (%) 4.5526 12.2631 4.8129 4.6188 4.5364 4.2550 4.5876
Export/sales (%) 7.1765 19.7142 8.6293 7.6260 6.8584 6.3458 6.2854
Advertisement/sales (%) 0.8232 3.2174 0.8244 0.8234 0.8212 0.8404 0.7980
Investment in unaffiliated firms
9.5445 7.9068 9.2477 9.2984 9.4227 9.8342 10.0214
over assets(%)
Investment in affiliated firms
3.2616 6.4997 2.7192 3.1708 3.5401 3.7283 3.0071
over assets (%)
Business group-affiliated firms 0.2725 0.4453 0.3051 0.3010 0.2767 0.2614 0.1999
Large 70 chaebol dummy 0.1223 0.3277 0.1381 0.1365 0.1225 0.1147 0.1068
Publicly traded firm dummy 0.2428 0.4278 0.2869 0.2685 0.2349 0.2118 0.1950
Number of observations 19,497 3,953 4,007 4,189 4,456 2,892

45
Table 3 Impact of ownership concentration on firm profitability
The dependent variables are ordinary income to assets and net income to assets. Panel A controls for
firm and time-fixed effects, and Panel B controls for industry and time-fixed effects. Each column
shows the results of within-unit estimation. All variables are measured as the deviation from the time-
series mean of each variable, per firm in Panel A and per industry in Panel B. Large chaebol is a
dummy variable that takes 1 when a firm belongs to the largest 70 business groups. T-statistics are in
parentheses.
Panel A:
Ordinary Ordinary Ordinary
Net income to Net income to Net income to
income to income to income to
assets assets assets
assets assets assets
Ownership 0.0139 0.0133 0.0166 0.0161 0.0186 0.0180
concentration (2.63) (2.35) (3.16) (2.88) (4.14) (3.67)
3.9637 4.1161 1.9651 2.1624
Log (Assets)
(15.15) (14.81) (8.74) (8.80)
0.2682 0.2622
Equity ratio
(71.50) (63.95)

Firm dummies Included Included Included Included Included Included

Year dummies Included Included Included Included Included Included

# of observations 19,497 19,497 19,497 19,497 19,497 19,497

R2 0.7030 0.6365 0.7079 0.6423 0.7874 0.7247

Panel B:
Ordinary Net Ordinary Net Ordinary Net Ordinary Net
income income to income income to income income to income income to
to assets assets to assets assets to assets assets to assets assets
Ownership 0.0128 0.0095 0.0066 0.0043 0.0148 0.0128 0.0180 0.0153
concentration (5.44) (4.21) (2.69) (1.81) (5.94) (5.31) (7.37) (6.47)
Large 70 chaebol -2.3167 -1.9574 -3.7379 -3.4183 -3.0995 -2.8973
dummy (-9.02) (-7.90) (-13.57) (-12.88) (-11.51) (-11.09)
1.0927 1.1232 0.9429 1.0010
Log (Assets)
(13.85) (14.78) (12.23) (13.38)
0.0729 0.0594
Equity ratiot-1
(31.58) (26.54)

Industry dummies Included Included Included Included Included Included Included Included

Year dummies Included Included Included Included Included Included Included Included

# of observations 19,497 19,497 19,497 19,497 19,497 19,497 19,497 19,497

R2 0.1007 0.0947 0.1045 0.0976 0.1134 0.1077 0.1571 0.1393

46
Table 4 Yearly estimation of the impacts of ownership concentration on firm profitability
The dependent variables are ordinary income to assets and net income to assets. The other explanatory variables are firm characteristics, including the
affiliation status to the large 70 business groups, firm size, and financial structure. In each regression, 4-digit industry dummies are included. Large 70
chaebol is a dummy variable that has a value of 1 when a firm belongs to the largest 70 business groups. T-statistics are in parentheses.

1993 1994 1995 1996 1997

Ordinary Ordinary Ordinary Ordinary Ordinary
Net income Net income Net income Net income Net income
income to income to income to income to income to
to assets to assets to assets to assets to assets
assets assets assets assets assets

0.5002 0.6569 0.4319 0.3942 0.8335 0.8989 1.2885 1.2722 1.2331 1.4114
Log (Asset)
(2.88) (3.85) (3.12) (2.70) (5.48) (6.20) (8.16) (8.56) (4.99) (6.16)

0.0047 0.0034 0.0760 0.0241 0.0921 0.0791 0.1097 0.1028 0.1591 0.1741
Equity ratiot-1
(1.06) (0.78) (19.12) (5.75) (20.03) (18.04) (20.22) (20.13) (17.36) (20.52)

0.0158 0.0119 0.0071 -0.0001 0.0118 0.0146 0.0209 0.0173 0.0480 0.0456
Ownership concentration
(2.81) (2.15) (1.66) (-0.03) (2.49) (3.24) (4.14) (3.65) (6.05) (6.20)

-2.7577 -3.0489 -2.0985 -1.8918 -2.9028 -2.0098 -3.3187 -3.0606 -2.3838 -2.7783
Large 70 chaebol dummy
(-4.74) (-5.33) (-4.63) (-3.96) (-5.49) (-3.98) (-5.78) (-5.66) (-2.51) (-3.15)

Industry dummies Included Included Included Included Included Included Included Included Included Included

# of observations 3,953 3,953 4,007 4,007 4,189 4,189 4,456 4,456 2,892 2,892

R2 0.1066 0.1098 0.2293 0.1199 0.2302 0.2137 0.2386 0.2344 0.3066 0.3510

47
Table 5 Determinants of firm profitability
This table shows two panels of results on determinants of profitability. The dependent variables are
ordinary income to assets and net income to assets. Panel A uses ownership concentration as an
explanatory variable. Panel B uses the difference between control rights and ownership rights
(control–ownership) as an explanatory variable. The three pairs of columns in each panel show the
results using data from (a) all sample firms, (b) only publicly traded firms, and (c) firms affiliated
with business groups. Controlling for industry and time-fixed effects, within-unit estimation is used.
All variables are measured as the deviation from the time-series industry mean of each variable. Large
70 chaebol is a dummy variable that has a value of 1 when a firm belongs to one of the largest 70
business groups. T-statistics are in parentheses.

Panel A:
Business group affiliated
All sample firms Publicly traded firms only
firms
Ordinary Ordinary Ordinary
Net income Net income Net income
income Income income
to assets to assets to assets
to assets to assets to assets
0.6108 0.7343 0.1321 0.2406 0.6036 0.8164
Log(Asset)
(7.11) (8.81) (0.89) (1.70) (4.26) (5.57)
0.0723 0.0589 0.0272 0.0235 0.0865 0.0527
Equity ratiot-1
(31.42) (26.37) (7.41) (6.69) (19.69) (11.59)
Ownership 0.0176 0.0151 0.0506 0.0465 0.0288 0.0213
concentration (7.23) (6.39) (9.87) (9.48) (5.46) (3.90)
Large 70 chaebol -3.2539 -3.0300 -1.3560 -1.2592 -0.4191 -0.7662
Dummy (-12.11) (-11.62) (-3.54) (-3.43) (-1.19) (-2.10)
0.0033 0.0075 0.0128 0.0113 0.0185 0.0180
Export/sales
(0.80) (1.88) (2.16) (1.99) (2.40) (2.25)
-0.1739 -0.2117 -0.0341 -0.1670 -0.3942 -0.5096
Advertisement/sales
(-6.87) (-8.62) (-0.47) (-2.39) (-6.01) (-7.50)
Market share in 4-digit 0.1166 0.0956 0.0683 0.0491 0.1071 0.0762
industry (9.53) (8.06) (3.95) (2.96) (5.58) (3.84)
2.9301 3.0677 2.2524 2.2163 3.1592 3.5635
1993 Year dummy
(11.65) (12.57) (5.31) (5.46) (5.98) (6.52)
2.9949 3.3263 2.5072 2.3472 3.8991 4.3317
1994 Year dummy
(11.99) (13.72) (5.94) (5.81) (7.41) (7.94)
1.9971 2.2535 2.2572 2.2995 2.7728 3.1214
1995 Year dummy
(8.08) (9.40) (5.30) (5.64) (5.23) (5.69)
1.0126 1.3196 1.1658 0.9898 1.1422 1.3847
1996 Year dummy
(4.15) (5.57) (2.73) (2.42) (2.16) (2.53)

Industry dummies Included Included Included Included Included Included

# of observations 19,497 19,497 4,702 4,702 5,314 5,314

R2 0.1632 0.1458 0.3671 0.3160 0.2875 0.2146

48
Panel B:
Publicly traded firms Business group affiliated
All sample firms
only firms
(a)
(b) (c)
Ordinary Net Ordinary Net Ordinary Net
income income income income income income
to assets to assets to assets to assets to assets to assets
0.4728 0.6163 -0.3219 -0.1671 0.4094 0.6794
Log(Asset)
(5.64) (7.58) (-2.25) (-1.22) (2.92) (4.68)

0.0717 0.0583 0.0258 0.0222 0.0855 0.0520

Equity ratiot-1
(31.11) (26.09) (6.95) (6.26) (19.43) (11.43)

Difference between control -0.0014 0.0003 -0.0204 -0.0140 -0.0086 -0.0048

rights and ownership rights (-0.56) (0.13) (-3.96) (-2.83) (-1.90) (-1.03)

-3.5456 -3.3184 1.5786 -1.5616 -0.6811 -0.9970

Large 70 chaebol Dummy
(-13.01) (-12.55) (-3.99) (-4.13) (-1.90) (-2.69)

0.0033 0.0075 0.0122 0.0108 0.0178 0.0173

Export/sales
(0.80) (1.87) (2.04) (1.89) (2.30) 92.17)

-0.1720 -0.2103 0.0121 -0.1435 -0.3794 -0.4980

Advertisement/sales
(-6.78) (-8.55) (-0.16) (-2.04) (-5.77) (-7.33)

Market share in 4-digit 0.1184 0.0974 0.0805 0.0607 0.1117 0.0800

industry (9.67) (8.20) (4.62) (3.64) (5.81) (4.02)

2.8634 3.0180 2.2279 2.2143 3.1332 3.5640

1993 Year dummy
(11.36) (12.35) (5.20) (5.40) (5.89) (6.48)

2.9491 3.2917 2.4691 2.3245 3.8975 4.3461

1994 Year dummy
(11.78) (13.56) (5.80) (5.70) (7.36) (7.93)

1.9796 2.2425 2.3312 2.3776 2.8269 3.1760

1995 Year dummy
(8.00) (9.34) (5.42) (5.78) (5.31) (5.77)

1.0031 1.3139 1.2326 1.0569 1.2006 1.4380

1996 Year dummy
(4.10) (5.54) (2.86) (2.56) (2.26) (2.62)

Industry dummies Included Included Included Included Included Included

# of observations 19,497 19,497 4,702 4,702 5,314 5,314

R2 0.1610 0.1440 0.3357 0.3037 0.2838 0.2124

49
Table 6 Robustness tests with different profitability measure and a narrow classification of large chaebols
In Panel A, profitability is measured through ordinary income to sales and net income to sales, and size is controlled by the log value of sales. In Panel B, large chaebol
means the top 30 chaebols identified by the government in 1995. The three pairs of columns in each panel show the results data from (a) all sample firms, (b) only publicly
traded firms, and (c) firms affiliated with business groups. Controlling for industry and time-fixed effects, within-unit estimation is used. All variables are measured as the
deviation from the time-series industry mean of each variable. T-statistics are in parentheses.

Panel A:
All sample firms Publicly traded firms only Business group affiliated firms only
Ordinary Net Ordinary Net Ordinary Net Ordinary Net Ordinary Net Ordinary Net
income income income income income income income income income income income income
to sales to sales to sales to sales to sales to sales to sales to sales to sales to sales to sales to sales
0.1280 13.5506 0.1212 12.8907 0.0357 3.3364 0.0292 2.7610 0.1210 12.3869 0.1142 11.7132
Log (Sales)
(30.89) (31.49) (29.83) (30.55) (13.76) (12.28) (11.35) (10.25) (11.09) (12.12) (10.45) (11.44)
0.0008 0.0545 0.0008 0.0510 0.0005 0.0499 0.0005 0.0466 0.0006 0.0013 0.0006 -0.0035
Equity ratiot-1
(5.82) (3.85) (5.58) (3.60) (6.34) (6.08) (5.85) (5.65) (1.45) (0.03) (1.34) (-0.09)
0.0011 0.1094 0.0010 0.0875 0.0014 0.1424
Ownership concentration
(7.82) (7.38) (8.99) (7.81) (2.82) (3.07)
Difference between control -0.0004 -0.0281 -0.0006 -0.0473 -0.0003 -0.0248
rights and ownership rights (-2.61) (-1.89) (-5.26) (-4.12) (-0.65) (-0.62)
-0.2239 -23.1761 -0.2388 -24.8605 -0.0538 -5.3144 -0.0581 -5.8104 -0.1414 -14.0521 -0.1591 -15.9292
Large 70 chaebol dummy
(-14.45) (-14.40) (-15.28) (-15.32) (-6.74) (-6.35) (-7.13) (-6.82) (-4.28) (-4.54) (-4.76) (-5.09)

Industry dummies Included Included Included Included Included Included Included Included Included Included Included Included

Time dummies Included Included Included Included Included Included Included Included Included Included Included Included

R-square 0.0900 0.0940 0.0874 0.0916 0.1648 0.1523 0.1551 0.1441 0.0892 0.1436 0.0879 0.1421

50
Panel B:
All sample firms Publicly traded firms only Business group affiliated firms only
Ordinary Net Ordinary Net Ordinary Net Ordinary Net Ordinary Net Ordinary Net
income income income income income income income income income income income income
to assets to assets to assets to assets to assets to assets to assets to assets to assets to assets to assets to assets
0.8570 0.9320 0.7016 0.7996 0.3873 0.3796 -0.0448 0.0019 1.0005 1.1007 0.8199 0.9817
Log (Asset)
(11.15) (12.49) (9.41) (11.06) (3.12) (3.20) (-0.37) (0.02) (8.09) (8.60) (6.61) (7.65)

0.0738 0.0603 0.0733 0.0598 0.0277 0.0241 0.0261 0.0227 0.0879 0.0541 0.0870 0.0535
Equity ratiot-1
(31.99) (26.92) (31.71) (26.67) (7.51) (6.84) (7.01) (6.38) (19.91) (11.85) (19.66) (11.70)

0.0196 0.0167 0.0531 0.0480 0.0292 0.0207

Ownership concentration
(8.08) (7.10) (10.40) (9.83) (5.55) (3.79)

Difference between control -0.0049 -0.0028 -0.0227 -0.0155 -0.0098 -0.0053

rights and ownership rights (-2.02) (-1.18) (-4.39) (-3.13) (-2.18) (-1.14)

-2.8850 -2.8294 -3.1313 -3.0765 -1.0510 -0.9741 -1.2111 -1.2360 -0.3741 -0.8984 -0.6352 -1.1228
Large 30 chaebol dummy
(-8.77) (-8.87) (-9.42) (-9.55) (-2.42) (-2.35) (-2.70) (-2.88) (-0.91) (-2.11) (-1.52) (-2.60)

Include
Industry dummies Included Included Included Included Included Included Included Included Included Included Included
d

Include
Time dummies included Included Included Included Included Included Included Included Included Included Included
d

# of observations 19,497 19,497 19,497 19,497 4,702 4,702 4,702 4,702 5,314 5,314 5,314 5,314

R2 0.1547 0.1373 0.1520 0.1351 0.3634 0.3123 0.3510 0.2992 0.2764 0.2020 0.2727 0.1999

51
Table 7 Time varying effects of ownership concentration
The dependent variables are ordinary income to assets and net income to assets. Panel A reports how
the macroeconomic condition affects the effects of ownership concentration on firm profitability.
Table B reports the results without the restriction that ownership concentration yields the same slope
in each year. The other explanatory variables are firm characteristics. The three pairs of columns in
each panel show the results using for data from (a) all sample firms, (b) only publicly traded firms,
and (c) firms affiliated with business groups. Controlling for industry and time-fixed effects, within-
unit estimation is used. All variables are measured as the deviation from the time-series industry mean
of each variable. Large 70 chaebol is a dummy variable that has a value of 1 when a firm belongs to
the largest 70 business groups. T-statistics are in parentheses.

Panel A:
Publicly traded firms Business group affiliated
All sample firms
only firms

Ordinary Net Ordinary Net Ordinary Net

income income income income income income
to assets to assets to assets to assets to assets to assets
0.6163 0.7385 0.1345 0.2402 0.6040 0.8168
Log(Asset)
(7.18) (8.87) (0.91) (1.70) (4.27) (5.57)

0.0721 0.0587 0.0272 0.0235 0.0864 0.0526

Equity ratiot-1
(31.33) (26.30) (7.40) (6.69) (19.66) (11.56)

0.0594 0.0472 0.0687 0.0435 0.0556 0.0482

Ownership concentration
(5.75) (4.70) (3.24) (2.14) (2.48) (2.08)

Ownership concentration -0.0060 -0.0046 -0.0025 0.0004 -0.0037 -0.0037

* GNP growth rate (-4.17) (-3.30) (-0.88) (0.16) (-1.23) (-1.19)

-3.2642 -3.0379 -1.3519 -1.2599 -0.4249 -0.7720

Large 70 chaebol dummy
(-12.15) (-11.65) (-3.53) (-3.43) (-1.21) (-2.12)

0.0030 0.0073 0.0127 0.0113 0.0184 0.0178

Export/sales
(0.74) (1.82) (2.14) (1.99) (2.39) (2.23)

-0.1740 -0.2118 -0.0357 -0.1667 -0.3947 -0.5101

Advertisement/sales
(-6.87) (-8.62) (-0.49) (-2.39) (-6.01) (-7.51)

Market share in 4-digit 0.1167 0.0956 0.0683 0.0491 0.1072 0.0762

industry (9.54) (8.06) (3.95) (2.96) (5.59) (3.84)

Industry dummies Included Included Included Included Included Included

Time dummies Included Included Included Included Included Included

# of observations 19,497 19,497 4,702 4,702 5,314 5,314

R2 0.1640 0.1463 0.3672 0.3160 0.2877 0.2148

52
Panel B:
Business group affiliated
All sample firms Publicly traded firms only
firms

Ordinary Ordinary Ordinary

Net income Net income Net income
income income income
to assets to assets to assets
to assets to assets to assets
0.6197 0.7425 0.1370 0.2441 0.5969 0.8088
Log(Asset)
(7.22) (8.91) (0.93) (1.73) (4.21) (5.52)

0.0722 0.0588 0.0271 0.0234 0.0864 0.0526

Equity ratiot-1
(31.37) (26.34) (7.37) (6.64) (19.66) (11.56)
Ownership 0.0194 0.0145 0.0494 0.0376 0.0324 0.0226
concentration in 93 (3.88) (3.00) (5.66) (4.50) (3.33) (2.24)
Ownership 0.0091 0.0064 0.0443 0.0462 0.0294 0.0201
concentration in 94 (1.88) (1.36) (5.06) (5.51) (3.13) (2.06)
Ownership 0.0094 0.0111 0.0480 0.0449 0.0238 0.0188
concentration in 95 (2.00) (2.44) (5.05) (4.93) (2.49) (1.90)
Ownership 0.0160 0.0132 0.0581 0.0579 0.0178 0.0089
concentration in 96 (3.53) (2.99) (5.83) (6.06) (1.86) (0.90)
Ownership 0.0399 0.0352 0.0644 0.0543 0.0573 0.0576
concentration in 97 (7.30) (6.64) (4.38) (3.87) (3.87) (3.76)
Large 70 chaebol -3.2777 -3.0537 -1.3486 -1.2576 -0.4367 -0.7875
dummy (-12.20) (-11.71) (-3.52) (-3.43) (-1.24) (-2.16)
0.0030 0.0073 0.0126 0.0112 0.0182 0.0176
Export/sales
(0.73) (1.82) (2.13) (1.98) (2.36) (2.20)
-0.1725 -0.2101 -0.0353 -0.1664 -0.3949 -0.5098
Advertisement/sales
(-6.81) (-8.55) (-0.48) (-2.38) (-6.02) (-7.50)
Market share in 4-digit 0.1165 0.0955 0.0680 0.0486 0.1074 0.0764
industry (9.53) (8.05) (3.93) (2.93) (5.60) (3.85)

Industry dummies Included Included Included Included Included Included

Time dummies Included Included Included Included Included Included

# of observations 19,497 19,497 4,702 4,702 5,314 5,314

R2 0.1643 0.1467 0.3674 0.3165 0.2883 0.2158

53
Table 8 Non-linearity of ownership effects on profitability
The dependent variables are ordinary income to assets and net income to assets. Panel A tests for ownership's nonlinear effects using its linear,
squared, and cubed terms as explanatory variables. Panel B tests for ownership's nonlinear effects using piecewise linear splines from 0-5%, 5-25%,
and 25-100%, as in Morck, Shleifer, and Vishny (1988). The three pairs of columns in each panel show the results using data from (a) all sample
firms, (b) only publicly traded firms, and (c) firms affiliated with business groups. Controlling for industry and time-fixed effects, within unit
estimation is used. All variables are measured as the deviation from the time-series industry mean of each variable. Large 70 chaebol is a dummy
variable that has avalue of 1 when a firm belongs to the largest 70 business groups. T-statistics are in parentheses.
Panel A:
All sample firms Publicly traded firms only Business group affiliated firms only
Ordinary income Net income to Ordinary income Net income to Ordinary income Net income to
to assets assets to assets assets to assets assets
0.6673 0.7934 0.1735 0.2804 0.6260 0.8507
Log (Assets)
(7.73) (9.48) (1.17) (1.98) (4.38) (5.75)
0.0721 0.0587 0.0269 0.0234 0.0866 0.0529
Equity ratiot-1
(31.39) (26.33) (7.31) (6.65) (19.70) (11.62)
-0.5311 -0.5522 -0.3440 -0.3459 -0.1529 -0.2107
Ownership concentration 0 to 5%
(-5.75) (-6.16) (-3.21) (-3.37) (-1.11) (-1.48)
0.1350 0.1332 0.1623 0.1360 0.0748 0.0641
Ownership concentration 5 to 25%
(5.53) (5.63) (5.65) (4.94) (1.92) (1.59)
0.0174 0.0156 0.0390 0.0415 0.0266 0.0236
Ownership concentration above 25%
(4.88) (4.51) (5.31) (5.90) (3.11) (2.66)
-3.3195 -3.1115 -1.3220 -1.3098 -0.4297 -0.8193
Large 70 chaebol dummy
(-12.16) (-11.74) (-3.40) (-3.52) (-1.21) (-2.22)
0.0035 0.0077 0.0131 0.0115 0.0185 0.0178
Export/sales
(0.85) (1.93) (2.22) (2.03) (2.39) (2.23)
-0.1772 -0.2152 -0.0480 -0.1758 -0.3975 -0.5110
Advertisement/sales
(-7.00) (-8.76) (-0.66) (-2.52) (-6.04) (-7.50)
0.1185 0.0976 0.0678 0.0487 0.0774
Market share in 4-digit industry (5.62)
(9.70) (8.23) (3.92) (2.94) (3.89)
Industry dummies Included Included Included Included Included Included
Time dummies Included Included Included Included Included Included
R2 0.1648 0.1476 0.3694 0.3183 0.2877 0.2151

55
Panel B:
All sample firms Publicly traded firms only Business group affiliated firms only

Ordinary income Net income to Ordinary income Net income to Ordinary income Net income to
to assets assets to assets assets to assets assets
0.6618 0.7849 0.2193 0.3189 0.6405 0.8587
Log (Assets)
(7.68) (9.39) (1.48) (2.24) (4.48) (5.80)
0.0721 0.0587 0.0269 0.0234 0.0865 0.0528
Equity ratiot-1
(31.36) (26.30) (7.31) (6.65) (19.69) (11.60)
-0.0809 -0.0857 -0.0116 -0.0348 -0.0241 -0.0489
Ownership concentration
(-4.40) (-4.80) (-0.40) (-1.25) (-0.67) (-1.31)
0.0030 0.0030 0.0027 0.0029 0.0020 0.0023
(Ownership concentration )2
(6.14) (6.35) (3.25) (3.73) (1.87) (2.10)
-0.00002 -0.00002 -0.00002 -0.00002 -0.00001 -0.00002
(Ownership concentration) 3
(-6.37) (-6.53) (-3.78) (-4.04) (-2.03) (-2.14)
-3.3072 -3.1012 -1.3185 -1.3056 -0.3959 -0.7857
Large 70 chaebol dummy
(-12.11) (-11.70) (-3.39) (-3.51) (-1.11) (-2.13)
0.0032 0.0074 0.0128 0.0112 0.0182 0.0175
Export/sales
(0.79) (1.86) (2.17) (1.98) (2.36) (2.19)
-0.1755 -0.2134 -0.0366 -0.1657 -0.4010 -0.5136
Advertisement/sales
(-6.94) (-8.69) (-0.50) (-2.38) (-6.10) (-7.54)
0.1169 0.0959 0.0630 0.0445 0.1066 0.0761
Market share in 4-digit industry
(9.56) (8.09) (3.64) (2.68) (5.56) (3.83)
Industry dummies Included Included Included Included Included Included
Time dummies Included Included Included Included Included Included
# of observations 19,497 19,497 4,702 4,702 5,314 5,314
2
R 0.1651 0.1477 0.3700 0.3188 0.2881 0.2153

56
Table 9 Expropriation in publicly traded firms

The dependent variables are ordinary income to assets and net income to assets. Panel A shows
that financial investment to affiliated firms affects firm profitability. Panel B shows that
controlling shareholder expropriation is more serious in publicly traded firms than privately
held firms. The public firm dummy and its interaction terms with financial investment to
affiliated firms are included as explanatory variables. Controlling for industry and time-fixed
effects, within unit estimation is used. All variables are measured as the deviation from the
time-series industry mean of each variable. Large 70 chaebol is a dummy variable that has a
value of 1 when a firm belongs to one of the largest 70 business groups. T-statistics are in
parentheses.
Panel A
Publicly traded firms Business group affiliated
All sample firms
only firms
Ordinary Ordinary Ordinary
Net income Net income Net income
income income income
to assets to assets to assets
to assets to assets to assets
Log (Assets) 0.7700 0.8847 0.1975 0.2808 0.7721 0.8716
(9.66) (11.48) (1.48) (2.20) (5.92) (6.54)
Equity ratiot-1 0.0779 0.0642 0.0287 0.0251 0.0909 0.0568
(33.64) (28.65) (7.85) (7.18) (20.23) (12.38)
Ownership concentration 0.0192 0.0168 0.0488 0.0443 0.0297 0.0216
(7.93) (7.17) (9.88) (9.38) (5.70) (4.05)
Large chaebol dummy -3.1891 -2.8755 -1.4610 -1.3195 -0.6687 -0.9485
(-11.88) (-11.08) (-3.94) (-3.73) (-1.91) (-2.66)
Investment/asset in -0.0367 -0.0261 -0.0599 -0.0641 -0.0678 -0.0443
affiliated firms ( -2.99) ( -2.20) ( -2.61) (-2.92) (-2.88) ( -1.84)
Investment /asset in 0.0541 0.0560 0.0980 0.0760 0.0061 0.0441
unaffiliated firms (5.42) (5.81) (5.22) (4.24) (0.31) (2.17)
Export/sales 0.0033 0.0077 0.0138 0.0124 0.0189 0.0197
(0.79) (1.94) (2.37) (2.21) (2.43) (2.48)
Advertisement/sales -0.1604 -0.1922 -0.0142 -0.1190 -0.3367 -0.4411
( -6.43) ( -7.96) ( -0.22) (-1.94) (-5.45) ( -6.99)
Market share in 4-digit
0.0576 0.0381 0.0406 0.0322 0.0663 0.0576
industry
(7.07) (4.84) (3.15) (2.61) (4.59) (3.91)
Industry dummies Included Included Included Included Included Included
Time dummies Included Included Included Included Included Included
R2 0.1245 0.1122 0.3542 0.3043 0.2120 0.1556

57
Panel B:
All sample firms Business group affiliated
firms
Ordinary Ordinary
Net income Net income
income income
to assets to assets
to assets to assets
Log (Assets) 0.9574 1.0119 1.0139 1.0489
(10.48) (11.45) (6.66) (6.74)
Equity ratiot-1 0.0782 0.0643 0.0908 0.0572
(33.74) (28.71) (20.11) (12.39)
Ownership concentration 0.0120 0.0103 0.0237 0.0151
(4.66) (4.14) (3.92) (2.44)
Ownership concentration* Public firm dummy 0.0473 0.0434 0.0197 0.0202
(7.69) (7.30) (1.91) (1.93)
Public firm dummy -2.2156 -1.4266 -2.7186 -1.2427
(-5.04) (-3.35) ( -4.03) ( -1.80)
Large chaebol dummy -3.0742 -2.7556 -0.8228 -1.0509
(-11.42) (-10.58) ( -2.31) ( -2.89)
Investment/assets in affiliated firms -0.0154 -0.0040 -0.0521 -0.0152
(-1.14) (-0.30) (-1.86) (-0.53)
Investment/assets in unaffiliated firms 0.0460 0.0524 -0.0339 0.0349
(4.20) (4.95) (-1.40) (1.41)
Export/sales 0.0040 0.0085 0.0186 0.0198
(0.97) (2.14) (2.39) (2.49)
Advertisement/sales -0.1632 -0.1942 -0.3436 -0.4450
(-6.55) (-8.06) ( -5.56) (-7.05)
Market share in 4-digit industry 0.0572 0.0381 0.0651 0.0576
(7.03) (4.84) (4.51) (3.90)
Investment/assets in affiliated firms -0.0762 -0.0928 -0.0334 -0.0818
* public firm dummy (-2.45) (-3.09) (-0.67) (-1.61)
Investment/assets in unaffiliated firms 0.0344 0.0045 0.1284 0.0348
*public firm dummy (1.35) (0.18) (3.15) (0.83)
Industry dummies included Included Included Included
Time dummies included Included Included included
# of observations 19,497 19,497 5,314 5,314
R2 0.1278 0.1153 0.2151 0.1570

58