Directors' and officers' liability insurance and corporate risk-taking

Full text

(1)

0

Directors' and officers' liability insurance and corporate risk-taking

Joon Ho Hwang

1

Byungmo Kim

2

ABSTRACT

This study examines the effect of directors' and officers' liability (D&O) insurance on the

firm’s risk-taking behavior and firm value. We utilize a sample of period in Korea during

which the disclosure of D&O insurance information was mandatory and there was a

significant cross-sectional variation in the firms’ choice to purchase D&O insurance. We find

that D&O insurance leads to a more risk-taking behavior of firms and also increases firm

value compared to non-insured firms. We find that increases in risk-taking behavior and firm

value are pronounced for firms with greater growth opportunities. These results imply that

D&O insurance can be beneficial to firms by mitigating the risk-averseness of managers,

particularly for high-growth firms that can benefit more from such change in behavior.

Keywords: Directors’ and officers’ liability insurance, risk-taking, firm value

1Joon Ho Hwang: Korea University Business School, Anam-dong, Seongbuk-gu, Seoul, KOREA

Email: joonhwang@korea.ac.kr Tel: +82-2-3290-2830

2Byungmo Kim (Corresponding Author): College of Business and Economics, Dankook University,

(2)

1

We examine the effects of directors' and officers' liability insurance (hereafter referred to as "D&O insurance") on corporate risk taking behavior and firm value. D&O insurance covers corporate directors and officers against claims arising from their activities as representatives of the firm. Being covered by D&O insurance and how much is covered can affect the incentives of directors and managers, and these incentives can in turn affect corporate decision makings. Recent studies that examine the effects of D&O insurance on corporate behavior primarily focus on the agency cost perspective, such as increasing the opportunistic behavior of directors and officers, from being covered by the insurance (Chalmers et al., 2002; Lin et al., 2011; Chen et al., 2012). This study examines another important role of D&O insurance: altering the general risk attitudes of officers and directors. Without D&O insurance, directors and officers can become overly averse to taking risks due to fears of litigation. Since D&O insurance insulates their liabilities from such litigation, we conjecture that D&O insurance can mitigate the problem of excessive risk-averseness in corporate investment. To verify our hypothesis, we exploit an ideal period of time in Korea during which firms were required to disclose their D&O information and there was also a significant divide between firms that were covered by D&O insurance and firms that were not covered by the insurance. We find evidence that D&O insurance offers directors and officers the incentive to invest in riskier assets and further find that firm value is higher for these firms relative to firms that do not have D&O insurance. We also discover that these results are especially pronounced for high-growth firms, suggesting that the increased risk-taking effect of D&O insurance works to mitigate the underinvestment problem in risky, yet value-adding projects. Our results are robust to various treatments for the potential endogeneity of D&O insurance usage with respect to firm risk and firm value.

When firms are faced with lawsuits, many of them, typically under an indemnification arrangement, reimburse directors and officers for the costs of lawsuits. D&O insurance in turn reimburses the firm for these costs. Therefore, a firm purchases D&O insurance in order to cover directors and managers for legal liability on behalf of the company. D&O insurance usually provides both corporate and personal coverage. Corporate coverage reimburses the firm when the firm indemnifies directors or officers for the costs of a suit. Personal coverage provides direct payment to

(3)

2

directors or officers when the firm is not able to indemnify them for legal reasons or because of financial distress.1 Typically, D&O insurers will pay the claim so long as directors and officers have acted in good faith.2 (Ferris et al., 2007)

As D&O insurance changes the liability risk of managers and directors, D&O insurance can affect their behaviors in corporate decision making. For example, Lin et al. (2011) examine the effect of D&O insurance on the outcomes of M&A decisions in Canada and find that acquirers with higher levels of D&O insurance coverage experience lower announcement-period stock returns, pay higher acquisition premiums and exhibit lower synergies. Chalmers et al. (2002) use a proprietary data of 72 IPO firms and find that firms with greater coverage of D&O insurance have worse three-year post-IPO stock returns. The general theme of these studies is that considering that D&O insurance insulates directors and officers from the cost of litigation, these people are more likely to act in a self-interested and opportunistic manner.

The other strand of literature examines the factors affecting the decision to purchase D&O insurance and finds results that are generally consistent with the managerial opportunism argument. As the purchase of the insurance typically does not need shareholder approval, directors and officers can purchase D&O insurance in order to serve their own interests. Core (1997) finds that for Canadian firms, corporate governance and managerial entrenchment affect the decision to purchase D&O insurance. Specifically, firms with greater insider voting control are more likely to purchase D&O insurance and carry higher coverage. Core (2000) observes that these firms also pay higher insurance premiums. According to Zou et al. (2008), the demand for D&O insurance in China has a positive

1 More specifically, a typical D&O insurance policy (i) provides litigation costs for claims made

against individual directors and officers for their wrongful acts to the extent which indemnification does not apply (personal coverage or ‘‘A-Side Coverage’’), (ii) reimburses the firm for its indemnification payments (corporate reimbursement coverage or ‘‘B-Side Coverage’’) and (iii) provides optional coverage for the corporation’s own liability (Entity Securities Coverage or ‘‘Insuring Agreement C’’).

2 In many cases, litigations are settled without trial, and directors and officers are presumed to

(4)

3

relationship with respect to the degree of conflict between controlling shareholders and minority shareholders, which suggests that D&O insurance is used to expropriate small shareholders. Boyer and Stern (2012) find that at the time of an IPO for Canadian firms, income trusts pay more for the same D&O insurance coverage than common equity firms. They argue that because income trusts have a more opaque governance structure, the D&O insurance premium reveals both litigation and governance risks.

Whereas the above recent studies suggest that D&O insurance is associated with managerial opportunism, earlier studies have also argued for various benefits to the firm when it carries D&O insurance. For example, D&O insurers can provide valuable monitoring services to the firm, such as scrutinizing the firm’s governance structure during the underwriting process (Bhagat et al., 1987; Holderness, 1990; O’Sullivan, 1997; Core, 2000); have a comparative advantage in providing claims adjudication or settlement services to the firm (Mayers and Smith, 1982); help attract competent and talented directors and officers and reduce the compensation necessary for these people if D&O insurance serves as a substitute of conventional form of compensation (Mayers and Smith, 1982)3; and lower the firm's likelihood of bankruptcy (Core, 1997; Zou and Adams, 2008). An empirical study supporting the benefit-side argument of having D&O insurance is the study by Bhagat et al. (1987), who report a positive stock price response to 25 New York firms around the announcement of a D&O insurance purchase.

However, what the previous literature on D&O insurance has yet to examine is another potentially important effect of D&O insurance: altering the firm's risk-taking behavior. In the absence of D&O insurance, company officers and directors can be reluctant to take on a risky project because the downside risk stemming from a project failure can be very large. These risk-averse directors or officers can behave in a less risk-averse manner if they are insulated from the costs of lawsuits that can occur based on their corporate decision makings. We hypothesize that D&O insurance can induce managers to behave in a less risk-averse manner and reduce the underinvestment problem. Parry and

(5)

4

Parry (1991) and Core (1997) refer to the notion that risk-averse directors and officers require D&O insurance as a condition of their service. Specifically, Core (1997) argues that the cost of having D&O insurance coverage can be a concern for firms with low-growth opportunities and large free cash flow; however, for high-growth firms, the benefit of having D&O insurance can outweigh this cost. Although these studies mention the possible effect that D&O insurance can create an environment for managers to become less risk-averse, there has not been a formal test of whether D&O insurance coverage actually induces such change in behavior. We attempt to fill the gap in the literature by examining how D&O insurance affects a firm’s risk-taking behavior.

Also, considering that the empirical evidence so far does not provide a clear answer to whether D&O insurance acts as a detrimental or a value-adding mechanism to the company, we explore how D&O insurance affects future firm value.4 Theoretically, the relationship between the degree of risk taking and its effect on firm value is not evident. First, there is ambiguity on whether risk-taking leads to investments in positive net present value but risky projects, or whether it leads to overinvestments in negative net present value projects. Second, there is a confounding effect on how increased risk-taking affects the firm’s cash flow and the cost of capital. However, since it is not possible to directly observe the project’s ex-ante net present value, we examine the effect on firm value in order to determine whether increased risk-taking benefits or hurts the firm. In addition to this test, we examine the different implications of increased risk-taking with respect to the growth opportunities of the firm. For high growth firms, it is more likely that underinvestment in positive but risky projects can be detrimental to firm value, and increased risk-taking can benefit the firm. On the other hand, for mature

4 For example, Lin et al. (2011, p.508) notes “…while our study identifies one potential cost of

D&O insurance in the M&A setting, it cannot address whether D&O insurance unambiguously increases or decreases firm value.” Other related studies find insignificant results with respect to the changes in director liability. Janjigian and Bolster (1990) study the impact of Delaware's decision to allow companies to eliminate director liability and find no significant difference between the performance of firms in Delaware and other states. Brook and Rao (1994) examine the effect of the firm's adoption of provisions aimed at limiting director liability and find insignificant stock price reactions. These studies suggest that there must be both benefits and costs to the changes in directors' liability coverage.

(6)

5

and low growth firms, too much risk-taking and overinvestment is more likely to hurt firm value. We use data on D&O insurance from a large sample of Korean companies in order to study the above question. Korea provides an ideal opportunity to examine the effect of D&O insurance on corporate policy because (1) firms were required to disclose the details of their D&O insurance policies in their annual reports between our sample period of 2002-2008 and (2) there was a significant cross-sectional variation in the firms’ decision to have D&O insurance coverage during the sample period. These features are in contrast to the environment in the U.S., as firms in the U.S. are not required to disclose information on D&O insurance, including the coverage amount or the insurance premium.5 The typical source of D&O insurance data for firms in the U.S. is the survey of Tillinghast-Towers Perrin.6 However, relying on the survey data invokes concerns over the sample selection bias toward large firms. On the other hand, because Korean firms were required to disclose their D&O insurance purchase data during our sample period, we are able to study the determinants and the effects of D&O insurance without worrying about the sample selection bias of the survey data. Likewise, some studies (for example, Core, 1997, 2000; Chung and Wynn, 2008; Wynn, 2008; Lin et al., 2011; Boyer and Stern, 2012) rely on Canadian data because following the 1992 Dey Report, firms need to disclose information on D&O insurance purchases in their annual proxy statements. However, one major limitation of using firms in Canada (as well as in the U.S.) is that almost all firms are covered by D&O insurance.7 For this reason, the aforementioned studies of D&O insurance using Canadian data, with the exception of Core (1997), resort on examining the coverage amount or the insurance premium information instead of analyzing the effect of the binary decision of carrying D&O

5 In general, firms also do not disclose this information voluntarily through 13K filings or through

the IPO prospectus.

6 These surveys are now conducted by Towers Watson. Another way of conducting a study of

D&O insurance is utilizing a proprietary set of data, such as Chalmers et al. (2002), who use proprietary information on 72 IPO firms in the U.S.

7 Chen et al. (2012) report that in 2001, 97% of U.S. firms and almost 90% of Canadian firms were

(7)

6

insurance on corporate policy.8 The uniqueness of the research setting provided by our sample, in terms of the mandatory disclosure requirement of D&O insurance policies and a significant cross-sectional variation in the firm’s decision to have D&O insurance coverage, makes our empirical analyses a meaningful contribution to the study of D&O insurance.9

Our empirical analyses are categorized into three parts. First, in order to examine the implications of a firm’s choice of D&O insurance coverage, we need to account for the possibility that firms with D&O insurance have firm-specific characteristics that affect both the decision to obtain D&O insurance coverage and the risk-taking attitudes of firms. To this end, we first study the determinants of D&O insurance of our sample of firms. This process also serves the purpose of examining whether the determinants of D&O insurance in Korea, which is under a civil law jurisdiction, are different from the previous studies that examine firms from a common law jurisdiction. As studies in law and finance find that the legal origin of the country can have a varying impact on corporate decision making (for example, La Porta et al., 2000, 2002), we make additional contribution to the literature by analyzing D&O insurance in a civil law based country.10

8 Core (1997) examines the decision to purchase D&O insurance in Canada during the period of

1993 and 1994. The analysis was possible because the proportion of firms carrying D&O insurance in their sample of firms at that time was 63%. However, this number has increased significantly in recent decades to 90%.

9 A by-product of our research setting is that our sample is much larger than other studies on

D&O insurance; we use 2,571 firms (814 firms with D&O insurance and 1,757 firms without D&O insurance). By comparison, Core (1997) uses a dataset of 222 firms in Canada, O'Sullivan (1997) uses a sample of 366 companies in the U.K., Chalmers et al. (2002) use proprietary data on 72 IPO firms in the U.S. and Zou et al. (2008) use 53 approvals of D&O insurance purchases for firms in China.

10 One study which investigates D&O insurance in a civil law environment is Zou et al. (2008),

which examine the demand for D&O insurance by Chinese listed companies. However, their study only looks at the determinants of D&O insurance demand and does not examine the effect of D&O insurance on corporate policy. Other limitations of the study are that the data used in the study comes from the shareholders' approval of D&O insurance purchase rather than the actual purchase data, and that it has a relatively small sample size of 53 firms which had D&O insurance approvals in the shareholders' meeting.

(8)

7

Second, we explore the causal impact of D&O insurance on corporate risk-taking behavior. Since high risk investments will increase the volatility of a firm's future cash flow, which in turn makes the firm’s stock returns more volatile, we measure the degree of corporate risk-taking by examining the company’s stock return volatility.11 We also use the amount of R&D investment as an additional measure of risk-taking behavior. R&D expenditures are considered a risky form of investment because they have a higher degree of uncertainty regarding their future economic benefits.12 We find that, controlling for the endogeneity of D&O insurance coverage, firms that are covered by D&O insurance show greater future stock return volatility and R&D intensity, and this relationship is more pronounced for high-growth firms. This suggests that D&O insurance increases the firm’s risk-taking behavior, especially for firms which can benefit more from doing so.

Lastly, we examine the relationship between having D&O insurance coverage and firm value. We follow the convention in the literature and use Tobin’s Q to measure the firm’s market valuation.13 We observe that firms that have D&O insurance experience higher market value for the next three years, and that this effect is stronger for high-growth firms. Collectively, our evidence suggests that D&O insurance alleviates the risk-averseness of decision makers. Further, D&O insurance can benefit the firm by alleviating the underinvestment problem in high-growth firms rather than creating an overinvestment problem in low-growth firms. Our finding is consistent with Bhagat et al. (1987), who finds a positive market reaction to D&O insurance purchases. Our study provides a channel by which D&O insurance can engender a value-increasing effect; firms can behave in a less risk-averse manner if they are insured by D&O insurance. In reconciliation with the previous studies, which argue for the

11 Some studies which use stock return volatility to proxy for the firm's risk-taking behavior are

Guay (1999), Rajpopal and Shevlin (2002), Coles et al. (2006), Low (2009), Cassell et al. (2012), Hirshleifer et al., (2012), and Kini and Williams (2012).

12 Some studies that use the amount of R&D spending to measure the degree of risk-taking are

Bhagat and Welch (1995), Kothari et al. (2002), Eberhart et al. (2004) Coles et al. (2006), Cassell et al. (2012), Hirshleifer et al. (2012), and Kini and Williams (2012).

13 Examples of studies which use Tobin's Q to measure the firm's market value are Morck et al.

(1988), Servaes (1996), Yermack (1996), ((Cummins, Lewis, and Wei, 2006JBF)), and Jin and Jorion (2006).

(9)

8

opportunistic behavior incurred by D&O insurance, we believe that in an environment in which the probability of litigation is high, (for example, IPO as in Chalmers et al. (2002) and M&A in Lin et al. (2011)) D&O managers may elect to use D&O insurance in order to shield their potential damage. However, when business is as usual or under a low-litigation environment, such as in a civil law jurisdiction, D&O insurance can benefit the firm in terms of reducing the risk-averseness of the decision makers.

The paper is organized as follows. In Section I, we describe the D&O insurance market in Korea and report the descriptive statistics of our sample firms. In Section II, we examine the causal impact of D&O insurance coverage on the firm’s risk-taking behavior. In Section III, we explore the effect of D&O insurance coverage on future firm value. We provide our concluding remarks in Section IV.

I.

D&O insurance in Korea and Descriptive Statistics

The sample of firms in our study comes from the companies listed on the Korea Stock Exchange, the data of which is available on the Korea Listed Companies Association database. Our sample period is between 2002 and 2008. The starting year of 2002 is selected because many firms decided to have D&O insurance coverage subsequent to the Asian financial crisis in the late 1990s. One highly publicized incidence which sparked companies’ interest in D&O insurance was the shareholder derivative lawsuit case of Korea First Bank (currently Standard Chartered Bank Korea Limited) in 2000.14 We employ a cutoff point of 2008 because the disclosure of D&O insurance coverage was mandatory in Korea until 2008. From the initial sample of firms that are listed on the Korea Stock Exchange during our sample period of 2002-2008, we require that (a) the company has a fiscal year end of December 31; (b) the firm does not belong in the financial industry; and (c) the company has equity greater than zero. We collect stock price information from the Dataguide of FnGuide and

14 In this lawsuit, the plaintiffs claimed 40 billion Korean Won in compensation from the former

president and directors of the bank, alleging that they had instructed staff employees to extend loans to Hanbo Group when bankruptcy of Hanbo Group was imminent. The Seoul District Court ruled in favor of the minority shareholders.

(10)

9

balance sheet items from the TS-2000 of Korea Listed Companies Association (KLCA). We hand-collect the variables related to D&O insurance from each company's annual reports on Data Analysis, Retrieval and Transfer System in the Financial Supervisory Service (http://dart.fss.or.kr).

The specific terms of D&O insurance in Korea is similar to those in the U.S., and many contracts are underwritten by U.S. insurers. A few minor differences in the provisions of a typical D&O insurance contract is as follows: (1) the “Excess Side A” coverage, which provides additional coverage to directors and officers for lawsuits in which a firm does not provide indemnification, such as lawsuits involving environmental pollution and lawsuits related to Employee Retirement Income Security Act, is typically not included in D&O insurance in Korea. (2) The severability clause, which separates the contract's valid parts or people that remain effective and enforceable from those which are declared void or innocent by the court, is applied to a smaller number of cases in Korea than in the U.S.15

The major difference in the D&O insurance market in Korea compared to the U.S is that the market is relatively young in Korea, as insurance coverage for corporate directors and officers first became available in 1991, which was underwritten by AIG (Yook, 2010). At the time, although corporations were not permitted to indemnify their directors and officers for legal costs, directors and officers did not perceive great litigation risk, and the market for D&O insurance coverage remained insignificant. The earliest known shareholder lawsuits were brought up by PSPD in 1997 against Korea First Bank and in 1998 against Samsung Electronics (Kim, 2006; Black et al., 2011). After the Asian financial crisis of 1997-1998, Korea introduced various reforms which facilitated shareholder lawsuits. Examples of such reforms in Korea include introducing fiduciary roles for company directors, reducing the shareholding requirements for initiating a derivative suit and lowering the shareholding requirements in order to access company records.16 In 2004, Korea enacted the

15 The severability clause is typically included in a D&O insurance contract in order to protect

directors or officers who have not taken part in the wrongful act (for example, misrepresentation or omission of a material fact) alleged in the lawsuits.

(11)

10

Securities Class Action Act of 2004, under which the class action procedure became available for small shareholders of large companies (asset size greater than 2 trillion Korean Won) beginning in 2005, and for all public companies from 2007.17 With the increasing number of claims against corporations, along with the threat that directors and officers themselves, and not only corporations, could face significant personal liability, D&O insurance has become a crucial source of protection for directors and officers. Panel A of Table I presents that in response to these changes in the litigation environment, there has been a steady increase in the demand for D&O insurance during the sample period. Table I also shows that on average, 31.7% of the sample firms are covered by D&O insurance. This is in contrast to more than 90% of firms that are covered by D&O insurance in the U.S., according to surveys by Tillinghast-Towers Perrin. Therefore, the Korean market provides us with a nice environment to examine the effect of D&O insurance due to the cross-sectional divide between firms that are covered by D&O insurance and firms that are not.

[Table I]

Table II illustrates the trend in the D&O coverage amount and D&O premium. A substantial upward trend is evident for the raw coverage amount. This reflects the need for firms to have higher coverage in response to the changes in the litigation environment over time. The mean coverage amount is KRW 12,765,356,000 or approximately US$11.61 million.18 This coverage amount is about one-third of the mean coverage amount for firms surveyed in the U.S. and about twice the coverage amount for IPO firms in the U.S.19 If we normalize the raw coverage amount by the lagged asset size

17 We control for this event in our later analyses of D&O insurance. 18 We assume an exchange ratio of KRW 1,100 per U.S. Dollar.

19 Cao and Narayanamoorthy (2011) find that according to the Tillinghast's D&O Insurance

Surveys of 2001 and 2002, the mean (median) coverage amount is US$37.68 ($20.00) million for their sample of 297 firm-year observations. For a sample of Canadian firms, the average (median) coverage amount is C$46.5 million (C$20.0 million), according to Lin et al.’s (2011) study of 709 completed mergers and acquisitions; and C$54 million (C$25 million) in Chung and Wynn (2008), who examine 224 firms during the period of 1998 to 2004. Chalmers et al. (2002) report a mean (median) coverage amount of US$6.61 ($5.00) million for their sample of 72 IPO firms issued between 1992 and 1996. Care must be taken in generalizing these amounts to all firms because

(12)

11

of the firm, we observe a slight upward trend for the mean ratio of the coverage amount to the lagged asset. There is also a large clustering of the coverage amount at 5 billion and 10 billion Korean Won.

As for the insurance premium, the mean (median) premium is KRW 181,679,000 (62,031,000) , which is approximately US$165,000 ($56,400). These are smaller than the corresponding figures in the U.S. and Canada. By comparison, Cao and Narayanamoorthy (2011) find that in the U.S., the mean (median) premium is US$480,000 ($390,000). In Canada, Chung and Wynn (2008) report a mean (median) premium of C$330,000 (C$80,000). We observe a downward trend in the median insurance premium during our sample period. Once we normalize the premium amount by the lagged asset size of the firm, the downward trend in this ratio is evident for both the mean and the median values of the ratio. Various sources of Korean financial press cite the following reasons for the decline in insurance premiums. In the earlier part of our sample period, the highly publicized governance problems of Korean firms in the aftermath of the Asian financial crisis influenced the high insurance premium priced by D&O insurers, most of whom were U.S. insurers. This concern gradually dissipated over time as governance issues improved for Korean firms. Also, the increase in market competition in the D&O insurance market contributed to the decline in insurance premiums.

For our sample of firms, D&O insurance provides coverage of 164.39 (median of 83.05) times the premium amount. This number is smaller than the corresponding figure in Canada, as Chung and Wynn (2008) report the mean (median) coverage-to-premium ratio as 657 (260).20 The reciprocal of the coverage-to-premium ratio is the unit price of D&O insurance for its coverage, or "rate-on-line" in the language of insurance underwriters. In our sample of firms, the mean (median) “rate-on-line” is 0.6 (1.2) cents. By comparison, Cao and Narayanamoorthy (2011) report the mean (median) “rate-on-line” of 2 cents in their sample of U.S. firms, and Chalmers et al. (2002) show a mean (median) of 4.0

as previously mentioned, studies that use the survey data incur a possible sample selection bias toward large firms.

20 The coverage-to-premium ratio of our sample firms in Korea is slightly larger than the ratio of

129 for the first (smallest) quartile of firms sorted by firm size in the sample of Canadian firms in Chung and Wynn (2008).

(13)

12

(3.8) cents for the sample of IPO firms in the U.S, which reflect the higher litigation risks of young IPO firms. Therefore, D&O insurance in Korea exhibits a lower premium for its coverage amount than in the U.S., but a larger one compared to Canada. This difference reflects the differences in the legal environment and the likelihood of litigation in different countries. Although U.S. and Canada are both under a common law jurisdiction, U.S. firms face greater likelihood of litigation than Canadian firms. Insofar as the "rate-on-line" proxies for the litigation risk, as argued by Cao and Narayanamoorthy (2011), our results imply that the average litigation risk in Korea is between that of U.S. and Canada.

[Table II]

Table III shows the descriptive statistics of the variables used in our sample. Definitions of all the variables are provided in the Appendix. The univariate analysis provides us with some simple insight regarding the differences between firms which are covered by D&O insurance and those that are not covered by the insurance. The mean difference test between the two groups of firms is based on t-statistics, and the median difference test is based on Wilcoxon statistics. With respect to the association between D&O insurance coverage and the firm’s riskiness, the descriptive statistics portray conflicting results. On the one hand, firms that are insured by D&O insurance show greater R&D expenditure, which implies greater firm risk. But on the other hand, firms covered by D&O insurance are more likely to have lower stock return volatility than firms that are not covered by D&O insurance. As for the other variables, firms that are covered by D&O insurance are likely to be larger firms, firms with higher market valuation as measured by Tobin’s Q, higher leverage, and higher profitability measured by ROA. While we defer the theoretical prediction and empirical relationship between these variables and the D&O insurance decision to the next section, the results of the univariate tests are (with the exception of the stock return volatility variable) generally consistent with the findings in the previous literature. There are also some differences between insured firms and non-insured firms with respect to the firm governance characteristics. For example, non-insured firms have a greater proportion of outside directors, less controlling shareholder ownership, greater ownership disparity and greater foreign shareholder ownership. Chaebol firms, cross-listed firms and firms

(14)

13

audited by auditors affiliated with the five largest auditing firms in the U.S. are also more likely to carry D&O insurance.21 Overall, since firms that are covered by D&O insurance and those that are not covered by insurance show heterogeneous characteristics with respect to several dimensions, we account for these firm-specific factors in our analyses of the association among D&O insurance coverage, risk-taking behavior and firm value.

[Table III]

II.

Determinants of D&O insurance

The main research objective of this study is to examine how D&O insurance coverage affects a firm’s risk-taking behavior and firm value. However, as seen from the descriptive statistics, we observe different firm characteristics between firms that are insured by D&O insurance and those that are not insured. In order to account for the potential endogeneity of D&O insurance coverage, we first examine the factors that affect the purchase of D&O insurance through a probit regression of D&O insurance purchase.22 This analysis also serves the purpose of discovering what factors determine the company’s decision to purchase D&O insurance in a country in which the D&O insurance market history is at a relatively young stage and in which the legal origin is based on civil law.

In the multivariate analysis of the determinants of D&O insurance coverage, the dependent variable is an indicator variable with a value of one if the firm reported that it carried D&O insurance in its annual report and zero otherwise. This specification is possible due to the mandatory disclosure requirement of D&O insurance coverage in Korea during our sample period of 2002-2008.As for the explanatory variables of D&O insurance, many of them are adopted from the studies of Core (1997), O'Sullivan (1997) and Zou et al. (2008), who examine the determinants of D&O insurance for firms in

21 The five largest auditing firms in the U.S. during our sample period were PwC (Pricewaterhouse

Coopers), Deloitte Touche Tohmatsu Limited, KPMG, Ernst & Young and Arthur Andersen. Detailed explanations for this and other variables are mentioned in the next section.

22 Earlier studies, such as Core (1997) and O’Sullivan (1997), use a logit regression, whereas more

recent studies, such as Chung and Wynn (2008) and Zou et al. (2008), use probit regression. We follow the latter approach. Results are similar if we use logit regression.

(15)

14

Canada, UK and China, respectively.23 Many explanatory variables have confounding effects on the decision to purchase D&O insurance. Moreover, previous studies present mixed results with respect to these variables due to different sample periods and different sample pools. In addition to the control variables used in previous studies, we consider the factors which are unique to the business and litigation environment in Korea. Equation (1) indicates the specification of the multivariate regression for the determinants of D&O insurance.

, 10 12 0 , , 1 11 18 , , 13 0 1 , 1 2 , 1 3 & ( ) ( ) ( ) i t k k i t l l i t k l m m i t i t m i t i t D O insurance

litigation control governance control

Korean business environment control

stock return stock return volatility stoc

β

β

β

β

ε

β β

β

β

= = = − − = + ⋅ + ⋅ + ⋅ + = + ⋅ + ⋅ + ⋅

, 1 4 , 1 5 , 1 6 , 1 7 , 1 8 , 1 9 , 1 10 , 1 11 ' i t i t i t i t i t i t i t i t k turnover Tobin s Q firm size earnings smoothing

risky industry dummy retail industry dummy leverage cash holdings proportion of outside director

β

β

β

β

β

β

β

β

− − − − − − − − + ⋅ + ⋅ + ⋅ + ⋅ + ⋅ + ⋅ + ⋅ + ⋅ , 1 12 , 1 13 , 1 14 , 1 15 , 1 16 , 1 17 , 1 18 i t i t i t i t i t i t i t

s controlling shareholder owernship

ownership disparity chaebol dummy foreign shareholder ownership big audit firm dummy cross listing dummy class a

β

β

β

β

β

β

β

− − − − − − − + ⋅ + ⋅ + ⋅ + ⋅

+ ⋅ + ⋅ − + ⋅ ction act dummyi t, 1 +

ε

i t,

(1)

One of the categories among the determinants of D&O insurance purchases includes variables which influence the likelihood of future litigation.24 Examples of such factors include stock return performance, stock return volatility, stock turnover, growth opportunity, firm size, degree of earnings management and industry factors (Core, 1997; Chung and Wynn, 2008; Gong et al., 2008; Zou et al., 2008; Cao and Narayanamoorthy, 2011). Studies find that firms face greater litigation risk when they

23 We also refer to studies which examine the determinants of D&O insurance coverage en route

to answering their main research question. Such studies include Chung and Wynn (2008) and Cao and Narayanamoorthy (2011).

24 Cao and Narayanamoorthy (2011) use the D&O insurance premium as a proxy of litigation risk

(16)

15

exhibit lower stock return performance, higher stock return volatility and higher turnover in trading. Firms with greater growth opportunities have more volatile and intangible growth options, which make them more susceptible to litigation. Following the convention in the literature (for example, Pilotte, 1992; Berger and Ofek, 1995; Lang et al. 1996; Yermack, 1996), we measure firm’s growth opportunity as Tobin’s Q, which is defined as the sum of market value of equity and book value of debt divided by the size of total assets. However, previous studies offer mixed evidence on the effect of this variable. Core (1997) reports a positive but non-robust relationship, whereas Chung and Wynn (2008) find a negative relationship.

The effect of firm size on D&O insurance coverage decision is ambiguous. On the one hand, larger firms can be targeted more often in shareholder lawsuits, leading to a greater demand for insurance. Chung and Wynn (2008) report results that are consistent with this conjecture.25 On the other hand, Mayers and Smith (1982), Core (1997), and Boyer and Stern (2012) note that whereas large firms tend to be equipped with in-house legal staff in order to defend against litigations, small firms are more likely to demand insurance coverage due to real-service efficiencies brought by the insurance, and because bankruptcy costs are proportionately higher. As for earnings management, Zou et al. (2008) find that firms which engage in more earnings manipulation are more likely to consider D&O insurance. Since the measure used in Zou et al. (2008) is applicable only in China,26 we follow Leuz et al. (2003) and John et al. (2008) and use the degree of earnings smoothing in order to proxy for earnings management. Earnings smoothing measure is defined as one minus the ratio of the standard deviation of (operating income/lagged total asset) over the past three years to the standard deviation of (operating cash flow/lagged total asset) over that past three years. We also control for the

25 O'Sullivan (1997) also finds a positive relationship between firm size and the likelihood of

carrying D&O insurance for their sample of firms in the U.K. However, their interpretation of the positive relationship is that as firm size increases, the need for D&O insurance as a monitoring mechanism increases.

26 The measure of earnings manipulation in Zou et al. (2008) is an indicator variable if the firm’s

reported return on equity is between 6% and 7%, because 6% is the profitability threshold stipulated by the China Securities Regulatory Commission for rights issues.

(17)

16

industry factor, as firms in certain industries are more prone to be targets of litigation than firms in other industries. Field et al. (2005) present evidence that firms with greater proportion of intangible assets are more likely to be targets of litigations. Following their study and Chung and Wynn (2008), we include an indicator variable for firms in biotechnology, computer and electronics industries, which are industries that have the most intangible assets, as well as another indicator variable for the retail industry, which is at the other end of the spectrum.

Firms with greater probability of financial distress are likely to purchase D&O insurance because they have greater litigation risk; further, being covered by the insurance also lowers their expected bankruptcy costs (Core, 1997; Zou et al., 2008). This additional effect due to the possibility of financial distress is captured by financial leverage and cash holdings. However, these variables can also have confounding effects on the firm’s decision to purchase D&O insurance. As for financial leverage, the typical argument is that more debt can lead to greater probability of financial distress, which can increase the demand for insurance. However, if external debtholders have an incentive to monitor the firm's management, such monitoring can act as a substitute for the monitoring services provided by D&O insurance. The effect of cash holdings is also twofold. On the one hand, firms that hold a large amount of cash and short-term securities are less exposed to the possibility of financial distress and bankruptcy, suggesting a negative relationship between cash holdings and the likelihood of D&O insurance purchase decision. However, according to Chung and Wynn (2008), firms with excess cash are more likely to purchase D&O insurance since insurance is cheaper than cash itself for indemnification as a form of self-insurance. According to this argument, there can be a positive relationship between cash holdings and the decision to purchase D&O insurance.

We also control for various governance structures of the firm, including those that are specific to the business environment in Korea. The proportion of outside directors can affect the likelihood of carrying D&O insurance. On the one hand, less outside directors increases the chance for the board to make decisions at the expense of minority shareholders, which can lead to greater demand for D&O insurance due to the increased likelihood of litigation. On the other hand, if outside directors require D&O insurance as part of their compensation package, then firms that have greater proportion of

(18)

17

outside board members can have greater likelihood of carrying D&O insurance. Studies such as Core (1997), O’Sullivan (1997), Chalmers et al. (2002), and Zou et al. (2008) find evidence consistent with the latter argument. The ownership of the controlling shareholder can also affect the likelihood of carrying D&O insurance due to various reasons. Zou et al. (2008) argue that firms with greater power of controlling shareholders, especially in countries or firms with greater agency problems, are more likely to be challenged by other shareholders, which lead to greater demand for insurance. Mayers and Smith (1982) explain the positive relationship between the two variables based on the insiders’ risk-aversion. On the other hand, Core (1997) argues that the incentive alignment effect, which acts as a substitute for D&O insurance, can result in a negative relationship between the two variables. O’Sullivan (1997) and Chung and Wynn (2008) also report a negative relationship between the two variables. We measure the control shareholder ownership as the percentage of shares held by the controlling shareholders and his/her relatives.

We include the following control variables which are shown to influence the litigation environment specifically in Korea. Firms with more conflict of interests between the controlling shareholders and minority shareholders are more likely to be involved in shareholder lawsuits (Shleifer and Vishny, 1997). Specifically, studies such as Claessens et al. (2000) and Carney and Child (2013) find a large divergence between cash flow rights and control rights for many East Asian companies and Baek et al. (2004) note that in Korea, owner-managers of large firms exercise substantial control over cross-ownership among affiliated firms. If owner-managers use equity ownership in affiliated firms in order to strengthen their control rights, we expect that these firms face greater likelihood of litigation due to the heightened agency problem. Following Claessens et al. (2002), Joh (2003), and Black et al. (2006), we measure the ownership disparity, sometimes referred to as "wedge" in the literature, as the ownership by all affiliated shareholders subtracted by the direct ownership of the controlling shareholder.

We also control for firms that belong to large Korean business groups, which is referred to as chaebols. Chaebol firms have extensive cross-holding ownership structure, which can be conducive to

(19)

18

shareholder expropriation.27 Baek et al. (2004) note that the cross-shareholding practices in chaebols can prevent outside shareholders from effectively monitoring the member firms. Bae et al. (2002) find evidence of tunneling for chaebol firms when they engage in mergers and acquisitions. These evidences suggest that chaebol firms, which carry many governance problems, are more susceptible to litigation in Korea. On the other hand, another characteristic of chaebol firms is that due to their internal capital markets and cross-debt guarantees among member firms, they are subject to fewer financing constraints than other firms (Shin and Park, 1999; Baek et al., 2004). This feature can have an effect on lessening the probability of financial distress. To control for this special form of business, we use the chaebol indicator variable, which takes a value of one if the firm belongs to one of the chaebol firms as listed in the Korea Free Trade Commission’s Annual Statistics.

The effect of foreign shareholder ownership can also affect the litigation environment for firms in Korea. On the one hand, firms with greater foreign shareholder ownership can face higher probability of litigation because foreign investors are, in general, more sophisticated and are more accustomed to engendering lawsuits. On the one hand, studies such as Baek et al. (2006) reveal that foreign investors have a strong incentive to monitor managers and discourage them from engaging in tunneling.

Studies on Korean firms, such as Nah and Choi (2003) and Kwon and Ki (2011), find that firms that are audited by auditing firms affiliated with the five largest auditing firms in the U.S. face less likelihood of accounting-related accusations, and further argue that these auditing firms provide higher audit quality. This argument is based on DeAngelo (1981), who indicates that the larger audit firms in the U.S. provide higher audit quality because auditors with a greater number of clients have ‘more to lose’ by failing to report a discovered breach. We therefore proxy for the audit quality by using a big audit firm variable, which indicates if a firm was audited by an auditor that is affiliated

27 Because direct interlocking ownership (for example, firm A owns firm B, and firm B owns firm A)

is illegal, chaebol firms use complex pyramidal or multilayered patterns of interlocking ownership (Joh, 2003).

(20)

19 with one of the five largest audit firms in the U.S.28

We also include an indicator variable for Korean companies cross-listed on the London, U.S. and Luxembourg exchanges, which are the three main stock exchanges for cross-listing of Korean firms (Kim, 2003). The effects of cross-listing on the decision to purchase D&O insurance can be approached from different perspectives. Studies which document the bonding hypothesis of cross-listing (e.g., Doidge et al., 2004; Lel and Miller, 2008; Karolyi, 2012) imply that cross-listed firms are better governed firms than those that do not cross-list. According to this argument, there may be less need for D&O insurance for cross-listed firms because they are better-governed firms. Alternatively, the litigation environment in the country whose stock is cross-listed can have an effect on the firm’s insurance purchase decision. Lang et al. (2003) note the perceived litigation risk as an influential factor in firms’ decision to cross-list, and studies such as Chung and Wynn (2008), Gong et al. (2008), and Cao and Narayanamoorthy (2011) find evidence that supports this argument.

The last control variable that is specific to the firm governance and litigation environment in Korea is the shareholder class action act dummy. In 2004, Korea enacted the Securities Class Action Act of 2004, under which the class action procedure became available for small shareholders of large companies (defined as asset size greater than KRW 2 trillion) beginning in 2005, and for all public companies from 2007. Therefore, to control for the higher likelihood of litigation environment for these firms, we include a dummy variable which equals one if the firm’s asset size is greater than KRW 2 trillion for the years between 2005 and 2007, and for all firms after 2007.

[Table IV]

Table IV presents the results of a probit regression of the determinants for D&O insurance coverage decision. All firm-specific independent variables are lagged one year prior to the

28 As aforementioned, the five largest auditing firms in the U.S. during our sample period were

PwC (Pricewaterhouse Coopers), Deloitte Touche Tohmatsu Limited, KPMG, Ernst & Young and Arthur Andersen. The big audit firm indicator variable also has an effect of controlling for the firm’s degree of earnings management (in addition to our earnings smoothing measure), as Nah and Choi (2003) find a strong negative relationship between this variable and the discretionary accruals for firms in Korea.

(21)

20

measurement of the dependent variable. The results are estimated with heteroscedasticity-robust standard errors and corrected for the clustering of observations at the firm level. In Panel A, we use variables following the previous literature on the determinants of purchasing D&O insurance. In Panel B, we include the variables that are specific to the corporate governance and litigation environment in Korea. We report both the coefficients and the marginal effects, which measure the change in the probability of the firm carrying D&O insurance for a unit change in the relevant variable from its mean value and all other variables remain at their mean values.

The result indicates that firms with lower performance, as measured by the previous year’s market adjusted stock return, are more likely to purchase D&O insurance. This result is consistent with that of previous studies and suggests that firms with bad performance are typically faced with a greater likelihood of litigation; thus, they are more likely to purchase D&O insurance. Our main research question addresses the possibility that the coverage of D&O insurance has an effect of making these firms engage in more aggressive investments (risky but value-adding projects) in the future. The coefficient of the stock return volatility is negative but insignificant. This result is different from the findings of previous studies, which reveal that firms with greater stock return volatility are more likely to carry D&O insurance because they have a greater likelihood of being the target of shareholder litigation. Our evidence suggests that, insofar as stock return volatility proxies for the firm’s risk-taking behavior (Guay, 1999; Rajpopal and Shevlin, 2002; Coles et al., 2006; Low 2009; Cassell et al., 2012; Hirshleifer et al., 2012; Kini and Williams, 2012), firms that take on higher risk are not necessarily more likely to purchase D&O insurance. In relation to our main hypothesis that D&O insurance coverage has an effect on increasing the risk-taking behavior, while we control for the endogeneity issue in our later analyses, the results in Table II and Table IV provide simple evidence that helps alleviate the reverse causality problem.

The coefficient of Tobin’s Q is positive and significant. The result implies that firms with greater growth opportunity are more likely to purchase D&O insurance. We posit that if D&O insurance shields managers from their risk-taking activities in a positive way, then firms with D&O insurance coverage will experience higher future growth and market value. This effect, if any, will be more

(22)

21

beneficial for high growth firms in terms of reducing their underinvestment problem. In the later sections, we formally examine whether the effects of D&O insurance on the firm’s risk-taking behavior and the subsequent changes in the firm value are different with respect to the firm's growth opportunity.

Firm size is positively related to the likelihood of carrying D&O insurance. In our sample, the firm size variable is measured as the log of asset size and the mean log of asset size, the average of which is 19.317. The marginal effect of this variable implies that when asset size increases from

e19.317×1,000=KRW 245 trillion to e20.317×1,000=KRW 666 trillion (172% increase in firm asset), the probability of carrying D&O insurance increases by 19.4%, according to Model 1. The positive relationship between firm size and the likelihood of carrying D&O insurance is consistent with the findings of O’Sullivan (1997) and Chung and Wynn (2008), but is different from that of Boyer and Stern (2012). The explanation provided by Boyer and Stern was that large firms carry alternatives to D&O insurance, such as having an in-house team of jurists. We find that the conventional argument that larger firms have greater demand for insurance because they can be targeted more often in shareholder lawsuits holds in our sample of Korean firms. We do not find evidence that firms in industries that are more prone to litigations are more likely to carry D&O insurance. A weak evidence of the industry effect is that once we control for the factors specific to the Korean business environment in Model 2, the result shows that firms in the retail industry, which typically has the least proportion of intangible assets, are less likely to carry D&O insurance. The coefficient of the leverage variable is positive but statistically insignificant, which is consistent with the confounding effects of debt (increasing both the bankruptcy risk and the debtholders’ monitoring effect) and the result of Boyer and Stern (2012).

As for the variables which measure the firm’s governance structure, firms are more likely to be insured if they have a greater proportion of outside directors, if the controlling shareholder ownership is low, for chaebol firms, and for non-cross-listed firms. The positive relationship between the ratio of outside directors and D&O insurance purchase decision is consistent with the risk-averse attitude of

(23)

22

the outside directors (Core, 1997; O'Sullivan, 1997; Zou et al., 2008). The negative coefficient of the controlling shareholder ownership variable is consistent with the findings of Core (1997), O’Sullivan (1997), and Chung and Wynn (2008), suggesting that there is an incentive alignment effect of insider stock ownership which can act as a substitute for D&O insurance. The positive coefficient of the chaebol indicator variable implies that the demand for D&O insurance created by the chaebol’s governance problem, which makes them susceptible to litigation, outweighs the effect of chaebols having a lower likelihood of financial distress. The coefficient of the cross-listing dummy variable is negative and significant. We note that in the univariate test, as shown in Table III, cross-listed firms are more likely to carry D&O insurance than firms that do not cross-list their shares. However, once we control for factors, such as firm size and chaebol firms, the marginal impact of cross-listing is negatively associated with the firm's likelihood of having D&O insurance. This result does not support the argument that firms feel a greater need to be covered by D&O insurance when they are exposed to an environment which is more conducive to shareholder activism. Rather, it supports the bonding mechanism of cross-listing, in which cross-listing firms have better corporate governance as they are exposed to a more stringent business environment.

Overall, we find that in Korea, some previously documented variables which determine the D&O insurance purchase decision in a common law jurisdiction also applies to Korea in the same manner. Such variables are firm size, firm performance, Tobin’s Q and the proportion of outside directors. On the other hand, we find that some variables, such as stock return volatility, stock turnover and degree of earnings management, which have been documented to affect the D&O insurance coverage decision in other countries do not affect the D&O insurance purchase decision in Korea. As for factors that are specific to the business environment in Korea, we find that chaebol firms are more likely to purchase D&O insurance, and firms that cross-list are less likely to have D&O insurance while controlling for other factors. Finally, the coefficient of the controlling shareholder ownership suggests that the incentive alignment effect outweighs the entrenchment effect in Korea.

(24)

23

III.

D&O insurance and corporate risk-taking behavior

As previously mentioned, changes in the liability risk imposed by D&O insurance can affect directors’ and managers’ decision making behavior in the firm. This section provides the empirical test for our hypothesis, which asserts that D&O insurance alters the corporate risk-taking behavior. High risk investments will increase the volatility of the firm's future cash flow, which in turn makes the firm stock returns more volatile. As such, we use the daily stock return volatility as a measure of the riskiness of the firm, as in Guay (1999), Rajpopal and Shevlin (2002), Coles et al., (2006), Low (2009), Cassell et al. (2012) Hirshleifer et al. (2012), and Kini and Williams (2012).29 We also use the R&D expenditure as an additional measure of corporate risk-taking behavior, as in Coles et al. (2006), Cassell et al. (2012), Kini and Williams (2012), and Hirshleifer et al. (2012). This measure is based on the reasoning that compared to capital expenditures, which involve investments in tangible assets, R&D investments are more uncertain and regarded as riskier investments (Bhagat and Welch, 1995; Kothari et al. 2002; Eberhart et al. 2004; Cassell et al. 2012; Kini and Williams, 2012).

In our attempt to analyze the relationship between D&O insurance purchase and corporate risk-taking, the decision to purchase D&O insurance is not an exogenous event and is affected by various factors, as shown in Equation (1). That is, firm-specific characteristics which affect the decision to purchase D&O insurance can in turn drive the differences in firms’ risk-taking attitudes. We deal with this endogenous nature of the insurance purchase variable in two ways. First, we employ a Heckman selection model (Heckman, 1979) in order to account for the possibility that firms select to purchase D&O insurance according to the extent of their risk-taking behavior. In this estimation model, we first examine the determinants of the decision to purchase D&O insurance using a probit regression, as presented in Model 2 of Table IV.30 The second step in the Heckman selection model involves an

29 A direct test of the validity of this measure is provided in Rajpopal and Shevlin (2002), who

examine a sample of oil and gas producers and their choice of riskiness of exploration projects. They find that the exploration risk (the risk of uncertain success in exploring for new oil and gas reserves) is significantly positively associated with stock return volatility.

(25)

24

OLS regression in order to determine what variables explain the firm’s risk-taking behavior, as measured by the stock return volatility (shown in Equation (2)) and R&D expenditure (shown in Equation (3)).31 In this step, the inverse Mills ratio from each of the first-stage probit regressions is included in the second step equation that predicts the degree of firm’s risk-taking activities. The inverse Mills ratio is a self-selection parameter and represents whether the error terms in the first-stage probit regression of D&O insurance coverage decision and the second-first-stage regression of stock return volatility are correlated with each other.32

, 0 1 , 2 , 1 3 , 1 4 , 1 5 , 1 6 , 1 7 , 1 8 , 1 9 ,

&

'

i t i t i t i t i t i t i t i t i t i t

stock return volatility

D

O insurance

firm size

Tobin s Q

stock return

stock return volatility

leverage

ROA

stock turnover

cash holdings

β

β

β

β

β

β

β

β

β

β

− − − − − − −

=

+ ⋅

+

+ ⋅

+

+

+

+ ⋅

+

+

1 10 , 1 11 , , i t i t s s t t i t s t

sales growth

inverse Mills

industry

year

β

β

β

β

ε

+

+

+

+

+

(2) of Model 2. We choose Model 2 since it has a richer set of explanatory variables of D&O purchase decision in Korea because it incorporates the factors that explain the firm's governance structure unique to Korea.

31 The instrument variables for the first-stage regression of the determinants of D&O insurance

coverage are litigation risk proxies such as the degree of earnings management, litigation-prone industry dummies and governance proxies such as the ratio of outside directors, controlling shareholder ownership, ownership disparity, chaebol dummy, foreign shareholder ownership, cross-listing dummy and the class action act dummy variable. These instruments have a direct theoretical association with the first-stage regression of D&O insurance coverage decision, as explained in the previous section. On the other hand, these variables do not have a clear theoretical relationship with the stock return volatility or R&D intensity; thus, they are likely to affect the second-stage dependent variable through its effect on the first-stage predicted likelihood of having D&O insurance coverage.

32 Specifically, the inverse Mills ratio is defined as the probability density function divided by the

cumulative density function of the estimated likelihood of purchasing D&O insurance for each insurer and time period.

(26)

25 , 0 1 , 2 , 1 3 , 1 4 , 1 5 , 1 6 , 7 , 1 8 , 1 9 , 1 10

&

&

'

&

&

i t i t i t i t i t i t i t i t i t i t

R

D

D

O insurance

firm size

Tobin s Q

leverage

prior change in R

D

industry change in R

D

cash holdings

free cash flow

ROA

change in CAP

β

β

β

β

β

β

β

β

β

β

β

− − − − − − −

=

+ ⋅

+

+

+ ⋅

+

+

+ ⋅

+

+

+

, 1 11 , 1 12 , , i t i t i t s s t t i t s t

EX

sales growth

inverse Mills

industry

year

β

β

β

β

ε

− −

+

+

+

+

+

(3)

In an alternative way to address the endogeneity of the D&O insurance coverage, we implement a two-stage regression in which the first stage is a probit regression for the determinants of D&O insurance, as shown in Model 2 of Table IV; then, we use the predicted value from the first-stage probit regression as an explanatory variable in the second-stage regression of stock return volatility (shown in Equation (4)) and R&D expenditure (shown in Equation (5)). Further, as a robustness check, we use the natural logarithm of the insurance coverage amount instead of the binary variable of the insurance coverage as the proxy for D&O insurance.33 This method is used in many previous studies that utilize U.S. or Canadian data, in which there is little cross-sectional variation in the decision to have insurance coverage (Core, 1997; Chalmers et al., 2002; Lin et al., 2011; Chen et al., 2012).

, 0 1 , 2 , 1 3 , 1 4 , 1 5 , 1 6 , 1 7 , 1 8 , 1 9

'

i t i t i t i t i t i t i t i t i t

stock return volatility

prob(D & O insurance)

firm size

Tobin s Q

stock return

stock return volatility

leverage

ROA

stock turnover

cash holdi

β

β

β

β

β

β

β

β

β

β

− − − − − − −

=

+ ⋅

+

+ ⋅

+

+

+

+ ⋅

+

+

, 1 10 , 1 , i t i t s s t t i t s t

ngs

sales growth

industry

year

β

β

β

ε

+

+

+

+

(4) , 0 1 , 2 , 1 3 , 1 4 , 1 5 , 1 6 , 7 , 1 8 , 1 9 , 1 10

&

'

&

&

i t i t i t i t i t i t i t i t i t i t

R

D

prob(D & O insurance)

firm size

Tobin s Q

leverage

prior change in R

D

industry change in R

D

cash holdings

free cash flow

ROA

chang

β

β

β

β

β

β

β

β

β

β

β

− − − − − − −

=

+ ⋅

+

+

+ ⋅

+

+

+ ⋅

+

+

+

, 1 11 , 1 , i t i t s s t t i t s t

e in CAPEX

sales growth

industry

year

β

β

β

ε

− −

+

+

+

+

(5)

A. D&O insurance and stock return volatility

Table V provides the empirical results for the relationship between D&O insurance coverage and

33 The results of the first-stage regression of the log of the coverage amount are qualitatively

(27)

26

risk-taking behavior, as proxied by stock return volatility. The dependent variable is the annualized standard deviation of the daily log stock return, and the right-hand side variable of interest is the D&O insurance variable. In specifications using the Heckman two-step method in Models 1 and 2, and 5-7 with the interaction terms included, the D&O insurance variable equals one if the firm carries D&O insurance. In specifications using the two-stage least squares method, this variable is the predicted likelihood of carrying D&O insurance in Model 3, and the predicted log coverage amount in Model 4.

The control variables of stock return volatility are taken from the aforementioned studies which use stock return volatility as a measure of firm’ risk-taking activity (Guay 1999; Rajpopal and Shevlin, 2002; Coles et al. 2006; Low, 2009; Hirshleifer et al., 2012). These control variables, which are all measured with a one-year lag of the dependent variable, are firm size, Tobin’s Q, stock return performance, past stock return volatility, leverage, ROA, stock turnover, cash holdings and sales growth.34 Smaller firms are more likely to make riskier investments (Pastor and Veronesi, 2003). Also, because high growth firms are more likely to make riskier investments, we control for investment and growth opportunities of the firm using Tobin's Q and sales growth. Lagged stock return volatility is included to control for the past riskiness of the firm. We include the past stock return and ROA in order to account for past firm performance, the leverage variable to capture the firm's financing decisions, the stock turnover variable to control for trading induced volatility, and the amount of cash holdings to measure the available funds to invest in new projects. Year dummies are included in order to account for the trend in macroeconomic factors and industry dummies based on three-digit SIC

34 While some studies also control for the information in managers’ stock option compensation

scheme, which can create incentives for risk-taking, we do not include this in our study because in Korea, the number of firms that grant stock options is very small and is concentrated on firms in the financial industry. According to reports from the KRX (Korea Exchange) and Kim and Sul (2010), the number of stock options grants for each year during our sample period of 2002-2008 was 49, 29, 25, 31, 35, 22 and 13, respectively. After excluding firms in the financial industry, the corresponding numbers are 31, 10, 9, 10, 19, 10 and 3, respectively. Since these are the number of cases, the number of firms would be even smaller than the above numbers. Further, Kim and Sul (2010) find no significant relationship between stock option grants and stock return volatility in Korea, possibly due to a small number of stock option grants.

Figure

Table I. Trend of firms with D&O insurance

Table I.

Trend of firms with D&O insurance p.47
Table II. Trend of D&O insurance information

Table II.

Trend of D&O insurance information p.48
Table III. Summary statistics

Table III.

Summary statistics p.49
Table IV. Determinants of D&O insurance

Table IV.

Determinants of D&O insurance p.50
Figure 1. Changes in firm value between insured firms and non-insured firms

Figure 1.

Changes in firm value between insured firms and non-insured firms p.59

References

Updating...

Outline : CONCLUDING REMARKS