行政院國家科學委員會專題研究計畫 成果報告
訴訟風險，董監事責任險，與公司經營績效
研究成果報告(精簡版)
計 畫 類 別 ： 個別型 計 畫 編 號 ： NSC 95-2416-H-002-024- 執 行 期 間 ： 95 年 08 月 01 日至 96 年 07 月 31 日 執 行 單 位 ： 國立臺灣大學會計學系暨研究所 計 畫 主 持 人 ： 李書行 計畫參與人員： 大學生-兼任助理：韓建婷、劉俐萱、李柏霖 處 理 方 式 ： 本計畫可公開查詢中 華 民 國 96 年 12 月 19 日
行政院國家科學委員會補助專題研究計畫
■ 成 果 報 告
□期中進度報告
訴訟風險，董監事責任險，與公司經營績效
Li
t
i
ga
t
i
on
Ri
s
k,
Di
r
e
c
t
or
s
’
a
nd
Of
f
i
c
e
r
s
’
I
ns
ur
a
nc
e
,
a
nd
Fi
r
m’
s
Pe
r
f
or
ma
nc
e
計畫類別：■ 個別型計畫
□ 整合型計畫
計畫編號：NSC 95－2416－H－002－024－
執行期間：
95 年 08 月 01 日 至 96 年 07 月 31 日
計畫主持人：李書行
共同主持人：
計畫參與人員： 韓建婷 、 劉俐萱 、 李柏霖
成果報告類型(依經費核定清單規定繳交)：■精簡報告
□完整報告
本成果報告包括以下應繳交之附件：
□赴國外出差或研習心得報告一份
□赴大陸地區出差或研習心得報告一份
□出席國際學術會議心得報告及發表之論文各一份
□國際合作研究計畫國外研究報告書一份
處理方式：除產學合作研究計畫、提升產業技術及人才培育研究計畫、列
管計畫及下列情形者外，得立即公開查詢
□涉及專利或其他智慧財產權，□一年□二年後可公開查詢
執行單位：台大會計系暨研究所
中
華
民
國
96
年
月
日
計畫中文摘要
當投資人投資失敗後，控告公司及其董監事似已蔚為風氣。為降低董監事
之潛在財務風險，公司逐漸依賴董監事責任險來對當事人提供進一步之保障。
尤其在安隆事件之後，董監事更面臨前所未有之注意及巨大之責任風險。就會
計而言，董監事責任險之購買資料或許能提供投資者相當有價值之資訊，以判
斷公司之潛在風險及未來之經營績效。為更充分瞭解董監事責任險之資訊意
涵，我們需先嘗試回答有關董監事責任險之二個基本問題：(1)、那一類型的公
司在購買此保險，為何購買？(2)、董監事責任險究竟是提供當事人好的激勵誘
因或促進不好的道德危機，而進一步又如何影響公司經營績效？本研究以台灣
證券交易所的資料並透過實證方法發現以下結果。(1) 有購買董監責任險的公
司比沒有購買的公司有較高的法律風險。（2）是否購買董監責任險對經營績效
並無顯著影響。
關鍵字：
董監事責任險、訴訟風險、公司經營績效、道德危機、代理理論
Litigation Risk, Directors’and Officers’Insurance, and Firm’s Performance
Abstract
It has become a trend that the investors sue the firms and their directors and officers when the investments go under. To reduce liability risk and judgment-proof problem, D&O insurance is a primary tool used in modern society by the firms to indemnify directors and officers for the potential loss arising from possible litigations. Especially after Enron scandal, directors and officers face unprecedented scrutiny and liability exposure. From accounting point of view, the D&O insurance purchase data may provide investors valuable information to assess the underlying risk of the firm and its future performance. The answers to two fundamental questions on D&O insurance will help us better understand the information implications of D&O insurance: (1) who is buying the D&O insurance and why do they buy? (2) does D&O insurance create moral hazard or provide incentives to directors and officers, and subsequently affect the firm’s performance? We empirically test the data by using companies listed in Taiwan Security Exchanges and find that: (1) the firms purchase D&O have higher risk than the firms without purchasing; (2) but the operation performances are not different between the two groups.
Key Words: Director’s and Officers’Insurance; Litigation Risk; Firm Performance; Moral Hazard; Agency Theory
Litigation Risk, Directors’
and Officers’
Insurance, and Firm’
s Performance.
1. Introduction
It has become a trend that the investors sue the firms and their directors and officers when the investments go under. To reduce liability risk and judgment-proof problem, D&O insurance is a primary tool used in modern society by the firms to indemnify directors and officers for the potential loss arising from possible litigations. Especially after Enron scandal, directors and officers face unprecedented scrutiny and liability exposure. Therefore, a comprehensive study on directors and officers’insurance is important to academic research. Especially the effect of D&O insurance on firm’s performance is lacking in the literature.
Most of D&O research in the literature concentrates on the demand or determinant for D&O insurance, or its underwriting process. Core (1997) used data of Canadian firms to test three sources of demand: (1) demand for D&O personal coverage from contract requirement for risk averse directors and officers; (2) demand for D&O corporate coverage from an efficient corporate insurance decision; and (3) the demand for D&O insurance from managerial entrenchment. Similar research could be found in O’Sullivan (2002). Parry and Parry (1991) and Mayers and Smith (1982) provided explanation for why the firms purchased insurance, although their studies did directly address D&O insurance. Redington (2005) discussed the underwriting implications of Section 404 of Sabanes-Oxley Act of 2002 on D&O insurance. Mayers and Smith also investigated the
underwriting problems for corporate insurance.
Apparently the effect of D&O insurance on corporate performance does not exist in the literature, with Chalmers, Dann and Harford (2002) as an exception. Chalmers, Dann and Harford (2002) used 72 IPO firms and found a significant negative relation between the three-year post IPO stock price performance and the D&O insurance purchases in conjunction with the IPO. They concluded that D&O insurance decisions revealed managers’opportunistic behaviors. This type of research does provide policy implications for information disclosures on D&O insurance purchase details. Although it is an important accounting issue, almost none research can be found in the accounting literature.
The purposes of this study are two folds: (1) refine the test on D&O insurance purchase and litigation risk faced by the firms; (2) provide a direct observation on how D&O insurance
purchases affect the performance of firms. Our findings suggest that the firms purchase D&O are in higher risk measured with our estimation model; but the operation performances are not different between the two groups. Although we can not find the support for D&O improving operation performance, it shall not be interpreted that D&O has no effect on the firm operation. To answer this, a more detailed analysis is required.
2. Literature and Hypotheses Development
It is a common notion that the firms with higher exposures to the litigations will have stronger incentives in buying D&O insurance to indemnify their board of directors and officers. When the firms buying the D&O insurance, the premiums paid for insurance, under the fail rate principle, are determined based on the requested coverage and the probability of being sued and lost. Therefore, the total premium paid by a particular D&O insurance is a composite of litigation probability and requested coverage. Chalmers, Dann and Harford (2002) used the coverage being purchased as the proxy for manager’s perceived damages and the per dollar coverage insurance premium as the proxy for the perceived probability of risk by insured. Prior Enron era, D&O insurance makes directors and officers immune from personal liability for corporate failure. Given the information asymmetry, firms determine the optimal purchase of D&O insurance according their self beliefs of the risk exposures and the estimated claims need to be paid in case of loss in litigation. However, this practice has been changed, according to Alles, Datar and Friedland (2005), when Enron and WorldCom directors are forced to pay settlements out of their personal assets by court. Therefore, Redinton (2005) and Baily (2005) both predicted that the trend on D&O insurance may cause the unavailability of adequate D&O coverage desired by the firms and the increase on the premium of such insurance after Enron and WorldCom scandal.
On the other hand, we may find a group of the firms electing not to purchase the D&O insurance. The firms with no D&O insurance may be either extremely high risk or extremely low risk according to Core (1997). When the firms are in a very high risk category, they may not able to find the carriers who are willing to underwrite their insurance. In contrary, the firms with very limited risk may choose not to purchase the D&O insurance in order to save the premiums. In summary, we derive the following hypothesis:
H1: The litigation risk for firms with D&O insurance is significantly higher than those without D&O insurance.
People usually concern whether D&O insurance will result in moral hazard of the insured (directors or officers) in reducing their efforts on protecting investors, which eventually affects firm’s performance. The academic studies about the impact of liability insurance on the insured’s preventive effort are controversial. Shavell (1982) suggested that the injurer’s incentive of care was not affected by liability insurance once the insurer can observe the prevention activity and thus governmental intervention in insurance market is not desirable. However, Sarath (1991) showed that liability insurance may dilute the incentive of care when there is uncertainty in litigation. The empirical study by Sloan, Reilly and Schenzler (1995) on the other hand showed that driver’s automobile liability insurance, especially compulsory insurance, discouraged binge drinking
behavior. Gutierrez (2003) used a principal-agent model and showed that the use of liability insurance could be optimal for a more efficient litigation strategy. Holderness (1990) suggested that the purchase of D&O insurance would result in the improvement of managerial behaviors because of the monitoring from the insurance company. Therefore, two competing hypotheses can be inferred from the literature:
H2(a) The moral hazard hypothesis: The purchase of D&O insurance creates moral hazard
problems and results in a poor performance.
H2(b) Optimal incentive and monitor hypothesis: The purchase of D&O insurance aligns the
incentives of directors and officers with shareholders and results in a better performance.
3. Sample and Research Design
3.1.Data and sample
In this study, we use Top 500 firms under Taiwan Common Wealth Magazine 2004 Annual Survey as the sample base. We then delete the non-public traded companies so that we can access the public available data base to collect necessary data items. The reason for using 2004 annual survey enables us to have the prior survey 2-year and post survey 1-year financial data. Therefore, we have the data from 2002 to 2005 for the public traded companies in Top 500 list.
We search the D&O insurance data, including the coverage amount, coverage dates, the premium paid, and the insurance companies through questionnaire, because the D&O insurance data are not available from the annual reports or any public data base.
3.2. Litigation risk and D&O insurance test
Although the positive relation of litigation risk and D&O insurance demand has
been empirically documented, the tests are done indirectly or partially. O’Sullivan (2002) used UK companies and showed that the D&O insurance purchase was associated with weak corporate governance. Similarly, Core (2000) documented that that D&O premium are positively related excess CEO compensation by using Canadian companies as sample. In that study, the excess CEO compensation implies weak corporate governance. In both O’Sullivan (2002) and Core (2000) studies, they implicitly assume that weak corporate governance is equivalent to high litigation risk. Therefore, their tests are indirect. Core (1997) used ten possible litigation risk proxies and tested each individual proxy’s association with D&O purchase by applying logit model. Although Core (1997) claimed that firms with higher litigation risk are more likely to purchase D&O insurance
and carry higher coverage, his conclusion was only partially and weakly supported by his tests.
In this study, we refine Core’s litigation risk hypothesis test by using a summary measure of litigation risk proposed by Shu (2000). Shu (2000) constructed a summary measure to serve as a proxy for litigation risk faced by an auditor. In general, the investors usually sue the auditors, directors, and officers at the same time. The litigation exposure of the auditor can be assumed to be equal to that of the directors and officers. Shu (2000) used the following logit model to estimate the litigation risk measure for each firm in the sample:
Y = B0 + B1 Size + B2 Inventory/lagged assets + B3 Receivables/lagged assets + B4 Return on assets + B5 Financial leverage + B6 Sale growth + B7Current ratio+B8 Stock return + B9 Beta +
B10 Stock turnover +B11 Qualified opinion dummy,
where Y is one for firms with litigation with investors and zero for non-litigation firms. All
variables for the litigation sample are measured in the year associated with litigation. In this study, we will use 2002 and 2003 data to form the litigation risk summary measure for each sample firm.
To test H1, we will compare the mean of the summary measure of litigation risk for the firms with D&O insurance purchase in 2004 with that for the firms without purchasing D&O insurance in 2004.
3.3. D&O insurance and firm’s performance test
By the agency theory of Jensen and Meckling (1976), there are conflicts of interests between shareholders and managers. Fama and Jensen (1983) further state that the board of directors have a fiduciary duty to exercise care monitoring management on behalf of all shareholders under the separation of ownership and control. The lower agency cost, the better is the corporate
performance, which will eventually enhance the shareholder values. One way to reduce the agency costs is to design a right managerial incentive contract. Does D&O insurance provide the right incentives to directors and officers and help to improve the firm’s performance? Does D&O insurance, in opposite, reduce the incentives of directors and officers, because it free the personal liabilities of directors and officers?
In this study, we will use three performance measures to test hypothesis 2: (1) stock return; (2) net income; and (3) return on assets. We will group our sample into two sub-samples, one with D&O insurance and one without D&O insurance, based on 2004 D&O insurance data. We will compare the performance difference of the two groups for each individual performance measure. To make the performance comparison meaningful, we use ROE as the performance measure for
comparison:
4. Empirical Results
4.1. The litigation risk estimation model
In order to test the relation between litigation risk and D&O insurance, we first construct a summary measure to serve as a proxy for litigation risk. According to Shu (2000), we regress the sample data in year 2001 and 2002 with the logit model to estimate the litigation risk. The model and the predictions are based on previously documented litigation-related variables. They include firm size, current ratio, financial condition, growth rate, accounts receivable, inventories and stock turnover (Alexander, 1991; Stice, 1991; Lys and Watts, 1994; Jones and Weingram, 1996). We also include qualified opinions for the lawsuit years, since the events leading to qualified opinions might also give rise to auditor litigation.
Multinomial logistic regression
Number of obs = 741 Wald chi2(11) = 40.00 Prob > chi2 = 0.0000 Pseudo R2 = 0.2948 Log pseudolikelihood = -54.434248 litigation risk Predicted sign
Coef. Std. z P>|z| [95% Conf. Interval]
lnasset + 0.764122 0.404186 1.89 0.059 -0.02807 1.556312
inventory + -3.49E-07 2.25E-07 -1.55 0.121 -7.90E-07 9.17E-08 receivable + -6.33E-07 5.21E-07 -1.21 0.225 -1.66E-06 3.89E-07
Current ratio - -0.040194 0.013138 3.06 0.002 0.014445 0.065944 Financial leverage + 0.01202 0.006031 1.99 0.046 0.000199 0.023841 Sales growth + 0.01165 0.007859 1.48 0.138 -0.00375 0.027053 Stock return - -0.00069 0.006747 -0.1 0.919 -0.01391 0.012537 beta + -0.81155 0.8491 -0.96 0.339 -2.47575 0.852658
Sock turnover + 0.00214 0.001087 1.97 0.049 1.02E-05 0.004269 Cpa opinion + 1.312079 0.765291 1.71 0.086 -0.18786 2.812023
_cons ? -17.5342 6.610739 -2.65 0.008 -30.491 -4.57742
As the result shows, the firms with large size(measured with log of total assets) will have higher litigation risk, even though the inventory and receivables are insignificant, respectively. Return on asset is negatively related to litigation risk at a significant level. The result concludes that litigation is related to the variablesthatincreaseplaintiffs’incentivesto bring lawsuits:firm size (measured as the log of total assets), firm structure, stock turnover, and closeness to financial distress (stock return, financial leverage).
After the construction of litigation risk model, we use it to estimate the expected litigation risk for our D&O data which is obtained by questionnaire. In order to test the incentive for purchasing D&O insurance, we divide the D&O data into two subgroups. One is the firms purchasing D&O insurance and the other one is those without. The two D&O subgroups data are summarized as follow:
4.2. D&O insurance and the litigation risk
Descriptive Statistics for firms with D&O insurance
Variable Obs Mean Std. Dev. Min Max
lnasset 32 16.23689 1.305939 14.44682 19.21568
inventory 32 2270703 3390011 15134 1.38E+07
receivable 32 4929233 1.05E+07 208901 5.52E+07
ROA 32 8.424375 9.952626 -27.44 24.06 Current ratio 32 44.01406 11.942 15.61 65.6 financial leverage 32 4.600937 20.51341 -0.36 117 sales growth 32 25.99938 30.53898 -36.34 141.85 stock return 32 0.4521031 0.2723099 -0.0418 0.9682 Beta 32 2.425625 40.89042 -67.0667 111.1884 stock turnover 32 54.75594 82.85485 2.59 449.62 Cpa 32 0.65625 0.4825587 0 1 ROE 32 14.13781 19.21519 -63.7 37.24
estimated litigation risk 32 17.11404 32.95229 -75.79962 68.52764 * the estimated litigation risk is shown as the percentage.
Descriptive Statistics for firms without D&O insurance
Variable Obs Mean Std. Dev. Min Max
D&O 49 0 0 0 0
inventory 49 1615700 1926226 97554 8741797 receivable 49 1744901 1963524 222711 1.24E+07 ROA 49 7.657755 8.515531 -8 32.39 Current ratio 49 43.50102 15.81472 16.18 77.03 financial leverage 49 5.231837 15.33214 -6.36 107.6 sales growth 49 24.75224 34.19029 -62.96 123.99 stock return 49 1.279141 0.199725 0.4681 1.5686 beta 49 31.09696 46.33825 -71.8902 131.9323 stock turnover 49 426.6669 272.148 53.23 1050.93 cpa 49 0.4489796 0.5025445 0 1 ROE 49 11.39224 15.10981 -27.3 47.38
estimated litigation risk 49 -10.26559 37.68125 -92.03281 72.16824 * the estimated litigation risk is shown as the percentage.
As can be seen, firms without D&O insurance tend to be small. The mean of the firms with D&O sample is 16.24 in log of the total assets, as compared to 15.9 for the mean in the non D&O sample. The differences are highly significant using the t-test. Taken at face value, this is intuitive with the D&O purchasing incentive since previous work has documented a positive relation between firm size and D&O insurance.
The non D&O group has higher beta and stock turnover (the proportion of shares traded at least once in the year prior to resignation, as defined below) than the other. A stylized fact is that class-action lawsuits tend to be precipitated by large drops in stock prices. The more volatile the stock price, the greater the probability of a substantial price drop, leading to higher legal exposure. Stock turnovercould also relateto plaintiffs’incentivesto initiate lawsuits because shareholder damages typically increase in the number of shares traded at the allegedly wrong price. Turnover is calculated as[1−Πt(1−volumetradedDayt/totalsharesDayt)].
The non D&O sample has higher stock return than the other sample. The mean return is 1.28, as compared to 0.45 for the with-D&O firms. In addition, the non D&O sample is closer to
financial distress as reflected by the lower ROA, lower current ratio, and higher leverage. Firms’ financial distress can affect the incentive for purchasing D&O indirectly via its relation with litigation risk since financial distress has been shown to increase the probability of litigation.
To test H1(a), we will compare the mean of the summary measure of litigation risk for the firms with D&O insurance purchase in 2004 with that for the firms without purchasing D&O insurance in 2004.
Group(D&O) Obs Mean Std. Err. Std. Dev. [95% Conf. Interval] 0 49 -10.2656 5.383036 37.68125 -21.0889 0.557726 1 32 17.11404 5.825197 32.95229 5.233473 28.99461 combined 81 0.551052 4.236947 38.13253 -7.88074 8.982846
diff -27.3796 8.159489 -43.6207 -11.1386
t-statistics degrees of freedom Pr(|T| > |t|)
-3.3556 79 0.0012
The p-value are statistically significant which shows that the mean of estimated litigation risk in two groups(with vs. without D&O insurance) are significant different. Therefore, out hypothesis 1 is supported that the demand for D&O insurance is to hedge the risk.
4.3. D&O insurance and firm performance
Our next question is to test whether the firms purchasing D&O insurance have higher operation performance. We use return on equity (ROE) as the performance measure to test hypothesis 2. We group our sample into two sub-samples, one with D&O insurance and one
without D&O insurance, based on 2004 D&O insurance data. We then compared the performance difference of the two groups with respect to their ROE.
Group(D&O) Obs Mean Std. Err. Std. Dev. [95% Conf. Interval] 0 49 11.39224 2.158544 15.10981 7.0522 15.73229 1 32 14.13781 3.396797 19.21519 7.209999 21.06563 combined 81 12.47691 1.865482 16.78934 8.764486 16.18934
diff -2.74557 3.827585 -10.3642 4.87305
t-statistics degrees of freedom Pr(|T| > |t|)
-0.7173 79 0.4753
The p-value are not significant which shows that the mean of the performance in two groups (with vs. without D&O insurance) are not in significant difference. This finding does not
hypothesis 2. However, we shall not interpret this result as the proof that D&O insurance has no effect on the firm performance.
In fact many factors can affect the firm performance. In order to draw a meaningful conclusion, we need to further study the board of directors structure and the experiences.
5. Conclusion
In this paper, we refine the risk assessment test and find that the demand for D&O insurance is driving by the risk concerns of the firms. This result is consistent with the previous literature. Our contribution is to introduce a new risk estimation method and find the result is robust. However, we can not find the performance difference between the two groups, with and without D&O insurance. This result may be subject to estimation error, because the firms buying the D&O are experiencing higher risks. The D&O insurance may encourage better board members and subsequently improves the operation risk and performance, comparing with its own. In this paper, we derive the conclusion based on the group comparison which may be problematic.
In the future research, we many need to look at the board structure to infer the effect of D&O insurance on the firms operation. The performance measures may be required to make the risk adjustment before using for comparison.
Bibliography
Alles, M., S. Datar, and J. Friedland, 2005, “Governance-Linked D&O Coverage: Leveraging the Audit Committee to Manage Governance Risk,”International Journal of Disclosure and
Governance, June, 114-129.
Bailey, D., 2005, “D&O Liability in the Post-Enron Era,”International Journal of Disclosure and
Governance, June, 159-176.
Chalmers, J., L. Dann, and J. Harford, 2002, “Managerial Opportunism? Evidence from Directors and Officers’Insurance Purchases,”Journal of Finance, April, 609-636.
Core, J., 1997, “On the Corporate Demand for Director’and Officers’Insurance,”Journal of Risk
and Insurance 64, 63-87.
Core, J., 2000, “The Directors’and Officers’Insurance Premium: An Outside Assessment of the Quality of Corporate Governance,”Journal of Law and Economics, October, 449-477. Fama, E. and M. Jensen, 1983, “Separation of Ownership and Control,”Journal of Law and
Economics, June, 301-325.
Gutierrez, M., 2003, “An Economic Analysis of Corporate Directors’Fiduciary Duties,”Rand
Journal of Economics 34, 516-535.
Holderness, O., 1990, “Liability Insurers as Corporate Monitors,”International Review of Law and
Economics, 10, 115-129.
Jensen, M. and W. Meckling, 1976, “Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure,”Journal of Financial Economics 3, 305-360.
Mayers, D. and C. Smith, 1982, “On the Corporate Demand for Insurance,”Journal of Business 55, 281-296.
O’Sullivan, N., 2002, “The Demand for Directors’and Officers’Insurance by Large UK Companies,”European Management Journal, October, 574-583.
Parry, M. and A. Parry, 1991, “The Purchase of Insurance by a Risk-Neutral Firm for a Risk-Averse Agent,”Journal of Risk and Insurance 58, 31-46.
Redington, W., 2005, “D&O Underwriting Implications of Sarbanes-Oxley,”International Journal
Sarath,B.,1991,“Uncertain Litigation and Liability Insurance,”Rand Journal of Economics,
218-231.
Shavell, S.,1982, “On Liability and Insurance, ”Bell Journal of Economics, 120-132. Shu, S., 2000, “Auditor Resignations: Clientele Effects and Legal Liability,”Journal of
Accounting and Economics 29, 173-205.
Sloan, F., B. Reilly, and C. Schenzler, 1995, “Effects of Tort Liability and Insurance on Heavy Drinking and Drinking and Driving,”Journal of Law and Economics, 38, 49-77.
