Directors’ and officers’ liability insurance and the trade-off between real and accrual-based earnings management^*

* Accepted by Hong Hwang upon recommendation by Yong Gyu Lee.

Abstract

In this study, we examine whether directors’ and officers’ legal liabilities affect their choices of different means of earnings management. We focus on listed firms in Taiwan, where information on directors’ and officers’ liability insurance is publicly available. Consistent with prior studies, we find that insured firms are more likely to use accrual-based earnings management. However, we demonstrate that the level of overall earnings management is marginally lower for insured firms because they rely less on costly real earnings management. Our findings suggest that carrying directors’ and officers’ liability insurance may, to some extent, mitigate costs resulting from earnings management.

Keywords:

• Directors’ and officers’ liability insurance
• real earnings management
• accrual-based earnings management

Acknowledgements

We thank Conrad Chang, Yan-Shing Chen, Sheng-Min Hung, Chan-Jane Lin, Hunghua Pan, Yanzhi Wang, and workshop participants at the 2014 Accounting Theory and Practice Conference in Taipei and at the Seminar on Finance at National Taiwan University. The authors are willing to share the data.

Notes

* Accepted by Hong Hwang upon recommendation by Yong Gyu Lee.

1. Real earnings management does not involve the direct violation of regulations as long as its outcomes are properly disclosed (Chi, Lisic, and Pevzner 2011). Because there is no clear division between real earnings management and the arrangement of ordinary business courses, real earnings management is relatively easy for managers to justify. Managers also consider managing earnings by using real earnings management to be more ethical than employing accrual-based earnings management (Bruns and Merchant 1990).

2. Graham, Harvey, and Rajgopal (2005) indicate that firms have great difficulty in assuring stakeholders that there is no accrual-based earnings management in their financial statements. They report the comment of one survey respondent who expressed ‘a corporate fear that even an appropriate accounting choice runs the risk of an overzealous regulator concluding ex post that accounting treatment was driven by an attempt to manage earnings’ (36).

3. Around 2004, a series of accounting scandals also exploded in Taiwan. In the aftermath of these scandals, Taiwan’s government introduced strict regulations for auditor hiring, auditor tenure, management responsibilities on financial reports, the function of audit committees, and the expertise of board members.

4. Taiwanese insurers provide highly similar D&O insurance policies. Moreover, D&O insurance indemnifies directors and officers only when the sued directors and officers have acted honestly and in good faith. Claims associated with any fraudulent act by insured directors and officers are thus excluded from indemnification in Taiwan. Therefore, we mainly focus on accrual-based earnings management manifested within the bounds of GAAP. However, insured directors and officers might believe that D&O insurance will provide indemnification for their fraudulent acts, as suggested by Lin et al. (2013), who report that firms with high D&O insurance coverage exhibit a high incidence of intentional financial restatements.

5. Directors and officers do not necessarily have the same interests under some circumstances, although theoretically they are all agents of shareholders. However, firms acquiring insurance for their directors generally acquire it for their officers as well. Following prior studies, we treat directors’ and officers’ liability insurance as a whole and control for board and managerial characteristics in our empirical model.

6. Core (2000) suggests that insurance premiums could measure firms’ risk and enhance understanding of firms’ real demand for D&O insurance. Unfortunately, insurance premium information is not disclosed in Taiwan. Thus, D&O insurance coverage may not accurately capture the level of litigation risk and may ignore firms’ real demand for D&O insurance. The possibility of an endogenous relationship between D&O insurance and earnings management may mean that firms engaging in accrual-based earnings management decide to purchase D&O insurance or carry high insurance coverage. We control for firms’ self-selections with regard to carrying D&O insurance in our main tests.

7. ZSCORE equals 0.3(net income/total assets) + 1.4(retained earnings/total assets) + 1.2(working capital/total assets) + 0.6(market value of equity/total liabilities).

8. Net operating assets is shareholders’ equity minus cash and short-term investments plus total debt.

9. Cash holdings are measured by cash and short-term investments divided by total assets at the beginning of the fiscal year. The determinants include ASSETS, MB, LEV, OUTBLOCK, HIGHTECH, CASHFLOW, and NWC, where ASSETS, MB, LEV, OUTBLOCK, and HIGHTECH are defined as those in Equations (5(5) )–(7(7) ). CASHFLOW is operating income before depreciation and amortization less the sum of interest, taxes, and common dividends, divided by total assets at the beginning of the fiscal year. NWC is noncash working capital divided by total assets.

10. Because D&O insurance is an element of corporate governance, we follow Liao and Wang (2015) to control for the quality of corporate governance. Specifically, we perform a principle component analysis and extract eight components from twenty-one corporate governance variables. Each component has an eigenvalue greater than one, and the cumulative percentage of explained variance is 69.39%. We use the Promax oblique rotation and add these composite components in Equations (5(5) ) and (6(6) ). This additional test generates qualitatively similar results (untabulated).

11. Long auditor tenure enhances auditor competence but impairs auditor independence; thus, the effect of auditor tenure on audit quality is still debatable. Because auditor industry expertise, compared with audit firm size and auditor tenure, may be a more pronounced measure of audit quality (Reichelt and Wang 2010; Chi, Lisic, and Pevzner 2011), we add auditor industry expertise in Equations (5(5) ) and (6(6) ). This additional test generates qualitatively similar results (untabulated).

12. We require that the propensity score difference between insured and matched uninsured firms is less than 0.01.

13. Effectively adopted earnings targets also include analyst forecasts and management forecasts. However, these forecasts are not common in Taiwan.

14. Zang (2012) requires that the yearly change in earnings per share before extraordinary items be less than $0.02, which equals approximately NT$0.6.