Family firms and the incentive contracting role of accounting earnings

Abstract

We investigate how family controls affect the sensitivity of the variable pay of top management to the firm’s accounting-based performance (PPS) in order to deal with both the traditional agency problem between owners and managers, and the central agency problem between controlling and minority shareholders. Using five-year data from Taiwan-listed firms, our results show that PPS is stronger for both non-family managers employed by family firms, and family managers, and is weakest for managers in non-family firms. Additionally, PPS is more pronounced in family firms with potential central agency problems, when the CEO is not a family member than when the CEO is.

Keywords:

• contracting role of accounting earnings
• family control
• central agency problem

JEL codes:

• G34
• J33

Notes

1. The incentive contracting role of accounting earnings, also known as the pay-for-performance sensitivity (PPS) for accounting earnings, refers to how sensitive variable pay is to accounting earnings (e.g. Bushman, Engel, and Smith 2006; Banker, Huang, and Natarajan 2009; Banker, Darrough, Huang and Plehn-Dujowich 2013, 11). In this study, this is measured by the coefficient of accounting earnings in the variable compensation function. Fundamentally, PPS refers to the change in a CEO’s payoff, which is associated with the performance of the firm that she or he leads (Murphy 1985). Researchers following this line of thought use Change Model in order to examine the association between a change in CEO variable compensation and changes in the value of the stock-price and earning performance (e.g. Sloan 1993). Practically, for the purpose of contracting, the principal exploits the sensitivity of performance measures (such as earnings and stock-price) to managerial effort in order to create incentives (Bushman, Engel, and Smith 2006). Researchers also utilize Level Model in order to investigate the association between the level of CEO variable compensation and the levels of stock price and earnings performance (Banker, Huang, and Natarajan 2009; Banker et al. 2013). Our view is consistent with those in the latter group, and focuses on whether earnings are incrementally useful in the variable compensation contracts of various managers (family managers vs. non-family managers employed by family firms vs. managers in the non-family firm) after considering stock-price performance.

2. We are grateful to the valuable comments by an anonymous referee. In the earnings management literature, the well-known bonus hypothesis posits that earnings-based compensation may induce earnings management by discretionary accruals (Healy 1985; Watts and Zimmerman 1986). Although CEOs can use discretionary accruals to manipulate current yearly income in order to increase their wealth, we would expect that managerial entrenchment will be minimal in family firms with a non-family CEO. Family owners typically have the power, incentive, and knowledge to run the business well (Anderson and Reeb 2003), and in turn closely monitor a non-family CEO, thereby reducing the non-family CEO’s incentive to exploit information asymmetry. In Section 4.5, we further analyze the link between the performance-based pay of a non-family CEO and earnings management.

3. The definition of family firms in our study follows that of Anderson and Reeb (2003), who define family firms in which the founder, or a member of his or her family by either blood or marriage, is an officer, director, or blockholder, either individually or as a group. Additionally, parents have control of power if they can appoint or remove the majority of the members of the board of directors, or can cast the majority of votes at meetings of the board of directors or an equivalent governing body, and control of the entity is by the board or body (IAS No. 27 “Consolidated and separate financial statements”). If family members hold more than 50% of the seats on the board, then family members have control of power in the firm.

4. The results remain unchanged when the variable capturing the family firm dummy, FMDM, is replaced by another indicator variable; family member ownership greater than 20%, FSH_DM (La Porta, Lopez-De-Silane, and Shleifer 1999; La Porta et al. 2000; Villalonga and Amit 2006). We also use alternative classifications of family firms in which family members hold at least 5 and 10% of the common stock of the firm (Tsao, Chen, and Chen 2013). Our results remain similar with these revised classifications.

5. It is worth noting that the available information on some managers’ compensation includes only the gross amount, which does not distinguish cash bonus from stock-based pay. Following Lin and Hu (2003), we first calculate the per-month change in the manager’s stock holdings. Because the change in the manager’s stock holdings occurs almost only on stock ex-rights dates or dividend pay-out dates, we can arrive at the amounts, which are possibly due to stock bonuses granted by adjusting the stock dividends. Next, we are able to obtain the market value of stock-based pay by utilizing the original stock price on the shareholder meeting date, multiplied by the proportion of total increasing shares to original total shares outstanding.

6. Previous studies examine the relationship between pay and firm performance. The causal direction of the relationship between compensation and performance is unclear (Kole 1996). Core, Guay, and Larcker (2003) suggest specifying structural and reduced-form equations to solve this issue. This study uses Core, Guay, and Larcker (2003) to examine hypotheses with previous performances, but some studies use current performance to examine the impact of performance on pay (e.g. Yermack 1995; Hung and Wang 2008). This study also uses current performance to replace past performance, and the results are similar.

7. We will suffer from small sample size with the traditional PPS method. Therefore, this research follows Murphy (1993) and Lippert and Porter (1997) to calculate PPS, which is simpler and less likely to measure with error, calculated as follows: .

8. When the PPS is measured as ΔCOMP /ΔROA, we find the negative denominator or numerator. The results remain unchanged when we exclude the negative denominator or numerator and re-run Equation (4(4) ).

9. Regarding H3, we also use the traditional PPS method. The PPS in Change Model is defined as the estimated slope coefficient (λ₁) in the below regression equation. Our inferences are robust to implementing the traditional PPS method.

10. We use EM to replace abRPS and re-run Equation (4(4) ). We empirically test whether earnings-based compensation induces earnings management in family firms. Our measure of earnings management (EM) is based on Dechow and Dichev (2002), , where ∆WCA is the increase in accounts receivable plus an increase in inventory plus a decrease in accounts payable and accrued liabilities plus a decrease in taxes accrued plus an increase (decrease) in other assets (liabilities), scaled by average assets; and CFO is the cash flow from operations, scaled by average assets. The results of earnings management are similar to those of related-party transactions.

11. The results of sensitivity test are not present in the table due to the page limitation.