경영자 보상계약에서 회계성과지표의 사용: 시계열적 변화를 중심으로

We examine the use of accounting information, in particular earnings, as performance measures in executive compensation contracts. This study is partly motivated by the argument that the usefulness of accounting information has declined in recent years (Lev and Gu 2016; Lev 2018). Such argument is often supported by empirical evidence of declines in the relation between stock returns and earnings (i.e., value relevance) in recent years. However, value relevance does not tell whether the usefulness of accounting information in contracts has also decreased. This is an important issue given the extensive use of accounting information in firms’ contracts, such as debt contracts and executive compensation contracts. To test whether the usefulness of accounting information in executive compensation contracts has declined, we examine the time-series changes in the use of accounting information in CEO’s bonus contracts in the U.S. Specifically, we obtain the detailed data on CEOs’ annual bonus contracts from Incentive Lab, including specific performance measures used in determining annual bonus. We use 11,331 firm-years for the 2006-2019 period for our empirical analyses. We intend to provide descriptive evidence on the types of specific performance measures used on CEOs’ bonus contracts. We are particularly interested in providing evidence on the time-series changes in the use of performance measures in bonus contracts. The empirical findings can be summarized as follows. First, we document that the proportion of bonus in total compensation (e.g., 26% on average) has not changed significantly during the sample period, which suggests that annual bonus provides important incentives to top executives (Guay et al. 2019). Second, when we classify performance measures used in bonus contracts into accounting, stock price, and other information, we find that almost all firms (i.e., 98.6%) use accounting performance measures in CEOs’ annual bonus contracts. We also note that the proportion of firms using stock prices is very low at 1.7%. Third, over 90% of the contracts use some forms of earnings as performance measures, and the proportion of firms using earnings in bonus contracts has not substantially changed during the sample period. The proportion of firms using sales or cash flows as a performance measure has increased steadily. Fourth, when we classify earnings-related measures into six categories based on specific items on the income statement, we find that the use of EPS (earnings per share) or net income has declined sharply, whereas the use of EBITDA (earnings before interest, taxes, depreciation, and amortization) has tripled during the period. Fifth, we estimate the logistic regression to examine whether there are significant time-series changes in the use of individual performance measures in bonus contracts. We define a variable to measure the time trend, and we control for firm- and manager-specific characteristics in the model. Empirical results from the logistic regressions indicate that the use of sales and EBITDA has increased significantly during the sample period, while the use of EPS and net income has declined significantly. These results are generally consistent with the time-series patterns observed from the annual means. In additional tests, we examine time-series variations in the performance measures used in performance-vesting stock grants. We find that accounting information is widely used for performance-vesting stock grants, which is also observed for cash-based bonus pay. In terms of time-series changes, we find that the use of EPS has declined, while the use of other earnings (e.g., ROE (return on equity)) and EBITDA has increased during the sample period. We also examine the time-series variation in the performance measures used in cash bonuses based on the “relative” performance evaluation. The proportion of firms using accounting information is around 59%, while the proportion of firms using stock prices is 43.6%. We also provide evidence of the differences in the performance measures employed by financial and non-financial firms. Overall, our results suggest that, despite the increasing doubts about the usefulness of accounting information, it provides useful information for executive compensation contracts. In addition, our analyses of the various performance measures and their time-series changes over the last decade offer important implications for practitioners, academics, and regulatory bodies. As a result, this study complements prior studies on the cross-sectional variation in the use of performance measures in compensation contracts (Curtis et al. 2021; Bloomfield et al. 2021). Finally, our results provide evidence on the variation in the use of specific earnings-related items. Thus, future studies might need to reconsider the common approaches of aggregating several earnings-related items into one category or using only specific types of earnings (e.g., EPS).