نظارت بر دارنده بلوک و بهره وری تعیین معیار عملکرد پرداخت

In this article I examine how the performance sensitivity of CEO compensation is related to the level and turnover of outside block ownership. Separating firm performance into firm-specific (Skill) and exogenous (Luck) components, I find that pay sensitivity to Luck increases with blockholder turnover, whereas pay sensitivity to Skill increases with blockholding size. Furthermore, when blockholder turnover is higher, CEO pay increases more with positive Luck but does not decrease as much with negative Luck; also, excess CEO compensation is larger. Thus, the rent accruing to CEOs via asymmetric pay sensitivity to Luck is partly explained by short investment horizons of large shareholders.

Recent research on executive compensation has focused on the incentive aligning efficiency of pay arrangements by more closely examining how CEO pay is benchmarked to firm performance. In a seminal study, Bertrand and Mullainathan (2001) document that CEOs get paid for exogenous shocks that affect firm performance (luck). Following this study, Garvey and Milbourn (2006) find that CEOs get paid for luck via asymmetric benchmarking of pay to performance; that is, pay increases when performance goes up because of exogenous shocks, but it does not decrease as much when performance goes down because of exogenous shocks. Although these regularities do not provide prima facie support for the efficient provision of incentives, recent theoretical and empirical studies argue that this asymmetry in benchmarking is optimal (Celentani & Loveira, 2006 and Gopalan et al., 2010). In this article, I examine the asymmetric benchmarking of CEO compensation by relating it to the monitoring role of large shareholders with heterogeneous investment horizons. Large-block shareholders (henceforth blockholders) are prevalent in U.S. corporations (Dlugosz et al., 2006 and Holderness, 2009) and have been shown to play significant roles in corporate governance. Blockholders do this through a combination of intervening in firms' operations (Shleifer & Vishny, 1997, Bethel et al., 1998, Kahn & Winton, 1998 and Gorton & Kahl, 2007) and trading on private information (Edmans, 2009). I capture these aspects of blockholders' roles in firm governance by using two dimensions of block ownership: its level and turnover. If asymmetric pay-performance benchmarking is associated with inefficiencies in setting managers' compensation, more pronounced asymmetric benchmarking should be expected when block ownership is lower and blockholder turnover is higher. The traditional contracting view of pay-performance benchmarking suggests that rewarding CEOs for exogenous changes in firm value (i.e., changes that are beyond management control) is suboptimal because this would make contracts riskier and costlier but would not create incremental alignment of incentives (Holmstrom, 1979 and Holmstrom, 1982). Therefore, following Bertrand & Mullainathan, 2001 and Garvey & Milbourn, 2006, I split firm performance into two components: firm-specific performance (Skill) and firm performance driven by exogenous factors such as industry performance (Luck). I relate the level and the turnover rate of block ownership to the sensitivity of CEO pay to overall firm performance, each of the Luck and Skill components of performance, and the asymmetry in the relation between pay and Luck. Following the literature, I measure outside block ownership as the sum of ownership stakes that individually constitute at least 5% of the voting rights and that are not affiliated with firm management. The extant literature examining ownership turnover focuses on institutional ownership and captures the effect of turnover by classifying institutional owners based on the frequency with which they turn over stocks in their portfolio (Bushee, 1998, Bushee, 2004, Gaspar et al., 2005 and Chen et al., 2007). By examining the turnover of block ownership at the investee firm level, I extend the literature on investor horizons to this important group of investors. I calculate block ownership turnover as the proportion of total block ownership that changes hands annually, thus developing a firm-specific measure that allows for heterogeneity in a particular blockholder's behavior across firms.1 A greater rate of block ownership turnover indicates a shorter average investment horizon. All else equal, blockholders with longer investment horizons are assumed to be more active in monitoring, and thus turnover rates of block ownership are used as an inverse proxy for blockholder monitoring.2 I start my examination by relating the pay-performance sensitivity (PPS) of CEO compensation to outside block ownership and find that the sensitivity of CEO compensation to firm performance is positively associated with both the level and the turnover rate of outside block ownership. However, the sources of the increased PPS due to each of these dimensions of block ownership are different. When I split total firm performance into Skill and Luck, I find that higher blockholder turnover (i.e., shorter average investment horizon) is associated with increased pay sensitivity to Luck, whereas higher level of block ownership is associated with increased pay sensitivity to Skill. These results suggest that CEO compensation is more sensitive to Skill than Luck when firms have better monitoring by blockholders as measured by their larger stakes and longer investment horizons. I next test whether large shareholders reduce asymmetric pay-performance benchmarking by examining the effect of block ownership and blockholder turnover on PPS when firm performance is conditioned on the sign of Luck. I find that the level of block ownership is not differentially associated with PPS based on whether Luck is positive or negative. However, greater blockholder turnover makes CEO pay more sensitive to Luck only when Luck is positive. When Luck is negative, greater blockholder turnover leads to smaller PPS to Luck, preventing CEOs from being penalized for poor Luck.3 This result suggests that shorter investment horizons of blockholders can exacerbate the asymmetry in pay-performance benchmarking. I further examine the effect of blockholder turnover on the asymmetric sensitivity of pay to Luck. First, I test whether the exacerbating effect of blockholder turnover on this asymmetric sensitivity is mitigated by the existence of a director blockholder. I expect that a director blockholder compensates for limitations in monitoring arising from short horizons of other blockholders. Consistent with this expectation, I find that the presence of a director blockholder considerably decreases pay sensitivity to Luck in general and increases the sensitivity of pay to Luck when Luck is negative, thus making CEO pay sensitivity to Luck more symmetric. Second, I test whether pay sensitivity to Luck and Skill varies across the following blockholder types: (1) banks, insurance companies, investment companies, and investment advisors; (2) pension funds; (3) corporations; and (4) activists and individual investors. Although in the current sample these four groups of blockholders are generally similar in terms of their blockholding size, their turnover rates differ from one another. The average turnover rate is 24.89% for banks, insurance companies, investment companies, and investment advisors; 13.43% for pension funds; 6.14% for corporations; and 4.99% for activists, and individuals. Consistent with this variation in turnover rates, pay sensitivity to Skill generally increases in the level of block ownership when the block is owned by pension funds or by activists and individual investors. In contrast, pay sensitivity to Skill decreases in the level of corporate block ownership, suggesting that the motivations of corporate blockholders differ from those of activist or individual blockholders. When the block is owned by banks, insurance companies, investment companies, or investment advisors, an increase in pay sensitivity is observed only for cash compensation. More importantly, the relation between blockholder turnover and asymmetric pay sensitivity to Luck continues to hold even after accounting for the effect of block ownership type, suggesting that blockholder turnover and blockholder type, despite their similarities, are not perfect surrogates for each other. Finally, I test whether greater asymmetry in pay-performance benchmarking due to higher blockholder turnover results in greater excess CEO compensation. I expect it will because CEOs of high-blockholder-turnover firms are more likely to receive higher pay when Luck is positive and smaller penalties when Luck is negative, thus resulting in higher pay on average. I find that total CEO compensation and equity-based compensation are larger in firms with higher blockholder turnover, whereas cash-based compensation is not. The overall results support the argument that monitoring by shareholders leads to increases in the pay-performance sensitivity of CEO compensation. However, these results could be noisy because of the endogeneity of blockholding characteristics to other determinants of CEO PPS. I address the simultaneity issue by including a comprehensive set of control variables, by examining the role of director blockholders and effects of blockholder type, and by testing the relation between block ownership and excess compensation. Regarding the issue of reverse causality (i.e., turnover and block ownership are determined by PPS), it is unclear ex ante why blockholders would choose firms with high asymmetric sensitivity of CEO compensation to Luck. The current results are relevant to the ongoing debate over the relative performance evaluation puzzle (Antle & Smith, 1986, Aggarwal & Samwick, 1999a, Aggarwal & Samwick, 1999b and Celentani & Loveira, 2006). Gopalan et al. (2010) argue that asymmetric benchmarking may be required to incentivize the CEO to strategically change the firm's exposure to industry factors (i.e., Luck in the present context). Thus, asymmetric benchmarking could be optimal. In contrast, Harford and Li (2007) document an increase in asymmetric sensitivity to Luck for acquiring firm CEOs in the postmerger period and conjecture that acquisitions are an important managerial tool for CEOs to achieve asymmetry. The current results suggest that asymmetric benchmarking is, at least partially, due to relatively poor governance of firms with smaller block ownership and blockholders with shorter investment horizons. The results also extend the literature on block ownership. Contrary to the notion of widely dispersed ownership structures in U.S. corporations, a recent study by Holderness (2009) reports the prevalence of block ownership in U. S. corporations. His findings empirically underscore the call for a richer understanding of the governance role of large shareholders (e.g., Bethel et al., 1998 and Demsetz & Villalonga, 2001). By examining both the turnover and the level of block ownership, the current research responds directly to this call. The rest of this article is organized as follows. Section 2 reviews the relevant literature and develops hypotheses. Section 3 discusses the sample and presents descriptive statistics. Section 4 discusses the empirical methodology and reports the results. Section 5 reports robustness tests, and Section 6 concludes.

In this article I introduce the investment horizon of blockholders as an additional dimension of block ownership and examine how heterogeneous investment horizons of blockholders, in addition to block ownership size, are associated with firms' policies on executive compensation. I find that larger block ownership is associated with improvements in PPS of CEO compensation via increased sensitivity of CEO pay to Skill, where Skill is defined as the firm-specific component of stock performance during the fiscal year. Similarly, longer investment horizons of blockholders measured by lower turnover rates of block ownership are associated with improvements in pay-performance benchmarking via decreased pay sensitivity to Luck, where Luck is defined as the component of stock performance due to the industry and the market. I also find that lower turnover rates of block ownership mitigate the asymmetric benchmarking of pay to Luck documented by Bertrand & Mullainathan, 2001 and Garvey & Milbourn, 2006. The association between blockholder turnover and asymmetric pay sensitivity to Luck is robust to various alternative explanations, such as other governance mechanisms, blockholders' investment strategies as measured by blockholder type, and prior firm performance. These findings make two contributions to the literature. First, the results suggest that investment horizon is an important determinant of monitoring activities even for shareholders with large ownership stakes. Demsetz and Villalonga (2001) argue that as a monitoring mechanism, ownership structure should be represented by more than just its concentration. To my knowledge, this is the first study that explicitly links blockholder investment horizons to firm policies. Second, I add to the literature on the efficiency of pay-performance benchmarking. Responding to the suggestion of poor corporate governance implied by evidence of asymmetric benchmarking of pay to luck (Bertrand & Mullainathan, 2001 and Garvey & Milbourn, 2006), recent studies provide optimal contracting explanations of this asymmetry as an alternative (Celentani & Loveira, 2006 and Gopalan et al., 2010). By documenting systematic associations between pay-performance benchmarking and block ownership and turnover, I argue that at least a part of the inefficiencies in pay-performance benchmarking arise from poor governance.