Relative Pay and CEO Turnover
Eric A. Fong
University of Alabama in Huntsville
Xuejing Xing[*]
University of Alabama in Huntsville
Abstract
Despite the fact that the majority of CEO turnover incidents can be classified as voluntary, little research has explored whether CEOs’ internal incentives to voluntarily leave their position affect CEO turnover. We fill this gap by investigating empirically the impact of CEO incentives on CEO turnover. Using relative CEO payas a proxy for CEOincentives, we find that these incentives can be significant determinants of CEO turnover.
Relative Pay and CEO Turnover
I. Introduction
Previous studies typically focus on firm performance in explaining CEO turnover because the board of directors or the market for corporate control would discipline poor-performing managers. However, empirical evidence shows that firm performance explains little of the cross-sectional variation in CEO turnover (see Brickley (2003) for a review of the literature). This weak connection between CEO turnover and firm performance has prompted researchers to suggest that “[w]e will have to consider other less-explored issues to increase our understanding of CEO turnovers and replacements” (Brickley, 2003).
In this paper we examine a potentially important, but largely overlooked, issue in explaining CEO turnover, that is, CEOs’ internal incentives to voluntarily leave their position. These internal incentives could play an important complementary role in driving CEO turnover because as many as 85% of all CEO turnover incidents can be classified as voluntary (Goldman, Hazarika, and Shivdasani, 2003). If voluntary turnovers make up a large proportion of all CEO turnovers, then CEOs’ internal incentives should be an important factor for these events. In addition, even forced turnovers could be related to CEOs’ internal incentives because these incentives affect CEO effort and thus firm performance.
As a proxy for the CEO’s internal incentives, we use the CEO’s pay relative to the market wage rate for the CEO, or relative CEO pay. In a competitive CEO labor market where firms compete for CEO talent, if a firm provides higher pay relative to what other similar firms offer, then the CEO of the firm wouldhave less incentive to leave the current position. To measurerelative CEO pay, we regress total CEO pay on a set of known determinants of CEO compensation, following Core, Holthausen, and Larcker (1999) andMurphy (1999). The residuals from this regression indicatethe approximate percentage difference between the actual CEO pay and the predicted fair market wage rate, or the extent to which CEOs are under- or overpaid relative to the executive labor market wage rate.
Using the hierarchical generalized linear modeling (HGLM) techniqueto minimize the potential influence of data non-independence, we find in a sample of 3,359CEO-year observations that relative CEO pay is a significant predictor for CEO turnover: the lower the relative pay, the higher the probability for CEO turnover to occur in the next period. We thus provide direct evidence on the importance of the CEO’s internal incentive variables such as relative pay in affecting CEO turnover.
II. Data
The Sample
We randomly select 800 (about 10% of the average annual number of U.S. firms on Compustat) U.S. firms from the 1995 Compustat files. For these 800 firms, there are a total of 6,076 firm-year observations on Compustat from 1991 to 1999. For these 6,076 firm-year observations, we manually collect CEO human capital data such as age, tenure, and experience from proxy statements, 10-K reports, annual reports, and Dun Bradstreet’s Reference Book of Corporate Management. We obtain other CEO data from Execucomp and Compact Disclosure, and all the accounting data items from Compustat. After dropping 2,717 firm-year observations for which there are no sufficient CEO data, we are left with an initial sample of 582 firms, 860 CEOs, and3,359CEO(firm)-year observations covering the period 1991–1999.
We use this initial sample to construct the relative CEO pay variable for each and every CEO-year observation. We then lag the key variables by one year as required by the test below. This lagging creates 860 CEO-year observations with missing lagged variables. Finally, we drop these 860 observations along with additional 346 observations with other missing variables and end up with afinal sample of 547 firms, 632 CEOs, and 2,153CEO-year observations.
Definition of CEO Turnover
As in Mikkelson and Partch (1997) and DeFond and Park (1999), we define CEO turnover as a dummy variable that takes on the value of one if there is any change in the identity of the individual who holds the CEO position.
Control Variables
CEO ageis the biological age (in years) of the CEO. CEO tenureis the number of years since the CEO was appointed to the position. CEO market competitionis the number of firms in the same two-digit SIC industry, as inBushman, Dai, and Wang (2007). Outside director ratiois the number of outside directors divided by the total number of directors. CEO/Chairmanduality is a dummy variable equal to one if the CEO is the chairman of the board and zero otherwise. Firm size is the natural log of the book value of total assets. Firm-specific riskis the standard deviation of residuals from a regression of daily stock returns on CRSP value-weighted index returns over the fiscal year. Firm age is the number of years since the firm was first covered by CRSP. As a general measure of firm performance we use Tobin’s Q, because better-performing firms should receive higher market valuations. We compute Tobin’s Qas the market value of assets divided by the book value of assets. The market value of assets is thebook value of assets plus the market value of common stock less the sum of the book value of common stock and the balance sheet-deferred taxes.
Measurement of Relative CEOPay
Similar to Murphy (1999)and Core, Guay, and Larcker (2008),we estimate the following CEO pay modelto determinerelative CEO pay:
Log(Total CEO pay) = β0 + β1Firm size + β2Firm-specific risk +β3Tobin’s Q
+ β4GDP growth rate+ β5Block shareholder + β6Owner-managed
+ β7Outside director ratio + β8CEO/Chairman duality
+ β9CEO age + β10Inside hire + β11 CEO experience
+ βi Location dummies+ βjYear dummies+ βj Industry dummies+e, (1)
where Total CEO pay consists of cash, incentive pay, stock options, stock grants, deferred pay, fringe benefits, and pension accruals. We value stock options using a modified version of the Black-Scholes (1973) method, allowing for the inclusion of dividend payments as suggested by Murphy (1985). Block shareholderis a dummy variable which is equal to one if there is a private owner other than the CEO with at least 5% ownership in the firm and zero otherwise. Owner-managed is a dummy variable equal to one if the CEO owns at least 5% of the firm and zero otherwise. We measure CEO experienceusing a 5-point scale with oneindicatingprior position (e.g., a CEO in another firm) most relevant and five suggesting prior position (e.g., a scientist or an engineer) least important to being a CEO. Inside hireis a dummy variable equal to one if the CEO is appointed from inside the firm and zero otherwise. We include 17 corporate-headquarter-location dummy variables to allow for the differences in pay levels across different regions. To account for the differences in CEO pay across industries, we include a dummy variable for each industry as defined by four-digit SIC codes. Year dummies are included to account for differences across time.
Table I reports the results of estimating the CEO pay model. It is clear that Firm size, Tobin’s Q, Outside director ratio, and CEO/chairman duality are significantly and positively related to CEO pay and that Owner-managed,CEO age, and Insider hire are significantly and negatively associated with CEO pay, consistent with prior research (e.g., Conyon and Murphy (2000), Wade, O’Reilly, and Pollock (2006), Cao and Wang (2007)). The CEO pay model appears to perform well, explaining more than 47% of the variance in CEO pay.
Similar toCore, Guay, and Larcker (2008), we take the residuals as our measure of relative CEO pay from the total pay regression. The magnitude of the residuals tellsthe approximate percentage difference between the actual CEO pay and the predicted fair market wage rate, or the extent to which the CEO islower- or higher-paid compared to his fellow CEOs. A positive sign on the residual indicates that the CEO is paid higher than his peers, and a negative sign suggests the opposite.
III. Results
To investigate the influence of relative CEO pay on CEO turnover, we follow previous studies (e.g., Parrino, 1997; DeFond and Park, 1999; Bushman, Dai, and Wang, 2007) and control for a set of variables known to affect CEO turnover: the age, size, and performance(Tobin’s Q)of the firm;the age and tenure of the CEO;and CEO market competition. We add outsider director ratio because it is more likely for the board to discipline the CEO if there are more outside directors on the board. Weisbach (1988) provides empirical evidence that the ratio of outside directors does affect CEO turnover. We also control for firm-specific risk because under-diversified CEOs (who typically hold a large position in their own firms) are more likely to consider a job in another firm with a lower level of risk. We include a dummy variable for CEO/chairman duality because it might be more difficult for the board to discipline the CEO if he also assumes the chairman position on the board (Goyal and Park, 2002). Thus, the baseline model we use in explaining CEO turnover is:
CEO turnovert= β0 + β1 Relative CEO payt-1 + β2 Tobin’s Qt + β3 Firm sizet
+β4 Log (Firm aget) + β5 Firm-specific riskt
+ β6 Outside director ratiot + β7 CEO aget + β8 CEO tenuret
+ β9 CEO/Chairman dualityt +β10 CEO market competitiont + e. (2)
Because the observations in our sample are not independent but nested in firms, which in turn are nested in industries, OLS is not appropriate. To correctly account for data non-independence, we take advantage of the HGLM technique. This technique also allows use to appropriately test binary (CEO turnover) outcomes using a Bernoulli sampling model and logit link (Raudenbush and Bryk, 2002).
Table II reports the HGLM results. We find that relative CEO pay is significantly and negatively related to CEO turnover, thereby suggesting that higher CEO pay relative to the executive labor market rate helps retain CEOs. We also find that the outside director ratio is significantly and positively associated with the probability of CEO turnover and that CEOs holding the chairman position are less likely to depart from the CEO position. In addition, the results suggest that both firm size and risk are significantly and positively correlated with CEO turnover, consistent with Bushman, Dai, and Wang (2007). The relation between CEO tenure and CEO turnover is significant and negative, suggesting that longer tenured CEOs are less likely to experience CEO turnover. Consistent withHarrison, Torres, and Kukalis (1988),Brickley (2003), and others, we find no significant relation between CEO turnover and firm performance.
IV. Conclusion
In this paper we examine whether CEOs’ internal incentives to voluntarily leave their position can explain CEO turnover. In doing so, we construct a sample of 3,359 CEO-year observations from 1991 to 1999. Using relative CEO pay as a proxy for the CEO’s internal incentives, we find that the lower the relative CEO pay the higher the probability for CEO turnover to occur in the next period. Our results thus suggest that the CEO’s internal incentive variables can be important determinants of CEO turnover.
References
Black, F. and M. Scholes (1973) Pricing of Options and Corporate Liabilities,Journal of Political Economy, 81, 637-654.
Brickley, J. (2003)Empirical Research on CEO Turnover and Firm-performance: A Discussion,Journal of Accounting and Economics, 36, 227-233.
Bushman, R., Z. Dai, and X. Wang (2007)Risk and CEO turnover, Unpublished working paper,University of North Carolina at Chapel Hill.
Cao, M. and R. Wang (2007)Search for Optimal CEO Compensation: Theory and Empirical Evidence, Unpublished working paper,YorkUniversity and College of William and Mary.
Conyon, M. and K. Murphy (2000) The Prince and the Pauper? CEO Pay in the United States and United Kingdom,The Economic Journal, 110, 640-671.
Core, J., W. Guay, and D. Larcker (2008)The Power of the Pen and Executive Compensation,Journal of Financial Economics, 88, 1-25
Core, J., R. Holthausen, and D. Larcker (1999) Corporate Governance, Chief Executive Officer Compensation, and Firm Performance,Journal of Financial Economics 51 (March), 371-406.
DeFond, M. and C. Park (1999)The Effect of Competition on CEO Turnover,Journal of Accounting and Economics,27, 35-56.
Goldman, E., S. Hazarika, and A. Shivdasani (2003)What Determines CEO Turnover? Unpublished working paper,University of North Carolina at Chapel Hill.
Goyal, V. and C. Park (2002) Board Leadership Structure and CEO Turnover,Journal of Corporate Finance 8 (January),49-66.
Harrison, J., D. Torres, and S. Kukalis(1988)The Changing of the Guard: Turnover and Structural Change in Top-management Positions,Administrative Science Quarterly, 33, 211-232.
Mikkelson, W. and M. Partch (1997) The Decline of Takeovers and Disciplinary Managerial Turnover,Journal of Financial Economics, 44, 205-228.
Murphy, K. (1985)Corporate Performance and Managerial Remuneration,Journal of Accounting and Economics, 7, 11-42.
Murphy, K. (1999)Executive Compensation, in Handbook of Labor Economics, Vol. 3, (Eds.) O. Ashenfelter and D. Card, North Holland, Amsterdam.
Parrino, R. (1997)CEO Turnover and Outside Succession: A Cross-sectional Analysis,Journal of Financial Economics,46, 165-197.
Raudenbush, S. and A. Bryk (2002)Hierarchical Linear Models: Applications and Data Analysis Methods, 2nd ed., Thousand Oaks, CA, Sage Publications, Inc.
Wade, J., C. O’Reilly,and T. Pollock(2006)Overpaid CEOs and Underpaid Managers: Fairness and Executive Compensation,Organization Science, 17, 527-544.
Weisbach, M. (1988) Outside Directors and CEO Turnover,Journal of Financial Economics20 (January-March), 431-460.
Table I
Estimation Results of the CEO Pay Model
Coefficient estimate / Standard errorConstant / 4.66*** / 0.19
Firm size / 0.37*** / 0.01
Tobin’s Q / 0.16*** / 0.01
GDP growth rate / -0.02 / 0.07
Firm-specific risk / -0.93 / 1.12
Block shareholder / 0.02 / 0.08
Owner-managed / -0.30*** / 0.09
Outside director ratio / 0.24** / 0.11
CEO/Chairman duality / 0.14*** / 0.04
CEO age / -0.01*** / 0.00
Inside hire / -0.19*** / 0.04
CEO experience / -0.03 / 0.03
Number of observations / 3,359
R2 / 0.47
* *p < .05; ***p < .01, two-tailed.
1
Table II
The Impact of Relative CEO Pay on CEO Turnover: HGLM Results
Coefficient estimate / Standard errorConstant / -2.28*** / 0.09
Relative CEO pay / -0.45*** / 0.09
Tobin’s Q / 0.01 / 0.05
Firm size / 0.19*** / 0.04
Log (Firm age) / -0.13 / 0.09
Firm-specific risk / 20.30*** / 4.12
Outside director ratio / 1.16*** / 0.38
CEO age / 0.01 / 0.02
CEO tenure / -0.36*** / 0.03
CEO/Chairman duality / -0.43*** / 0.08
CEO market competition / 0.00 / 0.00
χ2 / 454.86***
***p < .01, two-tailed.
1
[*] Corresponding author. Department of Accounting and Finance, University of Alabama in Huntsville, Huntsville, AL35899. Phone: 256-824-6493. Fax: 256-824-2929. Email: .