CEO compensation and corporate risk: Evidence from a natural experiment

Abstract

This paper examines the two-way relationship between managerial compensation and corporate risk by exploiting an unanticipated change in firms' business risks. The natural experiment provides an opportunity to examine two classic questions related to incentives and risk-how boards adjust incentives in response to firms' risk and how these incentives affect managers' risk-taking. We find that, after left-tail risk increases, boards reduce managers' exposure to stock price movements and that less convexity from options-based pay leads to greater risk-reducing activities. Specifically, managers with less convex payoffs tend to cut leverage and R&D, stockpile cash, and engage in more diversifying acquisitions. Disciplines AbstractThis paper examines the two-way relationship between managerial compensation and corporate risk by exploiting an unanticipated change in firms' business risks. The natural experiment provides an opportunity to examine two classic questions related to incentives and risk-how boards adjust incentives in response to firms' risk and how these incentives affect managers' risk-taking. We find that, after left-tail risk increases, boards reduce managers' exposure to stock price movements and that less convexity from options-based pay leads to greater risk-reducing activities. Specifically, managers with less convex payoffs tend to cut leverage and R&D, stockpile cash, and engage in more diversifying acquisitions.  Convex managerial payoffs and firm risk are known to be correlated, but the direction of causality remains unclear. We exploit a change in risk to show how boards adjust incentives and how these incentives affect managers' risk-taking. We find that firms reduce managers' exposure to stock price movements after left-tail risk increases. Because the change in risk is unanticipated, we can also identify compensation's effect on risktaking in the new environment. Our results show that convexity from options can encourage risk-taking in corporate investment and financing decisions.

Management Risk Incentives and the Readability of Corporate Disclosures

“…Options on company stock rose to account for almost half of all CEO pay before corporations began expensing option grants in 2004 (Gormley, Matsa, and Milbourn, 2013) and continues to account for about one-fourth of CEO compensation (Frydman and Jenter, 2010). CEO options-based compensation has increased over time.…”

Background: Cost-of-Capital in the Finance Literature

“…As Karake-Shalhoub and Petty (2002, p. 239) point out, the diverging interests between principals and agents also "get translated into differences in risk preferences". The specific risk behaviour of managers-and hence the extent of the divergence problem-depend on different factors such as their compensation (Gormley et al 2013). In the following two paragraphs, two examples of diverging risk preferences and their consequences for capital investment and financing decisions are discussed.…”

CEO Tenure and Risk-Taking

“…A large strand of literature examines the relationship between managerial compensation incentives and risk taking (e.g., DeFusco et al 1990;Tufano 1996;Schrand and Unal 1998;Guay 1999;Cohen et al 2000;Chen et al 2006;Coles et al 2006;Chakraborty et al 2007;Chava and Purnanandam 2010;DeYoung et al 2010;Gormley et al 2011;Brick et al 2012). More recently, scholars begin to shift their attention to the effect of the innate attributes of managers on their risk-taking incentives, such as sensation seeking, overconfidence, education, military background, depression-era life experiences, religious belief, and political affiliations (Bertrand and Schoar 2003;Hilary and Hui 2009;Benmelech and Frydman 2010;Graham et al 2010;Hirshleifer et al 2010;Hutton et al 2011;Malmendier et al 2011;Cain and Mckeon 2012).…”

“…The risk-reducing effect of vega, however, is harder to reconcile. In contrast, most empirical studies document a positive effect of vega on risk-taking (e.g., Coles et al 2006;Chakraborty et al 2007;Low 2009;Gormley et al 2011). Subsequently we show the importance of including the lagged dependent variable in the effect of vega on risk-taking, which may explain the difference between the results.…”