Print
E-MailShane Johnson, Professor of Finance at Texas A&M University. This post is based on an article authored by Professor Johnson; Timothy Campbell, Assistant Professor of Finance at Miami University of Ohio; and Neal Galpin, Associate Professor of Finance at the University of Melbourne. Related research from the Program on Corporate Governance includes Executive Pensions by Lucian Bebchuk and Robert J. Jackson Jr.
Four decades ago, Jensen and Meckling (1976) provided the first analysis of a hypothetical compensation contract that included both equity and debt for a CEO. But until Bebchuk and Jackson (2005) and Sundaram and Yermack (2007) provided early analyses based on then newly available data on CEO pensions, the dominant view was that debt-like claims held by CEOs were theoretically interesting but not empirically relevant. With the new data and evidence, a natural question arose: What is the optimal mix of CEO debt and equity?
The optimal contract in Jensen and Meckling (1976) is “one size fits all,” setting CEO leverage equal to the firm’s leverage regardless of CEO or firm characteristics. This solution strikes a balance between aligning the manager too closely with shareholders who prefer more risk, thereby making debt more expensive, and aligning the manager too closely with debtholders who prefer excessive safety, thereby making the firm too safe. Edmans and Liu (2011) develop a more complicated model in which the CEO can also exert effort with payoffs that differ in solvent and insolvent times. The optimal CEO leverage can be higher or lower than the firm’s leverage depending on the impact and value of the CEO’s effort. For example, if the value of CEO effort is higher in solvent times, the CEO’s leverage should optimally be lower than the firm’s leverage, i.e., tilted more toward equity incentives which pay off in solvent times.
In our paper, Optimal Inside Debt Compensation and the Value of Equity and Debt, which was recently featured in the Journal of Financial Economics, we use the optimal contracting framework in Edmans and Liu (2011) to create a target CEO leverage ratio based on various firm and CEO characteristics. We then show that as firms move toward the target leverage based on theory, stock and bond values rise. Prior literature such as Wei and Yermack (2011) already suggests that reducing a too-high managerial leverage ratio would help shareholders. We contribute to the literature by showing that approaching the theoretically motivated target CEO leverage—whether increasing or decreasing CEO leverage—increases firm and shareholder value.
Our empirical approach is easy to implement and should be useful in other settings as well. We regress CEOs’ leverage ratios on CEO and firm characteristics that theoretically determine optimal CEO leverage. The difference between the CEO’s actual leverage and the estimated target level indicates whether a given CEO’s leverage is too high or too low. If the deviation from the target is less positive one year than the prior year, firm value should rise if the target really is optimal. Likewise, if the deviation from the target is less negative one year than the prior year, firm value should also rise if the target is optimal. We therefore examine stock and bond price reactions to the announcement of the CEO’s leverage and how they relate to changes in the magnitude of the deviation. We first show that the average firm adjusts the CEO’s leverage to move towards the target level. We then show that such adjustments lead to increases in both shareholder and bondholder value.
We also run a “horse race” between the simple one-size-fits-all target prescribed by Jensen and Meckling (1976) and our more complex target based on Edmans and Liu (2011). We that show moving toward the more complex target increases stock prices while moving toward the simple target does not. Importantly, the complex target’s success in the horse race does not mean that the Edmans and Liu theory is a complete description of the optimum. More work is needed to fully understand the trade-offs involved in compensating a manager with debt versus equity.
Finally, and perhaps most surprisingly, we find little role for governance in estimating a value-relevant target. The results that we describe above are generally similar whether or not we include variables to measure firm governance when we estimate the target optimal leverage ratio. One case does not hold, however: adjusting a too-high leverage downward does not help shareholders if we include governance variables when estimating the target. Because our concern is whether shareholders benefit from the adjustments, we take this as evidence that the optimal contracting variables (firm and CEO characteristics) do a better job of generating a good target optimal leverage than do governance variables. We acknowledge the possibility that governance might matter, but the importance of governance may be difficult to measure using the governance variables in our analysis.
Our research has three main implications. First, the extent of debt-like claims in a CEO’s compensation package matters for firm value. Second, it is possible to estimate an optimum target leverage ratio that outperforms the “one-size-fits-all” level prescribed decades ago. Third, optimal contracting variables based on firm and CEO characteristics suggested by recent theory are more important in constructing an optimal leverage target than are a firm’s governance variables.
The full paper is available for download here.
