There is considerable debate over the level of executive pay. On one side, Bebchuk and Fried (2004) and others argue that weak governance allows executives to effectively set their own pay while disregarding market forces and shareholder value. On the other side, Gabaix and Landier (2008) and others argue that executive pay is determined in a competitive labor market, so executives have limited influence on their own pay.
In the paper, CEO Wage Dynamics: Estimates from a Learning Model, forthcoming in the Journal of Financial Economics, I use CEO wage dynamics as a laboratory for exploring this debate. Specifically, I examine how learning about a CEO’s ability affects the level of his or her pay. For example, suppose that after a year of high profits we update our beliefs about a CEO’s ability, and as a result the CEO’s perceived contribution to next year’s profits increases by $10 million. If the CEO obtains a $5 million raise for the following year, then the CEO captures half of the $10 million surplus and shareholders pocket the rest.
This paper’s goal is to measure these surpluses from learning, and also measure how CEOs and shareholders split them. The estimates have important implications for the nature of the executive labor market, CEO compensation contracts, corporate governance, and shareholder value. A CEO’s ability to capture a positive surplus depends on both the CEO and the firm’s outside options in the labor market, on the costs of renegotiating, and on governance strength. Also, a CEO’s perceived ability can drop over time, which raises the question of whether it is the CEO or shareholders who bear the cost of bad news about ability.
Measuring the surpluses from learning presents a challenge, since we cannot directly observe perceived CEO ability, our best signal of CEO ability (stock returns) depends endogenously on perceived ability, and the level of CEO compensation evolves endogenously. These challenges lend themselves to a structural estimation approach, which identifies the quantities of interest from endogenous patterns in the data. I estimate a model in which the CEO and shareholders gradually learn the CEO’s ability, defined as the CEO’s contribution to firm profitability. Agents learn by observing the firm’s noisy profits and an additional, latent signal. Stock prices, return volatility, and changes in the level of CEO pay respond endogenously to news about CEO ability. The CEO’s share of positive surpluses (from good news) and negative surpluses (from bad news) are key parameters that I estimate. I estimate the model using the simulated method of moments (SMM) and data on excess stock returns and changes in total compensation for Execucomp CEOs.
The paper’s main result is that the level of pay responds asymmetrically to good and bad news about CEO ability: the average CEO completely avoids the negative surplus resulting from bad news, whereas he or she captures roughly half of the surplus from good news. Since CEOs capture roughly half of the positive surplus from good news, CEOs and firms appear to have roughly equal bargaining power, on average. Since the level of CEO pay does not fall after bad news about ability, CEOs have downward rigid pay. It is therefore shareholders, not the CEO, who bear the negative surpluses from bad news about the CEO’s ability.
Downward rigid pay does not necessarily imply weak governance. Harris and Holmstrom (1982) show that offering the CEO downward rigid pay can be optimal for shareholders. The reason, according to their model, is that CEOs are risk averse and would like insurance against bad news, and the firm can cheaply provide this insurance in the form of a long-term contract promising that pay will never drop. Consistent with downward rigid pay reflecting optimal contracting rather than weak governance, I find that pay is downward rigid even in the subsample with high institutional ownership, a proxy for strong governance. Also, pay is slightly more downward rigid, although not significantly so, when the CEO has an explicit contract. Downward rigid pay is pervasive across subsamples formed on firm size, industry, calendar year, and ten other characteristics.
According to Harris and Holmstrom (1982), CEOs will accept lower average pay in return for the insurance provided by downward rigidity, which saves the firm money. To quantify these savings, I ask how much more a firm would have to pay its CEO if pay were not downward rigid. I find that the savings are quite large if the CEO is sufficiently risk averse. For example, if the CEO’s relative risk aversion is as high as 4.0, then in return for removing downward rigid pay, the firm would have to increase the CEO’s first-year pay by 1,081%, and the net present cost to the firm of five years of CEO pay would increase by 706%.
The full paper is available for download here.