Print
E-MailPierre Chaigneau is Associate Professor & Commerce ’77 Fellow of Finance at Queen’s University, Alex Edmans is a Professor of Finance, Academic Director, Centre for Corporate Governance at London Business School, and Daniel Gottlieb is a Professor of Managerial Economics and Strategy at the London School of Economics and Political Science. This post is based on their recent paper. Related research from the Program on Corporate Governance includes Paying for Long-Term Performance (discussed on the Forum here); Executive Compensation as an Agency Problem; Pay Without Performance; The Unfulfilled Promise of Executive Compensation, all by Lucian Bebchuk and Jesse M. Fried; and The Growth of Executive Pay by Lucian Bebchuk and Yaniv Grinstein.
Standard models of executive compensation assume that a manager cares about pay only because it allows him to consume more. He will only improve performance if doing so increases his pay, and the utility from the resulting extra consumption exceeds the cost of the effort required to improve performance. Such models have been highly influential and inspired a stream of empirical research.
However, it is not clear that consumption utility is the only, or even the most important, driver of CEO pay in practice, given that CEOs are typically wealthy and nearly all of their consumption needs are already met. A recent paper by Edmans, Gosling, and Jenter (2022) surveys directors and investors on how they set pay contracts. Both sets of respondents highlight how pay is driven not only by the desire to provide consumption incentives, but also the need to ensure the CEO feels fairly treated. This is consistent with prior research suggesting that pay is a hygiene factor — pay above a certain level provides limited additional motivation, but pay below that level is a strong demotivator.
The respondents also suggest that firm value is an important determinant of what directors, investors, and the CEO view to be a fair level of pay. If firm value has increased due to CEO effort, they believe that it is fair to reward the CEO for this increase. If firm value has increased (decreased) due to luck outside the CEO’s control, they believe the CEO should share in this good (bad) luck.
Our new paper studies how to set CEO pay when the CEO is motivated by both traditional consumption utility and fairness concerns. We model fairness concerns by specifying a perceived fair wage that is increasing in the firm’s output; output in turn depends on both CEO effort and luck. The CEO suffers disutility if his wage falls below the fair wage, the magnitude of which is increasing in the discrepancy. In our baseline model, the fair wage is linear in output, i.e. the CEO believes he should receive a certain percentage of output, and the disutility is linear in the discrepancy.
It may seem that fairness concerns should lead to the CEO always receiving a fair wage, but this turns out not to be the case. The optimal contract involves a threshold below which the CEO is paid zero, and above which he receives the fair wage. This contradicts the intuition that fairness concerns will lead to the CEO always receiving a fair wage. Instead, fairness concerns mean that unfairness can be a powerful motivator. If output is sufficiently low that it is a bad signal about CEO effort, the firm pays him the most unfair possible wage of zero. Only if output exceeds a lower threshold is the CEO paid the fair wage. Depending on parameter values, there may also be an additional upper threshold above which the CEO is paid the firm’s entire output.
The contract resembles performance shares, where the CEO receives shares that are forfeited if performance falls below a threshold. Even though they are frequently offered in reality, standard models, such as Holmström (1979), do not predict discontinuous contracts. Innes (1990) predicts a sharp discontinuity where the CEO’s pay increases from zero to the entire output, once output crosses a threshold, but such sharp discontinuities do not exist in reality. Our paper obtains milder and thus more realistic discontinuities — when performance crosses a threshold, the wage jumps from zero, but not to the entire output. Intuitively, performance shares provide fair wages if performance is good and unfair wages (zero) if performance is bad. This is an efficient way to motivate good performance.
We then extend the model to a non-linear one, where the fair wage is increasing but not necessarily linear in output, the utility loss is increasing but not necessarily linear in unfairness, and utility is increasing and concave in the wage if it is fair. The key features of the model continue to hold.
In all variants of the model, pay is increasing in output even if the CEO does not need financial incentives to work hard, for example due to intrinsic motivation. This is because fair wages are instead needed to persuade the CEO to accept the contract. If the CEO knows that he will not be rewarded for good performance – i.e. he does a good job yet is not recognised for this afterwards – he will not accept the contract to begin with. As a survey respondent in Edmans, Gosling, and Jenter (2022) remarked, no talented person would stay at a firm where he does not feel appreciated. Thus, observing that CEOs are paid for performance in reality need not automatically imply that they need financial incentives; instead, pay-for-performance can be driven by fairness concerns rather than incentive concerns.
A frequent criticism of performance-related pay for CEOs is that it should not be necessary – the CEO should be intrinsically motivated to exert effort, and/or the board should monitor CEO effort. Our model demonstrates that performance-related pay may be optimal not to induce effort, but to attract or retain a CEO with fairness concerns. CEOs are paid for performance not because they would refuse to work without the carrot of additional consumption, but because it is human to want to be recognised for a job well done.
The complete paper is available for download here.
