Subjective Evaluations: Discretionary Bonuses and Feedback Credibility

Subjective Evaluations: Discretionary Bonuses and Feedback Credibility We provide a new rationale for the use of discretionary bonuses. In a setting with unknown match qualities between a worker and a firm and subjective evaluations by the principal, bonuses are useful in order to make the feedback from the firm to the workers credible. This way workers in good matches are less inclined to accept outside offers. JEL Classification: D82, D83, D86, M5 Keywords: discretionary bonuses, feedback, signalling Corresponding author: William Fuchs Haas School of Business University of California Berkeley 545 Student Services Building Berkeley, CA 94720-1900 USA E-mail: [email protected] * An earlier version of this paper was titled “Subjective Evaluations: The Bonus as a Signal of Performance.” It is based on my dissertation work at Stanford GSB. Special thanks are owed to Andy Skrzypacz, Tom Sargent, Ed Lazear and Jon Levin. I also benefited from comments by Simon Board, Ben Hermalin, John Morgan, Aniko Ory, Paul Oyer, Alessandro Pavan and Pablo Spiller. Support from the Ayacucho Fellowship and the National Science Foundation is gratefully acknowledged. 1 Introduction Why do we see bonus pay?1 The literature has mainly focused on two main answers to this question: 1) as a result of moral hazard on the agent’s side, contingent pay is needed to motivate the worker to exert e¤ort; 2) the agent, typically a CEO, has private information about his …t for the job and accepts contingent pay as a way to signal he is a good match for the task. Neither of these models provide a good explanation for why employees who neither produce a clear measurable output nor are in top management positions should receive contingent pay. In this short paper, we propose a di¤erent explanation that applies to agents further down in the …rm’s hierarchy and whose output cannot be easily and objectively measured. In such environments, it is natural for the agent to be unaware of how well her superiors believe she is performing. The principal pays the agent a bonus conditional on good performance to signal to the agent that she is indeed a good …t for her job and has a bright future with her current employer. The bonus is not used to motivate the agent’s e¤ort or to have productive agents ex-ante self-select into the job, but rather its role is to make the principal’s feedback credible. To illustrate the problem, think of a business analyst at a consulting …rm or a junior investment banker. These workers do not produce an easy to measure product. Moreover, they have little and hard to measure impact on the bottom line of the consulting …rm or bank. Furthermore, it would be hard to argue that, in their case, accepting contingent pay is a way for them to signal something to their employers. On the contrary, partners in the …rm can see the analysts working and soon form an opinion on their long run potential with the …rm. In addition, a main concern of employees in these jobs is to determine their likelihood of making it to partner, that is the big carrot motivating long hours and weekends working hard. Many of these talented individuals receive outside o¤ers from industry. Such o¤ers are typically seen as a lower e¤ort lower pay alternatives, but preferable if, indeed, one would not make it to the top. If the partners in the …rm were not making any rents (or quasi-rents) out of their junior employees, then they would be indi¤erent about whether those employees stayed or took an outside o¤er. Hence, they would be happy to truthfully reveal to them their prospects in the company and see the less promising ones leave. On the other hand, if rents or quasi rents are being captured by the partners, then they would have incentives to hold on to their employees, even the less promising ones, for longer and would give all of them positive feedback and false hopes of their prospects of success in the …rm. Employees would then forgo most of their outside options. Clearly, if the employees are aware of the partners behavior, this cannot be an equilibrium: they would not believe the feedback and would indeed be more receptive to outside o¤ers. The partners are not interested in having their talent …shing for outside options and hence would be interested in making their feedback credible.2 Discretionary bonuses, raises or anticipated promotions allow them to do exactly that. By putting their money where their mouth is they can credibly signal to the junior employees their prospects with the company and convince those that are a better …t not to leave the …rm. We could also think of the …rst period as the interview phase and the bonus as a sign in bonus. Again, the 1Although I refer to bonus pay this stands in general for discretionary contingent pay. For example the bonus can be rolled over into next period’s wage. 2 In contrast, Oyer (2004) and Oyer and Schaefer (2005) analyze the use of option grants as a way to have agents receive higher compensation in states where the outside options are likely to be higher. Importantly, these contingent payments are not ex-post discretionary and are simply used to make the compensation adjustable to the overall market and do not convey any match speci…c information.