A central question for managers in organizations concerns how to effectively motivate workers to provide greater effort. This challenge stems from the classical moral hazard problem that arises when effort is either difficult to observe or cannot be enforced by a third party (like an arbitrator or the court system). If workers are paid a fixed wage, and effort is unobservable or non-contractible, workers have an incentive to shirk rather than provide costly effort.
The typical solution to this problem is to introduce incentives that tie overall compensation to performance. For instance, in addition to wages and salaries, firms increasingly offer compensation packages that rely on performance-based incentives like a bonus. While some bonuses are written into contracts, others rely on subjective assessments of performance. These kinds of bonuses are best described as “discretionary”, since they are made at the discretion of the manager.
In a joint-project with Dr. Sebastian Goerg (a former colleague at FSU, now TU Munich), generously funded by a Planning Grant from the FSU Council on Research & Creativity, we examine the effects of discretionary bonuses on worker performance. Our overarching objective is to better understand how exactly such bonuses work and when they improve employees’ performance.
Both conventional wisdom and research in management practices suggest two key channels through which bonuses operate. The first is as a reward for good performance. The idea is that when a bonus scheme is introduced, workers may increase effort in hopes of earning the bonus as a reward. The second is as a signal of trust, from which the manager aims to elicit a reciprocal response by the worker. That is, by paying a discretionary bonus, the manager may activate in the worker some inclination to reciprocate the manager’s generosity (or demonstration of trust) by working harder.
The most natural setting in which to study the impact of bonuses on worker performance is, of course, inside a real-world organization. Unfortunately, there are many confounding factors that make it difficult to draw precise conclusions using observational data from the field.
Enter experimental economics.
Over the last 40 years, experimental economics has grown in popularity, becoming commonplace within economics departments at major universities all around the world. We are fortunate to have a large, productive group of experimental economists (along with political scientists and accounting researchers) working together as part of the Experimental Social Sciences (xs/fs) Cluster at Florida State. The key advantage of conducting an economics experiment is that it affords the researcher considerably greater control over the decision environment, allowing for clean identification of cause and effect. In this way, the experimental laboratory serves as an excellent test bed for both classical and behavioral economic theory.
In our experiment, subjects were matched up in pairs and assigned to a role, as either a manager or worker, then asked to perform simple yet laborious computerized tasks. Every correctly completed task generated revenue for the manager, but no additional earnings for the worker. Rather, the worker received a fixed salary from the manager, independent of task performance. The interaction lasted for two periods, each 20 minutes long. At the same time, workers could switch from working on the task to browsing the Internet. The availability of the browsing activity provided a real leisure option for the workers to choose instead of working. In this way, our experiment was designed to create the same shirking incentives that exist in many real world organizations.
To explore the effects of a discretionary bonus, we gave managers the option to pay a fixed, one-time bonus to the worker. In treatments, we varied the timing of the bonus decision. In the Start treatment, the decision was made before the work task began. Thus, only the trust channel could be influential on worker effort. In contrast, in the End treatment, the decision was made after the two periods of work were completed. Thus, only the reward channel would be relevant. Finally, in the Middle treatment, managers made their bonus decisions between the two periods of work, allowing the reward and trust channels to reinforce one another.
Our main finding is that only when both channels (reward and trust) are operational, do workers substantially improve their performance. In terms of the timing of the bonus decision, shutting down either one of the channels generates no benefit, on average, over a baseline treatment without bonuses.
The forces that drive these results are quite remarkable. For instance, when the decision whether or not to pay the bonus is made before the work task begins, choosing not to pay has a significantly negative impact on worker performance. In fact, workers who are explicitly denied the upfront bonus perform substantially worse than those in the control treatment where the bonus was never even introduced. This finding highlights the flip side of the trust channel – when managers do not trust enough to pay the worker an upfront bonus, it can lead to negative reciprocity by the worker. Similarly, when the bonus decision is made at the end of the interaction, workers are typically pessimistic about their chances of being paid a discretionary bonus. Indeed, only workers who performed at an exceptionally high level were rewarded with a bonus in the experiment. In contrast, when the bonus decision was made in the middle of the interaction, worker performance increased substantially in the first period, as the workers tried to signal to the manager that they are both productive and trustworthy.
A key take home message that emerges from our study is that reciprocity is a double-edged sword. While a bonus can help to motivate through positive reciprocity, workers may come to expect a bonus in future interactions. Failing to pay an anticipated bonus may have counter-productive effects through negative reciprocity or reduced work morale. On the other hand, if managers are committed to employing discretionary bonuses transparently and in good faith, the twin forces of trust and reward may together create a powerful tool to help solve a fundamental managerial challenge.
Luke Boosey is an Assistant Professor in the Department of Economics. His research uses theoretical and experimental economics to study behavioral aspects of cooperation and competition in a variety of strategic economic settings.
The featured image is from Shutterstock.com.