Marie-Josée Ledoux


The effect of bonus-sharing and cohesion on team behaviour and performance

Read the full thesis (PDF).

Providing workers with a bonus pool based on team performance is a growing practice. Team-based rewards are believed to promote collaboration and, therefore, to increase the success of a team. However, when a bonus pool is equally shared, two workers can receive the same bonus, independent of their contribution to the pool or work effort compared to co-workers. Loss of motivation can result from such an undifferentiated bonus sharing method due to the classic free-rider problem. This paper investigates how different rules for sharing the bonus – equal sharing (ES), sharing based on individual performance (IP), and sharing based on relative performance (RP) – affect the level and mix of helping versus individualistic effort in teams. Team cohesion is incorporated as a mediator between rules and efforts.

This paper develops an agency model of team performance in which each member is accountable for a particular task and in which helping effort increases production and improves efficiency. The model shows the ES rule induces an efficient mix of effort but low levels of both helping and individualistic effort due to free-rider problems. Sharing based on IP induces an efficient level of individualistic effort but low levels of helping effort. Sharing based on RP results in more help than the individual performance case and higher individualistic effort than the equal sharing case. Prior research discourages attempts to differentiate the contributions of various individual team members because it can undermine collective effort. In my model, however, the reduction in collective effort is counterbalanced by the increase in total effort, causing the IP and RP rules to outperform the ES rule.

I incorporate a cohesion parameter in the model; the basic assumption being that cohesion increases the level of effort response among team members. The model shows that as cohesion increases, the total level of effort and the team performance increases under all three sharing rules, which is consistent with the resources invested by firms to stimulate team spirt. The cohesion parameter increases effort levels and/or improves effort mix, depending on the sharing rules considered i.e. there is an interaction between cohesion and sharing-rule. In particular, low levels of cohesion are associated with great differences in outcomes – mix and level of effort as well as performance. With high levels of cohesion, however, all bonus-sharing schemes produce similar (optimal) levels of effort and performance.

Hypotheses from the model are subsequently tested with a between-subjects experiment in which teams of students are randomly assigned to a specific bonus-sharing condition. The experiment was conducted with the participation of 487 undergraduate students from a large public university. Each team performed a computerised task for which each member had a distinct knowledge and could interact with other team members. As predicted, the experimental results suggest that individual effort, total effort, and performance are greater under the RP rule than under the ES rule. However, the results suggest no differences (1) in the levels of effort or performance between the ES and IP rules; and (2) in the mix of effort across all three sharing rules. Only under the RP rule is cohesion positively related to performance. The differences across rules are not reduced as cohesion increases as hypothesized. On the contrary, the experiment shows that the RP rule outperforms both rules when cohesion is high.