No gifts returned: Surprise bonuses reduce productivity in a natural field experiment

The gift exchange hypothesis posits that workers reciprocate above-market wages with increased productivity. This paper tests this hypothesis in a natural field experiment where one or both workers in a pair received a discretionary bonus after an initial round of data-entry tasks. Bonuses were assigned based on one of three criteria: (i) relative productivity in the initial round, (ii) economic need, or (iii) an arbitrary decision. Two conditions where neither or both workers received a bonus served as the baseline. Contrary to the gift exchange hypothesis, we found a significant decline in post-bonus productivity, especially when both workers received the bonus. This result suggests that workers interpreted the bonus as a signal of employer contentment, allowing them to reduce their effort. We conjectured that the post-bonus productivity decline may result from either (a) a lower perceived risk of repercussions from slacking, such as early dismissal, or (b) a reduced sense of obligation to reciprocate the employer's kindness. A follow-up experiment replicated the primary result, providing moderate evidence for the explanation based on reduced fear of dismissal. The main effect of bonuses on productivity was substantial, with effort reductions of 15.1% in the first experiment and 8.4% in the follow-up relative to baseline. In cases where only one worker received a bonus, non-recipients' inequality aversion appeared to decrease productivity markedly in the economic need treatment, while status-seeking behavior slightly increased productivity by bonus recipients in the productivity treatment.