The Ratchet Effect: Theory and Empirical Evidence

Abstract

We use a dynamic agency model to examine the implications of different target-setting processes for managerial incentives and for the interpretation of empirical tests used to make inferences about the ratchet effect. We assume that firms cannot commit to any specific future targets but they can commit to a process that generates future targets as a function of observed performance. Our findings suggest that commonly used empirical tests based on the target ratcheting model often falsely reject the null hypothesis of no ratchet effect. Alternative tests based on the serial correlation in performance relative to target perform better but can still lead be inconclusive about the incentive effects of target revisions. Next, we show that the ratchet effect can be identified as an effect of past performance on changes in perceived target difficulty. We use panel data from nine annual 2011–2019 surveys to implement this test. Similar to prior studies, we find strong evidence of target ratcheting in that targets are revised upward following good performance. Nevertheless, the estimation approach motivated by our model rejects the ratchet effect hypothesis because we find that good performance in one period is also associated with a decrease in perceived target difficulty in the next period.