Association for Public Policy Analysis & Management

Signals of quality are routine in many professions, yet we are still learning about how individuals respond to their own signals when making supply-side employment decisions. In this study, we measure the employment response of public school teachers to well-defined signals of quality before and after those signals are credibly tied to demand-side employment considerations. Specifically, we study whether teachers receiving a signal of their quality decide to exit a public school system immediately following a credible threat that tenure protection is ending. We do this by exploiting continuous measures of performance that are used to determine sharp designations as a low-, medium-, or high-performing teacher, along with the passage of a law ending tenure protection in North Carolina. This allows for a granular difference-in-differences approach along the dimension of the continuous quality signal. Preliminary estimates find that both low- and high-performing teachers are more likely to exit the public school system, but there are important differences in response. Of those with less than three years experience, high-performing teachers are more likely to leave. Across grade levels, high-performing high-school teachers are more likely to leave while low-performing elementary school teachers are more likely to leave. Given the widespread implementation of tenure protections in the United States, these findings provide valuable insight for agencies looking to restructure teacher tenure laws. In addition to the topical contributions, this work highlights important considerations for research related to measuring responses to repeated signals that are tied to sharp designations. Specifically, we discuss how responses to these signals deviate from traditional conceptions of regression discontinuity designs, and demonstrate how the granular difference-in-differences approach allows for the identification of behavioral responses when agents can respond to both continuous and discrete signals of quality.