Association for Public Policy Analysis & Management

In the early 21^stcentury, No Child Left Behind shined a spotlight on the nation’s lowest-performing schools, pressuring many schools, districts, and states to implement turnaround reforms aimed at improving the trajectory of these schools. Through School Improvement Grants and Race to the Top grants, the federal government provided funding to facilitate these turnaround initiatives, which has led to reforms ranging from intrusive state takeover to more collaborative partnerships between districts and schools. Many of these turnaround models have relied on a key element in their theory of change: recruiting and hiring high-performing teachers (USDOE, 2009; 2010). For example, many states and districts have attracted high-performing teachers to difficult-to-teach environments by offering financial incentives. While many low-performing schools have benefitted from these teachers, little attention has been given to an unintended consequence of these turnaround efforts – the effects on schools those high-performing teachers left.

In this analysis, we utilize a statewide, student-level, longitudinal dataset to study the unintended consequences of teacher recruitment into the Memphis Innovation Zone (iZone). The Memphis iZone has been regarded as one of the most successful turnaround initiatives aimed at raising student achievement (Gonzales, 2016; Kebede, 2016; Tillery, 2017; Zimmer et al., forthcoming), and a prominent strategy for iZone schools has been to hire highly effective teachers (“iZone,” 2017). Henry and colleagues (2017; Zimmer et al., forthcoming) show iZone schools have successfully recruited effective teachers; however, many of them came from local districts and even schools within the same district. In this analysis, we ask: To what extent has iZone schools’ practice of recruiting high quality teachers affected the achievement of students in the schools from which they came? Using value-added measures of student performance and focusing on the grade in which the teacher taught in the year prior to moving to an iZone school, we examine the changes in student test score gains after the teachers left for the iZone. In order to approach plausible causal estimates, we employ multiple fixed effect specifications similar to those used in two recent studies (Ronfeldt, Loeb & Wyckoff, 2014; Henry & Redding, 2017). First, we use a school-by-year fixed effect which allows us to exploit variance between grades within a school year to adjust for unobserved school shocks such as hiring a new principal. In this specification, we compare the effects of teacher turnover that occurred in one grade due to a teacher transferring to an iZone school to teacher turnover in other grades that did not lose teachers to the iZone. As an alternative specification, we examine variation over time using school-by-grade fixed effects. In this specification, the within-school differences in student achievement gains before and after teachers transferred to an iZone school are used to estimate the effects from losing a teacher to the iZone. Taken together, if these estimates are consistent, we can consider them as plausibly causal effects with direct implications for the short-run unintended consequences of incentivizing teacher transfers and better understanding the supply of effective teachers within an urban school district.