The Short and Long Run Impacts of Centralized Clearinghouses: Evidence from Matching Teach For America Teachers to Schools

Many labor markets have adopted centralized clearinghouses using variants of the deferred acceptance algorithm (DAA) in hopes of improving market outcomes. Despite the prevalence of these clearinghouses, evidence of their impacts is limited. This paper presents the first evidence from a field experiment of the impacts of adopting the DAA on short- and long-run match outcomes from a new application of market design. I worked with Teach for America (TFA) to match high school teachers to schools in Chicago using the DAA at a series of âinterview days,â while keeping its original âFirst Offer Mechanismâ (FOM) unchanged for elementary school teachers. I show that the FOM gives almost no autonomy to teachers and promotes strategic early hiring in theory and in practice. In contrast, the variant of the DAA I implemented allows teachers to pick their most preferred offer and I show that it is non-manipulable via preferences by either teachers or schools. I use a difference-in-difference strategy - comparing changes in outcomes over time for TFA high school teachers to the change among TFA elementary school teachers in Chicago - to measure the impact of adopting the DAA on short-run outcomes, like matches, hiring, sorting across schools, and preferences over matches, and on three long-run outcomes: teachersâ retention, satisfaction, and performance. I find that adopting the DAA reduces attrition through teachersâ two-year commitment to TFA by 9.9 percentage points (pp). This effect is driven by a 7.2 pp reduction in attrition prior to TFAâs initial summer training. While teachers were happier, they were not more productive. Using TFAâs preferred performance metric, I estimate that teachers who could have been matched with the DAA were 0.3 standard deviations less effective in their first year, but were nearly equally effective in their second year. However, up to two thirds of the decline in first-year performance can be attributed to the DAA shifting teachers to schools with lower average performance. The decline is not explained by higher retention among lower performing teachers. Revealed preference suggests the DAA was better for TFA Chicago: it was used at all of its interview days in 2015 and 2016.