Prolonged unemployment carries serious economic, health, and wellbeing costs. With federal support, most U.S. states now operate a Reemployment Services and Eligibility Assessment (RESEA) program to help Unemployment Insurance (UI) claimants return to work faster. We report results from a large (N = 23,549) preregistered randomized controlled trial (RCT) evaluating Rhode Island's RESEA program from February 2022 to September 2023. We estimate that selection into the program increased annualized wages by \$1,153, increased reemployment by 1.5 percentage points, and reduced UI duration by nearly two weeks. The vast majority of these wage and reemployment effects appeared within two quarters of claimants' first pay dates and persisted through at least the following year, and we estimate that each dollar spent on the program saved the state \$2.64. Using causal forests, a machine learning technique for estimating heterogeneous treatment effects (HTE), we also conduct an exploratory analysis to investigate if there are differential effects of selection into the RESEA program. We find that all participants experienced positive wage benefits from RESEA selection, with particularly large effects for older and lower-income workers. Finally, we improve upon prior RESEA evaluations by explicitly controlling for the week of treatment assignment -- a methodological refinement absent from several existing RCTs of job-training programs that is important to eliminate confounding bias. We also discuss ways to harvest precision gains from baseline covariate adjustment without introducing large-sample bias.
🌊
