Reemployment Services and Eligibility Assessments (RESEA) in Maryland— Formative Evaluation, Program Year 2019

Abstract

Unemployment insurance (UI) exists to provide temporary partial wage replacement during periods of involuntary unemployment while beneficiaries are actively seeking reemployment. The reemployment effort required of UI beneficiaries, which balances the work disincentive of income replacement, ensures that UI is social insurance rather than social welfare. In 2017, Congress appropriated funding to provide reemployment services and eligibility assessments (RESEA) to UI beneficiaries. The legislation also required that states receiving RESEA conduct annual evaluations to produce causal evidence that reemployment services and eligibility assessments are effective. In this formative evaluation, we produce the first causal effect estimates of the Maryland RESEA program for participants in program year 2019. Using a comparison-group design and administrative microdata, we find that participation in RESEA, relative to participation in Worker Profiling and Reemployment Services (WPRS), reduces UI benefit year compensation by 0.62 weeks, reduces the probability of UI benefit exhaustion by 3.1 percentage points, and decreases the proportion of benefits received by 2.3 percentage points. We also find that RESEA increases the probability of employment in the quarter following the benefit year begin date by 1.9 percentage points but does not affect medium-run employment and earnings outcomes. Results suggest that Maryland’s RESEA program successfully met its stated goal of reducing UI duration by increasing employment rates in the short term, but the program does not seem to offer a longer-term solution to improving UI beneficiaries’ labor market outcomes. Our evaluation design was driven by the available data, which include indicators of program participation but no information on referral to reemployment services programs. As in all states, Maryland assigns WPRS profiling scores, which measure the probability of UI benefit exhaustion, to all beneficiaries who are required to engage in an active search for reemployment. That is, UI beneficiaries who are neither union hiring hall members nor awaiting employer recall. Then, within each county, Maryland refers the 50 percent of UI beneficiaries determined most likely to exhaust their benefits to RESEA and the remainder to WPRS. We show, however, that distributions of profiling scores do not differ between RESEA and WPRS participants, and that observed proportions of UI benefits received are uncorrelated with profiling scores. In light of this, as a basis for this formative evaluation, we assume that assignment to RESEA or WPRS is as good as random, conditional on observable characteristics. We test the robustness of results to alternative specifications and matching models. We also