Citation
Benus, J., Poe-Yamagata, E., Wang, Y., & Blass, E. (2008). Reemployment and Eligibility Assessment (REA) Study: FY 2005 Initiative: final report. Columbia, MD: IMPAQ International. [Minnesota sample]
Highlights
- This study’s objective was to assess the effectiveness of the Reemployment and Eligibility Assessment (REA) initiative in Minnesota, a program that provided eligibility and reemployment case management services to Unemployment Insurance (UI) claimants, on UI benefits receipt, wages, and employment.
- The study randomly assigned 5,898 eligible claimants to one of two treatment groups, which received REA services of varying intensity, or the control group. Those in the treatment groups were required to participate in the services to retain their UI eligibility. The authors estimated the program’s effectiveness by analyzing state administrative records on UI payments and information from a follow-up survey of a subsample of study participants.
- The study found that UI claimants in the more-intensive treatment group claimed fewer weeks of UI benefits and were less likely to have an overpayment detected. There were no significant impacts on employment or earnings for either treatment group.
- The quality of causal evidence presented in this report is moderate because it is a randomized controlled trial with different selection processes for the treatment and control groups, but the authors controlled sufficiently for baseline characteristics in the analysis. This means we have some confidence that the estimated effects are attributable to REA, although other factors might also have contributed.
Intervention Examined
The Reemployment and Eligibility Assessment (REA) Initiative in Minnesota
Features of the Intervention
Since its inception in 2005, 40 states have implemented the REA initiative to encourage rapid reemployment of UI claimants through a combination of in-person eligibility reviews and employment-focused case management. Minnesota implemented two different dosages of REA: either one REA counseling session or up to four sessions on a monthly basis. Services were provided at 12 different One-Stop Career Centers (now known as American Job Centers) in the state, concentrated in the Minneapolis, St. Paul, Duluth, and St. Cloud areas.
The Minnesota REA initiative targeted participants at moderate risk of exhausting UI benefits; that is, those with profiling scores in the middle third of the distribution. To be eligible, UI claimants had to have an active UI claim and payment status and have no hold or denial on claims due to a nonmonetary determination.
Features of the Study
From April 2005 through March 2006, Minnesota randomly assigned 5,898 eligible UI claimants into one of two treatment groups, which could access REA services at different levels of intensity, or to a control group, which did not have access to REA but could access other existing services in the community.
The study analyzed UI benefits receipt outcomes using administrative records and employment and earnings information from a follow-up survey administered at least six months after the initial claim. For UI benefits receipt outcomes, the authors estimated multivariate regressions controlling for age, gender, race, ethnicity, education, veterans’ status, disability status, years in current occupation, number of eligible claim weeks allowed, and county-level quarterly unemployment in 2004 and 2005. For employment and earnings outcomes, the authors estimated similar models that also controlled for the UI profiling score.
Findings
Public benefits receipt
- The study found that UI claimants in the more-intensive treatment group, which could receive up to four REA sessions, had fewer UI weeks claimed and compensated than the control group, a statistically significant difference of 1.2 weeks. In addition, the proportion of claimants in the more-intensive treatment group with some overpayment detected was 3.8 percentage points less than the control group, a statistically significant difference.
Earnings and wages
- The study found no statistically significant impacts of either treatment group on earnings, either at the participants’ first jobs within six months of initial claim or their jobs held at the time of the follow-up interview.
Considerations for Interpreting the Findings
Although the study was designed as a randomized controlled trial, the treatment and control groups were selected using different methods. These differences undermine the key characteristic that supports causal inference in random assignment designs—that there are no systematic differences between research groups. The authors show that there were statistically significant differences between the treatment and control groups on several important characteristics, including previous earnings, which indicate the randomization was unsuccessful. However, the authors included sufficient control variables to mitigate this concern.
Causal Evidence Rating
The quality of causal evidence presented in this report is moderate because it is a randomized controlled trial with different selection processes for the treatment and control groups, but the authors controlled sufficiently for baseline characteristics in the analysis. This means we have some confidence that the estimated effects are attributable to REA, although other factors might also have contributed.