Machine learning identification of who benefited most from the Reemployment and Eligibility Assessment (REA) program

• In Event: New Experimental Evidence on the Reemployment Services and Eligibility Assessment (RESEA) Program

Abstract

This presentation provides new evidence on the long-term impacts of a program to support and monitor Unemployment Insurance (UI) claimants. The Reemployment and Eligibility Assessment (REA) Program was designed to reduce UI duration and improve employment outcomes by connecting participants to employment services and ensuring compliance with UI program rules. These mechanisms are also used in the Reemployment Services and Eligibility Assessment (RESEA) Program, the successor to REA that was codified through 2018 amendments to the Social Security Act. Abt Global evaluated REA in four states in 2015 and 2016 for the U.S. Department of Labor, finding favorable impacts on employment in the short-term and in the long-term, on average. The new results are based on an exploration of how REA affected the stability of claimants’ employment over an eight-year follow-up period, including four years before and four years after the pandemic downturn of 2020, and the subgroups of claimants who gained the most from REA over the follow-up period. The subgroup analyses use machine learning methods to identify claimants with the largest impacts after REA participation. Given the similarities between REA and its successor program, these results can inform development and assessment of new RESEA interventions—in line with the RESEA evidence-building aims laid out in Social Security Act.

Authors

• Andrew Clarkwest, Abt Global
  Presenting Author

• Correne Saunders, Abt Global
  Non-Presenting Co-Author

• Heinrich Hock, Abt Global
  Non-Presenting Co-Author