全球智库信息集成服务系统

The issue brief provides new statistics on the attributes of the jobs held by applicants to Social Security Disability Insurance (SSDI) before they applied to SSDI. Occupation data is collected from applicants when they seek SSDI benefits, but have not been used frequently in empirical research because they are collected in open-ended text fields. We use an innovative supervised machine-learning approach to map the open-ended text to a Standard Occupation Classification (SOC) code for a sample of SSDI applicants in 2007 and 2014. Using that information, we compare job attributes of the occupations held by SSDI applicants before they applied to SSDI with the job attributes of all workers in the same years. Our findings shed light on the types of policies that might be most beneficial to help workers with medical problems stay in the labor force instead of seeking SSDI benefits. The information in this brief highlights findings from a longer manuscript (Wu 2018).