About the Paper

Download Paper

In the paper, the researcher aimed to study the major predictors of disability onset for older workers in the United States and the role of various employer accommodations in retaining newly disabled workers in the workforce using nationally representative data. The researcher used machine learning (ML) as a primary tool to identify major predictors of disability onset for older workers due to its ability to predict unforeseen events in the future and generalizability achieved through cross-validation (Allen (1974), Stone (1974), Geisser (1975)), and propensity score matching (PSM) to measure the impact of employer accommodations on newly disabled workers.

Research Questions

• Are there significant differences in job characteristics, including occupational job requirements, working conditions and workplace-related benefits, as well as worker characteristics, by disability and accommodation status of workers near retirement in the United States?
• What are the major predictors of disability onset before normal retirement ages for workers near retirement?
• What is the impact of different types of employer accommodation on labor supply and disability claiming decisions of workers who experience disability onset before normal retirement ages?

Key Takeaways

• A descriptive analysis showed that the average number of years of education was significantly higher for individuals who never experienced a work disability (“No disability group”) than individuals who experienced a work disability before their normal retirement age (“Disability group”), and the “No disability group” also had fewer female respondents, a higher marriage rate, and significantly higher household assets and earnings.
• The results from the ML models suggest that relying on publicly available survey data such as Health and Retirement Survey (HRS), Occupational Requirements Survey (ORS), and O*NET and ML techniques can be useful in predicting disability onset of older workers in their later life.
• Workers who experienced disability onset and were accommodated by their employers were 13.6 percentage points more likely to stay in the labor force in 2 years after the disability onset, and 7.7 percentage points more by the next survey wave, i.e., up to four years after onset.

Citation

Lee, A. (n.d.) RAND Corporation. Role of Job Characteristics and Employer Accommodation in Labor Supply and Disability Benefit Claiming Decisions in Later Life: Analysis of Employee’s Decisions to Continue Working after Disability Onset or Claim SSDI/SSI Benefits Before Normal Retirement Age. Chief Evaluation Office, U.S. Department of Labor.

Download Paper   View Study Profile

This study was part of CEO’s Summer Data Challenge on Equity and Underserved Communities, and was produced outside of CEO’s standard research development process.