Trends in Work-Related Disability and What Drives Them: Evidence from the Current Population Survey and the National Health Interview Survey 1997-2014

This paper seeks to resolve a discrepancy between two literatures of economics that has arisen in recent years. The literature from public health finds that health is slightly declining and activity limitations are increasing among nonelderly adult Americans while the labor literature in general does not find that health and disability play a role in decreasing or stagnating prime-age labor force participation. The first reason the two subfields disagree in their results is that there is a disagreement in trends between their two major data sources, the Current Population Survey (CPS) and the National Health Interview Survey (NHIS). Although the two surveys sample from the same population and ask similar questions about health status and work-related disability, the NHIS generally reports steeper trends in health and disability in the nonelderly since 1997 than the CPS. The discrepancy holds across age and racial/ethnic groups.

Oaxaca decompositions of regressions of disability on demographics, health, and activity limitations that compare 1997-2000 with 2010-2014 in both datasets show that the two surveys agree that declines in health are contributing to increases in work-related disability. Decompositions that also include functional and activity limitations from the NHIS show that increases in those together with declines in health status and shifts in demographics can entirely account for the increase in work-related disability. When they are included, activity limitations contribute about twice as much to the increase in disability as health status. The largest contributions are made by increases in memory problems, trouble with instrumental activities of daily living, trouble walking, and trouble standing for two hours. The NHIS also asks what condition categories are causing the activity limitations; decompositions with the condition categories indicate that the largest contributions are made by increases in depression, anxiety, and emotional problems; musculoskeletal conditions; and neurological conditions.

Models using the NHIS that substitute two major risk factors, obesity and smoking, for health status and activity limitations show that the risk factors do not account at all for the rise in disability caused by declines in health status. The positive contribution made by rising obesity is completely offset by the negative contribution made by the decline in smoking. It is therefore unclear why activity limitations are increasing.

The two surveys also agree that the coefficient effect in the decompositions with only health status is negative, that is, that the underlying trend of work-related disability conditional on health status and demographics is declining. In the CPS, this effect is in the constant term and is therefore unexplained. The NHIS shows that the effect is explained by individuals who have a physical limitation being less likely to report in the later period that they have a work-related disability. The overall results are therefore consistent with the results found by both fields: health is decreasing and activity limitations are increasing but work-related disability is not.