Measuring work disability in the U.S.: estimates from a structural latent variable model

For the public sector to design effective and efficient programs for persons who have a work disability, there is a need for an accurate measure of work disability. To meet this need, we develop an easy to replicate measure based on a structural-equation model, which treats an individual’s disability as a latent variable that is related to: (1) the abilities and physical conditions of individuals and (2) the capability demands of employers. To empirically capture the individual or labor supply perspective, we use the 2008 panel of the Survey of Income and Program Participation (SIPP). To depict the demand side of the market, we use data from the Occupational Information Network (O*NET) database. In modeling the effect of disability on employment, our work disability estimates show a substantially smaller effect compared to estimates using self-reported work limitation (or health) as exogenous and a statistically larger effect compared to estimates using an objective disability index that excludes labor market information. Our work disability indexes identify individuals with severe work disability, and two sources of bias: (1) non-employed workers who rationalize their decisions not to work by citing work-related handicaps (or poor health) as reasons, and (2) employed workers who downplay their work limitations because they prefer to work.