emerged that the health of the elderly population, at least on this measure, has been improving since the early 1980s and that this improvement in health is reflected in a declining rate of disability (see, e.g., Manton et al., 1997; Crimmins et al., 1997; Schoeni et al., 2001; Crimmins, 2004; Freedman et al., 2004). Much of this literature has pointed to improvements in medical and assistive technologies for the treatment of disabled individuals, as well as the improved prevention of disabling conditions among the elderly population, as the leading causes of these declines.
These declines in disability came as a surprise to researchers because they reversed the trends of increasing rates of disability seen among elderly people in the 1970s. Then, researchers argued that increases in longevity amounted to extensions of the time spent in disability by elderly people (see, e.g., Crimmins et al., 1989; Waidmann et al., 1995). Thus, they predicted, increases in longevity would inevitably be accompanied by growth in the incidence of disability (see, e.g., Gruenberg, 1977). This concern has been replaced with the happy prospect of a compression of morbidity into shorter periods at the end of longer lives.
On the other hand, the news has not been all good over the past 20 years. For both the elderly and the working-age populations, there have been increases in the prevalence of some important chronic conditions, such as obesity and asthma. Working-age populations in the United States are more likely to claim disability benefits now than they were in 1990 (Autor and Duggan, 2001; Bound and Waidmann, 2000). In addition to increases in the size of the work-limited population, severe activity of daily living (ADL)-style disabilities have also risen in the working-age population since 1980 (Lakdawalla et al., 2003a,b; 2004).
The difference in the rates of changes in disability between the elderly and the working-age populations over the past 20 years is puzzling. If the explanation is the use of medical technology and preventive medicine, then the question is why these have not benefited younger populations in the same way that they have benefited the elderly population. This explanation presumes that at the root there have been real changes in the health of the populations under consideration that have led to the observed changes in the prevalence of disabilities. However, is this root assumption correct? Over this period there has been real deterioration in some measures of health in working-age populations, such as the dramatic rise in obesity rates. It seems implausible that such changes would have had no effect on disability rates, but how much is the effect?
The main aim of this paper is to describe the existing evidence on disability and chronic health trends in the population under age 65 years and to present new estimates of a decomposition of disability trends among working-age populations into two parts: the part of the trend explained by changes in the prevalence of chronic disease and obesity and the part of the