The Prevalence and Economic Consequences of Disability

With the U.S. Disability Insurance (DI) program serving an ever-increasing number of beneficiaries and the DI trust fund projected to be exhausted by 2016, the DI program is likely to be the object of increasing scrutiny by policy makers in coming years. A solid base of research will be needed to inform the evaluation and design of DI policies. Yet currently there are major gaps in our understanding of the economic consequences of disability.

One study that aims to help bridge these gaps is Disability, Earnings, Income and Consumption (NBER Working Paper 18869) by researchers Bruce Meyer and Wallace Mok. The authors note that designing an optimal DI system requires understanding the frequency of disability, the fall in consumption that results from disability, and the potentially negative effects of DI on work effort ("moral hazard"). While there is an extensive literature on the latter, there is less information on lifetime disability rates and the fall in consumption with disability.

The authors use over 40 years of data from the Panel Study of Income Dynamics (PSID) to estimate lifetime disability risk, as well as the changes in earnings, income, public transfer receipt, poverty, work, food consumption, and housing consumption that follow disability. In taking this wider view, the authors aim to obtain a better picture of the well-being of the disabled. Importantly, the authors' estimates account for underreporting of public transfers, a common problem in survey data. The authors also explore how the effects of disability vary with the severity and duration of the disability. In the analysis, respondents are characterized as disabled if they report that they have "any physical or nervous condition that limits the type of amount of work you can do."

A first key finding is that disability rates are high - a 50-year-old male household head has a 36 percent chance of having been disabled at least once during his working years. By age 50, nearly one in ten (9 percent) of men have begun a spell of disability that is both chronic (lasting at least four years) and severe (greatly limiting or eliminating the ability to work). These odds rise to one in four by age 60.

Second, the authors show that disability is associated with bad economic outcomes. Ten years after onset, respondents with a chronic and severe disability have on average experienced a 79 percent drop in earnings, a 35 percent drop in after-tax income, and a 22 percent drop in food consumption. Two-thirds of this group never returns to work. The declines in income and consumption for the chronically and severely disabled group are more than twice as large as those experienced by average disabled individuals.

Another key finding is that personal savings, family support, and private and government insurance only partially fill the drop in consumption that follows disability. One-sixth of families with a chronically and severely disabled household head fall below the poverty line in the long run, even after accounting for the value of in-kind transfers and the underreporting of benefits. The authors show that consumption begins to fall prior to the reported onset of disability, indicating that future disability status is somewhat (if imperfectly) predictable in the short run.

Some have argued that reported consumption may understand the true consumption of groups such as the retired and disabled because they get more for their money through increased shopping and home food preparation. Evidence from time-use surveys, however, suggests that the disabled do not spend more time on shopping or food preparation. Instead, it appears that they spend more time using medical services, watching television, relaxing, and sleeping. The authors also examine food surveys and find suggestive evidence that the diet of the disabled is worse than that of the non-disabled along a number of dimensions.

Finally, the authors use their findings to consider the optimality of current DI benefits. While there is necessarily some uncertainty in such calculations because of the difficulty of knowing certain key parameters (for example, how the generosity of DI benefits affects the number of people applying for and receiving DI), the authors "find that for a substantial range of plausible parameter values, current compensation for the most disabled appears to be lower than this standard model suggests is optimal. However, stronger statements require knowing preference parameters that have not been pinned down in the literature."

The authors conclude that there are many important questions remaining for future research, including the prevalence and consequences of disability for women, the effects of disabilities that begin at later ages, and the varying effects of disability by health condition.

The authors acknowledge financial support from the U.S. Social Security Administration through grant #10-M-98363-1-01 to the NBER as part of the SSA Retirement Research Consortium.