The Effect of Depression on Retirement and Disability Insurance Applications

Featured in print Bulletin on Aging & Health

Poor health has long been established as one of the most critical determinants of retirement behavior. Yet the existing literature has focused primarily on the role of physical illness. Since the early 1990s, mental illnesses have been the fastest growing source of claims for Disability Insurance (DI) and Supplemental Security Insurance (SSI) benefits, particularly among people aged 45-64. This suggests that poor mental health may also be an important factor in retirement decisions.

Understanding the effect of one important mental health condition, depression, on the labor supply of older workers is the subject of a new study by Rena Conti, Ernst Berndt, and Richard Frank, "Early Retirement and Public Disability Insurance Applications: Exploring the Impact of Depression" (NBER Working Paper 12237).

The authors hypothesize that depression may reduce labor supply directly by reducing individuals' ability to work and interest in employment and indirectly by magnifying the effect of physical illness and other life events. In their analysis, they focus on three outcomes: employment status, early retirement, and application for DI or SSI benefits. The data for the analysis comes from the Health and Retirement Study (HRS). They identify depression using the CESD, a self reported survey instrument designed to detect depression symptomology in the general population.

The authors first provide some summary statistics on depression. Among HRS respondents who were physically healthy and still in the labor force in 1994, about 6% of men and 12% of women were depressed at the 1994 baseline. Depression is more common among individuals who are Hispanic, poor, and lacking a high school education. Depressed individuals are more likely to experience a physical health shock in the future.

The authors' first empirical approach is to identify the direct effect of depression by comparing the subsequent employment outcomes of individuals who are depressed at baseline with those of individuals who are not depressed. They also interact baseline depression status with a subsequent health or other life shock to identify the indirect effect of depression on employment.

They find that depression increases the probability of DI/SSI application by 8 percentage points for men and 10 points for women. The magnitude of these effects are larger than the effects of physical health shocks. They also find evidence of an indirect effect of depression, as depression magnifies the effect of physical health and widowhood shocks.

In their analysis of labor supply, the authors find that depression increases the probability of retiring before age 65. The effect is smaller than that of physical health shocks but not small for men, depression increases the probability of early retirement by 27 percentage points. Depression does not amplify the effect of a physical health shock on retirement, but does magnify the effect of widowhood.

The authors' second empirical approach is to identify a subset of individuals with no history of psychiatric conditions and no significant symptoms of depression at baseline and to estimate the effect of a subsequent depression onset on labor market outcomes. Here they find that depression increases the probability of a DI/SSI application by 16 percentage points for men and 18 points for women. Depression and physical health shocks have similar and large effects on the probability of early retirement. Depression magnifies the effect of physical health shocks for both men and women and of widowhood shocks for men.

Overall, the results from both strategies indicate that depression has both direct and indirect effects on DI/SSI applications and early retirement, after controlling for other determinants of labor force exit. The magnitude of the effects is similar to that of physical illness alone.

These findings have a number of important policy implications. First, they suggest a significant role for mental illness in explaining recent trends in early retirement and public disability applications. Second, after accounting for the interaction between mental and physical illness, disability rates associated with physical illness may be decreasing faster than previous research suggests. This also implies a more nuanced interpretation of the official disability insurance statistics. Finally, effective and cost effective treatments for major depression are widely available. Insuring access to treatment and continuity of treatment for individuals suffering from a mental illness may have important effects on employment and public disability programs. As the authors note, "the impact of treatment receipt on individuals' lives may be substantial."

The authors acknowledge funding from the National Institute of Aging, the National Institute of Mental Health, and the Lasker Family Charitable Trust.