Health Insurance Availability and the Retirement Decision

Because individuals aged 55-64 face large and uncertain medicaJ expenditures without the guarantee of public insurance coverage provided by Medicare, the availability of post-retirement health insurance could be an important determinant in the retirement decisions of this group. We investigate the effect of health insurance on retirement by focusing on state and federal "continuation of coverage" mandates which grant the retiree the right to continue purchasing health insurance through a previous employer for a specified number of months after leaving the firm. We exploit variation in the timing and generosity of these laws to identify the effect of the availability of continuation coverage on retirement decisions, using data on 55-64 year-old males from the Current Population Survey and the Survey of Income and Program Participation. We find a sizeable and significant effect of continuation coverage on retirement; one year of mandated continuation benefits raises retirement rates by 20% The effect appears to be uniform at all ages rather than larger near the age of Medicare eligibility. There is also a large increase in the insurance coverage of individuals who would have retired in the absence of continuation benefits. Our findings have important implications for policies which change the insurance coverage of early retirees, such as national health insurance. The vast majority of Americans receive their health insurance coverage through one of two sources. Among the non-elderly, the primary source of coverage is their own or a family member's employment. For those over 65, Medicare covers a substantial fraction of medical expenditures. One group of individuals which often finds itself excluded from both of these systems are those who retire before the age of 65. Furthermore, medical expenditures for 55-64 year-olds are both high and uncertain, so that neither individual health insurance nor selfinsurance of medical expenditures may be very attractive options for early retirees. Therefore, it seems likely that the availability of post-retirement health insurance coverage could be an important determinant of the retirement decision for older workers. If the timing of retirement is very sensitive to insurance coverage, it could have major implications for the costs and consequences of a number of current policy proposals such as postponing the age of Medicare eligibility to 67 or providing government-sponsored health insurance to all individuals. Despite its likely importance, however, the role of health insurance coverage has, until recently, been largely ignored in the literature on retirement behavior. Given the sizeable evidence that health status is an important determinant of retirement, this is a surprising omission. In this paper, we attempt to address this gap by examining the effect of state and federal "continuation of coverage" mandates on the retirement decision. These mandates grant individuals the right to continue purchasing health insurance through a previous employer for a specified number of months after leaving the firm. Although individuals must pay the average employer cost of their group insurance, the price to the early retiree is typically well below that of a policy purchased in the individual market. This is both because of high loading factors on individual insurance, and because individual policies are generally age-rated while the cost of continuation coverage is not. By reducing the cost of health insurance after retirement, continuation of coverage benefits may increase the likelihood of early retirement for those individuals whose employers do not already provide post-retirement health insurance. The differences across states and over time in the availability and generosity of these laws provides the variation needed to assess the sensitivity of retirement behavior to this type of subsidized post-retirement insurance coverage. We model the retirement decision of 55-64 year-old males using data from two sources: the Current Population Survey (CPS) and the Survey of Income and Program Participation (SIPP). Our major finding, which is robust to a number of specification checks, is a sizeable and significant effect of continuation mandates on retirement decisions: one year of continuation coverage raises retirement rates by 20%. Furthermore, the effect appears to be uniform at all ages rather than disproportionately strong near the age of Medicare eligibility. We also find that continuation mandates had significant effects on the insurance coverage of early retirees. This effect is much larger than the effect on retirement, suggesting that the majority of older individuals with continuation coverage would have retired even in the absence of these benefits. The organization of the paper is as follows. Section I provides some background on retirement behavior and the health insurance coverage of early retirees. Section II discusses the details of the state and federal continuation of coverage mandates and their likely effects on retirement. The empirical results are presented in Sections III and IV. Section V then focuses on the welfare implications of continuation mandates by considering their impact on insurance coverage. The paper concludes in Section VI with a discussion of the policy implications of our