I examine the effects of public catastrophic insurance programs on enrollment and selection in private health insurance to supplement Medicare. Using variation over time in the availability and generosity of these programs I show that public catastrophic insurance crowds out private insurance coverage for individuals in worse health status, with little, if any, effect on healthier individuals. In addition, individuals for whom the public program is less generous are more likely to have private insurance coverage. The selective crowd-out of individuals in worse health induces advantageous selection in Medigap, one of two types of supplementary private insurance, and leads to lower insurance premiums. I also find that more cognitively able individuals are more likely to be crowded out if they are in bad health and that accounting for cognitive ability attenuates selection most strongly in states that provide public catastrophic insurance.
Ex ante, like ex-post, moral hazard reduces the efficiency of the insurance market. I use exogenous variation in risk protection associated with public catastrophic health insurance programs to test for ex-ante moral hazard. Using the introduction of and changes in these programs, I find large decreases in self-protection as the public program provides greater risk protection. Ex-ante moral hazard is stronger for women than for men and affects a variety of investments, including smoking, obesity, and cancer screening. Differences by gender in ex-ante moral hazard are consistent with greater returns to self-protection for women than for men.
I show that employees sort into firms on the basis of fringe benefits using laws mandating that employers provide a fringe benefit (insurance for mental illness) and a proxy for individual's demand for the benefit (mental distress). My estimates indicate that the mandate only affected less educated individuals, reduced labor demand, but increased labor supply for individuals who value mental health benefits; these results are driven by individuals switching into different jobs. I find no evidence for an aggregate welfare loss, but there is an increase in welfare for distressed individuals, at the expense of non-distressed individuals.