Three Essays in Health Insurance Coverage.

Abstract

Obtaining health insurance through an employer pools risk, but may lead to moral hazard, where employees with more coverage seek care valued below cost, and adverse selection, where the unhealthy choose more generous plans, driving up premiums. In the first essay, I model the decision to choose the more generous employer plan, which suggests testable empirical predictions. Using MEPS data from 1996-2001, I find that enrollees in non-HMO plans, but not HMO plans, spend more on medical care the more generous their plan, indicating moral hazard. Within HMO or non-HMO options, the more generous plan does not enroll the unhealthier employees, but adverse selection may lead to healthier employees choosing an HMO over a non-HMO.

The second essay, written with Catherine McLaughlin, examines the growing disparity between Hispanics and non-Hispanic Whites in health insurance coverage. While the percentage of non-Hispanic Whites without health insurance has fallen slightly since 1983, the uninsured rate has risen dramatically among Hispanics. Using 25 years of SIPP data, we find that differences in citizenship and education explain some of the divergence, but more than half of the increase in the gap, or more than one million extra uninsured Hispanics, remains unexplained by differences in any observable characteristic. While observable differences account for most of the divergence in public and employer coverage rates, much of the decline in family coverage remains unexplained.

In the final essay, I examine the decision to select a health insurance plan from the options offered by one's employer, considering not only the level of the plan's financial generosity, but also the variability of out-of-pocket spending under the plan. I also consider other elements that may differ between plans, including access to providers and supplemental coverage. I find that employees demand a plan with lower expected out-of-pocket costs, but that the variance of a plan's out-of-pocket spending has no effect on plan choice. Employees are only weakly responsive to their share of the premium, but do tend to choose the most popular plan at a firm, suggesting the importance of unobserved plan quality differences and/or the employees opting for the default plan.