UCLA

My first chapter explores the effect of income on health in the context of Social Security and Medicare. Income is a powerful predictor of health among the elderly, but existing research has struggled to identify a causal link. I estimate the causal effect of Social Security income on health care utilization and health outcomes among elderly men. Using Medicare administrative records and a regression discontinuity design, I exploit several changes in the Social Security benefit formula that vary abruptly by date of birth. This feature has been overlooked by prior research and causes workers born one day apart to receive positive and negative income shocks. Over my sample period, I estimate a $100 increase in monthly Social Security benefits leads to a $38 decline in monthly federal Medicare expenditures. To provide evidence the decline in spending is driven by improvements in health, I show income leads to reductions in diagnoses for chronic conditions and mortality. My results suggest cuts to Social Security benefits may have unintended social and fiscal costs. Overall, these findings highlight the importance of examining health outcomes when evaluating the costs and benefits of social insurance programs. My second chapter examines the Qualified Medicare Beneficiary (QMB) program---a means-tested benefit that exempts low-income Medicare beneficiaries from their cost-sharing obligations. Patients enrolled in the QMB program face zero prices for Medicare services, receive monthly premium exemptions, and subsidized Part D prescription drug coverage. However, because providers face administrative burdens and lower reimbursement rates, they may limit access for QMB patients. Given these offsetting features, it is unclear whether patients benefit from enrolling in the program. To examine the effect of QMB enrollment on patient access and outcomes, I study an expansion of QMB eligibility that occurred in Connecticut in 2009. By doubling the income limit and removing the asset test, the state more than doubled enrollment. Using a difference-in-difference design and Medicare administrative records, I estimate the effect of QMB enrollment on health care utilization and health outcomes. I find that the program appears to be successful at reducing beneficiary costs without limiting access to care. I show QMB enrollment leads patients to save $1,300 in annual outpatient cost-sharing liability without any change in the utilization of outpatient services. Overall, my results suggest that within the universe of Medicare, the program transfers a substantial share of producer surplus to consumers. My third chapter studies how cuts to Social Security retirement benefits affect Social Security disability enrollment. Specifically, I exploit a policy change which caused an abrupt decline in the generosity of retirement benefits for workers near certain date of birth cutoffs. Using various regression discontinuity designs, I do not find evidence that cohorts affected by the policy change have higher rates of benefit receipt. My preferred specification estimates a precise null effect. My confidence interval implies the increase in the full retirement age caused disability enrollment to only change between -0.25 and 0.28 percentage points. Relative to existing work, my paper provides evidence that disability enrollment is less sensitive to policy changes. I find there are fewer marginal SSDI applicants implying the program's moral hazard effects are more modest than prior work would suggest.