UC Berkeley

This dissertation studies the role of principal-agent problems as a barrier to market efficiency in the U.S. health care industry. The rise of health care as a percent of U.S. gross domestic product, as well as the documented dispersion in the productivity and price of health care, demand a policy response. In this dissertation I ask how demand-side (Chapter 1) and supply-side (Chapter 2) incentives can work to increase or decrease competitive forces and affect the efficient functioning of the market.

The first chapter, coauthored with Amitabh Chandra, Benjamin Handel, and Jonathan Kolstad, studies the role of principal-agent problems on the demand side. We study a large U.S. employer that changed their health insurance benefits from a comprehensive plan with no employee cost-sharing to a high-deductible health plan with significant cost-sharing, effectively increasing the price of health care to those employees. We find that, although employees do reduce their health care spending in response to this change, they do so in suboptimal ways. We find that (i) they reduce high-value care in similar proportion to low-value care; (ii) that reductions in spending come from reductions in utilization rather than substitution of care to lower-priced providers; and (iii) that employees do not understand the dynamic incentives embedded in high-deductible plans, so that even those employees whose effective marginal price of care has not changed cut back spending in response to deductibles. Our results suggest that although demand-side policies may reduce spending, they do so in highly inefficient ways, and may not solve the underlying causes of high U.S. health care spending growth.

The second chapter studies the role of principal-agent issues on the supply-side in determining the productivity of the U.S. health care industry. I document extensive vertical integration between primary care physicians (PCPs) and orthopedic joint surgeons. Using a stylized model, I show that this integration can increase productivity through technical production efficiencies, or lower productivity by distorting where PCPs refer their patients. I estimate that both of these effects are present, but that which effect dominates depends critically on the identities of the integrating parties. Echoing my first chapter, I find that demand is insensitive to price, lowering the potential distortionary effects of vertical integration on demand. I do find that the use of `global budget' capitation contracts, which make PCPs share the cost of patient care, do induce the reallocation of patients towards lower-cost orthopedists. This suggests that supply-side remedies may be more effective for future policy than demand-side remedies.