Essay 1 studies physician agency problems, which arise whenever physicians fail to maximize their patients' preferences, given available information. These agency problems are well documented, but the magnitude of their welfare consequences for patients---the losses from suboptimal treatment choice induced by agency---are unclear. I infer patient drug preference from their compliance decisions. I begin by showing that initial prescriptions respond to physician financial incentives to control costs and to pharmaceutical detailing, but compliance does not, pointing to agency problems. I then develop and estimate a model of physician-patient interactions where physician write initial prescriptions, but patients choose whether to comply. Fully eliminating agency problems increases compliance by 6.5 percentage points, and raises patient welfare by 22% of drug spending. Contracts that better align doctor and patient preferences can improve patient welfare, but attain only half the gains from eliminating agency completely. Although physician agency problems reduce patient welfare, eliminating them is thus likely difficult. Essay 2, co-authored with Alexander M. Gelber and Damon Jones, studies frictions in adjusting earnings to changes in the Social Security Annual Earnings Test (AET) using a panel of Social Security Administration microdata on one percent of the U.S. population from 1961 to 2006. Individuals continue to "bunch" at the convex kink the AET creates even when they are no longer subject to the AET, consistent with the existence of earnings adjustment frictions in the U.S. We develop a novel estimation framework and estimate in a baseline case that the earnings elasticity with respect to the implicit net-of-tax share is 0.23, and the fixed cost of adjustment is $152.08. Essay 3 studies the impact of health expenditure risk on annuitization. Theoretical research suggests that such risk can have an ambiguous influence on the annuitization decisions of the elderly. I provide empirical evidence on this linkage, by estimating the impact of supplemental Medicare insurance (Medigap) coverage on the annuity demand of older Americans. Medigap coverage has a strong impact on annuitization: the extensive margin elasticity is 0.39, the overall elasticity of private annuity income with respect to Medigap coverage is 0.56. These results are robust to controls for health, wealth, and preferences, as well as other robustness tests. They suggest that medical expenditure risk has a large impact on underannuitization.