Adverse selection in health insurance markets may reduce social welfare by leading some low-risk consumers to underinsure or too few consumers to purchase coverage, relative to the socially optimal level. I study the social welfare implications of policies that are designed to mitigate these effects of adverse selection, including (1) an individual mandate for purchasing insurance and (2) risk adjustment. The mandate addresses suboptimal enrollment in the market that results from adverse selection, while risk adjustment addresses underinsurance. I show that the welfare impact of the policies is theoretically ambiguous because there is a tradeoff in addressing the effects of adverse selection. I then assess how the mandate and risk adjustment affect social welfare in the Affordable Care Act (ACA) insurance exchanges. Using consumer-level data from the California and Washington exchanges, I estimate demand for insurance and obtain estimates of marginal cost that I relate to premiums to account for adverse selection. I compute equilibrium premiums under alternative scenarios and find that the mandate (relative to no mandate) modestly reduces premiums and risk adjustment (relative to no risk adjustment) increases premiums for less costly exchange plans. The mandate decreases consumer surplus by 2.6 percent because it makes underinsurance more prevalent and the ACA's price-linked subsidies limit consumer exposure to premium reductions. Conversely, risk adjustment increases consumer surplus by 3.9 percent because it addresses underinsurance and the ACA's price-linked subsidies limit consumer exposure to premium increases. I conduct simulations using the estimated model and find the impact of the two policies is sensitive to subsidy designs that expose consumers to premium changes. If ACA price-linked subsidies were converted to fixed subsidies or vouchers as proposed in some legislative alternatives to the ACA, the mandate would increase consumer surplus by 6.6 percent and risk adjustment would decrease consumer surplus by 4.2 percent.