Risk-adjustment policies, which transfer money from insurers with healthy consumers to those with sick consumers, are an important tool to contend with adverse selection in health insurance markets. While the steady-state properties of risk-adjustment have been studied extensively, there is less evidence on the transition phase of policy implementation. We study the introduction and removal of risk-adjustment at California Public Employees' Retirement System and show that these changes meaningfully impact premiums via plan differences in enrollee health status. Despite these premium differences, there is limited consumer resorting due to consumer inertia, though new active enrollees respond more fluidly. We show that, with inertial consumers, risk-adjustment changes have substantial distributional consequences, leading to worse outcomes for sicker consumers when removed and vice-versa when implemented. We estimate a model of plan choice with premium sensitivity, brand preferences, and inertia and use these estimates to study the interaction between riskadjustment policies and the strength of inertia.