We analyze the evolution of health insurer costs in Massachusetts between 2010-2012, a period in which the use of physician cost control incentives spread among insurers. We show that the growth of costs and its relationship to the introduction of cost control incentives cannot be understood without accounting for (i) consumers’ switching between plans, and (ii) differences in cost characteristics between new entrants and those leaving the market. New entrants are markedly less costly than those leaving (and their costs fall after their entering year), so cost growth of those who stay in a plan is significantly higher than average per-member cost growth. Cost control incentives were used by Health Maintenance Organizations (HMOs). Relatively high-cost HMO members switched to Preferred Provider Organizations (PPOs) while low-cost PPO members switched to HMOs. As a result, the impact of cost control incentives on HMO costs is likely different from their impact on market-wide insurer costs.