Price-Linked Subsidies and Health Insurance Markups

Abstract

Subsidies in many health insurance programs depend on prices set by competing insurers -as prices rise, so do subsidies. We study the economics of these "pricelinked" subsidies compared to "fixed" subsidies set independently of market prices. We show that price-linked subsidies weaken competition, leading to higher markups and raising costs for the government or consumers. However, price-linked subsidies have advantages when insurance costs are uncertain and optimal subsidies increase as costs rise. We evaluate this tradeoff empirically using a model estimated with administrative data from Massachusetts' health insurance exchange. Relative to fixed subsidies, pricelinking increases prices by up to 6% in a market with four competitors, and about twice as much when we simulate markets with two insurers. For levels of cost uncertainty reasonable in a mature market, we find that the losses from higher markups outweigh the benefits of price-linking. approach (Chetty, 2009), drawing on natural experiments in Massachusetts to estimate the key statistics that enter our first-order approximation to the pricing distortion. The main natural experiment is the introduction of the mandate penalty in December 2007. Using income groups exempt from the penalty as a control group, we estimate that each $1 increase in the relative monthly price of uninsurance raised insurance demand by about 1%. 7 Our second empirical method is to estimate a full structural model of demand and cost and use the estimated parameters to simulate equilibrium in the insurance exchange under price-linked and fixed subsidies. While this approach necessarily involves more assumptions, it lets us go beyond price effects to estimate welfare impacts and simulate the tradeoffs involved in the presence of cost uncertainty and under different market structures. It also takes into account strategic interactions and adverse selection, which are not in the reducedform estimation. An important strength of our structural estimates is that we use moments from the natural experiments to identify parameters that are key to determining the relevant demand substitution patterns. 8 Across both methods, we find three sets of results. First, we find that price-linked subsidies raise prices by a non-trivial amount. Using the sufficient statistics method, we find that price-linked subsidies would raise the pivotal plan's price by about 10% ($38 per month) in a market with similar demand and supply-side parameters as the Massachusetts exchange. The estimates from the structural model simulations are somewhat smaller (1-6% of prices under fixed subsidies, or $4-26 per month) but generally in the same range. Based on our 2011 estimates, switching from price-linked to fixed subsidies could achieve either the same coverage at 6.1% lower subsidy cost, or 1.3% greater coverage at the same cost. Although modest, these effects imply meaningful increases in government costs. For instance, the $24/month subsidy distortion (from our structural model simulations for 2011) would transla...

Subsidizing Health Insurance for Low-Income Adults: Evidence from Massachusetts

“…selection across contracts that vary in limited ways, rather than selection that causes a market to unravel, leaving open the possibility of larger welfare costs on this margin Einav, Finkelstein, and Levin (2010). Our work, however, finds evidence of significant adverse selection on the extensive margin of purchasing insurance versus remaining uninsured -a finding consistent with past work on the Massachusetts reform (Chandra, Gruber, and McKnight, 2011;Hackmann, Kolstad, and Kowalski, 2015;Jaffe and Shepard, 2017). But it also finds that adverse selection is not the primary driver of limited demand for formal insurance among low-income adults.…”

Selection in Health Insurance Markets and Its Policy Remedies

“…This type of subsidy is used in the state-level Marketplaces, where tax credits are benchmarked to the price of the second-lowest price "Silver" plan. Jaffe and Shepard (2017) show that this type of price-linked subsidy distorts insurer prices because insurers that have some probability of having the second-lowest price plan will distort their prices upward to increase the size of the subsidy. On the other hand, this type of subsidy has the potential benefit that it protects subsidized consumers from changes in insurer prices (due to changes in technology, adverse selection, or other features) that are not anticipated by the regulator and thus cannot be incorporated into a fixed subsidy.…”

Health Plan Payment in U.S. Marketplaces: Regulated Competition with a Weak Mandate

“…If a consumer buys a higher-or lower-price plan, they pay or save the incremental price, as long as this does not push their payment below zero. Jaffe and Shepard (2017) and Tebaldi (2016) analyze what this "price-linked" subsidy design means for competition, relative to a system in which policymakers set a "fixed" subsidy amount based on their best estimate of what prices will be. They show that pricelinking weakens price competition, since insurers that expect to be "subsidy pivotal" have a greater incentive to markup their plans' prices.…”