Association for Public Policy Analysis & Management

Research Objective: A health plan’s physician network is the set of physicians who deliver medical care to the plan’s enrollees. Plans with smaller “narrow” networks may be less costly for enrollees. However, they may limit healthcare access. As narrow networks become increasingly common, it is imperative to determine whether narrow network plans are the result of market failure from adverse selection or if, given the trade-off between cost and access, consumers prefer them. My research objective is to determine whether consumer preferences or adverse selection explain insurers’ narrow network offerings in the Health Insurance Marketplaces.

Study Design: I begin by measuring network size as the quantity of physicians available to Marketplace enrollees within CMS’s recommended driving distances per CMS’s 2017 Letter to Issuers. I calculate network size using physician-level network data and geographic information system software. I proceed by examining whether consumer preferences for network size are satisfied in the Marketplaces. To do so, I determine whether a positive association exists between variation in network size and variation in enrollee health preferences. I measure variation in network size as the coefficient of variation of network size among networks offered in a county. I measure variation in enrollee health preferences as the variance of target enrollee demographics (e.g., age, family size). Controlling demographics and insurers cost shifters, I use these measures to examine the association between variation in enrollee health preferences and variation in network offerings. Lastly, I determine if market failure is present by analyzing whether indicators of adverse selection (e.g., disease prevalence, proportion of target enrollees above age 50) are negatively associated with the insurers’ county-level log network size. I use the same controls above, as well as indicators for competition, insurer brand, and insurer type. I will estimate both equations with generalized linear models.

Data Collection: I obtained 2017 Marketplace physician network data from machine-readable network databases distributed online per CMS regulations. I augmented this data with plan characteristics from the 2017 Qualified Health Plan Landscape file and target enrollee demographics from the 2015 5-year American Community Survey.

Population Studied: Insurers operating in counties in 37 of the 39 states that used the Federally-Facilitated Marketplace (i.e., healthcare.gov) in 2017.

Principal Findings: Results, which will be available by June 2017, will clarify whether consumer preferences or adverse selection dominate insurers’ decisions to offer narrow networks. This distinction clarifies whether the proliferation of narrow network plans is due to market failure from adverse selection, which may be suggestive of a death spiral.

Conclusions and Policy Implications: Determining whether consumer preferences or adverse selection drive narrow network plan offerings is necessary to determine whether additional regulations are appropriate to address the proliferation of narrow networks. If narrow network plans are the result of consumer preferences, increased network adequacy regulations will not improve consumer welfare in the economic sense, though they may still improve healthcare access. If adverse selection drives narrow network plans, however, regulations such as high risk pools or increased risk adjustment payments may be justified to correct market failure.