VoxEU Column Health Economics Industrial organisation

The two-margin problem in insurance markets

Sicker consumers tend to exhibit higher demand for health insurance, which drives up costs. This column argues that this adverse selection takes place along two margins: whether to buy insurance at all and how much coverage to buy, It develops a new framework that incorporates both selection margins, and shows that policies aimed at addressing one margin can often exacerbate selection along the other. It is therefore vital for optimal policy to consider both margins simultaneously. 

Some of the most important problems in health insurance markets stem from adverse selection, or the tendency of sicker consumers to exhibit higher demand for insurance. Because these sick individuals cost more to insure, they tend to drive up costs for all insurers, and especially for insurers offering more generous plan options that are more attractive to these individuals. 

While adverse selection is generally thought of as a single concept, a deeper look at its role in health insurance markets reveals that there are in fact two problems, corresponding to two margins along which selection can occur. Adverse selection both leads healthier people to opt out of health insurance altogether – an extensive margin response – and to select into less generous plans within the market – an intensive margin response. 

As public insurance is increasingly provided through regulated competitive markets, policymakers have designed a number of policies to separately address these two selection margins (Gruber 2017, Geruso and Layton 2017). For instance, policies such as penalties for uninsurance and subsidies primarily target the extensive margin – the choice of whether or not to purchase health insurance. Such policies have been effective at inducing healthy individuals to take up insurance, thereby lowering costs in some regulated insurance markets (Hackman et al. 2015). Other policies, such as risk adjustment and reinsurance, are designed to affect selection on the intensive margin – the choice of which insurance plan within a given market to purchase. Such measures have been effective at lowering costs of more generous plans in settings such as Medicare Advantage (Newhouse et al. 2015).

In a recent paper, we argue that these two selection margins are in fact connected via their impact on market prices and competition (Geruso et al. 2019). Policy responses to one margin need to consider interactions with and unintended consequences for selection along the other margin. To demonstrate these interactions, we present a model of a market where individuals may choose between a generous plan, a cheap ‘skimpy’ plan, or go without insurance altogether. We use this model to show that policies designed to address selection on one margin can have unintended consequences on the other. In some cases, the unintended consequences are so severe that they can reverse the gains from fixing selection problems on the other margin.

Unintended consequences across margins

Uninsurance penalties, subsidies, and other policies designed to address selection on the extensive margin induce healthy, low cost individuals to enter insurance markets. This influx of healthy individuals is intended to lower insurer costs and premiums in that market. However, in markets where multiple plans are available, inducing the entry of healthy people can have unintended effects. To see this, note that the group of marginal enrollees induced to take up insurance are likely to (1) have lower health care costs and (2) choose the cheapest plans in the market (the skimpier offerings). This influx of low-cost individuals allows skimpy plans to further lower their premiums. This leads to a wider price gap between the generous and the skimpy plans, resulting in some marginal individuals switching away from more generous coverage. Thus, while extensive margin policies induce more people to enter the market, they will also tend to induce some people who would otherwise have chosen a generous plan to purchase less generous coverage. This introduces a trade-off between the uninsurance rate and the level of generosity of coverage purchased by consumers in the market.

Intensive margin policies designed to increase enrolment in more generous plans may also have unintended consequences if individuals can choose to go without insurance altogether. For example, a common intensive margin policy is risk adjustment, or mandated financial transfers from plans enrolling relatively healthy consumers to plans enrolling relatively sick consumers. Under risk adjustment, skimpy plans that attract relatively low-cost individuals will pay transfers to more generous plans with high-cost enrollees. These transfer payments to the more generous plans result in higher premiums for the skimpy plans and lower premiums for the generous plans, compressing the gap between the premiums of these two options. The smaller premium gap causes more individuals to enrol in more generous coverage, the intended effect of the policy. However, the smaller premium gap is often paired with a higher overall premium for the skimpy option. Because the skimpy plan must pay transfers to the more generous plan, the skimpy plan will have increased premiums. This causes some individuals who would have chosen the skimpy plan to opt out of the market altogether and go uninsured. Again, there is a trade-off between the level of generosity of coverage purchased by consumers in the market and the level of uninsurance in the market.

Other policies have similar effects. On the intensive margin, regulators may also consider implementing minimum requirements for the generosity of health insurance plans. Such benefit regulation has the intended effect of shifting healthy consumers, who would have otherwise enrolled in skimpy insurance plans, into generous plans, as skimpy plans are no longer allowed. However, by removing skimpy plans from the market, the price of insurance is now higher. In a market where consumers can opt out of insurance altogether, some individuals who would have purchased skimpy insurance may decide to be uninsured. 

These findings apply broadly to any policy aimed at one margin or the other, including plan benefits requirements, network adequacy rules, risk adjustment, reinsurance, subsidies, and behavioural interventions like plan choice architectures or auto-enrolment. Each involves a potential trade-off. Policies that aim to address intensive margin selection tend to worsen extensive margin selection, and vice versa.

First-order impacts on prices, enrolment, and welfare

How big, in practice, are these effects? We investigate this empirically by simulating welfare under a variety of policy combinations in a market similar to our model, where individuals may choose between a generous plan, a skimpy plan, or may go without insurance. Our simulations use demand and cost estimates from the Massachusetts individual insurance market, the Connector (Finkelstein et al. 2019; Hackman et al. 2015). We simulate equilibrium enrolment and welfare under varying uninsurance penalties (an extensive margin policy) and risk adjustment transfers (an intensive margin policy). 

We find that the size of the unintended cross-margin effects can be large enough to imply significant impacts on market shares. We find that a strong mandate sufficient to move all consumers into insurance – increasing enrolment by around 25 percentage points in our setting – can cause the market share of more generous plans to shrink by more than 15 percentage points, or 35% of baseline market share (see Figure 1). In the other direction, strengthening risk adjustment transfers to the point where the market ‘upravels’ to include only generous coverage (because the transfers are so large that skimpy options cannot be sustained) can substantially reduce market-level consumer participation – in our setting by as much as 15 percentage points or 60% of the baseline uninsurance rate. With the additional assumption that consumer choices reveal plan valuations, we find that the welfare impacts of the unintended cross-margin effects of these policies can be similarly large, and often first-order. The parameters here are specific to the setting that we study, but the general point is that this phenomenon is not merely an academic curiosity: It can have first-order impacts on prices, enrolment, and welfare. 

Figure 1

Optimal policy requires considering both the uninsurance problem and the underinsurance problem together

We find that, because of the interaction between the two margins of selection, the optimal extensive margin policy depends on the optimal intensive margin policy and vice versa. For high uninsurance penalties (i.e. strong extensive margin policies), risk adjustment transfers should be higher (i.e. stronger intensive margin policies). For lower uninsurance penalties, more moderate intensive margin policies are optimal (see Figure 2). 

Figure 2

While the specific optimal levels of each of these policies will depend upon the consumers, plans, and other factors present in the market, our simulations demonstrate that in markets with both intensive and extensive margin selection, different policies should be considered in combination with one another. 


Finkelstein, A, N Hendren and M Shepard (2019), "Subsidizing health insurance for low-income adults: Evidence from Massachusetts", American Economic Review 109(4): 1530-67.

Geruso, M and T J Layton (2017), "Selection in health insurance markets and its policy remedies", Journal of Economic Perspectives 31(4): 23-50.

Geruso, M, T J Layton, G McCormack and M Shepard (2019), "The two margin problem in insurance markets", Available at SSRN 3385492.

Gruber, J (2017), "Delivering public health insurance through private plan choice in the United States", Journal of Economic Perspectives 31(4): 3-22.

Hackmann, M B, J T Kolstad and A E Kowalski (2015), "Adverse selection and an individual mandate: When theory meets practice", American Economic Review 105(3): 1030-66.

Newhouse, J P, M Price, J Hsu, M McWilliams and T G McGuire (2015), "How much favorable selection is left in Medicare Advantage?", American Journal of Health Economics 1(1): 1-26.

630 Reads