I. The Purpose of Health Insurance

At a basic conceptual level the purpose of health insurance is for a community to provide mutual aid (in the form of healthcare interventions) to its needing members. Like all mutual aid schemes, health insurance systems distribute the burdens of individual risk across the community. The underlying normative impulse for such a system in healthcare comes from the notion that because sickness is an unchosen burden that affects individuals in a morally arbitrary manner, communities should pool resources so that the individual members may receive healthcare benefits as they need. If they do not need to draw from the pooled resources, i.e. if they do not need health care interventions, then they do not receive direct benefits from contributing to the mutual aid system. However, because it is impossible to know if and when an individual will require health care, it is “rational” for all individuals to want to participate in this mutual scheme. Moreover, insofar as individuals understand themselves as members of larger social communities to whom they are obligated, participation in such a mutual aid scheme can be understood as a moral duty.

Let us examine this idea in more concrete terms. We can identify two distinct functions of insurance. First, it acts as a cost spreading mechanism (across time) for the individual. Through insurance the customer pays the costs of his or her healthcare in small increments rather than choosing to take the risk of being burdened with a massive financial burden all at once.

Each month the consumer pays $X rather than paying, say, $400X in the event of serious illness or injury. Second, insurance pools risk across its policy holders. As Ronald Dworkin puts it: “the premise of all social insurance plans …is that inescapable risks should be shared across a political community between those more and those less at risk.”^{Reference Dworkin6} Thus, most of the burden of paying for an unexpected and expensive medical intervention is shouldered by the mutual aid community, rather than by the particular, unlucky individual.

The reality of health insurance systems tends to diverge significantly from the conceptual analysis of healthcare qua mutual aid scheme just presented. Consider, for example, the American context. Prior to the passage of the Affordable Care Act the distributive logic of the American healthcare system was to each according to willingness and ability to pay, i.e. “you get what you pay for.” In healthcare, this principle is known as actuarial fairness, and has been historically central to the American system.^{Reference Deborah7} On this view, in an ideal world each consumer pays premiums equal to the amount of healthcare resources that they use. Actuarial fairness directly undercuts the conceptual purpose of healthcare qua mutual aid scheme, by treating health as a good qualitatively equivalent to any other commodity. In lieu of valuing actuarial fairness, many societies favor some form of solidarity. These societies regard healthcare as a properly redistributive enterprise where healthy people subsidize sick ones.⁸ As such, they place the value of the community ahead of the private property rights of individuals, and tend to regard health as qualitatively distinct from other commodities. On this view, individuals are entitled to healthcare treatment as they need it rather than based on what they have input into the system.^{Reference Elhauge9}

If we accept the solidarity principle and understand the basic purpose of a health insurance system as the provision of mutual aid to group members who do not know if and when they will require medical care, then we must also understand high-risk pools as antithetical to our values. For high-risk pools — subdivisions of health insurance pools that divide members into homogenous groups based on their risk profile — are merely a particular institutional embodiment of the logic of actual fairness, and as such can ultimately be defended only by recourse to the principle that it is “unjust for the healthy to subsidize the sick” because individuals should be responsible for their own health needs. As Deborah Stone puts it, in practice actuarial fairness is an: “anti-redistributive ideology and a method of organizing mutual aid by fragmenting communities into ever-smaller, more homogenous groups, leading ultimately to the destruction of mutual aid.”¹⁰ And because high-risk pools follow this logic, they work precisely this way. When we separate one segment of society from another (in this case “high-risk” persons from “low or moderate risk” persons) we place them into distinct mutual aid schemes.

There are various efficiency based arguments which nonetheless defend high-risk pools.^{Reference Syed11} Crudely, these arguments claim that high-risk pools allow the rest of the health insurance market to operate efficiently, and that the resources saved in doing so can subsidize the high-risk pool. Presented in this way, it appears as though the two groups are indeed linked in a mutual cooperative schema. But this intentionally deceptive and ideologically loaded framing does little to allay the concerns of those actually affected. For in practice, high-risk pools are chronically underfunded and those who have the ill-fortune of being forced to participate in them are too often denied care in pursuit of greater profit margins.^{Reference Chollet, Achman and Chollet12} The attempt to cast high-risk pools as consonant with the logic of solidarity must, therefore, overlook the fact that such pools often have absolute caps on coverage per person, caps on the number of persons able to enter the pool, extended wait times even for those who are provisionally accepted, high deductibles, and a vast array of restrictions on the sorts of claims that will be covered. More importantly, even if we grant the social-scientific veracity of efficiency arguments there remain serious normative concerns that the use of high-risk pools creates, which outweigh any supposed benefit from increased market efficiency.

II. Social Recognition & Group Fragmentation

In this section we explore some social-recognitional problems that group fragmentation resultant from the use of high-risk pools creates.^{Reference Honneth13} But before doing so we should note that in practice the American health insurance market is already de facto fragmented along various lines. The introduction of high-risk pools thus represents an additional layer of fragmentation, but one that is especially concerning because it explicitly works to fragment groups on the basis of an unchosen trait. While there are normative concerns with all such divisions, the focus here is on high-risk pools as a particularly invidious mechanism.

To isolate the normative problems particular to high-risk pools let us abstract from the American context and assume that we have only two groups in our health insurance scheme: high-risk persons and everyone else. Here, the legal-institutional apparatus of the state recognizes two types of persons marked by a quantitative difference (their risk profile) that entitles them to qualitatively different benefit packages.¹⁴ We can distinguish two interconnected processes at play in this scenario. First, because this social insurance schema fragments members into two sub-groups, it creates legal categories based upon social differences that might not have previously been regarded as salient. We can examine this process from two perspectives: the individual and the structural. From the point of view of the individual, the categorization of persons as “high risk” lacks social salience until it is constructed by the state's legal apparatus, but once it has been constructed it takes on both a socio-legal and a social form. Whereas in a society that does not make use of high-risk pools I might encounter Frank as my friendly neighbor who unfortunately suffers from a chronic illness, in a society that does make use of such ratings I am more prone to encounter Frank as that high-risk person who lives in my building, talks too much, and consumes more resources than he contributes. In the latter scenario, it is not hard to imagine that I might grow to resent Frank for his “over” consumption of the pooled resources.

Let us now take a more bird's-eye, or structural, perspective. Assuming that Frank does indeed suffer from a chronic illness that requires expensive medical treatment, in a healthcare system that makes use of high risk pools (or any other form of actuarial rating) that characteristic moves from a private concern to a public one. While the effect of his chronic illness (i.e. need for, and cost of, more medical care) would be a public matter in any mutual aid scheme, the illness itself becomes salient in a system with high-risk pools. This is a subtle but significant difference. In the first scenario — in a system based upon solidarity — members who contribute to the mutual aid scheme know nothing about the persons who will benefit, bar the fact that those persons will need medical care. In the second scenario — in a system that is guided by the ideal of actual fairness and makes use of high risk pools — members who contribute to the mutual aid scheme know a priori that the persons who will benefit from their contributions are of a certain type. Put differently, in the first case members know that they will likely be subsidizing high-risk persons, but that effect is circumstantial rather than structural. In the second scenario, persons know that they will necessarily be subsidizing high-risk persons, and that effect is structural.

Second, the newly constructed category becomes part of a larger classification schema that has material repercussions. To see this, we return now to the individual perspective and assume Frank's point of view. The introduction of high-risk pools gives Frank a new socio-legal identity that separates him from the rest of society. He is now part of the high-risk group, and as such is subject to a different set of rules and regulations than the rest of the population. Like most invidious social classification schemas (e.g. racial caste systems) the relevant criterion that separates Frank from the rest of the low-risk population is unchosen. This unchosen characteristic, for which he is neither morally responsible nor able to change, forces him to accept a different socio-legal identity along with the inferior benefits and privileges that it entails.

Sometimes this characteristic is hidden from the social eye, but other times it is as obvious as one's race or sex. For example, Frank might suffer from Crohn's Disease, a medical condition which his friends and neighbors are wholly unaware of, or he might suffer from muscular dystrophy, a condition that tends to be quite visible. In either case, Frank's medical history and risk profile come to be constitutive factors of his socio-legal identity. The compound “socio-legal” is here key. If Frank's society does not make use of high risk pools and instead bases its healthcare scheme on a commitment to solidarity, then his social identity might certainly be affected by his medical condition (insofar as it is readily visible or otherwise known to others) yet it would have no immediate legal repercussions. However, in the high-risk pool scenario Frank's medical condition is socio-legal, because it identifies him as a particular type of citizen — namely, one who consumes more resources from a mutual aid scheme than he or she contributes — and makes him subject to a more stringent set of rules and regulations. Therefore, we can say that the use of high-risk pools creates a legal system where group membership is hierarchically tiered, and access to the higher form of membership is distributed in a morally arbitrary manner. Membership is hierarchical because participation in the high-risk pool (lower tier) comes with a restricted set of rights and privileges vis-à-vis the other cohort. The distribution of membership between these two groups is morally arbitrary, because one cannot be deemed morally responsible for their medical risk profile. This holds true for all cases except those in which one's risk profile is obviously affected by the decisions one has made. One could quite reasonably argue, for example, that smoking cigarettes or frequently skiing without proper protective gear increases one's risk profile in such a way that they can be held morally responsible for. In the vast majority of cases, however, one's risk profile is the result of luck or chance.

Beyond the material effects, social narratives that stigmatize Frank and his high-risk cohorts might also develop. This tends to follow from any form of group fragmentation, regardless of the characteristic that is made recognitionally salient. When something is made recognitionally salient it is brought to the attention of the social group, and given “objective” social weight — here objective means only that it is taken as established and authoritative within a community. Recognitional salience diverges from mere social salience because of this objective aspect. For example, liking David Foster Wallace novels might be a socially salient characteristic for Americans in the sense that such a disposition is considered to be an indication of good tastes, but this would not make it recognitionally salient. If, however, liking David Foster Wallace novels somehow becomes a necessary disposition to be treated as a fully functioning moral being who is worthy of respect within the community, then we could say that it has become recognitionally salient.

Though easiest to see in the case of race and the pejorative stereotypes that persistently emerge to justify segregating practices and institutions, this same process — the emergence of disparaging social narratives — occurs with most unchosen characteristics that are made recognitionally salient in this way, as, for example, with sex, sexual orientation, or disability. This in turn can lead to changes in the recognitional status of the marked group which might have further material implications. In other words, there can be domino effects of group fragmentation, which might begin innocuously with small differences in legal entitlements that are both spawned and reinforced by disparaging socio-cultural narratives, which then justify further legal differences with even more significant material consequences.

Anytime an unchosen characteristic is used to segment one portion of the population from another we should be deeply suspicious. This is especially relevant in cases where that social fragmentation entitles one group to different benefits (whether in quality or quantity) than the others. Even when those benefits are couched in ideological language that makes it superficially appear as though fragmentation is in the interest of all groups involved, more often than not the benefits disproportionately favor one subset over the others.

III. Fairness in Resource Access Tradeoffs

Having dealt briefly with some of the social-recognitional problems associated with high-risk pools and the group fragmentation that they necessarily entail, we move now to a concern that is more traditionally recognized as falling under the purview of distributive justice. I argue here that interpersonal resource-access tradeoffs — that is, decisions that determine interpersonal priority in access to the pooled resource — in any mutual aid scheme should be made against all members of a social group, and as such any fragmentation of that group is morally problematic. Furthermore, insofar as we are committed to ensuring that “each person's life … go as well as is possible and fair” the determination of interpersonal priority in resource access should aim to minimize the effects of factors that are morally arbitrary.^{Reference Rawls15} When resource-access tradeoffs are made only among one sub-group, as they are in any healthcare system that uses high-risk pools, individuals are denied fair access to a substantive benefit.^{Reference Rawls16} Therefore, high-risk pools are both procedurally and substantively unjust.

Prior to expanding on these claims we must clarify a crucial matter, namely that all healthcare systems need to make resource access tradeoffs between persons. Since a healthcare budget that is not fixed at some level will encompass the entirety of any state's budget, healthcare systems must have a mechanism for making such tradeoffs.^{Reference Elhauge17} It is therefore inevitable that some individuals will be denied health interventions that are too costly for too marginal a benefit. There is an ongoing discussion in the legal, social science, and philosophical literature about the best mechanism to make this tradeoff, but relevant for our purposes here is merely the need for such a mechanism.^{Reference Syed18}

The mechanism is used to make decisions that come in the form of determining whether person A, B, C, or D, will receive a given health intervention. In other words, it makes interpersonal tradeoffs between individuals who, in virtue of their membership in the given social community, must be considered fairly. Here, fairness requires that each person has an a priori equally legitimate claim to the pooled resources. This means that, taken in the abstract, each member of a mutual aid schema has a claim of equal weight to the resource. However, because in an operative distributive scheme we need to prioritize between such claimants, priority in resource-access determinations must be decided based upon the “background conditions” of the individual claimants — that is, based upon their particular life circumstances. Thus, the mechanism must take into account individual characteristics such as overall well-being and the marginal return — in terms of Q.A.L.Y.'s (Quality Adjusted Life Years) or some other measurement of the sort — of the health intervention, in order to determine access priority.¹⁹

The content of this mechanism is, in my view, the most important matter of distributive justice in any health care system, given that it determines who should get priority for access to the substantive resource (health intervention).^{Reference Syed20} For our purposes we can assume that we have agreed upon such a mechanism. Having done so, we turn our attention to the circumstances for a fair operationalization of the mechanism. Given that procedural fairness requires that the claims of any individual are weighed against the claims of all other individuals, a fair procedure must be one that utilizes a unitary risk pool. Any fragmentation of the risk pool will necessarily make it the case that individuals' claims to resources are not weighed against all others, but are weighed only against their respective sub-group. It follows, then, that because high-risk pools separate individuals into different subgroups, and in effect separate persons into different distributive justice systems, they violate the demands of procedural fairness.²¹

To illustrate this point let us consider a hypothetical scenario. Hank, Hallie, and Heather are all high-risk individuals and are separated into their own risk pool. Larry, Larissa, and Logan are low-risk individuals and are in their own risk pool. In this situation, the H's are in their own distinct mutual aid system, as their claims to any mutually pooled resources are weighed only amongst themselves. The same holds for the L's. This fragmentation of risk pools is problematic because it makes a qualitative distinction between the claims of the H's and the L's to the mutually pooled resources. Though Howard and Larissa might be neighbors, or even husband and wife, they are embedded in different cooperative social schemes. Thus, within a single social community the mutual aid schemes are fragmented so that one's claim to a resource is weighed only against the claims of other individuals grouped by an unchosen characteristic.

The normative concern here becomes obvious if we replace the characteristic separating the two pools with race, sex, or class. For example, it should seem flagrantly problematic to suggest that as citizens of the same juridical community Blacks and Whites should pool together their resources, but that the subsequent claims of group members to those resources should be weighed only against those within their racial group. It becomes even more problematic when we add a more stringent set of rules and regulations to the resource claims of Blacks, which White claimants are not subject to. This creates a scenario wherein both groups are contributing to the mutual aid scheme either proportionately to their risk profile or equally, but are benefiting differentially according to their race. If we find this problematic, then we ought to similarly find high-risk pools troubling. For one's medical risk profile, like one's race, is, for the most part, an unchosen characteristic.²² And if we divide access to mutual benefit schemes based on an unchosen characteristic, then we are punishing individuals for attributes that, from the point of view of justice, ought to be irrelevant.

Furthermore, the substantive distributive consequences of separating individuals into cohorts according to their risk profiles are deeply troublesome. Since the goal of risk pooling, as consonant with the logic of actuarial fairness, is to ensure that participants in each pool make appropriately weighted contributions to their pools, and the H's (i.e. persons in the high-risk cohort) will have much higher per person healthcare costs than the L's (i.e. persons who are in the low-risk cohort), the H's will be forced to contribute much more resources per person to the pool relative to the L's. And despite this greater contribution of resources per person, it remains likely that the high-risk pool will not have enough resources to cover the complete healthcare needs of its members.²³

Moreover, even if the high-risk pool is able to meet the health needs of its members, since the H's are contributing more of their resources to healthcare, it follows that with regard to other goods (e.g. education, food, housing) the H's will have increased access disparities vis-à-vis the L's, given that access to such goods depends on the same resources that the H's will have exhausted on healthcare. As such, high-risk pools do not merely create access disparities in the domain of health, they create disparities in access to goods across the board. Simply put, because high-risk pools make it so that high-risk individuals have to pay more for their health insurance than others, high-risk persons have fewer resources to expend on other vital goods. Differential resource burdens for individuals in the domain of healthcare, then, turn out to affect individuals' general ability to access goods. Therefore, insofar as we are committed to reducing disparities in resource access that are the result of morally arbitrary factors, high-risk pools are also substantively unjust.