In our country, access to insurance can be a matter of life and death, as well as financial security. Despite these great stakes, the cost and quality of insurance are often influenced by social factors like sexual orientation, age, gender, and race. Such discrimination, forbidden in other settings like employment, is forgiven in insurance, even seen as fairer, on the grounds of actuarial fairness. That is, insurance classifications are lawful so long as they are based on evidence that some groups are costlier to insure, with the understanding that others shouldn’t have to offset those expenses. This Article challenges this concept of fairness in insurance using a stigma-based critique. The sociological stigma literature describes a natural and frequent social tendency to seek out differences, stereotype, and create underclasses who enjoy less social standing and experience structural and individual discrimination. Considering insurance through the lens of stigma reveals that it is no more inoculated from social context nor human nature than any other part of our lives. Of course, some people will be costlier to insure, but stigma theory suggests that we may be incapable of determining this in an unbiased way free from harmful social constructs. To guard against unfair insurance underclasses, we should ban discrimination in insurance as in other contexts.
Introduction
In our country, access to insurance can be a matter of life and death, as well as financial security.[1] Despite these great stakes, the ability to get affordable insurance is often influenced by social factors. HIV and AIDS are frequently used as examples in this Article but age, race, sex, zip code, credit score, health status, genetics, weight, and domestic violence history have all also frequently been used to underwrite in the context of personal lines of insurance[2] and are in many cases still lawfully permitted.[3]
This type of conduct, forbidden in other settings like employment, is forgiven in insurance on the grounds of actuarial fairness.[4] That is, insurance classifications are lawful so long as they are based on evidence that some groups are costlier to insure, with the understanding that others shouldn’t have to offset those expenses.[5]
This Article advances a stigma-based critique to challenge this conception of lawful discrimination and fairness in insurance. The sociological stigma literature describes a natural and frequent social tendency to seek out differences, stereotype, and create underclasses who enjoy less social standing and experience structural and individual discrimination.[6] Considering insurance through the lens of stigma reveals that it is no more inoculated from social context nor human nature than any other part of our lives. Of course, some people will be costlier to insure, but stigma theory suggests that we may be incapable of determining this in an unbiased way free from harmful social constructs. Our natural tendency is to gravitate to the usual suspects, looking for differences where we assume they will be and ignoring other possibilities. This is troubling for the same reasons that stigma is wrong generally. Far from being ensconced in scientific objectivity, insurance is influenced by subjective social constructs. It robs people of meaningful and important life opportunities and generates inequality, based on unjust and biased reasons rather than legitimate differences.
Americans spend about 1.9 trillion dollars per year on insurance premiums, and insurance makes up about 2.5% of the GDP; insurance is clearly a product many value and this is for obvious reasons.[7] Insurance products line up fairly closely with those things that people tend to find most precious and most worth guarding: our bodies and health; our family’s economic security in the event we are seriously injured or die; and protections for our most expensive assets like our homes and cars. Insurance is a gatekeeper to important resources, like health care[8] or income.[9] Moreover, insurance is often not elective, or at least not if one wants to engage in valued social activities like driving or owning a home. Insurance discrimination can be costly and even deadly.
This Article advances a theory of stigma to reshape debates by policymakers, lawmakers, the courts, and scholars about what should constitute fair and lawful discrimination in insurance. Part I introduces the dominant theory of actuarial fairness that pervades insurance law. Part II reconsiders insurance classifications through the lens of stigma. The theory behind social stigma is explained and applied to specific practices in insurance. The wider policy harms of permitting an insurance system that is influenced by stigma are discussed. In Part III, the Article explores regulatory solutions to address stigma in insurance. Mutual aid, an insurance system where everybody is eligible and treated equally, is advanced as the best remedy. Bans on classifications are explored as a more modest proposal. Lastly, if insurance regulations remain at status quo, the Article suggests ways that regulators can work within the current legal framework to reduce the incidence of and harm to insurance underclasses.
In the American insurance system, insurers are typically free to classify risks with few legal restrictions.[10] For instance, insurers have used AIDS and HIV status,[11] weight,[12] health status,[13] genetics,[14] domestic violence history,[15] race,[16] sex,[17] zip code or area one lives,[18] credit rating,[19] and even dog-breed ownership[20] as grounds to alter the terms of insurance, whether by charging more, limiting certain benefits, or excluding people from coverage altogether. Such classification is justified under the theory of actuarial fairness that classification is necessary and fair.[21] To proponents of actuarial fairness, people should pay only their share and not have to shoulder the cost of others’ risks.[22] Insurers, then, should be free to find and sort according to risk. This vision of fairness has justified class-based insurance discrimination, even while such forms of discrimination are typically impermissible in other contexts like employment, education, or public accommodations.[23] This Part will explain the theory in support of actuarial fairness as well as the ways in which the law continues to protect this version of fairness in insurance.
A. The Theory Behind Actuarial Fairness
Personal lines of insurance function to insulate individuals and their families from financially damaging and, often, unforeseeable accidents and injuries.[24] A terrible car accident, significant property damage, an unexpected injury that ends one’s ability to work, or a cancer diagnosis are all examples of events that could bankrupt ordinary families absent some form of insurance.[25] Insurance is redistributive. Individuals pay in with the expectation that, though they may never need it, insurance will be there for them and others if and when they do.[26] People who pay into insurance accept that they may be paying more in than they will ever cash out, but they do so to guarantee protection in the event of misfortune.[27]
Insurers can tolerate risk better than individuals can, both because they can pool that risk across a larger population and because they are often insulated by safety mechanisms like reinsurance and other risk spreading systems.[28] Insurers seek enough premiums to offset the cost of any payouts, administrative fees, and profit.[29] Thus, insurers must be able to gauge risk at least across the pool of their insureds as a whole.[30]
Once insurers know the estimated risk of the group, then public policy questions arise as to how to allocate that risk. One approach, actuarial fairness, is to classify risk and then allocate cost according to that risk; the riskier pay higher fees or have fewer benefits than those who are less risky.[31] The alternative is to simply share that risk equally across all insureds, sometimes called mutual aid.[32] Both of these models will be discussed in detail later in the Article.
If actuarial fairness is the model, then classification is necessary. What a single person pays for insurance depends on a multi-step process. First, insurers generally develop base rates off of risk tables that take into account a number of common factors related to risk.[33] For example, insurers typically use driver age and gender, model year, and accident history for establishing what the base rates are for car insurance.[34] This data is generally historical in nature and may come from wider industry pools of data or a specific insurer’s internal historical data.[35]
Next, insurers engage in underwriting. In this process, they look at individual risks to determine whether a specific insured should be excluded altogether or their premium altered to offset specific risks.[36] Underwriting is often seen as more subjective than risk tables.[37] Risk tables are also more frequently publicly available, whereas underwriting criteria are frequently not (unless required by law).[38]
Last, insurance premiums may be adjusted by territory (zip code).[39] And, for some product lines, insurers may offer discounts to reward and incentivize certain good conduct, for example, lower insurance premiums if you have certain safety features in a home or a car.[40]
Having assessed the risk of an insured, an insurer can simply refuse to cover the person who is found to be too great a risk.[41] The insurer can also offer insurance but limit the benefits covered.[42] Or the insurer can engage in rate adjustments by charging higher premiums to individuals to offset any increased risks.[43] Lastly, the insurer can discriminate against whole communities in a process called redlining.[44] This essentially means that the insurer doesn’t offer certain lines of insurance to a geographic region because they find the entire community to be of too high of risk.[45]
Insurers argue that two insurance phenomena make the need for risk classification even more imperative: adverse selection and moral hazard.
Adverse selection theorizes that people are unwilling to buy insurance until they need it.[46] There is a significant information asymmetry in insurance where the consumer knows her individual risk far better than the insurer, so she might be able to buy insurance for much lower than what her actuarial risk is.[47] For example, absent some penalty, she might put off buying health insurance until she knows she is sick and only then enroll and at a much lower rate than what the insurer would charge her if it knew about the illness.[48] Without adjustments of some kind, insurance pools are more likely to be filled with people who need to use the insurance and not enough people who are paying in and not needing it.[49] This may lead to rising premiums, making it even less likely that lower risk people will be willing to purchase insurance. This can eventually lead to what is known as a “death spiral,” where insurance becomes increasingly too costly for anyone to afford and the market collapses.[50] Some scholars suggest, though, that death spirals are a largely theoretical problem, and that some insureds who really need coverage will always be willing to buy into the system.[51] Instead, some argue, the greater risk is reverse adverse selection.[52] That is, insurers use their knowledge about risk pools to weed out and avoid anyone who may demonstrate increased risk.[53] They do this to better compete against rival insurers by “cream-skimming” and therefore being able to offer the best premiums (which further may draw in customers).[54]
The other concept is moral hazard which dictates that, once insured, a person may be more willing to undertake risks or to use insurance.[55] “The curse of insurance is that its mere existence tends to favor an increase in . . . risk: by protecting an insured against the adverse consequences of her decisions, insurance weakens her incentives to reduce risk.”[56] For instance, an individual might be more willing to engage in a risky winter sport if she knows she has access to health coverage. Related, individuals may be more likely to overuse their benefits because they know they have adequate coverage.[57] Of course, insurance can be modeled to offset some of the harms of moral hazard, for instance by putting some of the expense of claims back onto the insured in the form of copays, coinsurance, or bans on coverage for intentional wrongdoing.[58]
Proponents of actuarial fairness argue that insurers must be allowed to treat individuals or groups differently as part of the business model of insurance.[59] Even classifications based on traditionally protected groups are typically viewed as acceptable if necessary to prevent lower risk insureds from having to subsidize the risks of others.[60] Without some adjustment for risk, very low risk individuals may wind up paying a great deal more to subsidize the much higher risks of a few.[61] This is viewed as unfair because each person should only have to pay her own way.[62] And it may reward individual responsibility.[63] Why must the modest sedan driver who always sticks to the speed limit have to pay to compensate for the lead-foot, sports car-driving-and-texting individual in the next lane?[64] And low risk people may be more likely to buy into insurance, even if they don’t believe they’ll need it, if it is affordable and their risk aversion is rewarded.[65]
Risk classification is also valuable because it can prevent loss, thus allowing the insurer to offer better premiums to customers.[66] And it’s viewed as necessary for insurers to better compete with rival insurers.[67] An insurer who is exposed to more risk may have to increase premiums, and may lose business as a result.[68] While proponents of actuarial fairness believe that insurers need not cover higher risks or should do so only at the expense of those individuals, they are often focused on private insurers and the competition therein. They can be more accepting of the government taking on these higher risk individuals.[69]
According to actuary scientists, insurers are careful to look for specific qualities that align with risk and ignore other qualities that do not tell us anything about risk.[70] For instance, the health of a person might tell us something about their longevity, but eye color will not.[71] Some of these qualities can be objective; for instance, it is fairly easy to look at a collection of mortality data and see that women, on average, live longer than men.[72] Other classifications have subjective elements. For instance, from a given medical diagnosis, the insurer has to estimate how sick that person will be and how long they might live with that particular disease.[73] Often, that information which is useful is gleaned from examining historical data or, in other words, that data that has been customarily kept by government, insurers, or others.[74] Advocates of actuarial fairness generally see all of these sources of evidence as useful, so long as it helps the insurer to better predict and offset risk.
Opponents of actuarial fairness generally see it as coming at too high of a price. Professors Abraham and Chiappori worry about “classification risk,” that is, the insured “not only faces the risk of suffering loss, but also the risk of being classified as at high risk of suffering” it.[75] This classification can occur both with intrinsic features (where the insurer will know at the outset of the risk and seek to offset it) and reclassification risk (where the insurer finds out later about a risk, and then adjusts benefits accordingly).[76] Others argue that some classifications, often those involving historically disenfranchised groups, such as people discriminated against on the basis of race or gender, should be off limits.[77] Still others argue for mutual aid, that we should all be willing to pool the risks of one another so as to protect all members of society from harm.[78] Some worry that actuarial fairness is actually harmful for competition.[79] Insurers could find alternative ways to compete, such as by lowering prices, by having a more efficient administrative system, or by offering better quality and coverage, or better customer service.[80] However, insurers are not incentivized to engage in these other forms of competition when they are permitted by regulators to simply classify.[81]
B. Insurance Regulation Typically Reflects Actuarial Fairness
Private insurers are typically allowed to freely discriminate, with little government intervention.[82] This is an embrace by the regulators of actuarial fairness and the notion that to limit classification is to inappropriately intrude in necessary and beneficial insurance competition.[83]
Professor Mary Heen keenly observes there are no federal laws that directly ban race or sex discrimination in private insurance, compared with other existing laws in employment, public accommodation, education, fair credit, and housing.[84] The only real, significant legal protections at the federal level are in health insurance where public trends have long suggested movement away from actuarial fairness and towards mutual aid culminating in the Patient Protection and Affordable Care Act (ACA).[85]
Most insurance classification protections are at the state level[86] and they are not robust. States frequently ban “unfair discrimination” in insurance practices without specifying whether any given use of classification is inappropriate.[87] “Unfair discrimination” typically means that an insurer must demonstrate that the classifier’s risk is reasonably different from other groups and that the data supporting this is credible.[88] The insurer typically needs only to justify that the group demonstrates an increased risk that is “reasonable,” not that the insurer’s response (whether increased premiums, benefit restrictions, or denial of coverage) is proportional.[89] An insurer who wants to limit benefits for a person with AIDS, say, would have to show that the condition was costly compared to other conditions, but it would be free to charge an outsized penalty beyond the heightened risk.
States sometimes adopt an antidiscrimination model and prohibit certain class-based underwriting.[90] In a comprehensive survey of all states (and Washington D.C.) for insurance antidiscrimination laws across various private insurance lines, prohibitions on discrimination line up relatively well with equal protection jurisprudence.[91] States are more likely to ban discrimination on the basis of race, national origin, and religion, moderately likely to do so on the basis of gender and sexual orientation, and least likely with other statuses like credit score or zip code.[92] There is significant under-regulation across most classes and most types of insurance, even for typically protected groups.[93]
Taken as a whole, laws tend to embrace the vision of actuarial fairness, permitting classification to abound. Few laws prohibit class-based insurance discrimination at the federal or state level, and when they do, they typically permit discrimination if the insurer can show that actuarial methods support it. Additionally, laws tend to only protect some of the classes who experience insurance discrimination, and then not comprehensively, while a number of other groups often go unprotected.
The law undergirds actuarial fairness for one primary reason—it thinks it is necessary to do so in order to protect insurers and their important role in society. But leading insurance scholars are not so sure that this is necessary. In a forthcoming article, Professor Tom Baker draws from the development of insurance runoff markets to suggest that insurers don’t need as much safeguarding as the law allows.[94] His work suggests that, in the face of great uncertainty in different times in history, insurers have found ways to make do.[95]
Indeed, the rise of runoff suggests that we may have learned exactly the wrong lesson from the property casualty insurance industry’s harrowing asbestos and environmental liability experience. The right lesson is not that insurance markets need legal certainty, but rather that insurance markets can handle even extreme, once-in-history legal uncertainties.[96]
Other scholars agree: “[A] high degree of scientific and technical uncertainty permeates the insurance industry, the very business that is charged with transforming uncertainty into risk. Insurers do not necessarily back off from a high degree of uncertainty. Rather, they respond with a range of creative and sometimes ingenious solutions.”[97]
According to Baker, the aim of insurance laws should not be to protect insurers and their historical practices; it should be to seek out equality for insureds and let insurers figure out how to get there.
It is time for legal thought to update its insurance ideas and metaphors, and its use of insurance practices, to this more realistic understanding of insurance. Perhaps ironically, this more realistic understanding of insurance markets may hold the greatest promise within legal thought more broadly for scholars whose ideas least take markets into account. If insurance markets always and everywhere trade in uncertainty, and if insurance organizations always and everywhere develop ways to transform that uncertainty into manageable risk, then legal thought can safely relax its concern about the impact of legal change on insurance markets . . . . [L]egal thought can safely focus more on identifying the just distribution of legal rights and obligations and less on the destabilizing impact that moving toward that distribution might have on insurance markets.[98]
With this bold suggestion in mind, we turn to how concepts of stigma can reshape the notions of fairness and, in turn, regulation in insurance.
Actuarial fairness relies on a system of determining who is costlier and who is less so, and then insuring accordingly.[99] But its fairness depends entirely on how objective it is in picking those winners and losers.[100] This Article takes issue with that objectivity. Insurance classification doesn’t occur in a vacuum; like anything else, it may be influenced by the society that it takes place in.[101]
This Part explores how social stigma may influence insurance classification. First, I describe how stigma gets generated and perpetuated according to sociologists; then, I consider how we can better understand insurance discrimination through this lens.
Stigma is a commonly studied sociological phenomenon of how and why societies have a natural tendency to create and abuse underclasses.[102] Professor Erving Goffman coined the term stigma and authored a foundational work on the topic, which has since seen decades of social science contributions, both theoretical and empirical.[103] To Goffman, the stigmatized are people who suffer from “spoiled identities,” or who are “reduced in our minds from a whole and usual person to a tainted, discounted one.”[104]
Societies construct stigmatized statuses and then exert power over the stigmatized, who are seen as inherently less worthy of acceptance, approval, status, and social resources.[105] Stigma is a social phenomenon we see across all sorts of social domains: in education, in employment, in health care, and in other contexts.[106] Insurance is unlikely to be invulnerable to it. Stigma is a general phenomenon, applicable to any society. But each society may vary in how stigma takes form.[107] For instance, gender nonconformity is a stigmatized trait in some cultures while other cultures glorify it.[108]
Professors Bruce Link and Jo Phelan, leading scholars in the field, have described a multi-step process by which stigma is generated: labeling, stereotyping, separation, status loss, and subordination/discrimination.[109]
Labeling, according to Link and Phelan, is a natural and inevitable human function that is not necessarily harmful.[110] Instead, it is the meaning we ascribe to these labeled differences that shapes whether a group will occupy a stigmatized class.[111]
The vast majority of human differences are ignored and are therefore socially irrelevant. Some of these—such as the color of one’s car, the last three digits of one’s social security number, or whether one has hairy ears—are routinely (but not always) overlooked. . . . But other differences, such as one’s skin color, IQ, sexual preferences, or gender are highly salient in the United States at this time.[112]
Once that difference is viewed as meaningful, it becomes hard to overlook it and society begins to see these labels as inherently meaningful.[113] “[T]here are black people and white people, blind people and sighted people, people who are handicapped and people who are not.”[114] Categorizations are also often over-simplified.[115] Skin color can vary significantly, as can sexual preference, yet we often paint people in terms of binaries—black or white, gay or straight, as examples.[116]
Once a label has attached and becomes socially salient, stereotypes are often established to entrench these labels.[117] Stereotyping is, of course, generally a concern in antidiscrimination law; some scholars argue this is the basic ingredient that renders discrimination unlawful and unfair.[118] While antidiscrimination scholars often focus on stereotyping, stigma scholars view stereotyping as necessary, but as only one of several components in the stigma process. The stereotypes can be positive, negative, or neutral attributes.[119] Stereotyping is implicit; the person need not think overtly about the stereotype but instead will implicitly associate the stereotype with the label.[120]
The next step in stigma is separation. Once a group has been negatively labeled and stereotyped against, the natural instinct is for other members of society to want to distance themselves from the tainted or lesser being.[121] The more an out-group is distinguished from in-groups, the easier it is to view those individuals as “other,” as irrevocably different, and even less human than themselves.[122] Historical treatment of immigrants and African Americans is exemplary of this point in America. Professor Deborah Hellman has raised similar ideas about stigma in her work on demeaning in the context of discrimination. According to Hellman, what makes discrimination problematic is when it treats some individuals as of less moral equivalency than others, or when it demeans.[123] This separation can be motivated by desires to “keep people down,” to continue to subjugate underclasses so stigmatizers can continue to maintain power.[124] Sometimes it’s motivated by a desire to “keep people in,” to police norms and keep people within certain social boundaries.[125] Lastly, stigma sometimes functions to “keep people away,” quarantining so-called dangerous and undesirable people from the rest of society.[126]
Discrimination and status loss are natural consequences of stigma.[127] Underclasses often suffer from reduced social standing, even when it’s hard to point to any single act of discrimination.[128] Take the example of how women and people of color are often given less opportunity to be leaders through a variety of forms of implicit bias and subjugation.[129] This is often the product of structural stigma, or widespread institutional disadvantages that accrue to the individual across many systems over time.[130] For example, schizophrenia is a stigmatized mental health condition, and people with that condition experience structural discrimination in the form of how treatment and research are funded, where treatment centers are located, who is willing to provide care, and other factors.[131] Interpersonal stigma is also possible, with stigma pervading one’s personal relationships. For instance, an employer engages in interpersonal stigma if it treats an employee more poorly than others because that employee is a member of a stigmatized group.[132] Lastly, individuals can self-stigmatize, or adopt the social views that they are less than others.[133] In doing so, they may be more accepting of their reduced social standing and less likely to fight for broader rights and respect.
Subordination is the final step. While stigma is sometimes seen as interchangeable to discrimination, discrimination is best viewed as one piece of a larger stigma process, and subordination is the key to understanding why.[134] Certain groups may experience labeling, separation, stereotyping, and discrimination, but ultimately, they are not stigmatized; this is typically because they have power and social capital. Lawyers and politicians are good examples; they may be stereotyped and disliked but they ultimately occupy too much power for it to harm them.[135] Power by the stigmatizers over the stigmatized is necessary to “ensure that the human difference they recognize and label is broadly identified in the culture . . . . [T]hat the culture recognizes and deeply accepts the stereotypes they connect to the labeled differences . . . . [to] have the power to separate ‘us’ from ‘them’ and to have the designation stick . . . . [to] control access to major life domains.”[136] Stigma research suggests that the law and other systems may frequently operate to protect this power; those in power do not want to cede their power easily, and systems may wind up promoting stigma rather than tearing it down.[137] Antidiscrimination scholars have raised similar concerns—that discrimination is problematic when it creates historically disadvantaged groups across time.[138]
B. Applying Stigma Concepts to Insurance
Other scholars have recognized and critiqued classification for being biased. Professor Deborah Stone has observed that the classifications insurers use “dovetail precisely with those identities that have formed our major social cleavages: race, ethnicity, class, and more recently sexual orientation and disability.”[139] Professor Regina Austin has likewise critiqued classification: “However much [insurers] plead happenstance, insurance ‘risk’ classifications correlate with a fairly simplistic and static notion of social stratification that is familiar to everyone.”[140] This Article seeks to add to this narrative through the use of stigma theory.
Insurance classification shares uncanny resemblances to the stigma process. This suggests that stigma may pervade or influence insurance policy and practices. Stigma can help explain the reason behind the broader observations that Professors Stone, Professor Austin, and others make that insurance discrimination is a mirror on broader societal discrimination. Why insurers isolate some factors but not others in underwriting; why we tolerate actuarial fairness despite the clear harms it poses to some groups; and why lawmakers continuously protect insurance discrimination as an exception to the general rule that classifications, especially certain types, are unlawful.
Insurers, through their underwriting practices, are constantly attempting to delineate differences. The process of underwriting is, first, identifying classifications within a population and labeling them as good, bad, or neutral with respect to risk.[141] This forms the classifications by which insurers may then adjust premiums, deny coverage, or engage in other forms of discrimination.[142] Actuarial fairness advocates suggest that math does the labeling and that actuaries find distinctions based on real risk differentials that are objectively true.[143]
Stigma theory can explain why it may be no accident that actuaries label in the same ways that we do in other social contexts: because actuarial science, short of being objective and scientific, is instead shaped by our society.
Take an example of AIDS discrimination by insurers. Insurers historically frequently imposed lifetime or annual limits on AIDS-related care,[144] or simply refused to enroll HIV-positive individuals into health plans.[145] This was not a phenomenon just in the small or individual group markets; insurers frequently refused to cover groups if any individual in that group was HIV positive or had AIDS.[146] Lest one thinks that HIV/AIDS discrimination was something that occurred only at the height of the AIDS epidemic, studies suggest that it was occurring as frequently as the early 2000s before the adoption and implementation of the ACA.[147]
In Doe v. Mutual of Omaha Insurance Co., two Does challenged a private health insurance company’s lifetime caps of one hundred thousand dollars and twenty-five thousand dollars on benefits for AIDS or AIDS-related conditions when most other conditions had lifetime caps of one million dollars.[148] Judge Posner upheld the caps as permissible under the Americans with Disabilities Act (ADA), despite any proof from the insurer that the treatment of AIDS was in fact costlier.[149] Indeed, evidence at the time suggested that, while AIDS was costly to treat, it was no more so than many other chronic conditions that were freely covered, like heart disease, cystic fibrosis, and cancer.[150]
Why did insurers select AIDS as a relevant label, but not cancer? Undoubtedly, at that time and today, AIDS has much more social salience than cancer. In the public’s imagination, AIDS was a ravaging condition, wrecking the lives of young people with deaths by AIDS very much at the forefront of the public mind.[151] AIDS might have been more equated in the public mind with risk, or blame, or “other” than cancer.[152]
Or perhaps actuaries gravitate to socially salient labels because that is where the data is.[153] Insurers often look to large population level historical data sets for patterns to inform pricing.[154] The easiest data to obtain is that which is already available in some other context, for instance data gathered by government agencies. Importantly, though, what we count and how we count it is informed by social saliency.[155] Life insurers began using sex as a classification only after the government began collecting morbidity data based on sex.[156] But the government only began collecting such data in the first place because sex was viewed as a socially salient category with respect to mortality.[157] Some other label may be equally relevant, like height or freckles or hairy ears, but has gone ignored because it lacks social salience.[158] Under this logic, actuaries may find distinctions because they look there, and might not find other distinctions where they do not look.
Once a classification is created, insurers rely on custom in continuing to use it.[159] Thus, the same groups who are viewed as socially salient in other contexts, get classified in insurance, and then the practice of underwriting based on this class continues until regulators end it. It was not until 2015 that the first major insurer announced that it would cover individual life insurance policies for people with HIV and AIDS.[160] This was largely due to medical advancements that have made HIV more chronic than acute, which had been invented and in use for some time prior.[161] Still, life, disability, and long-term care insurers continue to discriminate against people with HIV and AIDS.[162] In January 2019, a settlement was reached in a different Doe v. Mutual of Omaha Insurance, after the insurer refused to issue long-term care insurance to an individual who was taking pre-exposure prophylactic (PrEP) drugs.[163] PrEP is typically taken daily with a goal of preventing HIV infection.[164] The medication is recommended for anyone who is at a higher risk of acquiring HIV.[165] This includes people who engage in injection drug use, people who have partners that are HIV positive, or people who have unprotected sex with people of unknown HIV status.[166] PrEP is at least ninety percent effective in preventing HIV from sexual intercourse and even higher when other precautions like condom use are taken.[167] The drug is seventy percent effective in the case of injection drug use.[168] Insurers admitted their goal was “reducing the number of people with HIV among its insureds.”[169] Only after a lawsuit did Mutual of Omaha agree to revise its underwriting policy to no longer exclude coverage based solely on PrEP use and has since issued a policy to Doe.[170] PrEP use was a proxy for AIDS¾a factor insurers have relied on as customary evidence of increased risk, despite evidence suggesting that these risks have significantly shifted in the almost forty years since AIDS was first known.
Labeling in insurance is not inherently problematic. But it is worrisome that labeling in insurance so frequently coincides with other socially salient labels.[171] We must be careful what we label, and what we do not, as it has serious (and often long-lasting consequences) for the labeled. Actuarial fairness presumes that we are capable of fairly labeling. Stigma theory suggests otherwise.
Stereotyping has been a longstanding concern in insurance, with some opponents to actuarial fairness asking whether underwriting, short of being cold hard math, is often a byproduct of stereotypes.[172]
Professor Regina Austin provides an example of discriminatory car insurance premiums in urban areas because “insurance companies give credence to the popular image of the city as an area of blight, high crime, on-street parking, and narrow, congested thoroughfares . . . . [A]lso, city residents are supposedly of lower educational and economic attainment.”[173] We can see how this may over-generalize and not adequately represent all urban areas or communities. Another notable example is women being viewed as economically and physically inferior compared to men in nineteenth century life insurance policies[174] or life insurance policies in the same era that insured emancipated slaves at two-thirds the benefits provided to white policyholders.[175] These are clear examples of how stereotypes or biases about the population got baked into the math of insurance.
Another example of stereotyping, some insurers refused to cover menopausal women because menopause, according to them, disrupted all sorts of physical functions and created nervousness.[176] Life insurance policies have varied for bartenders versus clergymen, based on the thinking that the latter lead more temperate lives.[177] Some have viewed a reluctance on the part of insurers to cover birth control pills, while covering Viagra, as stereotypes about acceptability of sexuality across gender.[178] Other examples can be found in historical exclusions by insurers of specific ethnic minorities[179] and discrimination in disability insurance.[180]
These stereotypes can play in favor of or against the individual. The individual is, according to the insurer, a sum of many parts, for some of which the insurer may raise rates and for some of which it may lower.[181] In either case, stereotyping should be worrisome to those who advocate for actuarial fairness as it suggests that some classifications may not be based in objective truth alone.
The narrative of who is or isn’t insurable may feed our stereotypes about social worth and good versus bad people. The good group is framed as deserving of being rewarded for their lower risk, or at least not beholden to the higher risk group, suggesting there is some amount of virtue and control in risk category.[182] This can fuel further stereotypes, not only causing economic consequences, but creating expressive harms, for instance, that certain groups of people are more careless or are more of a drain on the public coffers.[183] By classifying and insuring based on stereotypes, insurers may confirm biases against certain groups and make labels more meaningful than they might otherwise be.[184]
Moreover, in the act of labeling and stereotyping, insurers boil off individual identity, concentrating insureds into a few groups.[185] This goes to fundamental concerns raised by stigma of creating underclasses where individual identity is lost. In insurance, this can have the unfair effect of burdening individuals with whatever characteristic the insurer can say of the group, even when the individual may not embody any of those traits. This was perhaps most famously stated by the Supreme Court in City of L.A., Department of Water & Power v. Manhart after female employees sued their employer for pension discrimination: classification “in terms of religion, race, or sex tend to preserve traditional assumptions about groups rather than thoughtful scrutiny of individuals”[186] and furthermore
[w]omen, as a class, do live longer than men. . . . It is equally true, however, that all individuals in the respective classes do not share the characteristic that differentiates the average class representatives. Many women do not live as long as the average man and many men outlive the average woman. The question, therefore, is whether the existence or nonexistence of “discrimination” is to be determined by comparison of class characteristics or individual characteristics.[187]
In the early stages of the HIV and AIDS epidemic, insurers went to great lengths to avoid enrolling insureds who may be infected with the virus, using stereotypes of who they thought most prone to be infected.[188] Being a young gay male became a virtually uninsurable trait, if the insurer could uncover this. Some insurers used marital status, beneficiary identities, and profession, all with an aim of identifying men who have sex with men and keep them out of the insurance pool.[189] Insurers stereotyped based on sexual orientation, rendering all gay men as alike and not as individuals.[190] Yet, of course, within the population of gay men of the time, as in the heterosexual population, there was tremendous variation in terms of risk level. Some individuals might be in monogamous relationships, take great caution with sexual activity, or be sexually inactive. Also, recall that HIV was not costlier to insure than many other conditions.[191] Moreover, many other people were also at risk of acquiring AIDS besides gay men, but they did not experience the same discrimination.[192]
Decades later, insurers are still stereotyping based on sexual orientation and AIDS risk. In the PrEP discrimination cases, insurers are not asking extensive questions of insureds to weed out those who engage in activities that increase the likelihood of acquiring HIV. Instead, they are only focusing on PrEP use.[193] PrEP use, for insurers, is a useful proxy for the sexual orientation of that user (as eighty percent of all prescriptions are taken by gay men).[194] Certainly, those taking PrEP are acknowledging that they are at an increased risk of HIV¾but they are also undertaking largely successful steps to reduce that very risk.[195] A far better effort to avoid HIV-positive people in one’s insurance pool would be to find people who engage in risky activities and do not take PrEP. Or, if insurers are unable to screen for that risk, enroll everyone and provide bonuses for those who take PrEP.
Having convinced ourselves that certain people are meaningfully different and having attached negative stereotypes to those differences, we begin to treat those groups as “other” and their interests as inapposite to our own.[196] Actuarial fairness, as a policy position, adopts this very framing.[197] Insurers suggest that, if they insure one group whose classification they deem as too risky, others will be left to pick up the tab, referred to as cross-subsidy.[198] This model emphasizes a zero sum game where the inclusion of any single group means harm to the others, even though rarely is an insurer called upon to demonstrate how real and how great that harm is.[199]
Professor Deborah Stone emphasizes how actuarial fairness causes this harmful focus on separation.[200] To her, individuals are not inherently insurable or uninsurable; permitting the categorization of people into risk groups at all is a deliberate policy choice made by a given society.[201] She notes that we could just as easily adopt a mutual aid model of insurance, where everyone has some responsibility to chip in for the welfare of others.[202] Actuarial fairness favors, to some groups’ harm, an “us” versus “them” dichotomy.[203] It feeds into a tendency of people to identify themselves with low risk groups and to look for outsiders that are a higher risk than they are. For instance, in health insurance debates, lawmakers have frequently singled out pregnancy as a category of people for whom risk need not be pooled, despite the fact that all sorts of medical conditions could prove equally or more costly to insurers.[204]
Insurers admit that they only classify or “separate” where it is socially acceptable.[205] For instance, insurers commonly discriminated on the basis of domestic violence history in health insurance until public outcry led to a general end to the practice and legislation opposing it.[206] But what makes one form of separation or classification more or less socially tolerable than another? If one really buys into actuarial fairness, it should not be about social acceptability, but instead about data. The public and regulators often accept this narrative of the “other” who is costlier, even though they often lack such data,[207] and when data exists, it’s often overly reductive.[208] The public and regulators are frequently left to trust that there are “others” and that insurers know how to find them. Stigma theory helps to explain why so many people may be comfortable with this scenario because insurance treats as “other” the exact groups who we view as “other” more broadly in society. Classifiers may well be tolerated if they follow successful negative labeling and stereotyping of groups as risky and costly, or of reduced social standing generally.[209]
Discrimination by insurer against AIDS, for instance, may be no accident as people with AIDS suffer stigma in a variety of contexts.[210] It can be understood as a function of status loss experienced by persons with HIV and by gay men more broadly across society. Groups who suffer wider social condemnation and subjugation are apt targets for unforgiving insurance practices. Some observed, for instance, after Posner’s decision in Doe v. Omaha, that the caps on AIDS treatment were done in part because they would not engender the same outrage and widespread condemnation as if they had been lodged against cancer patients or others.[211]
Through insurance discrimination, the individual experiences clear status loss and discrimination in the form of lack of insurability, reduced benefits, or higher premiums. Posner’s holding in Omaha stripped those two Does from access to acute and chronic medical care, including access to life-saving antiretrovirals.[212] Access to insurance also has important implications for economic security for families and individuals.[213]
Insurance discrimination can also have more structural effects, creating community-wide or group-wide harms. The allowance of caps on AIDs benefits affects the financial well-being and morbidity and mortality of the entire population of people with AIDS or at least those who have private insurance. Lack of insurance can also compound misfortune, adding financial and other difficulties for those already in distress from accident or injury.[214] Across time, insurance policies can drive population-level disparities. For instance, one study estimated that African Americans are subject to on average thirty percent higher car insurance premiums than white peers, based on data in four different communities, even when the white individuals live in communities that are equally, objectively at risk of car accidents.[215] This can affect communities generationally, cutting into household pay over time, and controlling whether people can afford to commute to educational or work opportunities outside of their neighborhood.
Insurance discrimination can also create structural barriers to larger public policy goals. In the PrEP denial cases, people were responding to insurance denials by quitting PrEP, posing serious threats to public and individual health.[216] Indeed, this conduct is not even good for insurers as a whole, over time, as it creates a reverse moral hazard where people are exposing themselves to greater risk to become eligible for insurance.
In stigma theory, law is often used by people in power to retain it and to keep others subordinated.[217] Insurance discrimination laws can be seen in this light, as a way to protect the status quo, benefiting some, despite the many harms they pose to others.
Other scholars have recognized elements of subordination in the insurance system. Professor Deborah Stone has observed that “the principle of actuarial fairness in all its institutional forms is a marvellously invisible way of creating and perpetuating a segregated society. It explains misfortune as the result of unalterable natural characteristics of individuals.”[218] And as Professor Stone reminds us “[i]nsurance underwriting, far from being a dry statistical exercise, is a political exercise in drawing the boundaries of community membership. That insurers always understood they were creating communities of privilege is very clear.”[219]
It’s notable that despite decades of scholarship taking issue with actuarial fairness, lawmakers continue to widely permit insurance classifications.[220] Sexual minorities remain unprotected from many forms of insurance discrimination, despite decades of evidence that gay men have been subject to actuarially-unjustified discrimination in insurance specifically. Yet, no federal law forbids the use of sexual orientation in denial of life, disability, or long-term care insurance. Health insurance discrimination was only forbidden on the basis of sex with the passage of the ACA and even then, rules implementing the statute did not clearly and explicitly forbid discrimination based on sexual orientation.[221] Use of sexual orientation as an insurance classification is only banned in six states,[222] and less than half of states ban sexual orientation discrimination in public accommodation statutes.[223]
C. Implications of Stigma Theory for Insurance Policy
A stigma approach to insurance fairness ultimately undermines actuarial fairness primarily because it suggests that we cannot be trusted in our classifications. Actuarial fairness demands that we label and categorize risk, but we are sharply limited, as humans, in our ability to do this in a way that isn’t tainted by social context. We look for differences, but we inevitably gravitate to those that are socially relevant, rather than empirically so.[224] Insurers must have somewhere to start in classification¾naturally, they turn to the “differences” that matter so much in other parts of life: age, race, gender, sexual orientation, disease status, wealth.[225] In doing so, they may over-emphasize those labels’ meaning and miss other labels that would also be significant if scrutinized. The burden of actuarial fairness may fall unfairly onto some and unfairly benefit others.
Once we have labeled, the stigma literature tells us that we dig in and entrench those labels. We stereotype, finding information to support those distinctions that we now are convinced are meaningful, rather than looking for other real differences.[226] Enough focus on our differences, and it’s inevitable that we will justify differential treatment, using power to subjugate others and to protect ourselves. The result is an underclass of people who are consistently treated badly by insurers. And because this process is not divorced from broader society, that underclass is typically discriminated against in many forms of insurance, and across the broader society too.
Stigma theory also helps us to explain why the narrative of actuarial fairness is so compelling. Others have observed how the cloaking of insurance classification in objective science and visions of fairness has insulated it from scrutiny.[227] Insurers, consciously or not, have leveraged a natural tendency in people to stigmatize.
Stigma theory asks us to approach insurance quite differently. It asks us to be skeptical of labels that are socially salient, to be on the lookout for and eliminate stereotyping, and to examine insurance classifications for hierarchies of power for who stands to gain and who stands to lose in a given situation.[228] And it allows for a scrutinizing of the entire insurance system, not any single classification. It helps to explain why actuarial fairness, as a concept, is problematic. At its broadest, it asks us to consider insurance policy and practices as not immune from, but instead driven by, social constructs.
III. Imagining an Insurance System that Redresses Stigma
Stigma theory provides a new framework to assess the policy underpinning our insurance system. It suggests that actuarial fairness, rather than being rooted in objective science, is prone to human error that unjustifiably harms certain members of society. Lawmakers and others need to revisit insurance practices and policy with stigma in mind.
This Part explores policy remedies to redress stigma in our insurance system. The Article promotes a mutual aid model of insurance as the best opportunity to tackle stigma in insurance. If mutual aid is not politically feasible, a more robust antidiscrimination model provides some relief. Lastly, if actuarial fairness is to remain the model for insurance policy, then transparency can help to alleviate some forms of stigma.
An insurance system that, at least in some personal lines of insurance, adopts a mutual aid model may be the best chance at eliminating stigma. A mutual aid model of insurance asks that we, as a collective society, share risks equally, regardless of individual levels of risk.[229] A number of scholars have advocated for this model in health and other forms of insurance.[230]
Proponents of mutual aid argue that it helps those who are less fortunate to gain access to the financial security of insurance.[231] Insurance becomes a form of social redistribution, where everyone carries some of the risk of accidents and injuries which individuals would be less able to shoulder on their own.[232] It reflects a society where we view our responsibilities to one another, and not just ourselves.[233] It also makes insurance more efficient. Under actuarial fairness, those who most need insurance are least likely to get it;[234] with mutual aid, insurance finally functions as the risk spreader it is intended as.[235]
From a stigma perspective, mutual aid gets rid of the most problematic function of insurance: classification.[236] Under a mutual aid model, all who buy into insurance would be enrolled and receive a similar package of benefits at the same rate. Insurers would no longer need to engage in the social- and values-laden enterprise of deciding who is more or less risky and who is more deserving of the best benefits.
Mutual aid is a common enough theory but rarely demonstrated in practice in America where actuarial fairness rules the day. The ACA is one noteworthy example of mutual aid at work. In the era before the ACA, access to health insurance (and by proxy, health care) was impossible, or nearly so, for many patients.[237] In a 2001 study, a “mock” HIV-positive individual phoned sixty health insurers, seeking coverage. He was denied coverage by every plan he contacted.[238] To many health insurers at the time, the eight hundred thousand to nine hundred thousand Americans living with HIV were simply uninsurable.[239] Some even promoted the testing of insureds for the virus so insurers could more proactively avoid inadvertently insuring HIV-positive individuals.[240] The health insurance system was a perfect model of actuarial fairness with health status being the predominant factor that insurers used to classify risk.
The level of discrimination in the health insurance market was decried by many, though. Health insurance was too costly or completely unavailable for those who needed it most, with fatal consequences and an uninsured rate drifting around seventeen percent of the population.[241] Several laws sought to establish some form of mutual aid and eliminate certain classifications in health insurance, generally on a more fragmented basis. HIPAA and ERISA imposed some limits on health status discrimination[242] followed by bans on genetic discrimination by the Genetic Information Nondiscrimination Act (GINA),[243] mental health and substance use discrimination in the Mental Health Parity and Addiction Equity Act (MHPAEA),[244] discrimination in employer health plans in Title VII of the Civil Rights Act,[245] disability discrimination by private insurers in Title III of the ADA,[246] and age discrimination in the Age Discrimination in Employment Act.[247]
The ACA was the nation’s most recent and most expansive effort to impose mutual aid onto the health insurance industry. Under the ACA, insurers are widely restricted from health status discrimination in premiums,[248] enrollment,[249] and benefits.[250] Section 1557 of the ACA also forbids discrimination on the basis of race, age, sex, and disability by entities receiving federal funds, which now includes most private insurers.[251] The ACA was designed largely to address uninsurance and to make it possible for everyone, even those with poor health status, to obtain affordable health insurance. Of course, such antidiscrimination measures might make insurance prohibitively costly, so the government provides subsidies for people to purchase insurance (and cover out-of-pocket expenses).[252] It also mandates that individuals purchase insurance to spread risk across the greatest pool possible.[253]
The ACA’s mutual aid model is not limitless; it still allows some forms of insurance discrimination.[254] And the system still relies on for-profit insurance companies, thus it has had to be carefully crafted and then monitored for compliance to discourage insurers from engaging in implicit or hidden classification.[255] Complaints sometimes persist that insurance companies continue to unfairly discriminate against the sick even after the ACA was adopted. For instance, insurers have placed AIDS-related medications on the high cost-sharing tiers in order to shift cost back onto patients and generally to discourage them from enrolling in the health plan, as one example.[256] Another, New York state officials have received complaints that insurers are imposing prior authorization requirements and improper denials of coverage for PrEP drugs, likely with an aim of discouraging PrEP users from enrolling in their plans.[257]
But largely the ACA is evidence of a widespread rejection of health-based classification in health insurance. Protections for people with preexisting conditions are now widely popular,[258] and any health reform aimed at repealing or replacing the ACA would face public challenge if it failed to meet these new expectations. This may be one reason why even greater models of mutual aid, like Medicare for All, appear to be growing in popularity among citizens[259] and some democratic candidates in 2020.[260] Such a model may be possible in health insurance in America, since health insurance has more closely hewn to mutual aid than any other form of private insurance. But, even in health care, such a change would be contentious, as the move towards mutual aid in the ACA was.
A fundamental question would be whether all forms of private insurance merit mutual aid or whether some justify actuarial fairness? Many scholars distinguish health insurance as special compared to other insurance because it involves access to life or death procedures.[261] For this reason, health care is often viewed as a right not a privilege, whereas other forms of insurance may occupy lesser importance. Yet, some other forms of personal insurance are almost or equally as high stakes. Long-term care insurance can directly affect one’s access to health-based resources. And other forms of insurance implicate financial health significantly and, as such, have important implications for individual and group well-being.
Health care also often involves brute luck where people have unexpected and terrible blows to their health through absolutely no fault of their own.[262] Certainly, though, many other forms of insurance cover brute luck, such as a car accident that was not that individual’s fault, or an unexpected early death in the family from cancer. This begs the question of how to handle insurance risks that are more fault-based, for instance a driver who causes an accident.[263] Some may support classification if the low risk is deserved,[264] for example safe driver discounts.[265] Yet, mutual aid does not wholly foreclose such options. The ACA fundamentally overhauled the insurance system to widely eliminate health status discrimination, but it did still allow some rewards for health status, namely penalties for tobacco use and age[266] and wellness plans that reward certain health outcomes or participation.[267] Some plans are also exempt from complying with the ACA’s extensive consumer protections.[268] One could envision a car insurance program, say, where everybody paid a base rate (not influenced by classifiers) but still received some safe driver discounts, and where grandfathered plans or exemptions were sometimes possible.
Lastly, any effort to impose mutual aid on other insurance markets would have to consider how to offset any financial costs associated with this. However, it’s important to note that it’s unclear how much costlier insurance would be. Less classification may mean that more people are allowed in the market at lower rates and this would have to be offset. But it also means that some people, outright excluded from the market, would now be let in. We have to really trust our actuarial methods to assume that the sum increased cost would be negative. If classifications, right now, are not always accurate, then there may be some low risks being precluded from the market. Moreover, there are ways to offset these expenses. In the ACA, this was government subsidies as well as a universal mandate to purchase insurance. It seems unlikely that the government would be willing to expend resources to prop up many of these other private insurance markets, but mandates may be possible (and already exist in some insurance markets like car insurance). The ACA also redistributed funds from “winning” insurers to “losing” ones which could also be a viable model in other personal insurance lines.[269] And there are other methods to combat adverse selection and moral hazard. For the former, individuals can be limited in when they can enroll in the market (the ACA also does this); for the latter, cost-sharing. Classification is not the only way.[270] Also, cost is often lodged as a defense to justify discrimination, and many scholars would argue that, if discrimination is harmful, we should eliminate it regardless of expense.[271] And, of course, some may object on fairness grounds, that if the system is costlier, their premiums should not go up. But this goes to the heart of the stigma argument¾in many cases, can any single individual be sure that they are the low risk, deserving of the better benefit, or might they be benefitting from a system that has historically protected them unjustly at the expense of others?
Greater consideration would need to go into this model in other personal lines of insurance. Does this model enhance equity or raise different and new challenges? For now, a movement towards mutual aid in other lines of insurance seems politically infeasible, for these and many other reasons, despite the fact that it may logically be the best way to eliminate stigma and insurance underclasses.
Short of mutual aid, increased bans on classification could also lead to some improvement in reducing the harms of stigma. The “anti-discrimination” perspective opposes the use of certain classifications, particularly those restricted by antidiscrimination law in other contexts.[272] This model differs from mutual aid in that it focuses only on certain protected groups, instead of protections for everyone.[273]
Insurance classification based on immutable traits like race or sex are particularly critiqued for efficiency reasons; such classifications do nothing to promote risk aversion.[274] It’s not always clear that some of these classifiers do a very good job at predicting loss; there may be better metrics available.[275] And some predictions may be flavored by stereotypes rather than objective measurements of risk.[276] Some worry that insurers overinflate concerns about adverse selection and moral hazard to advance whatever classification they want.[277] This approach is also supported by arguments that people should not be penalized for factors that are beyond their control, especially immutable traits that are typically protected otherwise.[278]
Certainly, from a stigma perspective, reduction in classification would be a good thing. A focus on this at both federal and state levels could go a long way towards eliminating the more egregious and common forms of discrimination in gender, age, race, zip code, health status, and other factors.[279] It eliminates, at least in some cases, that uncertain exercise in adversely labeling certain groups, when stigma tells us that we cannot be objective in doing so.[280] It cuts off labeling at the pass, thus we can avoid stereotyping and any impulse to separate or to discriminate. Greater anticlassification bans on AIDS-based or sexual orientation-based discrimination, for example, could have chilled much of the historical discrimination we have seen continue for the last several decades.
However, this is undeniably an incremental and inferior approach compared with mutual aid. For one, bans on classification for certain groups may lead insurers to find proxies for that trait, or other indirect ways to continue discriminating. For instance, in the PrEP discrimination case, outright discrimination based on sexual orientation and disability was banned in that particular state, but the insurer still found indirect ways to reach those categories of people.[281] Increasingly, this may be complicated by the bevy of knowledge about consumers that insurers and other companies have at their fingertips through the gathering and selling of metadata.[282] A recent study suggests that insurers can easily obtain information about “race, education level, TV habits, marital status, net worth” and other intimate aspects of our lives.[283] More information could mean more factors insurers can consider for classification, which could actually help to take the focus off of traditional classifications and put it on new categories of risk.[284] However, there is also the possibility that insurers may simply use this wealth of information to identify better proxies that correlate with but are not protected traits.[285] In this way, they could obscure protected class discrimination through what appears to be neutral terms: the wealth of a neighborhood one lives in, how educated an individual is, or income.[286] Additionally, while some may argue that big data and algorithms can help insurers to be more scientifically accurate, other scholars have suggested that metadata are subject to bias and manipulation too and may only double-down on the same groups of people who suffer historical discrimination.[287] Greater regulation could focus on these issues, but there would surely be significant regulatory burdens and litigation around what amounts to classification.
Moreover, a focus on classification demands that we choose which traits are more or less worthy of regulation. Short of major legal reforms, we are probably likely to only see protections for a handful of traditionally protected traits like race, gender, age, and disability. But the stigma literature suggests there may be as many as ninety-three stigmatized traits in our society.[288] Even if we select the more common traits used by insurers, it is still a significant amount of regulation, and there has been a notable unwillingness to regulate many of these categories historically. For example, obesity,[289] type of job,[290] and credit score[291] are all categories that might be stigmatized generally and that are commonly used in insurance, though they are not protected classes in other contexts. And some discrimination goes not just to traits but conduct too, for instance, smoking.[292] A fair system of classification would require regulators to scrutinize all forms of classification for fairness and stigma, rather than extending protections only in an ad hoc manner, as the system currently operates.
Insurers frequently rely on custom in their underwriting procedures and may be unwilling to change this.[293] However, as Baker advises, legal scholarship on insurance should focus more on distributive justice aims and less on impacts on markets, as insurers have historically shown an ability to adjust and innovate when regulated.[294] Again, though, stigma theory argues that this may need to be a necessary expense, pooled across insurers and the population, to make sure that insurers are not simply recycling unfair and unsubstantiated insurance underclasses.
C. Improving the Status Quo with Increased Transparency
The most likely option is that insurance regulation will remain as the status quo. Perhaps some classifications may be eliminated through federal and state law, but generally classification will persist, and actuarial fairness will rule.
If this is the case, there are still ways to improve classification within the existing legal structure to reduce the harms of stigma. Many states forbid unfair businesses practices, often those which are not actuarially justified, even if they do not ban specific forms of classification.[295] Additionally, a number of federal laws likewise ban discrimination that is not actuarially justified.[296] Insurance regulators can be more aggressive in enforcing these laws, in rooting out particular insurance practices and demanding justifications from insurers.
These laws embrace actuarial fairness, but fundamentally they ask regulators to pay attention to actuarial objectivity. This Article doesn’t disagree outright with the concept of actuarial fairness in theory (that some should pay more than others if they are costlier); instead, it argues that actuarial fairness may be impossible to achieve because we as a society cannot objectively evaluate risk classifiers in a way that isn’t influenced by stigma.[297]
If we cannot overhaul the entire insurance industry, we can at least be more thoughtful and transparent in holding to account the actuarial process¾and this could help with stigma.
What would an actuarial process that really considers stigma look like? First, regulators would need to ask whether the classification is against a group that is socially salient.[298] This could be a broader exercise than protected class determinations discussed in the prior Section. Obesity, for example, is typically not a protected trait, but it does have social salience in American society and is often the subject of discrimination.[299] Second, do the actuarial methods reveal anything that suggests the classification is rooted in stereotyping or assumptions about the group? Would you expect that the finding might confirm existing biases about that group? Is it based on old and antiquated, possibly biased historical data? Why was this classification chosen over other classifiers, and how expensive is the risk group comparable to others? Third, who stands to gain by this classification, and who stands to lose? Are similar classifications happening against a group who is also higher risk but typically not stigmatized?
While these factors are by no means comprehensive, they may provide a starting point for a dialogue between regulators and insurers about how a particular classification may be rooted in stigma.
Consider, again, our caps on benefits for patients with AIDS in Mutual of Omaha.[300] The legal challenge was that the caps violated the ADA.[301] The ADA has an insurance safe harbor which exempts insurers from discrimination on the basis of disability so long as it is actuarially fair and it is not a subterfuge for discrimination.[302] It was put in place with an aim, according to legislative history, of preventing employers from failing to hire disabled employees out of fear of increased costs to employer health plans (and thus employers).[303] The Equal Employment Opportunity Commission (EEOC) has interpreted the safe harbor to require that “conditions with comparable actuarial data and/or experience are treated the same way.”[304] In a recent EEOC rule about the safe harbor, the agency clarified an important point about what it meant by actuarial fairness: “The safe harbor provision . . . allows the insurance industry and sponsors of insurance plans, such as employers, to treat individuals differently based on disability (normally a prohibited practice under the ADA), but only if the differences can be justified by increased risks and costs ‘based on sound actuarial data and not on speculation.’”[305]
Such provisions, if enforced, would permit judges and state officials to put the onus on insurers to justify their classification practices. In Posner’s opinion, the insurer fails to appeal to the safe harbor.[306] Posner should have been more skeptical of this. The insurer failed to do so in part because it did not have data to support the claim of actuarial fairness.[307] If we were to focus more on stigma, and be more serious about actuarial fairness, this should have been enough to suggest that the classification was possibly rooted in disability discrimination. At the minimum, that mere speculation of increased cost should not be enough. Re-evaluating that case, a judge could ask where the evidence is to support the classification; the law allows for this expressly. The judge (or regulator) can ask why AIDS was treated differently than other similar conditions, like cancer. AIDS and HIV might never be treated like a head cold, but have they been treated in the same way as other similar serious and costly illnesses, like cancer or renal disease.[308] This could prevent the silo-ing and stigmatizing of certain groups. Insurers would be unlikely to discriminate against cancer, and so they also could not discriminate against AIDS.[309]
A more modern example, New York State demanded to see the actuarial data that supported long-term care insurers discriminating based on PrEP.[310] Upon reviewing the data, the state banned the practice, stating that
underwriting practices in which adverse underwriting decisions are applied to individuals who take PrEP to mitigate the risk of contracting HIV, but no adverse underwriting decisions are applied to individuals with the same level of potential exposure to HIV who do not take PrEP to mitigate the risk of contracting HIV, are neither based on sound actuarial principles nor related to actual or reasonably anticipated experience.[311]
Unfortunately, the state did not publish the data supporting its decision in its ban, but the state’s decision to ban insurance discrimination based on PrEP is still a major step towards scrutinizing insurers within the letter of the law.[312] It forced the insurers to go from mere speculation to proof, and when asked to do this, they did not have enough evidence to satisfy the classification.[313]
Of course, this proposal depends on insurance regulators and judges being willing to enforce these laws and to be more rigorous in assessing what constitutes unfair insurance practices. It also raises questions of whether there should be greater transparency to the public about classification practices. Engaging with some of these questions in a public manner could help dramatically improve the public’s understanding of how insurance classifications are created and whether they are, according to the public’s mind, appropriate. While insurers agree that classifications which are socially unacceptable should not be used,[314] there has simply never been enough information for the public to understand and engage with discriminatory insurance classifications.
Another challenge of this proposal I have also argued elsewhere is that some safe harbors may be evidence of bias and stigma.[315] The ADA’s safe harbor is an outlier compared to other antidiscrimination statutes that govern employer benefits or other insurance, as other insurance laws typically do not contain such a safe harbor.[316] What this suggests is that lawmakers saw something special about disability that foretold high insurance expenses: either that the lawmakers assumed this, or they expected employers to.[317] Stereotypes of disabled people as inherently unhealthy might inform this distinction.[318] Rulemaking that forgives actuarially-based discrimination in some cases but not in others may be inherently unfair. It may be rooted in suspicions, rather than evidence, about the cost of any given group.[319] And it certainly does not recognize the individuality of claimants.[320] But this is a narrow objection. It only applies to safe harbors that are carved out for one group and potentially rooted in stereotypes.[321] General state laws that forbid all versions of unfair insurance discrimination would not raise these same concerns.
Insurers may object to transparency in underwriting on the grounds that such data should be treated as proprietary and needs to be protected from competitors.[322] Insurers generally have a desire to protect the use of classifications, as it can require time and effort to find alternatives. Yet, there are mechanisms to obtain such information and times in which we ask for it. State agencies sometimes poll insurers for underwriting practices in certain cases, as was the case with PrEP discrimination in New York.[323] One way to mitigate these concerns, too, is to limit access to the information to state agencies. However, there may be times where public knowledge would also be beneficial.
Lastly, while I cite many examples where classifications did not hold up under sunlight, there may well be some classifications that are in fact actuarially fair after they are more closely scrutinized. Those who have been frequently stigmatized may well present higher risks in insurance.[324] In some instances, these differences may be truly innate, for instance, the fact that women live longer than men.[325] In other cases, it might reveal how deeply structural discrimination has harmed certain groups in insurance and more broadly.[326] Racial minorities, as a collective, are in poorer health, because they have historically lacked access to social goods that make others healthy (safe housing, adequate income, etc.) as a product of the stigma they long suffered in society.[327] This has been aggravated by the discrimination they have historically faced in health insurance, which has functionally denied them access to health care.[328] This may explain why so many socially salient groups are often subject to insurance discrimination; because the way we treat them in other contexts makes them have life circumstances that are less attractive to insurers.
Ultimately, if greater transparency of insurance practices reveals that much of this discrimination is actuarially fair, then this would function to highlight broader, systemic inequities that the insurance industry may be exacerbating, and we would then need to ask fundamental questions about the proper function of our insurance industry and social safety net.[329] All this would suggest is that the study of stigma and insurance becomes even more important. That evidence base could well be used to support my earlier arguments that mutual aid or, at the very least, bans on insurance classification are the only acceptable remedies.[330]
Discrimination is often permitted in insurance when it is prohibited otherwise. This is in part due to notions of actuarial fairness, the idea that some people are costlier to insure and they should shoulder that expense, not others. This cloaks insurance discrimination in matters of equity and objectivity. Yet, stigma theory suggests that we may be incapable of being purely objective in the highly social exercise of deciding whom we label as risky and not worthy of social pooling. Stigma in insurance can lead to intolerable harms and the creation of permanent underclasses in insurance and more broadly. At the minimum, we need to begin scrutinizing insurers for whether their methods are based in objective science or whether they are influenced by stigma. A superior option is to move the insurance system, or at least some lines of insurance, towards a mutual aid model where everyone shares equal risk, and classification is unnecessary.
[1] Health insurance can control access to necessary medical care, while other lines of insurance guarantee financial security during extended illness, a disability, death of a family member, or serious accidents like a house fire.
[2] The Article focuses on personal insurance, those insurance products individuals purchase to guard against their own risks, including life, disability, health, long-term care, and auto insurance. This Article does not address liability insurance that companies or professionals take out, as this may raise distinct issues about underwriting and discrimination. See infra notes 8–20 for examples of the categories of people frequently discriminated against in insurance.
[4] State law typically prohibits unfair insurance practices, with actuarial justifications being a defense. See infra notes 84–90.
[5] See, e.g., Deborah A. Stone, The Struggle for the Soul of Health Insurance, 18 J. Health Pol., Pol’y & L. 287, 293–94 (1993) (discussing actuarial fairness and mutual aid as competing models in insurance).
[6] Bruce G. Link & Jo C. Phelan, Conceptualizing Stigma, 27 Ann. Rev. Soc. 363 (2001) [hereinafter Link & Phelan, Conceptualizing Stigma].
[7] Raymond A. Guenter & Elisabeth Ditomassi, Fundamentals of Insurance Regulation: The Rules and the Rationale 1 (2017).
[8] See generally Sharona Hoffman, Unmanaged Care: Towards Moral Fairness in Health Care Coverage, 78 Ind. L.J. 659 (2003); see also Norman Daniels, Justice, Health, and HealthCare, 1 Am. J. Bioethics 2, 3 (2001) (asserting the underlying importance of health care is to participate in society and exercise fundamental rights of liberty and autonomy); Wendy K. Mariner, Health Reform: What’s Insurance Got to Do with It? Recognizing Health Insurance as a Separate Species of Insurance, 36 Am. J.L. & Med. 436, 442 (2010) (rejecting actuarial fairness in the context of health insurance because of the importance of access to health care).
[9] See, e.g., Jeffrey P. Kahn & Susan M. Wolf, Understanding the Role of Genetics in Disability Insurance, 35 J.L., Med. & Ethics 5 (2007) (arguing that genetic discrimination in the context of disability insurance is particularly harmful because it functions as income replacement for essential and basic needs).
[10] Regulation of insurance is discussed in more detail infra Section I.B and Part III.
[11] See infra notes 144–150, 160–170 for examples in health, life, disability, and long-term care insurance.
[12] Becca Rausch, Health Cover(age)ing, 90 Neb. L. Rev. 920, 921–22 (2012); Rebecca Puhl & Kelly D. Brownell, Bias, Discrimination, and Obesity, 9 Obesity Res. 788, 790, 794–95 (2001); Jay Bhattacharya & M. Kate Bundorf, The Incidence of the Healthcare Costs of Obesity, 28 J. Health Econ. 649, 657 (2009) (finding that “obese workers with employer-sponsored health insurance pay for their higher expected medical expenditures through lower cash wages. These wage differences are greatest among female workers, who have larger expected medical expenditure differences associated with obesity than male workers”).
[13] See generally Mary Crossley, Discrimination Against the Unhealthy in Health Insurance, 54 Kan. L. Rev. 73 (2005). For further examples of this practice in the insurance industry, pre-Affordable Care Act (ACA), see Gary Claxton et al., Pre-Existing Conditions and Medical Underwriting in the Individual Insurance Market Prior to the ACA, Kaiser Fam. Found. (Dec. 12, 2016), https://www.kff.org/health-reform/issue-brief/pre-existing-conditions-and-medical-underwriting-in-the-individual-insurance-market-prior-to-the-aca [https://perma.cc/9MFZ-4QU8]. The practice continues post-ACA in unregulated insurance markets. For examples of the many declinable conditions health insurers may use, see Blue Shield of Cal., Application Eligibility and Underwriting Process Guide: For Individual and Family Off-Exchange Plans and Medicare Supplement Plans 23–24 (2016), https://www.blueshieldca.com/bsca/bsc/public/broker/PortalComponents/StreamDocumentServlet?fileName=A16159_7-16.pdf [https://perma.cc/ZHD4-VUF5].
[14] Yann Joly, Ida Ngueng Feze & Jacques Simard, Genetic Discrimination and Life Insurance: A Systematic Review of the Evidence, 11 BMC Med. 25 (2013); Jessica L. Roberts, Preempting Discrimination: Lessons from the Genetic Information Nondiscrimination Act, 63 Vand. L. Rev. 437, 443 (2010).
[15] For foundational work on this topic, see Deborah Hellman, Is Actuarially Fair Insurance Pricing Actually Fair?: A Case Study in Insuring Battered Women, 32 Harv. C.R.-C.L. L. Rev. 355 (1997). For an extensive overview of how and when historically insurers have used this criteria in various states, see Terry L. Fromson & Nancy Durborow, Nat’l Health Res. Ctr. on Domestic Violence & Women’s Law Project, Insurance Discrimination Against Victims of Domestic Violence (2014), http://www.womenslawproject.org/wp-content/uploads/2016/04/Insurance_discrim_domestic_violence-1.pdf [https://perma.cc/Q7AN-ETES]. For a survey of where state laws stand on domestic violence as a criteria for insurance now, see Emily C. Wilson, Stop Re-Victimizing the Victims: A Call for Stronger State Laws Prohibiting Insurance Discrimination Against Victims of Domestic Violence, 23 Am. U. J. Gender, Soc. Pol’y & L. 413 (2015).
[16] For context on federal laws governing race-based redlining, see William E. Murray, Homeowners Insurance Redlining: The Inadequacy of Federal Remedies and the Future of the Property Insurance War, 4 Conn. Ins. L.J. 735 (1998); Willy E. Rice, Race, Gender, “Redlining,” and the Discriminatory Access to Loans, Credit, and Insurance: An Historical and Empirical Analysis of Consumers Who Sued Lenders and Insurers in Federal and State Courts, 1950–1995, 33 San Diego L. Rev. 583 (1996). For modern examples of car insurance discrimination based on race and related proxies, see Jeff Larson et al., How We Examined Racial Discrimination in Auto Insurance Prices, ProPublica (Apr. 5, 2017), https://www.propublica.org/article/minority-neighborhoods-higher-car-insurance-premiums-methodology [https://perma.cc/AQ6D-N895]; Julia Angwin, California Is Investigating Racial Discrimination in Auto Insurance Premiums, Pacific Standard (May 23, 2017), https://psmag.com/economics/racial-discrimination-in-auto-insurance-premiums [https://perma.cc/A8SN-KMSL].
[17] Rice, supra note 16; Naomi Naierman & Ruth Brannon, Sex Discrimination in Insurance, in 1 Discrimination Against Minorities and Women in Pensions and Health, Life, and Disability Insurance 473 (U.S. Comm’n on Civil Rights ed., 1978).
[18] Gary Williams, The Wrong Side of the Tracks: Territorial Rating and the Setting of Automobile Liability Insurance Rates in California, 19 Hastings Const. L.Q. 845 (1992).
[19] Lea Shepard, Seeking Solutions to Financial History Discrimination, 46 Conn. L. Rev. 993 (2014); Latonia Williams, African American Homeownership and the Dream Deferred: A Disparate Impact Argument Against the Use of Credit Scores in Homeownership Insurance Underwriting, 15 Conn. Ins. L.J. 295 (2008); Darcy Steeg Morris, Daniel Schwarcz & Joshua C. Teitelbaum, Do Credit-Based Insurance Scores Proxy for Income in Predicting Auto Claim Risk?, 14 J. Empirical Legal Stud. 397 (2017).
[20] Larry Cunningham, The Case Against Dog Breed Discrimination by Homeowners’ Insurance Companies, 11 Conn. Ins. L.J. 1 (2005).
[23] The term discrimination exudes negative connotations of unfair treatment. Yet, discrimination is merely the act of classifying based on an identifying feature. Most scholars agree that discrimination is sometimes permissible; for instance, we divide classrooms by age, or locker rooms by gender. Deborah Hellman, When Is Discrimination Wrong? 2–3 (2008). When I use the term “discrimination” in this Article, I mean it as a neutral term that an insurer has in some way treated that insured differently from others through premiums, benefits, or enrollment. I do not mean the term to suggest that the discrimination is necessarily unlawful or even unfair in any given context.
[24] Spencer L. Kimball, The Purpose of Insurance Regulation: A Preliminary Inquiry in the Theory of Insurance Law, 45 Minn. L. Rev. 471, 478 (1961) [hereinafter Kimball, Insurance Regulation] (“[Insurance] provides a degree of objective certainty in an uncertain world, it converts unpredictable risk to predictable cost, it smooths the path of economic activity.”).
[25] As just one example, the cost of certain kinds of chemotherapy for breast cancer ranged from $82,260 to $160,590, but patients with health insurance only paid out-of-pocket costs around $3,381 and $2,724. Sharon H. Giordano et al., Estimating Regimen-Specific Costs of Chemotherapy for Breast Cancer: Observational Cohort Study, 122 Cancer 3447, 3449–51 (2016).
[26] Kimball, Insurance Regulation, supra note 24, at 478; see generally Ronen Avraham, The Economics of Insurance Law—A Primer, 19 Conn. Ins. L.J. 29, 32–33 (2012).
[27] Spencer L. Kimball, Reverse Sex Discrimination: Manhart, 1979 Am. B. Found. Res. J. 83, 102, 106 (1979) [hereinafter Kimball, Reverse Sex Discrimination]; Kenneth S. Abraham, Efficiency and Fairness in Insurance Risk Classification, 71 Va. L. Rev. 403 (1985).
[28] Omri Ben-Shahar & Kyle D. Logue, Outsourcing Regulation: How Insurance Reduces Moral Hazard, 111 Mich. L. Rev. 197, 203 (2012); Kenneth S. Abraham & Pierre-André Chiappori, Classification Risk and Its Regulation, in Research Handbook on the Economics of Insurance Law 290, 292 (Daniel Schwarcz & Peter Siegelman eds., 2015) (Professors Abraham and Chiappori call this “diversification,” or the ability of the insurer to spread the risk across the pool of insureds.).
[29] Avraham, supra note 26, at 38; Leah Wortham, The Economics of Insurance Classification: The Sound of One Invisible Hand Clapping, 47 Ohio St. L.J. 835, 842–43 (1986) [hereinafter Wortham, Economics of Insurance Classification].
[30] Avraham, supra note 26, at 38; Wortham, Economics of Insurance Classification, supra note 29, at 843–44.
[31] Stone, supra note 5, at 290, 293.
[33] Geoff Werner & Claudine Modlin, Willis Towers Watson, Basic Ratemaking 13–14 (Casualty Actuarial Soc’y ed., 5th ed. 2016).
[36] For extensive discussion of how these processes work, see Ben-Shahar & Logue, supra note 28, at 205–08.
[37] See Werner & Modlin, supra note 33, at 15–16; Ben-Shahar & Logue, supra note 28, at 205–08.
[38] See Werner & Modlin, supra note 33, at 17.
[40] See Ben-Shahar & Logue, supra note 28, at 211–12, 224.
[41] See Sara Rosenbaum, O’Neill Inst. For Nat’l & Glob. Health Law, Insurance Discrimination on the Basis of Health Status: An Overview of Discrimination Practices, Federal Law and Federal Reform Options 6 (2009), http://www.rwjf.org/content/dam/farm/reports/reports/2009/rwjf36943 [https://perma.cc/UUN6-RF3R]. Rosenbaum provides a practical description of ways insurers can exercise discrimination. Id. at 6–7. Of course, there may be limits to this. For instance, health insurers must guarantee issue and renewability of insurance under the ACA. See infra notes 248–250.
[42] See Rosenbaum, supra note 41, at 6–7.
[44] See Murray, supra note 16, at 736–37; Rice, supra note 16, at 584–87.
[45] See Robert Works, Whatever’s FAIR—Adequacy, Equity, and the Underwriting Prerogative in Property Insurance Markets, 56 Neb. L. Rev. 445, 469, 471 (1977). For instance, insurers were historically reluctant to sell property insurance in some urban areas. Id. at 494. This was a matter the federal government tried to address with the Fair Access to Insurance Requirements (FAIR) program, a program where the government reinsured insurers who offered products to those communities. Id. at 446–48.
[46] See Peter Siegelman, Adverse Selection in Insurance Markets: An Exaggerated Threat, 113 Yale L.J. 1223, 1223 (2004); Tom Baker, Containing the Promise of Insurance: Adverse Selection and Risk Classification, 9 Conn. Ins. L.J. 371, 373, 375 (2003) [hereinafter Baker, Adverse Selection].
[47] See Tom Baker, The Shifting Terrain of Risk and Uncertainty on the Liability Insurance Field, 60 DePaul L. Rev. 521, 522 (2011) [hereinafter Baker, Risk and Uncertainty]; Wortham, Economics of Insurance Classification, supra note 29, at 844–45; Baker, Adverse Selection, supra note 46, at 373.
[48] See Siegelman, supra note 46, at 1223.
[49] See Baker, Adverse Selection, supra note 46, at 375; Wortham, Economics of Insurance Classification, supra note 29, at 844.
[50] For a summary on the evidence surrounding death spirals, see Siegelman, supra note 46, at 1254–58.
[51] See Abraham & Chiappori, supra note 28, at 312 (asking whether death spirals are plausible and discussing how “very risk averse people are unlikely to drop insurance coverage even if the price is high”). Other possibilities include that individuals may drop out of the market and self-insure or may demand of the government that rates be regulated. Ben-Shahar & Logue, supra note 28, at 204.
[52] See Siegelman, supra note 46, at 1253. Professor Baker argues that adverse selection is a “dual problem” that affects both sides of the insurance exchange; insureds who are low risk do not buy insurance, but insurers avoid those who are high risk. See Baker, Adverse Selection, supra note 46, at 374–79. Insurers’ response to this phenomenon ultimately makes them less useful in their original social function of spreading risk between parties. Id. at 376–79.
[53] Siegelman, supra note 46, at 1253.
[54] See Abraham & Chiappori, supra note 28, at 293; see also Siegelman, supra note 46, at 1253; Daniel Schwarcz, Regulating Consumer Demand in Insurance Markets, 3 Erasmus L. Rev. 23, 43 (2010).
[55] Baker, Adverse Selection, supra note 46, at 373.
[56] Abraham & Chiappori, supra note 28, at 296.
[59] Wortham, Economics of Insurance Classification, supra note 29, at 44; Avraham, supra note 26, at 37–42.
[60] Kimball, Reverse Sex Discrimination, supra note 27, at 103–08; Abraham, supra note 27, at 403–04.
[61] Avraham, supra note 26, at 44–49; Wortham, Economics of Insurance Classification, supra note 29, at 874–75.
[62] Leah Wortham, Insurance Classification: Too Important to Be Left to the Actuaries, 19 U. Mich. J.L. Reform 349 (1986) [hereinafter Wortham, Insurance Classification]; Abraham, supra note 27, at 429–31.
[63] Abraham, supra note 27, at 440–41.
[64] Some believe that insurance discounts can strongly induce safer conduct. “Insurers that can offer more coverage at lower premiums will attract customers, even when they require customers to modify their conduct in a costly way. As long as the standards imposed by the insurers are efficient, customers should be lured by the discounts.” Ben-Shahar & Logue, supra note 28, at 201–02.
[65] For instance, take this health insurance model proposed by Professors Baker and Siegelman. Young, healthy individuals are less likely to buy into the health insurance market because they frequently do not require costly medical services. This model proposes to entice them into the market by rewarding them if at the end of the year they did in fact not significantly use the insurance. Tom Baker & Peter Siegelman, Tontines for the Invincibles: Enticing Low Risks into the Health-Insurance Pool with an Idea from Insurance History and Behavioral Economics, 2010 Wis. L. Rev. 79 (2010).
[66] Abraham, supra note 27, at 413.
[67] Kimball, Reverse Sex Discrimination, supra note 27, at 135–36; Abraham, supra note 27, at 407.
[68] Kimball, Reverse Sex Discrimination, supra note 27, at 135–36; Abraham, supra note 27, at 408.
[69] This may be done through shifting costly risks onto public programs. See, e.g., Ruth E. Kim & Kimball R. McMullin, AIDS and the Insurance Industry: An Evolving Resolution of Conflicting Interests and Rights, 7 St. Louis. U. Pub. L. Rev. 155 (1988) (advocating for permissible AIDS testing in private insurance while establishing a risk pool arrangement by the states or an expansion of Medicaid to cover those who become ineligible for private insurance). Alternatively, private insurers may be compensated by the government for insuring certain risks. Works, supra note 45, at 446–47.
[70] Am. Acad. of Actuaries Risk Classification Work Grp., On Risk Classification 4–5 (2011), https://www.actuary.org/sites/default/files/files/publications/RCWG_RiskMonograph_Nov2011.pdf [https://perma.cc/2MCT-AUNT] [hereinafter Risk Classification Work Grp.].
[74] Id. at 5. For instance, life insurance variations in pricing according to sex first began in response to mortality tables established by the U.S. government. Another example, insurers began looking at tobacco use in life insurance after reports by the U.S. Surgeon General about risks to health. Charles L. Trowbridge, Actuarial Educ. & Research Fund, Fundamental Concepts of Actuarial Science 56–57 (1989), https://www.actuariayfinanzas.net/images/sampledata/Conceptos-fundamentales-de-la-Ciencia-Actuarial.pdf [https://perma.cc/AWX2-H764].
[75] Abraham & Chiappori, supra note 28, at 291.
[77] This model is discussed further infra Section III.B.
[78] This model is also discussed infra Section III.A.
[79] Wortham, Economics of Insurance Classification, supra note 29, at 882.
[82] Of course, if the conduct is by governor actors, the usual due process and equal protection laws would apply. Likewise, the states frequently have equal rights amendments or state constitutions that prohibit some forms of discrimination by state actors, for instance discrimination based on race or sex. However, much of the insurance market is private and out of reach of these laws. Jill Gaulding, Note, Race, Sex, and Genetic Discrimination in Insurance: What’s Fair?, 80 Cornell L. Rev. 1646 (1995).
[83] Wortham, Economics of Insurance Classification, supra note 29, at 883.
[84] Mary L. Heen, From Coverture to Contract: Engendering Insurance on Lives, 23 Yale J.L. & Feminism 335, 341–42 (2011) (exploring gender and marital discrimination in life insurance).
[86] Insurance regulation is typically left to the states in accordance with the McCarran-Ferguson Act. 15 U.S.C. §§ 1011–1015 (2018).
[87] Ronen Avraham, Kyle D. Logue & Daniel Schwarcz, Understanding Insurance Antidiscrimination Laws, 87 S. Cal. L. Rev. 195, 229 nn.119–122 (2014). The authors provide a breakdown of the types of state laws governing insurance. Thirteen states generally forbid unfair discrimination in insurance law. Id. at 232 n.124. Other states forbid unfair discrimination in a particular line of insurance, a particular classification, or a combination thereof. Id. at 232–33.
[88] Wortham, Insurance Classification, supra note 62, at 372.
[90] See generally Avraham, Logue & Schwarcz, supra note 87.
[91] Id. Researchers studied health, life, disability, auto, and property/casualty insurance for protections from discrimination on the basis of race/national origin, religion, gender, age, genetics, sexual orientation, credit score, and zip code. Id.
[93] Id. Only twelve states explicitly prohibit use of race and national origin for all five types of insurance; of those twelve, ten states do so for religion as well. Five states expressly permit gender discrimination, while a number of other states offer general or weak protections. For instance, all states but one allow gender discrimination in life insurance, and a surprising eighteen permit its use in health care. Six states ban use of sexual orientation for all insurance products. One state bans the use of genetics in each insurance product. A substantial majority of states expressly permit consideration of age in life and health insurance, and a handful of states also expressly allow its use in car insurance and disability insurance. Few states expressly prohibit the use of zip codes or credit scores in insurance and, notably, some states expressly permit zip code use by health insurers, and credit score use by life, disability, and health insurers. Id.
[94] Tom Baker, The Rise of Insurance Runoff (June 7, 2019) [hereinafter Baker, Insurance Runoff] (unpublished manuscript) (on file with author).
[96] Id. at 7–8 (internal footnotes omitted).
[97] Richard V. Ericson & Aaron Doyle, Uncertain Business: Risk, Insurance, and the Limits of Knowledge 5 (2004).
[98] Baker, Insurance Runoff, supra note 94.
[99] Wortham, Insurance Classification, supra note 62.
[100] Regina Austin, The Insurance Classification Controversy, 131 U. Pa. L. Rev. 517, 534 (1983).
[101] See Wortham, Insurance Classification, supra note 62; Austin, supra note 100.
[102] Erving Goffman, Stigma: Notes on the Management of Spoiled Identity (1963). Other fields study similar types of matters; for instance, in psychology, this process is considered in the field of prejudice. Scholars suggest that prejudice and stigma have overlapping and consistent themes. Jo Phelan, Bruce G. Link & John F. Dovidio, Stigma and Prejudice: One Animal or Two?, 67 Soc. Sci. & Med. 358 (2008).
[103] Goffman, supra note 102, at 2–3.
[106] See Mark L. Hatzenbuehler, Jo C. Phelan & Bruce G. Link, Stigma as a Fundamental Cause of Population Health Inequalities, 103 Am. J. Pub. Health 813 (2013) (providing an overview of how stigma occurs across various life domains for certain classes of people).
[107] Goffman, supra note 102, at 2.
[108] Michael D. Mink et al., Stress, Stigma, and Sexual Minority Status: The Intersectional Ecology Model of LGBTQ Health, 26 J. Gay & Lesbian Soc. Servs. 502 (2014) (observing how Native American cultures view gender nonconformity as mystical and laden with healing powers).
[109] Link & Phelan, Conceptualizing Stigma, supra note 6, at 367.
[116] Link and Phelan select the word “label” with care to reflect the fact that the demarcation is something society puts on the person, rather than something inherent in or part of that person. Id. at 368. The general idea is that we are prone to considering these distinctions as inherent and caused by nature, rather than socially created and structural in nature. The idea is also very similar to availability heuristics. Others have discussed how this is a natural, though controllable phenomenon. For example, in Timur Kuran & Cass R. Sunstein, Availability Cascades and Risk Regulation, 51 Stan. L. Rev. 683 (1999), the authors write about availability heuristics, that is, mental shortcuts that we all take where we associate a given thing with another in our mind. These heuristics can cascade and be shared across communities and can, in some instances, lead to availability errors, which are widespread and mistaken beliefs about things that are not accurate. This is similar to the stigma process by which we stereotype and then extrapolate that all parties in a given group share common characteristics.
[117] Link & Phelan, Conceptualizing Stigma, supra note 6, at 369–70.
[118] See Larry Alexander, What Makes Wrongful Discrimination Wrong? Biases, Preferences, Stereotypes, and Proxies, 141 U. Pa. L. Rev. 149 (1992). Alexander theorizes that stereotyping is harmful when it is based on inaccurate information, when it leads us to have confirmation biases, and when it tends to lead to behavior changes that may go against public policy. Id. at 167–73. “For example, if women are allowed to drink at an earlier age than men because they are generally more responsible drinkers, men might be reinforced in the attitudes that foster their relative irresponsibility.” Id. at 170; see also Charles R. Lawrence III, The Id, the Ego, and Equal Protection: Reckoning with Unconscious Racism, 39 Stan. L. Rev. 317 (1987) (emphasizing that implicit bias, stereotyping, and the psychology behind racism all must be understood and taken into account in antidiscrimination laws). But see Frederick Schauer, Profiles, Probabilities, and Stereotypes (2003) (arguing that stereotyping can be problematic, but we paint with too broad of a brush when we consider all stereotyping to be harmful).
[119] Link & Phelan, Conceptualizing Stigma, supra note 6, at 368–70.
[123] Hellman stresses context in assessing whether a particular action is demeaning and thus discriminatory. Hellman, supra note 23. For instance, demanding that young black children go sit in the back of a bus would have a uniquely more demeaning context in the United States than, perhaps, in other societies, she suggests. Id. at 26–28.
[124] Bruce G. Link & Jo Phelan, Stigma Power, 103 Soc. Sci. & Med. 24 (2014) [hereinafter Link & Phelan, Stigma Power].
[125] Id. Interestingly, this is akin to the basic function of insurance for Foucauldian scholars. François Ewald, Insurance and Risk, in The Foucault Effect: Studies in Governmentality 197, 201–02 (Graham Burchell et al. eds., 1991).
[126] Link & Phelan, Stigma Power, supra note 124; Link & Phelan, Conceptualizing Stigma, supra note 6, at 373.
[127] Link & Phelan, Conceptualizing Stigma, supra note 6, at 370–71.
[133] Id. at 373–74. “When you stigmatize someone, your aim is not merely to respond to a trait that you find undesirable but to mark that person in such a way that they find the trait undesirable.” Andrew M. Courtwright, Justice, Stigma, & the New Epidemiology of Health Disparities, 23 Bioethics 90, 91 (2009).
[134] See Link & Phelan, Conceptualizing Stigma, supra note 6, at 376; see also Scott Burris, Disease Stigma in U.S. Public Health Law, 30 J.L. Med. & Ethics 179 (2002).
[135] See Link & Phelan, Conceptualizing Stigma, supra note 6. A very recent example is efforts by Howard Schultz to argue that “billionaire[s]” are stigmatized and the label should be renamed to something like “people of means.” Arwa Mahdawi, Don’t Call Howard Schultz a Billionaire. He’s a ‘Person of Means,’ Guardian (Feb. 6, 2019), https://www.theguardian.com/commentisfree/2019/feb/06/dont-call-howard-schultz-billionaire-wealth-washing [https://perma.cc/F69X-4QWK]. For comical alternatives, see Alexandra Petrified (@petridishes), Twitter (Feb. 5, 2019, 9:14 AM), https://twitter.com/petridishes/status/1092833764030533632?lang=en [https://perma.cc/G5YT-DVUL] (suggesting, facetiously, “butler-adjacent,” “fiscally tinged,” or “silver spoon havers”). Any claim that a billionaire suffers stigma from name-calling would be uncompelling because of subordination. Certainly, we label billionaires as such and we may stereotype them, we may even discriminate against them. But they will never fully occupy stigmatized categories in society because they have tremendous wealth and control. The public’s dislike will never result in full subordination of this class of people.
[136] Link & Phelan, Conceptualizing Stigma, supra note 6, at 376.
[138] See Paul Brest, Foreword, In Defense of the Antidiscrimination Principle, 90 Harv. L. Rev. 1, 10 (1976).
[139] Stone, supra note 5, at 314.
[140] Austin, supra note 100, at 534.
[141] See Kimball, Insurance Regulation, supra note 24, at 496–97; Abraham, supra note 27, at 405.
[142] See Kimball, Insurance Regulation, supra note 24; Abraham, supra note 27.
[143] See Wortham, Economics of Insurance Classification, supra note 29, at 837.
[144] For an example of litigation related to such a cap, see Doe v. Mutual of Omaha Insurance Co., 179 F.3d 557, 558–59 (7th Cir. 1999).
[146] In a study of insurers in the early 90s in Texas, insurers frequently sought to determine if an enrollee had HIV or AIDS and would frequently refuse to cover groups as large as seventy-five if they included even one individual who was HIV positive or had AIDS. Id.
[147] Georgetown Univ. Inst. for Health Care Research & Policy & K.A. Thomas & Assocs., Henry J. Kaiser Family Found., How Accessible Is Individual Health Insurance for Consumers in Less-than-Perfect Health? (2001), https://www.kff.org/wp-content/uploads/2013/01/how-accessible-is-individual-health-insurance-for-consumer-in-less-than-perfect-health-report.pdf [https://perma.cc/RK2J-94YF] [hereinafter Kaiser Fam. Found.].
[149] Id. at 561. Judge Posner held that insurers could refuse to cover AIDS-related benefits for an insurer because
[t]he common sense of the statute is that the content of the goods or services offered by a place of public accommodation is not regulated. A camera store may not refuse to sell cameras to a disabled person, but it is not required to stock cameras specially designed for such persons.
Id. at 560. Posner’s opinion falls in line with earlier cases alleging disability discrimination in insurance. For instance, Alexander v. Choate, where the Supreme Court held that disability laws can only reach access to benefits but not their content; thus insurers need not create health plans that take into account and address disabled people’s health care. 469 U.S. 287, 304 (1985). Choate dealt with a challenge to Medicaid cuts under the Rehabilitation Act of 1973. Id.
[150] For instance, consider data from 1991 when HIV and AIDS were less well managed:
[w]hile the data clearly demonstrates that lifetime treatment costs of persons with AIDS are substantial, they are not unlike treatment costs incurred by persons with other serious diseases typically covered by health insurance policies. For example, lifetime treatment costs for a person diagnosed with AIDS in 1991 are about the same as the costs associated with an episode of care for a bone marrow transplant in the same year—about $150,000. The average lifetime costs of a person with AIDS somewhat less than those for a person receiving a kidney transplant ($90,000 per transplant in 1991, with total lifetime costs two to three times higher).
Robert A. Padgug, Gerald M. Oppenheimer & Jon Eisenhandler, AIDS and Private Health Insurance: A Crisis of Risk Sharing, 3 Cornell J.L. & Pub. Pol’y 55, 62 (1993); see also Samuel A. Marcosson, Who Is “Us” and Who Is “Them”—Common Threads and the Discriminatory Cut-Off of Health Care Benefits for AIDS Under ERISA and the Americans with Disabilities Act, 44 Am. U. L. Rev. 361, 414–17 (1994).
[151] HIV Stigma and Discrimination, Avert, https://www.avert.org/professionals/hiv-social-issues/stigma-discrimination [https://perma.cc/DUS4-RAH8] (last updated Oct. 10, 2019) [hereinafter HIV Stigma]; Anish P. Mahajan et al., Stigma in the HIV/AIDS Epidemic: A Review of the Literature and Recommendations for the Way Forward, 22 AIDS (Supp. 2) S67 (2008).
[152] HIV Stigma, supra note 151.
[153] Risk Classification Work Grp., supra note 70, at 38, 58–59; Trowbridge, supra note 74, at 60–61.
[154] Risk Classification Work Grp., supra note 70, at 5; Trowbridge, supra note 74, at 21.
[155] Works, supra note 45, at 460 (“Regardless of their numbers and the skill with which they are chosen, the selection of class-defining risk indicia produces a classification scheme that is inevitably imperfect and value-laden.”).
[156] Heen, supra note 84, at 336–37.
[157] Risk Classification Work Grp., supra note 70, at 58–59; Trowbridge, supra note 74, at 21.
[158] Risk Classification Work Grp., supra note 70, at 5; Trowbridge, supra note 74, at 21.
[159] Risk Classification Work Grp., supra note 70, at 6.
[160] Barbara Feder Ostrov, Major Insurer Says It Will Offer Individual Life Insurance Coverage to People with HIV, Kaiser Health News (Dec. 1, 2015), https://khn.org/news/major-insurer-says-it-will-offer-individual-life-insurance-coverage-to-people-with-hiv [https://perma.cc/JS2A-GY4S].
[161] Antiretrovirals, developed in the mid-90s, have made it possible for people infected with HIV to live dramatically longer and better-quality lives. James Myhre & Dennis Sifris, Antiretrovirals Overview: How Combination Therapy Renders HIV Powerless, Verywell Health, https://www.verywellhealth.com/what-are-antiretrovirals-and-how-do-they-work-49639 [https://perma.cc/7KCK-YNK8] (last updated Sept. 23, 2019).
[162] Plaintiff John Doe’s Memorandum in Support of Motion for Summary Judgment and Opposition to Defendant’s Motion for Summary Judgment at 2, Doe v. Mut. of Omaha Ins. Co., No. 1:16-cv-11381-GAO (D. Mass July 18, 2018) [hereinafter Plaintiff John Doe’s Memorandum in Support of Motion for Summary Judgment]; Donald G. McNeil Jr., He Took a Drug to Prevent AIDS. Then He Couldn’t Get Disability Insurance., N.Y. Times (Feb. 12, 2018), https://www.nytimes.com/2018/02/12/health/truvada-hiv-insurance.html [https://perma.cc/8E6G-2MXQ].
[163] AIDS has taken the lives of almost 675,000 Americans since its origins in the 1980s. CDC Fact Sheet: Today’s HIV/AIDS Epidemic, CDC (Aug. 2016), https://www.cdc.gov/nchhstp/newsroom/docs/factsheets/todaysepidemic-508.pdf [https://perma.cc/6PW9-VMKC]. While new transmissions have decreased, anywhere from 35,000–40,000 new cases of AIDS are still diagnosed each year. U.S. Statistics, HIV.gov, https://www.hiv.gov/hiv-basics/overview/data-and-trends/statistics [https://perma.cc/B7JT-PARS] (last updated Mar. 13, 2019). The suit alleged violation of the Massachusetts Public Accommodation statute which forbids discrimination on the basis of sexual orientation in public accommodations. HIV risk is not specific to the gay community and PrEP is indicated for many other groups of people, though gay men make up eighty percent of PrEP users. The claim alleges disability discrimination because while
Mutual regarded [Doe] as disabled in two different ways. First, the testimony . . . reveals that for purposes of assessing Doe’s risk for needing long-term care services, Mutual treated Doe as if he did, in fact, have HIV. Second, the undisputed facts demonstrate that Mutual denied Doe long-term care insurance because of its concern or fear that he would contract HIV in the future.
Plaintiff John Doe’s Memorandum in Support of Motion for Summary Judgment, supra note 162; see also Letter from Lisette Johnson, Bureau Chief, Health Bureau, to All Insurers Authorized to Write Accident and Health Ins. in N.Y State et al. (Dec. 1, 2017) [hereinafter Letter from Lisette Johnson], https://www.dfs.ny.gov/insurance/circltr/2017/cl2017_21.htm [https://perma.cc/44RR-AY6V].
[164] PrEP, CDC, https://www.cdc.gov/hiv/basics/prep.html [https://perma.cc/A56G-5FVH].
[167] Id. Risk in sexual intercourse is even lower when condoms are used. Id.
[169] Plaintiff John Doe’s Memorandum in Support of Motion for Summary Judgment , supra note 162, at 2.
[170] Doe v. Mutual of Omaha Insurance Company, GLAD, https://www.glad.org/cases/doe-v-mutual-of-omaha [https://perma.cc/86L9-THGN].
[171] Wortham, Insurance Classification, supra note 62, at 354–56; Austin supra note 100, at 534–35.
[172] See Austin, supra note 100, at 552.
[174] Heen, supra note 84, at 339–40.
[175] Mary L. Heen, Ending Jim Crow Life Insurance Rates, 4 Nw. J.L. & Soc. Pol’y 360 (2009).
[176] Stone, supra note 5, at 290–97 (citing Edward Hauschild, The Accident and Health Underwriter’s Guide 83 (1931)).
[177] Id. at 356 n.28 (demonstrating how life insurance policies based on occupation closely mirrored social class categories).
[178] Yet traditional stereotypes of men suggest that men are “meant to have erections and sexual pleasure,” and therefore Viagra merely aids what nature intended. On the other hand, traditional stereotypes of women say that women are intended to get pregnant, become mothers, and only tolerate sex. Thus, the traditional stereotype of women discourages use of “unnatural” contraception and even abortion. Rather than bind women to outdated and oppressive standards, insurance companies should recognize that both Viagra and contraceptives are medically necessary to the well-being and sexual health of both men and women. Lisa A. Hayden, Gender Discrimination Within the Reproductive Health Care System: Viagra v. Birth Control, 13 J.L. & Health 171, 172 n.5 (1999) (citing Janet Benshoof, By Covering Viagra, Insurers Show that Men’s Sexual “Well-Being” Is Still More Vital than Women’s, Chi. Trib. (June 7, 1998), https://www.chicagotribune.com/news/ct-xpm-1998-06-07-9806070369-story.html [https://perma.cc/J7CW-U6R9]).
[179] Wortham, Insurance Classification, supra note 62. “For example, a 1930 pocket manual for agents of the Northwest Union Life Insurance Company (1930) at 9 begins a section titled ‘Uninsurable Risks’ with the following statement: ‘Negroes, Chinese, Japanese, Mexicans and more than one-fourth blood Indians will not be considered.’” Stone, supra note 5, at 296.
[180] Elizabeth Pendo, Shifting the Conversation: Disability, Disparities and Health Care Reform, 6 Fla. Int’l U. L. Rev. 87 (2010) (describing differences in access to care for persons with disabilities as partly due to stereotypes).
[181] Austin, supra note 100, at 547 (“He is the sum of the many roles he plays as a result of being a member of many status groups. To an insurance company, the same individual may be an adult, a female, a divorcee, a parent, a lover, an executive, a debtor, a homeowner, a citizen, an urbanite, a commuter, a teetotaler, a lawbreaker, and a klutz. She is not a plenary, monolithic person. The company does not know her; it knows only the roles she plays. Although the multiplicity of roles may cause the individual to suffer normative conflict and uncertainty, role or status inconsistency does not impede insurers. In automobile insurance pricing, for example, each relevant role is assigned a numerical value; the individual insured is literally the multiplicative product of his parts.”).
[182] This is the very heart of the “fairness” aspects of actuarial fairness. Wortham, Insurance Classification, supra note 62.
[183] Anne C. Cicero, Note, Strategies for the Elimination of Sex Discrimination in Private Insurance, 20 Harv. C.R.-C.L. L. Rev. 211, 226 (1985) (“[T]he use of sex classifications perpetuates cultural stereotypes which may in turn contribute to gender role assignments in society.”); Deborah Hellman, What Makes Genetic Discrimination Exceptional?, 29 Am. J.L. & Med. 77, 79 (2003) [hereinafter Hellman, Genetic Discrimination] (describing how genetic discrimination can be considered wrong when it creates expressive harms).
[184] Cicero, supra note 183, at 211–12; Hellman, supra note 23, at 7.
[185] See Kenneth L. Karst, Why Equality Matters, 17 Ga. L. Rev. 245, 285 n.180 (1983) (“But the very concepts of stigma and stereotype are inseparable from the stigmatized or stereotyped individuals’ group membership; the victims are dehumanized precisely because they are denied their individuality and treated according to race, sex, etc.”).
[186] City of L.A., Dep’t of Water & Power v. Manhart, 435 U.S. 702, 709 (1978). The case arose after female workers were required to make larger contributions to pensions than male employees on the actuarially-based theory that women, on average, live longer than men and across their lifetime receive more pay out by the pensioner. In a 6–2 decision authored by Justice Stevens, the Court held that these practices violate Title VII based mainly on a textual reading of the statute. Because Title VII forbids individual discrimination that is sex-based, generalizations about how much most women cost the pension sorts based on class and does not treat women as individuals. Id. A few years later, in Arizona Governing Committee for Tax Deferred Annuity & Deferred Compensation Plans v. Norris, 463 U.S. 1073 (1983), the Supreme Court struck down a state voluntary pension plan that allocated less funds to women over the month than to men; this was also held to violate Title VII for similar reasoning.
[187] Manhart, 435 U.S. 702; Norris, 463 U.S. 1073 (a few years after Manhart).
[188] Katy Chi-Wen Li, The Private Insurance Industry’s Tactics Against Suspected Homosexuals: Redlining Based on Occupation, Residence and Marital Status, 22 Am. J.L. & Med. 477 (1996).
[189] Id. at 479 (“[P]rivate insurance companies have also developed and practiced methods to distinguish those who, because of their geographical location of residence, marital status, occupation or beneficiary selection, presumably have a higher risk for contracting HIV” despite the fact that “the majority of gay and bisexual men will not develop AIDS. Reports show that in many foreign countries, AIDS is largely a heterosexual phenomenon. In addition, the percentage of AIDS cases resulting from transmission through heterosexual contact is rising dramatically. Furthermore, groups such as racial minorities and intravenous drug users also constitute a large proportion of AIDS cases.” (internal footnotes omitted)).
[191] See Padgug, Oppenheimer & Eisenhandler, supra note 150, at 62.
[192] For instance, heterosexual individuals who engage in unprotected sex also have a risk of acquiring HIV/AIDS, as do people who use intravenous drugs.
[193] Plaintiff John Doe’s Memorandum in Support of Motion for Summary Judgment, supra note 162.
[196] Link & Phelan, Conceptualizing Stigma, supra note 6, at 370.
[197] Stone, supra note 5; Crossley, supra note 13, at 85 (observing how actuarial fairness creates an us versus them scenario); Schauer, supra note 118, at 34–39 (recognizing that insurance practices could reinforce “otherness”).
[198] See Abraham & Chiappori, supra note 28, at 290.
[199] Wortham, Insurance Classification, supra note 62.
[201] Id. at 298–99. Foucauldian scholars also emphasize that we have a choice in how we develop the institution of insurance, and that this reflects our broader societal priorities. Ewald, supra note 125, at 198 (“The particular form insurance technology takes in a given institution at a given moment depends on an insurantial imaginary: that is to say, on the ways in which, in a given social context, profitable, useful and necessary uses can be found for insurance technology.”).
[204] For example, House Representative John Shimkus (R-Ill.) complained that premiums were skyrocketing in his state because of mandates from the ACA, asking “What about men having to purchase prenatal care? . . . [S]hould they?” Heidi Stevens, A Few Reasons Why Men Should Pay for Prenatal Care, Chi. Trib. (Mar. 10, 2017, 8:33 AM), http://www.chicagotribune.com/lifestyles/stevens/ct-why-men-should-cover-prenatal-care-balancing-0310-20170310-column.html [https://perma.cc/Q3KB-MZBR].
[205] Risk Classification Work Grp., supra note 70, at 58; Trowbridge, supra note 74, at 60.
[206] See Hellman, supra note 15, at 355–56.
[207] Wortham, Insurance Classification, supra note 62, at 374–75; Austin, supra note 100, at 552.
[209] Link & Phelan, Conceptualizing Stigma, supra note 6.
[210] For instance, Lambda Legal describes discrimination against these individuals in the legal system, employment, health care, immigration, among other areas. Defending People Living with HIV, Lambda Legal, https://www.lambdalegal.org/issues/hiv [https://perma.cc/WC38-ZV34].
[211] Aviam Soifer, Disabling the ADA: Essences, Better Angels, and Unprincipled Neutrality Claims, 44 Wm. & Mary L. Rev. 1285 (2003).
[212] Doe v. Mut. of Omaha Ins. Co., 179 F.3d 557 (7th Cir. 1999). The cost of antiretrovirals, without insurance, is prohibitively expensive for many. U.S. Dep’t of Health & Hum. Servs., Guidelines for the Use of Antiretroviral Agents in Adults and Adolescents with HIV, AIDSinfo, https://aidsinfo.nih.gov/guidelines/html/1/adult-and-adolescent-arv/459/cost-considerations-and-antiretroviral-therapy [https://perma.cc/A6GZ-WBM6] (last updated November 26, 2018).
[213] The work of Professor Mary Heen particularly amplifies the issue that some groups (including racial minorities and women) have not been able to enjoy the same economic security for themselves and their families that insurance affords others. For examples, see Heen, supra note 84 (exploring gender and marital discrimination in life insurance); Heen, supra note 175 (discussing race discrimination in housing).
[214] Professor Mary Crossley used this as one reason why to eliminate health status-based discrimination in the context of health insurance pre-ACA. Crossley, supra note 13. Baker provides another example:
[T]he children of a parent refused life or disability insurance maintain a more tenuous grasp on their position as a result of the insurers having classified their parent as a high risk. Should the parent die or become disabled, the children’s resulting loss of social position derives not only from the death or disability, but also from the insurance risk classification.
Baker, Adverse Selection, supra note 46, at 8.
[215] Larson, supra note 16; Julia Angwin et al., Minority Neighborhoods Pay Higher Car Insurance Premiums than White Areas with the Same Risk, ProPublica (Apr. 5, 2017), https://www.propublica.org/article/minority-neighborhoods-higher-car-insurance-premiums-white-areas-same-risk [https://perma.cc/SNF5-72GP].
[216] McNeil, supra note 162. In one instance, a young physician was denied disability insurance, first because he took PrEP to prevent hospital acquisition of HIV and later because he continued to take it to prevent sexually-acquired HIV. He was denied a lifetime disability policy, but the insurer later gave him a shorter-term policy. In the meantime, he actually stopped taking PrEP in order to be eligible for the more comprehensive insurance policy. He eventually found another insurer willing to insure him while on PrEP. Id.
[217] Link & Phelan, Conceptualizing Stigma, supra note 6.
[218] Stone, supra note 5, at 314.
[220] Avraham, Logue & Schwarcz, supra note 87.
[221] Nondiscrimination in Health Programs and Activities, 81 Fed. Reg. 31376 (May 18, 2016) (to be codified at 45 C.F.R. pt. 92). Section 1557 of the ACA forbids discrimination by some health care entities, including providers and insurers based on sex, but regulators under the Obama Administration declined to include sexual orientation in the definition of sex discrimination. Instead, they opted to defer to the courts in determining what constitutes sex discrimination in other contexts like Title VII employment discrimination. The Obama Administration did define sex discrimination more expansively in other ways, by including discrimination based on gender identity. Id. Under the Trump Administration, a new rule has been proposed declining to provide any specificity for what constitutes sex discrimination, allowing for the greater possibility of such discrimination in health care. 84 Fed. Reg. 27846, 27857 (June 14, 2019).
[222] Avraham, Logue & Schwarcz supra note 87, at 252.
[223] Twenty-five states and the District of Columbia currently address sexual orientation. State Public Accommodation Laws, Nat’l Conference of State Legislatures (Apr. 8, 2019), http://www.ncsl.org/research/civil-and-criminal-justice/state-public-accommodation-laws.aspx [https://perma.cc/BEH9-H7TY].
[224] Link & Phelan, Conceptualizing Stigma, supra note 6.
[225] Wortham, Insurance Classification, supra note 62; Austin supra note 100, at 534.
[226] Link & Phelan, Conceptualizing Stigma, supra note 6, at 368–69.
[227] Jonathan Simon, The Ideological Effects of Actuarial Practices, 22 Law & Soc’y Rev. 771 (1988) (exploring how actuarial fairness pushes a narrative that insurance discrimination is different by offering neutral justifications of objectivity and science); Austin, supra note 100. Actuarial fairness, by situating discrimination in terms of objectivity and ideas of equity, has “neutralize[d] the moral charge” carried by unfair classifications and has led to less social mobilization and political action around change. Simon, supra, at 794. Thus, structural forms of discrimination can continue to play out, in part, because they are masked by visions that actuarial fairness, even when it is sometimes implicating otherwise protected groups, is fair and justified.
[228] It raises similar concerns to those scholars who worry about historical disadvantage in insurance. See, e.g., Heen, supra note 84, at 338. See also Brest, supra note 138, at 2, for the argument that a broader aim of antidiscrimination laws is to redress historical disadvantage more broadly.
[229] Stone, supra note 5, at 289.
[230] Id. at 287–91; Deborah A. Stone, Beyond Moral Hazard: Insurance as Moral Opportunity, 6 Conn. Ins. L.J. 11, 13 (1999) [hereinafter Stone, Beyond Moral Hazard]; John V. Jacobi, The Ends of Health Insurance, 30 U.C. Davis L. Rev. 311, 312–19 (1997); Mariner, supra note 8, at 200; Christopher C. French, Insuring Landslides: America’s Uninsured Natural Catastrophes, 17 Nev. L.J. 63 (2016); Christopher C. French, Understanding Insurance Policies as Noncontracts: An Alternative Approach to Drafting and Construing These Unique Financial Instruments, 89 Temp. L. Rev. 535 (2017). Social security and disability benefits may also be seen as a form of mutual aid.
[231] See Austin, supra note 100, at 527–28; see also Stone, supra note 5; Wortham, Insurance Classification, supra note 62, at 400 (“Consideration of the role of insurance in society, as well as the public choices that have shaped that role, leads one not only to a concern for perceived legitimacy of classifications but also to the more important issue of availability of coverage: Can people buy the insurance they need?”).
[232] Ewald, supra note 125, at 208. Ewald, in his analysis of insurance, suggests that insurance may even shape one’s understanding of risk from an act of God outside of our control to one that is society’s obligation to meet and address. Id. (“With insurance and its philosophy, one enters a universe where the ills that befall us lose their old providential meaning: a world without God, a laicized world where ‘society’ becomes the general arbiter answerable for the causes of our destiny.”).
[233] According to Stone, “[i]nsurance is a social institution that particularly invites moral contemplation about questions of suffering, compassion, and responsibility.” Stone, Beyond Moral Hazard, supra note 230, at 16 (discussing how Foucauldian scholars see insurance as norm-setting, as creating expectations for how we act in a given society and how we enforce those expected behaviors).
[234] See, e.g., Schauer, supra note 118, at 34–39 (using denial of health insurance to a group of women with genetic predisposition to breast cancer as an example of unjust activity because of the inherent importance of insurance to that group).
[235] Professor Baker argues that reverse adverse selection ultimately makes insurers less useful in their original social function of spreading risk between parties. Baker, Adverse Selection, supra note 46; see generally Richard A. Booth, The Economic Case for Gender-Neutral Life Insurance, 13 Conn. Ins. L.J. 267 (2007) (arguing against gender ratings on the basis of efficiency).
[236] Alfred E. Kahn, The Economics of Regulation: Principles and Institutions 188 (Mass. Inst. Of Tech. 1998) (1970) (noting that all classifications “involve complex distributional effects; all will be economically imperfect; and all will inevitably raise noneconomic questions about what is fair, politically acceptable, and so on”).
[237] Crossley, supra note 13 (detailing widespread health status discrimination across the market).
[238] Kaiser Fam. Found., supra note 147, at ii.
[240] Karen A. Clifford & Russel P. Inculano, AIDS and Insurance: The Rationale for AIDS-Related Testing, 100 Harv. L. Rev. 1806 (1987); Benjamin Schatz, The AIDS Insurance Crisis: Underwriting or Overreaching?, 100 Harv. L. Rev. 1782, 1784–86 (1987); Sandra Elizabeth Stone, HIV Testing and Insurance Applicants: Exploring Constitutional Alternatives to Statutory Protections, 19 Hastings Const. L.Q. 1163, 1170 (1992).
[241] The uninsured rate has declined to around ten percent post-ACA. Rachel Garfield, Kendal Orgera & Anthony Damico, The Uninsured and the ACA: A Primer—Key Facts About Health Insurance and the Uninsured Amidst Changes to the Affordable Care Act, Kaiser Fam. Found. (Jan. 25, 2019), https://www.kff.org/report-section/the-uninsured-and-the-aca-a-primer-key-facts-about-health-insurance-and-the-uninsured-amidst-changes-to-the-affordable-care-act-how-many-people-are-uninsured [https://perma.cc/N8XQ-HDKG].
[242] Rosenbaum, supra note 41, at 9–15.
[243] GINA specifically states that “[a] group health plan, and a health insurance issuer offering health insurance coverage in connection with a group health plan, shall not request, require, or purchase genetic information for underwriting purposes.” 42 U.S.C. § 300gg-4(d)(1) (2018). The rationale for this protection was stated in the findings:
Congress has collected substantial evidence that the American public and the medical community find the existing patchwork of State and Federal laws to be confusing and inadequate to protect them from discrimination. . . . [E]stablishing a national and uniform basic standard is necessary to fully . . . allay [the public’s] concerns about the potential for discrimination, thereby allowing individuals to take advantage of genetic testing, technologies, research, and new therapies.
Genetic Information Nondiscrimination Act of 2008, 29 C.F.R. § 1635. GINA also forbids employers from using genetic information in employment decisions. Id. § 1635.1.
[244] Paul Wellstone and Pete Domenici Mental Health Parity and Addiction Equity Act of 2008 (MHPAEA), Pub. L. No. 110-343, 122 Stat. 3765. Specifically, the law requires that limits imposed on mental health benefits must be equally imposed on other covered benefits, though it does not require insurers to cover any such services. For instance, if such an insurer restricted the number of covered days in a hospital for a mental health condition, they would need to have the same limitation for cardiac care or another inpatient procedure. For critiques of how this law fails to adequately protect the interests of patients with mental health diagnoses, see John V. Jacobi, Mental Illness: Access and Freedom, 16 Hous. J. Health L. & Pol’y 37 (2016) [hereinafter Jacobi, Mental Illness]. For a discussion of the broader need to end inequality in mental health benefits, see Sonja B. Starr, Simple Fairness: Ending Discrimination in Health Insurance Coverage of Addiction Treatment, 111 Yale L.J. 2321 (2002). The MHPAEA is largely overshadowed by the ACA which, through its essential health benefits provision, requires insurers to cover mental health and addiction services and extends these requirements to small group and individual insurers as well. ACA § 1302(b)(1), 42 U.S.C. § 18022 (2018).
[245] Title VII forbids sex discrimination in employer benefits by employers with more than fifteen employees. Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e (2018). Title VII was amended by the Pregnancy Discrimination Act of 1978, Pub. L. No. 95-555, 92 Stat. 2076, which also prohibits employer health plans from discriminating on the basis of pregnancy in their benefits. These legal challenges often ask whether benefits are comparable between the sexes, for instance, past failures to cover contraception in light of coverage of Viagra. Erickson v. Bartell Drug Co., 141 F. Supp. 2d 1266, 1271–72 (W.D. Wash. 2001) (“[W]hen an employer decides to offer a prescription plan . . . it has a legal obligation to make sure that the resulting plan does not discriminate based on sex-based characteristics and that it provides equally comprehensive coverage for both sexes.” “The special or increased healthcare needs associated with a woman’s unique sex-based characteristics must be met to the same extent, and on the same terms, as other healthcare needs. Even if one were to assume that Bartell’s prescription plan was not the result of intentional discrimination, the exclusion of women-only benefits from a generally comprehensive prescription plan is sex discrimination under Title VII.”). For commentary on this case, see Melissa Cole, Beyond Sex Discrimination: Why Employers Discriminate Against Women with Disabilities When Their Employee Health Plans Exclude Contraceptives from Prescription Coverage, 43 Ariz. L. Rev. 501 (2001). For broader discussions of the Viagra versus birth control debate, see Hayden, supra note 178.
[246] The ADA prohibits discrimination by public accommodations. 42 U.S.C. § 12182(a) (2018). Public accommodations include insurance offices. Id. § 12181(7)(F). There is some dispute as to whether this means insurers must only provide access to physical spaces, or also nondiscrimination in access to insurance and in benefits. See Parker v. Metro. Life Ins. Co., 121 F.3d 1006, 1010–11 (6th Cir. 1997) (holding that Title III only applies to physical spaces of public accommodations). But see Carparts Distrib. Ctr., Inc. v. Auto. Wholesaler’s Ass’n of New England, Inc., 37 F.3d 12, 19–21 (1st Cir. 1994) (holding that Title III also applies to the goods and services offered by a public accommodation).
[247] 29 U.S.C. § 621(b) (1967), amended by Older Workers Benefit Protection Act, Pub. L. No. 101-433, 104 Stat. 978 (1990).
[248] Insurers must charge all insureds the same premiums, with exceptions for tobacco use, age, and geographic rating; thus, premium discrimination based on once-commonly used factors like health status, race, and sex is forbidden. ACA § 2701, 42 U.S.C. § 300gg (2018). For theories on why this may still permit some forms of gender rating, see Theresa Joux Neisen, A Liberal Feminist Perspective on Gender Rating and the Patient Protection and Affordable Care Act—Is Limited Protection Enough?, 11 Loy. J. Pub. Int. L. 469 (2010) (critiquing the ACA for continuing discrimination in grandfathered-in health plans).
[249] The law bans discrimination based on preexisting conditions, ACA § 2704, 42 U.S.C. § 300gg-3, as well as a host of other health status factors, see id. § 300gg-4. The ACA also requires that insurance be guaranteed available, id. § 300gg-1, and guaranteed renewable, id. § 300gg-2.
[250] The ACA mandates that all insurers universally cover ten essential health benefits, and at the same premiums for all, regardless of health status or other criteria. Id. § 18022. The ACA also removes, to some extent, financial incentives to discriminate by attempting to redistribute funds from those insurers who covered fewer risks to those who covered more.
[251] ACA, 42 U.S.C. § 18116(a). In regulations, U.S. Department of Health and Human Services (HHS) clarified that discrimination on the basis of sex incudes sex stereotyping (which may encompass some claims of sexual orientation discrimination) as well as discrimination on the basis of gender identify. 45 C.F.R. § 92.206 (2016). Section 1557 does not mandate that insurers cover specific benefits, but its regulations do forbid covered entities from offering “coverage that operates in a discriminatory manner.” Id. § 92 (preamble). An example of unlawful discrimination includes a “plan that covers inpatient treatment for eating disorders in men but not women.” Id.
[252] Explaining Health Care Reform: Questions About Health Insurance Subsidies, Kaiser Family Found. (Nov. 20, 2018), https://www.kff.org/health-reform/issue-brief/explaining-health-care-reform-questions-about-health [https://perma.cc/3B4W-7XCH].
[253] See ACA § 1501(a)(2), 42 U.S.C. § 18091. The individual mandate has been zeroed out, in effect, by the Tax Cuts and Jobs Act of 2017, which terminated the enforcement of the penalties associated with failure to purchase insurance. Christine Eibner & Sarah Nowak, The Effect of Eliminating the Individual Mandate Penalty and the Role of Behavioral Factors, Commonwealth Fund (July 11, 2018), https://www.commonwealthfund.org/publications/fund-reports/2018/jul/eliminating-individual-mandate-penalty-behavioral-factors [https://perma.cc/AA78-PYH5].
[254] See infra notes 266–268.
[255] Abraham & Chiappori, supra note 28, at 304 (discussing the need for market regulation to address discrimination by insurers).
[256] Ryan Lee, HIV/AIDS Group: Insurance Companies Discriminating Against Georgians Living with HIV, Harv. L. Sch. Ctr. for Health L. & Pol’y Innovation (Nov. 29, 2017), https://www.chlpi.org/hivaids-group-insurance-companies-discriminating-georgians-living-hiv [https://perma.cc/7QEU-SWPF].
[257] In an insurance circular letter to state health insurers, New York state officials made clear that PrEP must be covered as part of the essential health benefits package for the state. Letter from Lisette Johnson, supra note 163.
[258] Seventy-two percent of the public believes insurers should not charge sick people more for health insurance. Poll: The ACA’s Pre-Existing Condition Protections Remain Popular with the Public, Including Republicans, as Legal Challenge Looms This Week, Kaiser Fam. Found. (Sept. 5, 2018), https://www.kff.org/health-costs/press-release/poll-acas-pre-existing-condition-protections-remain-popular-with-public [https://perma.cc/YW49-TP3E].
[259] Medicare for All polls are at seventy-percent popularity. Yoni Blumberg, 70% of Americans Now Support Medicare-for-All—Here’s How Single-Payer Could Affect You, CNBC (Aug. 28, 2018), https://www.cnbc.com/2018/08/28/most-americans-now-support-medicare-for-all-and-free-college-tuition.html [https://perma.cc/Y46N-3GMR].
[260] Jonathan Martin & Abby Goodnough, Medicare for All Emerges as Early Policy Test for 2020 Democrats, N.Y. Times (Feb. 2, 2019), https://www.nytimes.com/2019/02/02/us/politics/medicare-for-all-2020.html [https://perma.cc/3ZN9-4VK2].
[261] Mariner, supra note 8, at 441.
[262] For an example in the context of health insurance, see Allison K. Hoffman, Three Models of Health Insurance: The Conceptual Pluralism of the Patient Protection and Affordable Care Act, 159 U. Pa. L. Rev. 1873 (2011) (describing how insurance sometimes seeks to repay individuals who suffer brute luck).
[263] This of course is also sometimes perceived to be true in health care too, if, for example, the poor health is related to an unhealthy behavior.
[264] Hellman, supra note 15, at 398.
[266] The ACA established community rating in insurance but does permit increased premiums based on geography, tobacco use of a factor of 1.5:1, as well as age 3:1. ACA § 2701(a)(1), 42 U.S.C. § 300gg (2018). For a discussion of ethical issues, see David B. Resnik, Charging Smokers Higher Health Insurance Rates: Is It Ethical?, Hastings Ctr. (Sept. 19, 2013), https://www.thehastingscenter.org/charging-smokers-higher-health-insurance-rates-is-it-ethical [https://perma.cc/27GP-2GPD]; Alex C. Liber et al., Tobacco Surcharges on 2015 Health Insurance Plans Sold in Federally Facilitated Marketplaces: Variations by Age and Geography and Implications for Health Equity, 105 Am. J. Pub. Health S696, S696 (2015), http://ajph.aphapublications.org/doi/pdf/10.2105/AJPH.2015.302694 [https://perma.cc/Z6KW-FW9Y].
[267] See ACA § 2705(j)(3), 42 U.S.C. § 300gg-4(j)(3). Wellness plans can base premium discounts by as much as thirty percent on health status outcomes, for example not just participation in a tobacco abstinence program, but proof one actually quit tobacco. For criticism of wellness plans, see Jill R. Horwitz et al., Wellness Incentives in the Workplace: Cost Savings Through Cost Shifting to Unhealthy Workers, 32 Health Aff. 468, 468 (2013).
[268] Employer plans, ERISA self-funded plans, and grandfathered plans are exempt from certain requirements of the ACA. And, insurers can avoid ACA requirements by not offering plans on the exchange (though they will forgo some consumers who are eligible for federal subsidies if they purchase on the exchange). Given these limitations, discriminatory plans persist in some cases.
[269] Cynthia Cox et al., Explaining Health Care Reform: Risk Adjustment, Reinsurance, and Risk Corridors, Kaiser Fam. Found. (Aug. 17, 2016), https://www.kff.org/health-reform/issue-brief/explaining-health-care-reform-risk-adjustment-reinsurance-and-risk-corridors [https://perma.cc/P8FG-HP4Q].
[270] Wortham, Insurance Classification, supra note 62, at 876 (reminding us that classification is only one model of competition in insurance).
[271] For instance, we do not accept all pregnancy discrimination in employment, even if pregnant women are costlier to their employers. However, there are important exceptions to this, for example, disability discrimination is generally prohibited, but employers can defend against accommodations that are too costly, even though in either case the disabled person is equally disenfranchised. Samuel R. Bagenstos, “Rational Discrimination,” Accommodation, and the Politics of (Disability) Civil Rights, 89 Va. L. Rev. 825, 849–50 (2003). For broader conversations on the economic efficiency of antidiscrimination laws, in the context of Title VII, see John J. Donohue III, Is Title VII Efficient?, 134 U. Pa. L. Rev. 1411 (1986); Richard A. Posner, The Efficiency and the Efficacy of Title VII, 136 U. Pa. L. Rev. 513 (1987).
[272] Wortham, Economics of Insurance Classification, supra note 29.
[273] Stone notes the success of a variety of social movements in framing insurance discrimination around matters of inequality; for instance, the battle over birth control coverage versus Viagra, or the effort to seek parity between mental health and other health care benefits. Stone, Beyond Moral Hazard, supra note 230, at 40–42.
[274] Wortham, Economics of Insurance Classification, supra note 29.
[275] For a general overview of challenges of capturing risk in insurance, see Ericson & Doyle, supra note 97, at 5. For specific examples, see Baker, Risk and Uncertainty, supra note 47 (exploring inaccuracies in liability insurance); Timothy Alborn, Regulated Lives: Life Insurance and British Society 1800–1914 (2009) (examining life insurance in England). Some scholars argue that accuracy is somewhat moot; that it really comes down to what is politically tolerable. Austin, supra note 100, at 552 (“Analysis reveals that accuracy either cannot be defined in a neutral, apolitical consensual fashion or must be balanced against, and sometimes give way to, competing non-neutral considerations through a blatantly political process. The predictive accuracy of the classification system and its political acceptability are thus inextricably bound. Legitimacy does not follow accuracy; quite the reverse is the case.”).
[276] See supra notes 173–180.
[277] Tom Baker, On the Genealogy of Moral Hazard, 75 Tex. L. Rev. 237 (1996). Baker discusses extensively gaps in the argument that individuals engage in riskier behavior once insured. Among his findings is that “[t]here is no strong evidence that insurance reduces the level of care individuals take to prevent bodily injury,” citing examples in both no fault car insurance and workers compensation schemes. Id. at 284; see also Siegelman, supra note 46 (arguing that adverse selection’s threat has been exaggerated beyond that which empirical or normative arguments can support, and it is often used reflexively by the courts and policymakers to justify the course they would prefer to take anyway).
[278] Abraham, supra note 27. Rawlsian theory also raises objections to the use of fixed characteristics for which an individual does not have control. John Rawls, A Theory of Justice (1971). This notion that the immutability or fixed nature of characteristics should define equity and protection has come under scrutiny as discrimination is increasingly recognized against some arguably changeable traits, for instance weight. Jessica A. Clarke, Against Immutability, 125 Yale L.J. 2 (2015); see also Mary L. Heen, Nondiscrimination in Insurance: The Next Chapter, 49 Ga. L. Rev. 1, 9 (2014) (asking whether it is preferable to select “less invidious characteristics that might affect mortality or morbidity, such as smoking or other risky behavior or medical history” as compared with protected class discrimination).
[279] Avraham, Logue & Schwarcz, supra note 87.
[280] See supra notes 141–171.
[281] See Plaintiff John Doe’s Memorandum in Support of Motion for Summary Judgment, supra note 162.
[282] ProPublica along with NPR have been investigating partnerships between health insurance and data brokers that mine social media and other internet histories to obtain detailed information on consumers. Marshall Allen, Health Insurers Are Vacuuming Up Details About You—And It Could Raise Your Rates, ProPublica (July 17, 2018, 5:00 AM), https://www.propublica.org/article/health-insurers-are-vacuuming-up-details-about-you-and-it-could-raise-your-rates [https://perma.cc/Z5QR-VB3A]. For more on the impacts of big data on insurance, see Rick Swedloff, Risk Classification’s Big Data (R)evolution, 21 Conn. Ins. L.J. 339 (2014).
[287] Solon Barocas & Andrew D. Selbst, Big Data’s Disparate Impact, 104 Calif. L. Rev. 671 (2016).
[288] John E. Pachankis et al., The Burden of Stigma on Health and Well-Being: A Taxonomy of Concealment, Course, Disruptiveness, Aesthetics, Origin, and Peril Across 93 Stigmas, 44 Personality & Soc. Psychol. Bull. 451 (2018).
[289] See Puhl & Brownwell, supra note 12.
[290] Wortham, Insurance Classification, supra note 62, at 356.
[292] This is a matter that has received pushback, after the ACA established community rating in insurance, but does permit increased premiums based on tobacco use of a factor of 1.5:1. ACA § 2701(a)(1)(A)(iv), 42 U.S.C. § 300gg (2018). For a discussion of ethical issues, see Resnik, supra note 266.
[293] Risk Classification Work Grp., supra note 70, at 44–45.
[294] Baker, Insurance Runoff, supra note 94.
[295] Wortham, Insurance Classification, supra note 62, at 846.
[296] For example, the ADA prohibits insurance classification based on disability unless it is actuarially based. 42 U.S.C. § 12201(c) (“[T]his Act shall not be construed to prohibit or restrict—(1) an insurer, hospital or medical service company, health maintenance organization, or any agent, or entity that administers benefit plans, or similar organizations from underwriting risks, classifying risks, or administering such risks that are based on or not inconsistent with State law; or (2) a person or organization covered by this Act from establishing, sponsoring, observing or administering the terms of a bona fide benefit plan that are based on underwriting risks, classifying risks, or administering such risks that are based on or not inconsistent with State law; or (3) a person or organization covered by this Act from establishing, sponsoring, observing or administering the terms of a bona fide benefit plan that is not subject to State laws that regulate insurance.”).
[297] Of course, other scholarly work may continue to object to actuarial fairness on other grounds, for instance moral ones. See Stone, supra note 5.
[298] Link & Phelan, Conceptualizing Stigma, supra note 6, at 368.
[300] Doe v. Mut. of Omaha Ins. Co., 179 F.3d 557 (7th Cir. 1999).
[303] Jacobi, supra note 230 (describing legislative history of the ADA in detail and suggesting that it proposes an objective test of actuarial fairness).
[304] U.S. Equal Emp. Opportunity Comm’n, EEOC Compliance Manual, No. 915.003, at ch. 3 (2000).
[305] 29 C.F.R. § 1630 (2016) (alteration added).
[307] Id.; see also supra note 150.
[308] Jacobi, supra note 230, at 364–65.
[310] Letter from Lisette Johnson, supra note 163.
[311] Letter from James Regalbuto, Deputy Superintendent, Life Insurance, & Troy Oechsner, Deputy Superintendent, Health Insurance, to All Insurers and Fraternal Benefit Societies Authorized to Write Life Ins. or Accident and Health Ins. in N.Y. State (June 22, 2018) [hereinafter Letter from Regalbuto & Oechsner], https://www.dfs.ny.gov/industry_guidance/circular_letters/cl2018_08 [https://perma.cc/MR3G-532Q].
[314] Risk Classification Work Grp., supra note 70, at 58–59, 62; Trowbridge, supra note 74, at 60–61, 78.
[315] Valarie Blake, Rethinking the Americans with Disabilities Act’s Insurance Safe Harbor, 6 Laws 25 (2017).
[318] In reality, disability can be understood as a social issue; people are disabled because of the way we build the world and not any inherent limitation. Many people with disabilities cost insurers far less than other people without disabilities. Consider the person who experiences a car accident compared with someone who may simply need a hearing aid. For a good overview of the topic of medicalization and stigma against the disabled, see Mike Oliver, The Social Model of Disability: Thirty Years On, 28 Disability & Soc’y 1024, 1024, 1026 (2013); Tom Shakespeare, The Social Model of Disability, in The Disability Studies Reader 195, 199–202 (Lennard J. Davis ed., 5th ed. 2017).
[319] Blake, supra note 315, at 33.
[321] See supra notes 185–187 and accompanying text.
[322] Risk Classification Work Grp., supra note 70, at 16–18; Trowbridge, supra note 74, at 55.
[323] Letter from Regalbuto & Oechsner, supra note 311.
[324] For example, in the context of health care, stigma led to health disparities which were then reflected in discriminatory pricing.
[325] Heen, supra note 84, at 372.
[326] See Hatzenbuehler, Phelan & Link, supra note 106, at 814, for evidence that stigma causes structural harms and disparities.
[327] Steven H. Woolf & Paula Braveman, Where Health Disparities Begin: The Role of Social and Economic Determinants—And Why Current Policies May Make Matters Worse, 30 Health Aff. 1852, 1854–55 (2011); Ctrs. for Disease Control and Prevention, U.S. Dep’t of Health and Human Servs, CDC Health Disparities and Inequalities Report—United States, 2013, 62 Morbidity & Mortality Wkly. Rep. 1, 1, 9, 20, 27 (2013), https://www.cdc.gov/mmwr/pdf/other/su6203.pdf [https://perma.cc/N29P-MPER].
[328] For some examples of disparities, see Samantha Artiga et al., Key Facts on Health and Health Care by Race and Ethnicity, Kaiser Fam. Found. (June 7, 2016), https://www.kff.org/disparities-policy/report/key-facts-on-health-and-health-care-by-race-and-ethnicity [https://perma.cc/95NG-XGTD].
[329] See, e.g., Cicero, supra note 183, at 217–19, 225–27 (suggesting that if actuarial calculations demonstrate real differences about risk, then society ought to bear that burden rather than individuals).