Recently, the Senate of Delaware passed a bill that bans the use of gender, gender identity, or sex as a factor in automobile insurance premiums, making Delaware the seventh state in the US to ban the usage of sex/gender as an actuarial factor by insurers.
Public criticism of the usage of sex/gender as an actuarial factor is grounded on the principles of equality and anti-discrimination. However, insurers argue that the efficiency of the insurance market cannot be compromised, especially if there is statistical evidence that establishes a correlation between sex and the risk identified. Accordingly, the question at the heart of this debate is: Can certain constitutionally recognised grounds of discrimination such as sex be used as a risk classifier to set insurance premium rates if there is an indisputable statistical correlation?
The economics of insurance incentivises insurers to classify individuals into different risk pools to charge different premium rates. If not, the pool of high-risk individuals is blended with the pool of low-risk individuals, resulting in inefficiency and higher costs for insurers.
Let’s assume an insurer proposes to have a blended healthcare insurance pool with two sets of individuals: high-risk individuals, such as A, and low-risk individuals, such as B. In an insurance market with inaccurate risk classification, both A and B would be expected to pay x, the uniform average premium rate for an insurance plan. (This model assumes that the insurance market is competitive and that there is no information asymmetry in the market between insurers and the insureds.)
Here, high-risk individuals like A would tend to opt-in for the insurance plan at x premium (conversely, B would tend to opt-out). As such, the insurer operates sub-optimally given that high-risk individuals like A are expensive healthcare consumers. Therefore, to avoid disincentivizing B from opting-out (i.e., adverse selection), premium rates are evaluated and set based on statistical data and actuarial factors.
Pertinently, Obamacare in 2010 introduced community-rated premiums which equalized the premium rates for different pools of individuals regardless of the varying degrees of risk. This mergence of expensive individuals with low-risk individuals garnered criticism from several stakeholders, as well as from Republicans, who subsequently campaigned for the repeal and replacement of Obamacare.
However, this decision to classify individuals into different risk pools requires evaluation as a matter of socio-economic justice given that insurance bears significant social utility. Health insurance plays a crucial social function by expanding access to healthcare. As a matter of state policy, life insurance is also perceived as a tool to promote social security for the insureds and their dependants by cross-subsiding the risk among other individuals. Therefore, socio-economic institutions almost always invite public interest scrutiny, especially when there is a possibility of those institutions transgressing principles of equality and anti-discrimination.
The principles of justice and equality ground the notion that the state and institutions of public interest (such as insurers) should not discriminate between citizens unfairly. An individual cannot be subjected to discrimination solely based on one’s affiliation to a particular social class. A transformative understanding of the equality jurisprudence implies that subjecting an individual to vast generalisation is arbitrary and unjustifiable to that individual, since the generalisation might not be true to their specific case.
Insurance premium rates, however, are inherently discriminatory. The mechanics of insurance is premised on spreading financial risks (evenly) among many contributors to the insurance plan. Hence, given that insurance is based on risk diversification, insurance premium rates are not individual-centric. So, when the economics of insurance (i.e., adverse selection and risk diversification) is juxtaposed against the principle of anti-discrimination, there is an unavoidable tension.
To reconcile this tension, one can argue that as long as an (intelligible) risk classifier is rational, the risk classification would be legitimate. If the goal of the insurance sector is to avoid adverse selection and to improve efficiency, then sex as a risk classifier would pass this muster.
This argument, however, is problematic on several fronts. The practices of insurers cannot escape scrutiny from the lens of equality and anti-discrimination, given the socio-economic salience of insurance. To stand the rigours of reasonableness and equality, a mere statistical correlation would not be sufficient. Even if one were to assume that the statistical correlation is accurate, what makes the classification objectionable is subjecting individuals to a set of assumptions and statistical correlation that is not true to their specific case. Hence, even if there a convincing statistical study suggesting that young male drivers tend to be rash, charging a male driver a higher premium than a female driver is perverse. Similarly, a mere statistical correlation between mortality and sex is not sufficient to justify differential premium rates for life insurance or annuity plans.
A more prudent and justiciable approach would be to ask if the efficiency of the insurer would be gravely affected, if not for the usage of the grounds of discrimination such as sex. This burden not only rules out the possibility of using sex as a proxy factor, but also requires a higher standard of causal link (and not mere statistical correlation).
Often, sex is used as a prima facie risk classifier since it is an observable and cost-efficient classifier that does not require complex risk management tools. Moreover, in case of auto insurance metrics such as age, driving records, type of vehicle, engine capacity, and individual driving behaviour are rationally better metrics than sex for risk correlation, given they establish a tangible causal link. But the issue is more nuanced in health and life insurances. The subjection of individuals to generalization (applicable to auto insurance) is less objectionable in health and life insurances since people of different sexes are anatomically different. Moreover, women are more expensive healthcare consumers than men due to pregnancy and maternity. This provides a well-founded justification for insurers to set differential premium rates for different sexes from an efficiency perspective.
However, sex as a personal trait is non-controllable. From an egalitarian perspective, it would be unfair to impose economic costs on a set of individuals who cannot control that trait. This tangent of argument is usually resorted to in cases of discriminatory insurance pricing based on genetic factors. Further, a blanket ban on using sex/gender as an actuarial factor would in fact be contrary to the achievement of substantive equality.
The European Commission (EC), in its interpretation of a significant ruling by the Court of Justice of European Union recognized this need to balance efficiency and substantive equality. It observed that differential premium rates that are based on engine capacity must be permissible, even though statistically male drivers use cars with higher torque and horsepower. Further, sex and gender can also be used as a metric in association with other factors to set premium rates. For example, a female who is genetically pre-disposed to breast or ovarian cancer or male to prostate cancer would be expected to pay higher premium rates. The only exception to this rule, the EC opined, is labour and post-natal costs, which would not require additional premium payments from women, as a matter of social justice. A similar public policy exception can be traced in life insurance as well—women should not be expected to contribute more to annuity plans merely because women generally have higher life expectancy than men.
In a sector like insurance which garners the attention of several interest groups, policy considerations require reasoning and engagement of all stakeholders. A solution that is solely based on a legal justification may not practically work. Similarly, a sound insurance practice that breaches the ideals of justice and equality would be subjected to immense public interest scrutiny. Therefore, optimal efficiency needs to be realised in conjunction with substantive equality. The EC’s attempt to navigate the paradox by identifying public policy exceptions to general insurance practices seems to be the way forward in reconciling complex efficiency-fairness debates.
Raghav Harini N is a lawyer based out of Mumbai, India.