Gender Neutral

The Court of Justice of the European Union (ECJ) announced on Mar. 1, 2011 that gender-based pricing is inconsistent with the basic principles of EU law and that it will not be allowed as of Dec. 21, 2012. The ruling in the Test Achats case will have a fundamental impact on both the insurance and pensions industry in the United Kingdom, where differential pricing between men and women based on actuarial factors is widespread, especially in relation to life, motor and health policies. It also raises questions about any potential effects such a decision may have in Canada.

The ECJ’s Decision

The Transition

The ECJ’s decision calls for a transitional period of up to Dec. 21, 2012, at which time all insurance pricing must become unisex. When the EU’s Gender Directive was introduced in 2008, it required the use of the carve-out to be reviewed five years after implementation. Basically, the ECJ used this window of review to allow insurers time to adjust to unisex pricing. The insurance market had expressed concern before the judgment that the ECJ would ban gender-based pricing with immediate effect. It is of some comfort the court has allowed a transitional period of slightly shorter than two years within which insurers must adapt their policies and practices.

The Uncertainty

The judgment is quite short, and unfortunately it does leave some questions unanswered. For example, it is not actually clear what effect the removal of the carve-out from Dec. 21, 2012 will have on existing contracts. Contracts entered into before the Gender Directive came into effect should generally be safe from the ruling’s effect.
 
For U.K. contracts entered into after Apr. 6, 2008 and before Dec. 21, 2012, it seems likely they will be exempt from the new unisex requirements (provided they have complied with the requirements of the carve-out – for example, differences in premiums or benefits are proportionate to published actuarial and statistical data that do not relate to pregnancy or childbirth). However, it is possible to argue the judgment could catch some existing contracts, in situations in which new benefits are purchased after the end of the transitional period, for example, and make them subject to the unisex requirement. Unless the ECJ makes a further statement, the effect on existing contracts may not be known until the U.K. government specifies how it will change the Equality Act 2010 to reflect the judgment.

It may be that motor insurers are slightly better placed than life and health insurers to carry on using more risk-based assessments, since they have potentially more factors and data on which to base their assessment. They will, however, face the same problem as life and health insurers in terms of not being able to control the mix of male/female business they get. They will therefore need to be confident in their assumptions. Some gender-targeted business models will need to make changes to the way they operate. They also need to be clear the rating factors they are using are not a proxy for gender.

Insurers have in a sense been left holding the baby: the new rules should not apply to reinsurance, although reinsurers may want some comfort that direct insurers are complying with the new rules. Also, Solvency II will presumably require a realistic, risk-based assessment of insurance liabilities. This could involve using available statistical data on gender, thus leaving the insurer with a real-world view of what is going on – a view upon which it might have to rely for pricing. Unfortunately for consumers, it seems that they may end up bearing the cost of this.  

Finally, it is worth noting that a proposed EU Equality Directive is looming on the horizon. This would prohibit discrimination on grounds of religion or belief, disability, age or sexual orientation. As originally drafted, this proposed directive contained a carve-out for insurance similar to what has just been found to be invalid. It currently contains a statement that in relation to age and disability, where their use is a determining factor in the assessment of risk based on relevant actuarial principles, statistical data or medical knowledge, it shall not be deemed to constitute discrimination for the purposes of the directive. However, following the logic of the Test Achats case, pricing based upon these characteristics could also be considered to be incompatible with EU law. Therefore, it is possible this carve-out could be removed.

Could this Happen in Canada?

The Supreme Court of Canada ruled in Zurich Insurance v. Ontario (Human Rights Commission) in 1992 that higher automobile insurance premiums for single male drivers under the age of 25 were discriminatory under the Ontario Human Rights Code. But this scheme was permitted because the code allows discriminatory practices if there are “reasonable and bona fide grounds” for doing them.¹  

A majority on the court held a discriminatory practice is reasonable in the context of insurance so long as it is based on a sound and accepted insurance practice and there is no practical alternative. Setting higher premiums for young, single males was sound because it achieved the legitimate business goal of aligning price with risk. There was no practical alternative, because actuarially reliable data that would allow the insurance industry to set premiums based on non-discriminatory criteria were not available.

Two judges on the Supreme Court dissented in the Zurich case.  In their view, an insurer must not only prove a statistical correlation between a particular group and higher risk, but a causal connection. Also, the dissenting judges found the insurance industry should not be able to rely on its inaction in collecting data, nor should it be able to rely upon its contention that premiums have always been set based on criteria we now recognize as discriminatory under human rights legislation.   

On challenges to discriminatory insurance practices, Canadian court rulings since Zurich have largely followed the approach of the dissenting judges. For example, an individual was refused life insurance in a 1999 case because his wife was HIV-positive. The British Columbia Human Rights Tribunal held this amounted to discrimination on the basis of marital status.² The tribunal rejected the insurer’s contention that this discrimination fell under the exception in Zurich. Essentially, the tribunal found the insurer had not provided any evidence that it would be impossible to assess the risk involved using non-discriminatory criteria. It was not enough for the insurer to show, as it had in Zurich, that its current practice was widespread in the Canadian industry and that no data was currently available to assess risk otherwise.  

Coupled with the ECJ ruling in the Test Achats case, this caselaw suggests it may be difficult for the Canadian insurance industry to prove there are no reasonable and practical alternatives to setting premiums based on discriminatory grounds. Given that the industry in Europe and various jurisdictions in the United States do not – or will not, as of December 2012 – permit the use of sex as a rating variable, Canadian courts may look askance at attempts by Canadian insurers to claim that they cannot do likewise.

1  Zurich Insurance Co. v. Ontario (Human Rights Commission), [1992] 2 S.C.R. 321.
2  J. v. London Life Insurance Co., [1999] B.C.J.R.T.D. No. 35.  See also:  Battlefords and District Co-operative Ltd. v. Gibbs, [1996] 3 S.C.R. 566