The European Court of Justice (ECJ) has found that the pricing of insurance and other financial services on the basis of the sex of the customer amounts to sexual discrimination.
Equality has always been one of the key foundations of the EU. The scope of its application has expanded gradually but ineluctably over the decades, from equal treatment for different nationalities, through equal treatment for different sexes, to equal treatment for people of different sexual orientations, to name but a few. In this respect it may come as no surprise that the ECJ has finally caught up with actuarial practices that directly discriminate between the sexes, even if this discrimination is based on unequivocal statistical correlations between gender, and tendencies to make insurance claims and/or live longer.
On 1 March 2011 the ECJ found (following the opinion of Advocate General Kokott) that a rule, contained in a piece of secondary EU legislation, which enables EU member states to maintain an exemption from the rule of unisex premiums and benefits, works against the achievement of the objective of equal treatment between men and women. This rule will be invalid from 21 December 2012.
The ECJ reached this conclusion despite almost universal opposition from several EU member states, including the UK, as well as the Council of the EU and the European Commission.
The judgment highlights the tensions between anti-discrimination legislation and actuarial practice. Anti-discrimination legislation is generally concerned with preventing a person from being pre-judged on the basis of the general characteristics of certain groups to which they belong. Actuarial practice is based on the application of the general characteristics of the group to which an individual belongs to that individual, generating a specific risk profile for that individual. This risk profile is the key input for the price of the financial product. The specific tendencies of the individual are not taken into account.
It is now clear that the ECJ considers that the use of statistical correlations between groups identified as being protected by anti-discrimination law and their propensity to make insurance claims will be subject to very strict judicial scrutiny. In many cases (such as gender), the use of such correlations will be prohibited. Actuaries will need to seek out new ways of identifying groups that fall outside the scope of anti-discrimination legislation when assessing risk premiums.
JUDGMENT OF THE ECJ
The case concerned a reference to the ECJ for a preliminary ruling from the Belgian Constitutional Court in relation to the lawfulness of Article 5(2) of Council Directive 2004/113/EC (the Directive).
The general purpose of the Directive was to establish a framework for combating gender discrimination in the provision of goods and services. Article 5(1) of the Directive states that member states should ensure that, for all new contracts concluded after 21 December 2007, sex is not a factor having any influence on the calculation of premiums and benefits for the purposes of insurance and related financial services.
By way of derogation from the general rule established by Article 5(1), Article 5(2) allowed member states to:
‘Permit proportionate differences in individuals’ premiums and benefits where the use of sex is a determining factor in the assessment of risk based on relevant and accurate actuarial and statistical data.’
The Directive was silent on how long member states could continue to allow such differences to apply: insurers were able to continue these practices indefinitely.
The Belgian government, as well as several other member states, including the UK, chose to take advantage of this provision, and specifically permitted the use of sex as a determining factor in the assessment of risk.
In 2008 two private individuals and a consumer protection organisation (Association Belge des Consommateurs Test-Achats ASBL) brought an action against the Belgian government to the Belgian Constitutional Court. The claimants asserted that Article 5(2) of the Directive, as transposed into law by the Belgian government on 21 December 2007, was incompatible with the principle of equal treatment for men and women. In its reference to the ECJ, the Belgian Constitutional Court asked the ECJ whether Article 5(2) of the Directive is valid – in other words, whether it is compatible with the principle of equal treatment under EU law.
The ECJ noted that the application of this general principle of equal treatment derives from Article 6 of the Treaty on the Functioning of the European Union (TFEU), which sets out that fundamental rights, such as those set out in the Charter of Fundamental Rights of the EU (the Charter), ‘shall constitute general principles of the Union’s law’. The Charter prohibits ‘any discrimination’ on several grounds, including sex (and, incidentally, age). The Charter further stipulates that: ‘Equality between men and women must be ensured in all areas.’
The Directive represents an attempt by the European legislature to introduce a measure to combat discrimination in the field of insurance, as it was empowered to do by the TFEU. In accordance with Article 19(1) of the TFEU, such measures must be ‘appropriate’ for combating discrimination. The key question was whether Article 5(2) was appropriate in this context.
The ECJ noted that the principle of equal treatment requires that comparable situations must not be treated differently and different situations must not be treated in the same way, unless such treatment is objectively justified. In line with this, Recital 18 to the Directive expressly states that the use of sex as an actuarial factor must not result in differences in premiums and benefits for insured individuals: unisex rules on premiums and benefits should apply.
Several intervening parties argued that the modus operandi of insurers indicated that the levels of insured risk may be different for men and for women. While the ECJ noted that this was what the exemption in Article 5(2) was meant to provide for, it gave short shrift to arguments that this was in any way consistent with the principle of equal treatment. It considered that:
‘There is a risk that EU law may permit the derogation from the equal treatment of men and women, provided for in Article 5(2), to persist indefinitely.’
It considered that such an exemption, that applied without time limit, was contrary to the explicit purpose of the Directive (unisex premiums), as well as contrary to the higher EU law set out in the TFEU and the Charter.
As such, the exemption set out in Article 5(2) ought to be severed from the rest of the Directive, effective from 21 December 2012, to enable a period of transition.
REASONING OF THE ADVOCATE GENERAL
The ECJ’s reasoning was very blunt, an indication perhaps of the degree of confidence with which it reached its conclusions. The Advocate General (who typically provides an opinion to the ECJ with the objective of persuading the ECJ of a particular viewpoint) provided more nuanced reasoning, which is instructive both for gender discrimination and for other types of discrimination.
The Attorney General noted that insurance companies provide services to which it cannot be said, at the time the contract is concluded, if, when and to what extent the insured person will have recourse to them. It is obviously necessary for insurance companies to reach a prognosis of the likely recourse any particular individual will have to the contracted services, to calculate a premium for that individual and develop their products in a way that do justice to the risk that they will be used. The Advocate General referred to life assurance and pension insurance, in which the likely life expectancy of the insured person was integral to the likely degree of recourse to the service. In relation to motor vehicle insurance, the key metric was the likelihood of an individual causing an accident when driving, while with regard to private medical insurance, it was the likelihood of an individual making use of medical services.
Kokott noted that a prognosis of particular individuals is rarely carried out: instead recourse is had to ‘experiential values’, because exact statements to the insurance risk linked to particular individuals are impossible to make. She acknowledged that it is ‘perfectly legitimate with regard to risk evaluation to carry out a group examination’, rather than an individual examination. What matters is what comparison groups are used.
Direct discrimination on grounds of sex, as a method of group comparison, was not permissible, unless it could be:
‘Established with certainty that there are necessary differences between men and women which necessitate such discrimination.’
The Advocate General was adamant that there is no such certainty as regards statistical differences in the insured risk of men and women:
‘There is a sweeping assumption that the different life expectancies of male and female insured persons, the difference in their propensity to take risks when driving and the difference in their inclination to utilise medical services – which merely come to light statistically – are essentially due to their sex.’
The Advocate General considered that there were many other factors at play in the evaluation of such insurance risks: life expectancy is strongly influenced by economic and social conditions, as well as by the habits of each individual. She went so far as to note that, nowadays, both men and women engage in demanding professional, sporting and recreational activities, such that one cannot make assumptions about behaviour on the basis of sex. Such a direct challenge to actuarial practice in which statistics are blind to any prejudice about the roles of men and women in society underlines the strength of feeling that infuses the opinion.
In conclusion, the Advocate General acknowledges that it is ‘especially easy to implement distinctions on the basis of sex in respect of insurance products’, whereas the:
‘Correct recording and evaluation of economic and social conditions and of habits of insured persons is much more complicated.’
However, such practical differences alone cannot justify unequal treatment, so that the use of sex as a:
‘Kind of substitute criterion for other distinguishing features is incompatible with the principle of equal treatment for men and women.’
IMPLICATIONS FOR CUSTOMERS AND PROVIDERS OF INSURANCE-RELATED FINANCIAL SERVICES
As has been widely reported, the eventual removal of sex as a criterion in the assessment of risk will likely lead to an increase in premiums for safer female drivers, rather than a reduction in premiums for other classes of drivers. At the same time, the unavailability of the strong statistical correlation between sex and life expectancy in the calculation of pension annuities may (in the absence of alternative statistical correlations being used) result in lower returns for men.
Unsurprisingly, the insurance industry is deeply concerned about the ECJ’s decision. The removal of the Article 5(2) exemption will restrict the ability of insurers to model the risks associated with the policies they offer. While insurers may, over time, develop alternative non-discriminatory means of assessing the risks posed by prospective customers, there is likely to be a period during which insurers are less confident as to the level of risk associated with particular policies. In these circumstances it would not be surprising if insurers were to increase the average premiums they charge for their policies.
The transition period will lessen the disruption to the insurance market. A substantial proportion of insurance contracts, for example health insurance and car insurance, are concluded for a year. However, longer term policies, such as life assurance, may require revision. While it may be tempting for insurers to argue that the risk assessment that led to the calculation of the price and/or the returns for the life assurance product were carried out at the point of sale, so any gender discrimination is a matter of history, customers who have been disadvantaged by this process are unlikely to agree. If insurers do not change these longer term policies in time for the December 2012 deadline, they are likely to face a wave of litigation from their customers.
The industry may well need to brace itself for a reference to the ECJ in relation to the use of age as a factor in the calculation of risk. Such a reference would present a slightly different set of issues for the ECJ to consider.
The use of age as a factor to determine risk in the context of car insurance products may amount to the use of a proxy measure for other, more subtle factors at play (such as level of driving experience), and as such would be likely to lead the ECJ to the same conclusions as it has reached in this case.
As regards pension annuity products, it is possible to envisage stronger arguments in favour of the use of age as a determinative factor in calculating annuity payments. Is it absurd for a 50 year old with a lump sum to be entitled to the same annual pension payment as a 65 year old with the same lump sum? An appraisal of the physical condition of the two individuals may find that the 50 year old who smokes and drinks heavily is statistically likely to live for another ten years, whereas a very fit and healthy 65 year old is statistically likely to live for a further 15 years.
Clearly there are factors other than age at play in the appraisal of the value of a pension product for specific individuals. Nonetheless, as regards two otherwise fit and healthy individuals, age may be the only factor by which to discriminate between them. If the ECJ prevents this factor from being used, it is conceivable that a fit and healthy 50 year old will receive the same payments as the fit and healthy 65 year old, even though the 50 year old is statistically likely to live for longer.
Nevertheless, if the ruling results in a more detailed and forensic analysis of risk on criteria other than age and gender (and transparent pricing of the same) the change will undoubtedly be positive.
Holman Fenwick Willan LLP is currently assessing the legal impact of the ruling in respect of its clients’ products, including policy wordings, and the manner in which risk is assessed. It is fair to say that the commercial effects are still being assessed.