‘Equality’ equals unequal results (in insurance, anyway)
Much of the rhetoric observed on the internet and social media these days is saturated with the word ‘equality’. In much the same way that the word ‘liberalism’ has been distorted from its original meaning of personal freedom and responsibility, to social welfare and increasing state control of economic and personal matters; ‘equality’ has come to mean preferential treatment for some to obtain equal consumption for all, rather than equal treatment under the law or blind justice. The attempt to implement elements of such rhetoric pervades almost every facet of life these days.
It came as no surprise, therefore, when the extent of the crusade for equality became more apparent during one of my Contingencies lectures recently: insurance companies in the European Union are legally barred from using the sex of the insured person as a factor in pricing their products. In 2011, the European Court of Justice (ECJ) ruled that the use of sex in determining insurance premiums was ‘unfair discrimination’, and subsequently gave insurers until the end of 2012 to change their practices – all in the name of greater ‘equality’.
To those unfamiliar with the inner workings of insurance-type business, which includes general insurance, life insurance and pension annuities, some context is needed to understand the implications of the ruling.
Insurance is the transfer of risk from one party, the insured, to another, the insurer. Most people understand this intuitively – but what exactly is risk? Risk (in the financial sense) is made up of two components: uncertainty and value. Stated crudely, risk can be considered as follows:
Risk = Uncertainty x Value
Value is the more easily understood and quantifiable component of risk – it is simply the monetary amount an insurer would have to pay out for a given event (which could be such things as theft, a car accident, retirement, death or survival), taking into account the timing of the cash-flow. Uncertainty is much trickier, since it comprises the probability that a payment must be made to the insured in future. Another way to think about uncertainty is considering it as the likelihood that a claim event will happen. Actuaries, who are responsible for much of what happens in insurance businesses, use advanced statistical methods to arrive at quantified estimates of the uncertainty component.
There are many aspects of the uncertainty component that are related to the sex of the insured. In Europe, men are more likely to claim on their car insurance – perhaps they are more likely to be involved in car accidents. Women, generally, are likely to live longer than men, so companies paying out annuities (after retirement, for example), will likely have to pay more to women than to men – in this regard, women bear greater financial risk. Conversely, for life insurance, men are likely to die sooner, and so the payments of death benefits for insured men may take place sooner than for women – here, when the time value of money is taken into account, men bear more risk. There are many other cases. The important thing to take away here is that being able to take the insured person’s sex into account when setting their premium – which is, after all, a function of that person’s risk – is a useful and cost-effective way of accurately incorporating their underlying risk.
This can be illustrated and understood through the following example. Consider a short-term (1 year) car insurance policy that pays R1 000 (the ‘sum insured’) to the insured person if they are involved in an accident. Furthermore, let’s assume the probability of being in an accident is 10% for male drivers and 5% for female drivers. Applying the risk equation outlined above, we can conclude (rather crudely) that the financial risk associated with this policy for a male is R100, whereas the corresponding amount is R50 for a female. The insurer would add margins for expenses and profits to both of these amounts, but the final premiums paid by women will, as a function of their lower risk, be less than those for men.
So what, then, is the result of the ECJ ruling? Actuaries working for European firms can no longer take the insured person’s sex into account, so the probability of having an accident has to be ‘averaged out’ between the two categories of people. Continuing our example, let’s assume that the new probability of an accident comes to 7.5% for all drivers. Clearly, the financial risk associated with policies for everyone is now R75, and premiums will be charged on a similar basis. In summary:
Some of the readers may have noticed the ‘additional expenses’ associated with the ‘after’ case. The reality of the situation is that many insurers have had to incur additional business costs in order to adapt to their new constraints. Part of this has been making significant changes to existing marketing strategies, while another has been more stringent underwriting and additional data collection, and so, contrary to what some people believe, costs (and risks) are not only redistributed between consumers, but increase overall. Ultimately, this represents the fact that more resources have to be consumed in producing insurance products. The problem lies especially in the fact that these costs result from an arbitrary, legally-imposed decision, as opposed to a real relative increase in the scarcity of resources, and so everyone is worse-off as a result.
Although the above example is contrived, what it does serve to illustrate is that these sorts of regulations benefit one type of consumer at the expense of another – that is, preferential treatment is given to one over another. In the car insurance industry, reductions in men’s premiums have come at the expense of increases in women’s premiums, whereas in the annuities business, for example, the roles have been reversed. These outcomes are not purely theoretical, though, as the table below shows for car insurance in Europe. Furthermore, a presentation by the reinsurer Swiss Re to the Society of Actuaries in 2014 showed that, shortly before the ruling came into effect, there was a spike in the new business written for those who stood to lose out after the adjustment.
It is not that insurers arbitrarily decided that women bear less risk for car insurance, but more for annuities. These relationships are demonstrably true and are grounded in reality; those who are invariably committed to egalitarian rhetoric would have us believe that the differences in risks between sexes are solely the products of perception and punitive business practices.
The question remains: is it really unfair for the price that an insurer charges you to reflect risks to which you are most likely subject? Even if we do take this as unfair, proponents of government regulation still have to demonstrate why the government necessarily needs to intervene. If sex-based pricing is widely regarded as unfair, the insurers that do not engage in such ‘unfair discrimination’ would be favoured by consumers over insurers that do, and the market share of companies would adjust commensurately. A more fundamental question is this: if you bear greater risk, why should somebody else have to pay for it? After all, if we regard the notion of ‘redistributing risk’ as acceptable, nobody in the ‘equality’ camp should be outraged by the bailouts of risk-taking financial institutions in the US, for example.
In tying this back to what was said at the outset, it is quite clear that the distortion of the words ‘liberalism’ and ‘equality’ from their original meanings is no coincidence. In an attempt to generate greater ‘equality’ (as has been demonstrated in the European insurance industry) the freedom of businesses to contract with whomever they wish, on mutually-agreed terms, is being restricted. Generally, actuaries are viewed as being pessimists – but when it comes to government regulation, they are certainly being realistic in preparing for more of the same.
 Legutko, J. 2013. BUS1003H: Introduction to actuarial science [BUS1003H course reader]. Actuarial Science section, University of Cape Town. (Unpublished).
 In reality, the new, generally applicable probability measure would not simply be the average of those for men and women – it could, for example, be derived by weighting the previous rates by the proportion of male and female policyholders on the company’s books.