Home | Links

Archive for February 2011

Thursday, February 17th, 2011

Gender and insurance pricing posted by Ross Dawkins

 

There is a long tradition in the UK (and elsewhere) of using gender as a risk rating factor in the pricing of insurance.  This tradition may shortly be broken.

European institutions in conflict

In 2009, a Belgian consumer group challenged the provision in the EC’s Gender Directive that allowed the use of gender in insurance risk pricing.  They argued that such use was not compatible with the principles of equality guaranteed in European legislation.  A European Court of Justice Opinion was issued (September 2010) finding in favour of this challenge.  This has yet to be confirmed in a final judgment (expected sometime around 1st March) but if confirmed would not be subject to further appeal.

Let's leave to one side the question of the ECJ's status in this decision (because it's outside my sphere of competence albeit not my area of interest).  Instead, let’s focus on one area of insurance that would be affected — motor insurance. 

What are the likely consequences of this for consumers? 
The answer to that depends upon which sex the consumer is and in which country the insurance is being bought.  As part of a study on the European retail insurance market (on behalf of the European Commission) we conducted a pan-EU mystery shopping exercise in motor insurance.  We tested six profiles of consumer-car combinations.  This provides us with a useful source of evidence.
 
Our sample included several countries where gender differentiation is prohibited already (e.g. Greece, the Netherlands).  Of the countries where it is not prohibited, pricing differences were present and sometimes highly significant.  In a profile where a 22 year-old driver was seeking third party, fire and theft cover (or the local equivalent) we found a price difference between the lowest male and female quotes of about 40 per cent in Spain, 30 per cent in the UK, 20 per cent in Italy and 5 per cent in Germany (the percentages are calculated relative to the lowest female quote).  The pricing data that we collected did not always show pricing differences (at least not material ones) — under the Gender Directive, differentials need country-level actuarial evidence to support them and the data may not exhibit differences.  In the UK, for example, risk and hence price differences tend to be focused in the age range 17 through to about 30.

The motivation for using gender as a rating factor is the same as for all rating factors: to improve the efficiency of pricing.  So the mooted change will introduce some additional friction into the market.  Would the new premium in the UK (and Germany, Italy, Spain, etc) simply be an average of the two?  Perhaps, but one of the factors underlying the superior safety record of young female drivers that drives the pricing difference is believed to be a tendency as a group towards higher risk aversion, at least relative to males of the same age. A critical factor is a tendency to a less aggressive driving style. This implies that the price elasticity of demand for insurance for the average woman will be more inelastic than that of the average man. And this means that there is reason to anticipate insurers setting unisex rates closer to the pre-existing male premium. This is complicated a little by third party cover being compulsory.

In any event the result would be the subsidising of young male drivers by their female peers.  The affordability of driving will be eased for a riskier on average group at the expense of a less risky group.  The result could be a mix of (a) an increase in young men having cars coupled with a reduction in motorisation for women; (b) male drivers switching to more powerful cars whilst women downgrade (in a model where drivers have a fixed motoring budget and so compensate for the changed insurance cost); and (c) the number of male drivers would not change but more would be insured.  That is just possible in the UK where uninsured drivers are sadly not uncommon, but less likely elsewhere in the EU.  Both (a) and (b) are likely to result in an increased accident rate — (c) is more optimistic.  All in all, a momentous decision is about to be made.

A possible industry response

I have considered the possible consequences for consumers, what about the insurance industry — how might it respond?  As I have noted already, gender is used as a risk factor to model expected driving behaviour.  It’s inevitably a crude tool hence the ECJ’s objection, although it’s also clear that it is still powerfully useful at least at some ages and in some countries.

The efforts made by the industry to individualise risk — such as through Pay As You Drive (PAYD) — have so far met with limited success.  PAYD involves fitting a black box onto the car which then monitors the distance travelled and at what time of day (rush hour driving attracts a different per mile tariff relative to driving “off-peak”).  It is yet to be successfully exploited on a large scale in Europe.  That said numerous pilots and to a lesser extent commercial product launches have been seen, as indicated by recent work by the PTOLEMUS Group.  More sophisticated products are termed PHYD (Pay How You Drive) which monitor not just mileage but also the types of roads used (i.e. urban or rural) and even the level of acceleration and braking. 

Whilst these PHYD products are appealing in terms of the pricing efficiency on offer, the intrusiveness and complexity of the pricing models are unattractive features for many.  In addition, the cost of installing the necessary black box, which might also then tie you to a particular provider, is disliked. 

However, the economics of the investment should change for the not so small community of young women likely to be put at a disadvantage by the end of gender differentiation.  Although these products have not proven particularly appealing to their current target demographic to date the decision to invest money in a black box when faced with a definite price increase will be computed differently by a consumer than a potential gain without such a strong first-hand experience or indeed to the computation of a potential saving to past experience.  The result should be an increase in demand from these consumers.

However, even if I am right, the result is not going to be the mainstreaming of PAYD or PHYD — rather a more solid beachhead in the marketplace.  If manufacturers can find a feature that makes the black box more attractive to (safe) women drivers then all the better. 

Those insurers currently providing female-focused insurance are likely to be particularly affected. Often, these are brands within larger insurers.  These brands do not just offer competitive pricing but also emphasise product features such as child seat cover, handbag cover (i.e. car contents insurance) or confidential 24 hour counselling.  How sustainable this would be going forward is difficult to say (obviously the same pricing would have to be offered to male drivers) but if product features can be found that attract the drivers wanted and repel the less safe, then the pricing may remain lower.  That’s a lot easier said than done.  If pricing is not allowed to be differentiated by brand then such specialist firms are likely to be de-merged if their business models remain attractive.   As always, change will make opportunities for the nimblest operators. 

It’s not just motor insurance

I have focused on motor insurance because I happen to have evidence most readily to hand in that particular market.  However, there are other forms of insurance where gender is used as a rating factor.  Jim Murphy and Manuel Peraita identify the following countries as those that do not currently permit the use of gender as a rating factor:

– Critical illness (Belgium, Cyprus)
– Disability/ income (Belgium, Cyprus)
– Health (Cyprus, Netherlands)
– Accident (Bulgaria, Ireland, Netherlands)
– Long-term care (Cyprus)

One very major area where no-one in the EU seems to have prohibited gender as a rating factor is in life insurance.  Men live shorter lives than women.  As a consequence a man can translate a pension pot into a higher annuity in the UK market than an equivalent woman of the same age.  Estimates of the impact of gender-neutral pricing on annuity incomes for men are a 5+ per cent reduction compared to now (this time men would be the losers — particularly those either without a (female) partner or else whose partner’s pension pot is well below the level of theirs and near to retirement so that the options for compensating action would be limited). 

Will it all simply balance out?  Well, in aggregate, that is possible at least once the market has settled to a new competitive equilibrium.  But clearly many individuals may be worse off — and for those coming up to retirement — there will be very little that they can do about it.

Sources: Europe Economics, November 2009, “Retail Insurance Market Study”; Jim Murphy and Manuel Peraita, March 2009, “Implementation of the Insurance Gender Directive, Provisional Results of Groupe Consultatif Survey”.

 

05:29 PM | Permalink