Posted by Guest Blogger Extraordinaire, Ian Turvill, of Fair Isaac
A description of an innovative insurance offering crossed my desk the other day, and I felt that I had to reflect on its EDM-related implications. In a report issued on June 15, 2006, entitled "Pay-As-You-Drive: Dynamic Insurance Emerges in Europe", written by Ricardo Arruda of Forrester Research. In it, he describes the growing demand for technology-enabled methods of insurance pricing, which charge consumers according to the speed, location, and timing of their driving.
Ricardo summarizes it like this:
"Pay-as-you-drive schemes rely on GPS technology and mobile phone networks to track individual car usage and determine tailored motor insurance premiums. Insurers were initially reluctant to launch this product, as it challenged establish pricing models and involved high implementation costs. But pioneers like Progressive Casualty Insurance showed how pay-as-you-drive schemes could yield significant business and customer benefits. Led by the largest insurers and government authorities, pay-as-you-drive schemes are now appearing in European insurance markets from UK to Italy. These emerging pay-as-you-drive schemes will create a dynamic insurance market that gives consumers greater control of their own premiums and sets insurers powerful new challenges and opportunities."
In many ways, "Pay-As-You-Drive" (PAYD) Insurance is the logical extension of the growing use of micro-segmentation that many US insurers are already offering. This involves the use of a far broader range of variables, in conjunction with custom predictive analytics, to precisely rate the risk presented by individual consumers. The difference here is that, in addition to the static measures of risk, such as the driver's age, driving history, or the commuting distance, they are using dynamic measures, such as speed, time of day, and location, to give the best possible overall assessment. Under PAYD, insurers may not have 4 or 100, or even 4,000, underwriting bands: in effect, they have a band for every single policyholder they serve.
Ricardo argues that PAYD brings multiple benefits to drivers, insurance companies, and even to the public at large.
- Drivers are likely to perceive the insurance rates as fairer, since their payments are tied directly to their usage, and because they now gain control over the amounts paid for insurance - just as they have some influence over the level of gasoline consumption. Ricardo indicates that drivers operating under a PAYD scheme tend to drive less and pay up to 25% less in insurance than they did under a prior methods calculating rates. He doesn't, however, explain the basic underlying economic rationale for this, which is that insurance changes from a "fixed expense", all paid at the outset of the agreement, to a "variable expense". Thus, the cost of insurance becomes part of the "marginal cost" that driver faces when he/she sets out from their garage. And if their marginal cost is higher, while the marginal value of a trip remains, then - on the margin - car owners will drive fewer miles. See, Economics 101 was useful, after all!
- European insurers, according to Ricardo, profess that their customers become much more loyal under a PAYD underwriting method, because they sense a far higher level of price transparency. He also suggests that the data they collect on customers' driving patterns could be valuable in the development of new products or even in the creation of marketing and advertising programs. He appears, however, to overlook the most fundamental tenet of insurance: for every risk, there is a price. When insurers can more precisely charge premiums according to risk, they get a double whammy. First, they can offer lower prices to the people who deserve them, and who might otherwise be charged higher prices elsewhere. As a result, they can win greater share. Second, they can charge higher prices to people who actually do present a higher risk. Two things will then happen: either the customers attrite (which is fine, because the insurer was in reality losing money on them before), or the customers stay and make higher payments (which is, of course, also just dandy).
- We have already noted that car usage tends to drop when insurance payments are made "as you drive". Some - particularly environmentalists - would say this is, of itself, a good thing. But, in addition, PAYD systems can help governments better ration the use of the road. Since the risk of accidents is correlated with the volume of traffic, and therefore PAYD-rates are set higher during those periods, there is a disincentive for PAYD-equipped customers to drive at times of heavy road usage. Moreover, the very same system that powers PAYD insurance also makes PAYD road charges possible. For governments, such as in the UK, which intends to enforce fees for every mile a motorist travels by 2010 (yes, really), this means that they could simply piggyback off commercial systems.
So now we've established that PAYD has some good things going for it, let's think a little about the technical implementation. Ricardo emphasizes some of the basic sensing and tracking mechanisms that have to be used to make this approach work.
"A 'black box' is installed beneath the hood of a car and receives signals from global positioning system (GPS) technology to determine the vehicle's current position, speed, and time and direction driven. The black box then acts as a wireless modem to transmit theses inputs through standard mobile phone networks to the insurer. The insurer then processes the information in an operations center and can pass it on to the client via the Internet."
Once the information reaches the insurer, there seems to be a little case of "And then the magic happens" to this description. How exactly does the insurer process this information?
Well, I'll tell you: a Business Rules Management System.
The insurer can set up rules that define the level of risk a client presents, based on their speed, location, and time of day. These can be further complemented by conventional rules and analytics that dictate the driver's "static" level of risk.
The insurer can apply the BRMS to the incoming stream of client data to generate exact calculations of payments due and communicate this to billing systems accordingly.
No existing rating engine or policy administration is equipped to do this. Only a rules engine, with appropriate access to analytics and connections to the necessary data, has the flexibility and power to deal with rapidly evolving PAYD rules and the volume of calculations required. Yet one more application for Enterprise Decision Management!
A postscript: It's interesting that European consumers appear to be very much more amenable to this "Big Brother" kind of intrusion than those in the US. I have two alternative hypotheses why this may be the case.
The first - which seems a little prejudiced - is that Europeans are simply more pliant than Americans, and that they are more willing to have someone watching over their shoulder. While I have no doubt that Yanks are generally very sensitive to monitoring, I really don't think that Europeans are truly any less so.
The second hypothesis is more interesting from a public policy point-of-view, and it is that American companies may be shooting themselves in the foot by lobbying against stricter consumer data protection laws. The European Union is known to have quite stringent rules for maintaining customer databases. Did you know, for example, that in the UK even the local chapter of the Salvation Army would have to record its membership data with the country's database registrar in order to be fully compliant with their consumer data statutes?
Is it possible that "Old World" consumers are more willing to be monitored because they have a stronger belief that their data will not be misused. By contrast, in the US, imagine what fears would go through consumers' heads when offered a GPS unit to sit in their cars and track their every move?
It will be interesting to see Pay-As-You-Drive Insurance remains primarily a non-US phenomenon, or if it also takes off in the States, because the fundamental appeal should be the same, even if the policy and cultural environments differ.