Writing at National Underwriter’s PropertyCasualty360 site, Robert Regis Hyle, editor-in-chief of Tech Decisions magazine, sees the future of auto insurance dominated by “telematics,” those advanced modeling systems that incorporate GPS and on-board computers to measure driving behavior.
Pioneered stateside by the likes of Progressive, telematics has more recently been making an impression abroad, as well. As policyholders become more familiar and comfortable with these technologies, Hyle thinks smaller and mid-sized personal auto insurers will have a harder and harder time keeping up:
Soon, it will become impossible for those without this technology to compete for the good drivers of the world with larger carriers. Good drivers have long been saddled with less efficient underwriting tools that lump their rates together with less careful drivers, pushing the rates of better customers higher despite fewer driving mishaps.
To support his thesis, Hyle draws largely from a November 2011 Towers Watson survey of U.S. and Canadian P&C insurance executives, which found that 83% reported the use of predictive modeling tools had a positive impact on their rate accuracy; 76% said they had a positive impact on loss ratios; and 73% said they had a positive impact on profitability.
While personal lines carriers are, at first, simply dipping their toes in the telematics waters, focusing primarily on their ability to track who is driving a vehicle, how far, when and where, the Towers Watson survey shows intense interest in more expansive applications of technology, with 89% reporting plans to use telematics data proactively in rating and 83% reporting plans to use it to help insureds improve their driving.
I happen to agree with Hyle that telematics are the industry’s future, although there remain some significant issues to work out before the technology can be expected to gain broad acceptance. The most obvious ones relate to privacy concerns. The notion of some faceless corporation tracking one’s movements with a computer strikes many, naturally, as deeply disconcerting, and insurers who opt for telematics devices are going to have an uphill climb to prove that the data they collect is secure and will not be misused.
There are also more mundane logistics that remain unexplored. Much of the early benefit insurers report from introducing telematics devices may simply be the result of advantageous selection – those who would volunteer to have their driving monitored are likely to be unusually careful drivers. We remain a long way away from having credible data about how specific types of driving behavior properly match up with rating and underwriting criteria. Regulators are going to insist that insurers produce such data before permitting broad usage, and insurers themselves are going to have to discern whether, when employed a broad scale, the devices provide enough of a competitive advantage to justify their costs.
But there’s one very good reason to be hopeful about the growth of telematics from a public policy perspective: it might, fully and finally, lay to rest the tiresome and, frankly, intractable debate that has waged for decades over insurers’ use of credit scores.
Let’s first be clear. Credit-based insurance scoring has been a wonderful tool that has done a tremendous amount of good over the past 30 years. Among the wholesale changes it has wrought:
- Credit scoring has permitted insurers to segment risks in ways that have contributed enormously to the depopulation of state residual auto markets. Where once only a narrow band of customers could meet the stringent underwriting rules that standard market insurers required (with the rest of population shunted into high-cost, uncompetitive assigned risk markets) credit-based underwriting has allowed auto insurers to compete for virtually every risk in almost every state. Today, the states with large residual auto insurance markets (North Carolina, Massachusetts) are the notable exceptions, where once they were the rule.
- Credit scoring is probably the single most important factor in facilitating the growth of the direct on-line insurance sales market, as it has allowed insurers to construct software algorithms that offer customers instant quotes without the need for an agent or broker to serve as intermediary. This has wrung a tremendous amount of bureaucratic waste from the system, and forced the agent community to respond by offering better service as the only reasonable way to justify their own jobs.
- After taking deep dives into the troves of data that consumer credit bureaus collect, insurers have learned to use that information – in concert with other factors, like age, gender, driving record, claims history, zip code, miles driven, occupation, education level, persistency of coverage and so on – to construct proprietary rating formulas that assign a truly personalized rate for each applicant. In so doing, they have largely rendered irrelevant the old rating bureau system, in which companies would receive recommendations on what customers to accept and what rates to charge from centralized, industry-owned data miners. This has been a remarkable step forward for the industry, and for consumers.
With all that said, credit-scoring has no doubt presented its own set of issues, and the industry has never fully grappled with how to properly address them. Virtually every study on credit-scoring, including independent analysis from the Texas Department of Insurance and the Federal Trade Commission, have come to two primary conclusions:
- Credit-based insurance scores are a credible actuarial variables, and have proven reliably predictive of future claims; and
- The use of credit information in rate-setting and underwriting decisions has a disproportionate impact on minorities and other protected classes.
Naturally, the insurance industry has responded to such studies by emphasizing the first finding. Consumer advocates have responded by emphasizing the second. And because both groups come to the issue with such fundamentally different attitudes and priorities, the ensuing debate is one that likely can never be resolved.
From the insurance industry’s perspective, the fact that credit scoring may have a disproportionate impact on some groups is incidental. Lots of perfectly reasonable rating variables have disproportionate impacts on someone. What matters is whether those variables do a good job in predicting claims. If they do, then it is reasonable to make use of them, because predicting claims is what the business of insurance is all about.
From the consumer groups’ perspective, it is irrelevant that a given variable is predictive if its use hurts minorities and the poor. After all, race itself can often be a credible underwriting variable, and it wasn’t so long ago that insurers explicitly used race in their underwriting and rate-setting decisions, particularly in the area of life insurance. We have as a society determined that race is out-of-bounds for setting insurance rates, and so, if some other variable serves as a proxy for race, it should be similarly off the table.
Now, to be sure, it has never been clear that credit serves as a “proxy” for race. If it did, you’d be able to guess a person’s race from their credit score. Moreover, the observation that consumers with better credit are less likely to file claims than consumers with poorer credit holds WITHIN racial groups as well: good credit whites file fewer claims than poor credit whites, good credit blacks file fewer claims than poor credit blacks. And so on.
It is also not entirely clear that the so-called “objective” factors that consumer groups would prefer insurers use, such as driving record, are any better on the topic of racial bias. Numerous studies, such as this 2003 investigation from the Boston Globe, have found huge disparities in how traffic officers treat similar driving infractions based on the race, age and gender of the driver. In brief, Massachusetts police who stopped motorists for speeding gave minority men a ticket 62.4% of the time and a warning 37.6% of the time. They gave white women a ticket 44.7% of the time and a warning 55.3% of the time. Of course, only the tickets will show up on the driver’s record, and thus be reflected in their auto insurance rates.
Given these figures, it’s ironic that driving record is the rating factor most often trumpeted by consumer groups as objective and neutral, to the extent that in states like California, it is practically the only underwriting variable insurers are permitted to use. It’s hard to see what role racial bias would play in keeping an auto insurance applicant from paying his or her bills on time. It’s not hard at all to see the role it could play in adding a few points to his or her license.
But from a public policy perspective, the real difference between driving records and credit scores is not their actuarial value, or even the extent to which they have an impact on one group or another, but the narrative that attaches to them. A layman finds it easy to understand why someone with speeding tickets should pay more for auto insurance than someone with a clean record. It feels equitable. It feels fair.
It is less easy to understand why someone with a bad credit history should have to pay more. In the public imagination, the one thing has seemingly nothing to do with the other. And insurance companies have failed remarkably in constructing any sort of narrative that would convince people otherwise.
The best attempt the industry has mustered is to focus on “responsibility.” In other words, it may simply be that the sort of person who pays their bills on time and doesn’t take out too many loans is also the sort of person who is likely to maintain a proper speed and avoid distractions on the road. And this may very well be a good part of the explanation for why those with better credit scores file fewer claims. But it FEELS flimsy. Most of us can immediately think of too many anecdotal examples to the contrary: the thrifty spender who is also an absent-minded driver, or the extremely cautious mom who just happened to fall behind on her bills.
And the truth is, the insurance industry doesn’t actually KNOW why people with good credit scores are good insurance risks. Maybe it just so happens that, for whatever reason, those with better credit tend to have less treacherous commutes, live in less congested areas or face less risk of auto theft or vandalism. Maybe those with better credit are wealthier and are thus more able to pay small claims out of pocket; maybe they don’t file as many claims simply because they’re less likely to report them. Insurers know that credit scoring works and, more or less, that’s the end of the story for them. It isn’t nearly as important to come up with a definitive reason for why it works.
Now, it is also true that there are any number of seemingly irrelevant data points that actuarial tables, in fact, show are relevant for underwriting purposes. The entire market of “affinity” insurance programs is built on them. The Hartford, for instance, has an exclusive contract to provide preferred auto insurance rates to members of the AARP. USAA built its business by marketing to members of the armed services and their families. GEICO, which now offers discounts to people who work in a variety of occupational fields, got its start as the Government Employees Insurance Company, specifically targeting federal workers in and around Washington, D.C.
There are no obvious reasons why federal workers, military personnel or retirees should have better driving habits than the rest of us. But apparently, they do. Or at least, enough of them do to make such discount programs sustainable over a fairly long term.
Perhaps insurers could have swayed public perception of credit scoring long ago with better framing; if they had, for instance, marketed good credit “discounts” much like the good student discounts that many companies offer. But thoughtful public relations has never been one of the stronger suits of an industry characterized in the public imagination (fairly or unfairly) as run by dispassionate numbers-crunchers. In any case, the public’s reaction to the entire notion of credit-scoring has been overwhelmingly negative, and the public policy fights the industry must perpetually wage to ensure its continued use might be proving more costly than they are worth.
Telematics offers a way forward. The technology provides the potential for objective, personalized evaluations of the risks presented by each and every driver, along with the ability to provide real-time advice to help insureds curb the risks they are taking. To a layman, it certainly feels fair. The Big Brother concerns that undoubtedly will greet the technology’s growth will almost certainly fade over time, much like the introduction of blogs, video sites and social networking has led people to share details about their lives that would have horrified earlier generations.
At this point, it seems inevitable: the future belongs to TomTom and its ilk.