A few days ago saw the FTC issuing its report on insurers use of credit scores, finding nothing nefarious going on.
I wrote at the time that I was surprised at how similar the FTC study sounded to some industry claims, and I expected consumer advocates to make a bit of a fuss.
Here it comes. A press release, via Insurance News Net:
Representatives of the Consumer Federation of America, the National Fair Housing Alliance, the National Consumer Law Center, and the Center for Economic Justice said the FTC study is fatally flawed because the insurance industry controlled the data used in the analysis. Instead of requiring the submission of comprehensive policy data by a large number of insurers, the FTC used data handpicked by the insurance industry[...].
The FTC study also confirms that, despite growing reliance on credit-based insurance scores, scant evidence exists to prove there is a meaningful connection between a consumer’s score and auto insurance losses. Without the need to demonstrate such a connection, insurers could use any consumer characteristic, such as hair color, to price insurance products.
“Despite finding no explanation for the alleged connection between insurance scores and losses, the FTC report somehow concludes credit scoring is valid and good for consumers. This is not an impartial analysis, but simply advocacy for insurers,” said Birny Birnbaum of the Center for Economic Justice. Birnbaum, a former insurance regulator, has studied insurance scoring for over 15 years.
One of the fundamental differences in opinion between insurers and the consumer advocates who “love” them is the extent to which causality (or at least intuitive correlation) must exist in rating or underwriting.
I hate to point out that if a stringent test were applied as regards causality, many of the variables we’ve traditionally used as an industry could be forced by the wayside, even though strong relationships can be measured to exist between these variables and future losses.
We could ignore those relationships, but then those insurance products become less attractive to write, which in turn reduces competition and drives up rates for the public in general, over and above the effective discounts many folks see because we are able to identify them as better-than-average risks.
CL&P Blog reports that the National Consumer Law Center has released a report that in part slams the use of credit data by insurers.
Considering my background, it should be no surprise that I take issue with a few of the claims made…and which have been repeatedly made, despite the fact that they are unsupported or false.
(My background includes working with the research into, implementation of, and education about credit scoring at a personal lines insurer. That was an assignment that I took on partly because I initially had problems believing that credit reports and insurance experience could be correlated. Boy did I learn differently.)
Quoting the CL&P post (and with a disclaimer — the NCLC website is down as I draft this post, so I can only respond to the points raised at CL&P, rather than in the study itself):
The NCLC report includes an analysis by Birny Birnbaum of the Center for Economic Justice on how insurance credit scores cost consumers billions of dollars every year, potentially in the neighborhood of $67 billion. According to Birnbaum, “Insurers’ use of credit scoring - the introduction of many, many rate levels based predominantly on the insurance score - has contributed to a dramatic increase in profitability.”
You know, once upon a time, profit wasn’t a dirty word. The ability to acquire capital to help underwrite the risk and to build surplus sufficient to pick up the tab in the event of a catastrophe is one of the reasons that the insurance industry works.
True, there are various ways that an insurer can be organized which increase or decrease the demands on what an acceptable long-term rate of profit is, and there are pros and cons associated with those different vehicles… but at some level, the ability to attract and develop capital is one of the requirements for a healthy insurance market to exist.
That being said, I suspect that Birny’s point is a non-sequitur.
In prior decades, personal lines insurance (i.e., personal auto and homeowners) sucked. Insurers routinely paid out $1.02, $1.05, or more in losses and expenses for every dollar of premium they collected. They relied on high rates of return from relatively risky investment, or even accepted a loss in the hopes of wooing agents to write other, more profitable lines of business.
With a couple of downturns in the market, new requirements on the amount of risk tolerated in insurers’ investment portfolios by regulators, insurers finally realized that they needed to learn how to make their products sufficiently profitable, or get out of the market.
The rise of easy, relatively cheap access to computing power, new statistical tools, and a bumper crop of actuarial students gave rise to several innovations in ratemaking and underwriting, including new multitiered programs and the use of insurance scores.
It is correct to say that insurers profits rose as credit scoring was introduced. It is also probably correct to attribute the process which included the adoption of scoring to being a major reason for personal lines insurance profits rising to levels commensurate to the risk they represent.
However, it is completely wrong to imply that “credit scores” = “PROFIT!”
When properly implemented, the average premium in a program that relies on scoring should be the same as the average premium in a program that doesn’t rely on scoring, all other things being equal. True, some folks will pay different rates in the programs (about 2/3rds will pay less, and about 1/3rd will pay more). But on average, the adoption of credit scoring should be revenue-neutral to the insurance marketplace as a whole.
Insurers like scoring because it provides a way to identify individuals who have patterns of low-risk behavior, and therefore to offer a rate commensurate with that low risk. If we can’t use scoring, that’s fine — our lowest rates won’t be quite as low, and our highest rates won’t be quite as high, but we’re still going to try to get the same level of profit regardless.
Other issues addressed in the NCLC report include:
Consumers lack information about the use of credit scoring in granting or pricing insurance, with only 36% in one survey indicating they know about the practice.
I agree that the level of financial education in this country is abysmal. However, you’d think with all the public service announcements and informercials on the subject, folks would know by now that the permanent record your teachers always threatened you with can come back to haunt you.
Some of the factors used in insurance scoring models are questionable, such as penalizing consumers with fewer than 2 credit card accounts.
I fear that the author is not familiar with some of the principles of scoring.
There are two relatively easy ways to explain that phenomenon:
- The correlation between number of credit cards and future losses is such that folks with 3-4 cards really are less risky than those with 1-2, who are still far less risky than those with 14 or more cards. (Google unfortunately fails me in trying to find some publicly disclosed numbers in support of that. However, if you want to see some other univariate relationships, I recommend this paper by Jim Monaghan.)
- You’re looking at just one piece of the puzzle, and trying to fit the interrelationship of variables in the scorecard. For example, saying that you’ve never had a late payment on any of your credit cards is far more impressive when you have a wallet-full of plastic.
Besides, directing a reader’s attention to one particular variable is somewhat disingenuous given that it’s a variable that doesn’t appear in several insurers’ models. You can see, for example, Progressive’s model starting on the 313th page of this pdf of a Florida rate filing.
Credit scoring in general has been criticized for high rates of errors and data inconsistency between the major credit bureaus.
Considering the purposes for which credit data is being used these days, you’d think that there would be more emphasis placed on having entities that report data to the credit bureaus do so completely and accurately, rather than simply banning the use of that information.
Heck, the fact that scoring works so well in spite of the noise in the data is a testament to just how powerful it is. If credit bureau data were cleaned up, I can only imagine how much better the models that were built could be.
The “high rate of errors” comment is also a little misleading in that there’s no mention made of the rates of certain kinds of errors. Yes there is certain data that is known to appear in credit files and be incredibly prone to error — employer names and birthdates are the “best” examples I can think of off the top of my head — and therefore they aren’t used in scoring algorithms.
Besides, I don’t hear too many consumer advocates complaining about insurers’ use of MVR data (driving records), yet the quality of that data is crappy beyond belief.
The insurance industry alleges that credit scoring is predictive in forecasting which consumers will have higher losses yet they have not offered a satisfactory explanation as to why this correlation exists.
Here we have credit scoring being held to a different standard than other, less controversial rating/underwriting variables. Insurers aren’t generally obliged to explain the correlation, just to prove that correlation exists and isn’t being double-counted with other elements of a rating and underwriting plan.
Think about it — have you heard an insurer give a concise-yet-complete explanation for the correlation of (say) good student discounts to future loss experience?
USA Today is carrying an article today highlighting Measure 42 in Oregon.
On Nov. 7, Oregonians will become the first voters in the USA to decide whether to bar insurers from setting premiums based on such factors as credit history, debt load and bill-paying habits.
The insurance industry, which opposes the measure, is pumping millions of dollars into an ad campaign to defeat it. The outcome will be closely watched by other states that could come under pressure to take similar steps if the Oregon ballot measure succeeds.
That sets the general tone of the article, which seems decidedly biased in favor of Measure 42.
There’s two passages in particular that merit comment:
The industry argues that eliminating credit-based scoring would likely mean that people with good credit would end up paying higher insurance rates. But [Consumers Union attorney] Garcia notes that in California, insurance rates have dropped since the use of credit scores was banned.
The thing is…credit scoring has never been permitted in California auto insurance rates, under Prop 103. Recent rate drops have been more a result of regulatory prodding to get the industry to recognize recent trends and reduce rates despite their reluctance to do so in a challenging regulatory environment.
Insurers also argue that people with low credit scores are likelier to file insurance claims. “People who manage their finances well tend to also manage other important aspects of their lives responsibly, such as driving a car,” the Insurance Information Institute says.
Consumers Union says there’s no proof of that. A review of how credit scores are used to set rates in Texas found that the scores have more to do with economic status than with personal responsibility, says Birny Birnbaum, a former Texas insurance regulator who is executive director of the Center for Economic Justice, a consumer advocacy group.
The correlation between scoring and future losses has been demonstrated in hundreds of rate filings around the country and several published studies, including this one commissioned by the Texas Department of Insurance and performed by the University of Texas.
From Insurance Journal comes this head-shaking bit of news:
On March 27, CFA and CEJ wrote to the National Association of Insurance Commissioners raising questions about recent upgrades in the RMS wind models that the groups maintain would lead to “unjustified increases in homeowners and other property casualty insurance rates.”
The letter, signed by CFA’s J. Robert Hunter and CEJ’s Bernie Birnbaum, called for state regulators to increase regulation of RMS and other third-party organizations, including credit-scoring firms, whose work impacts insurance rates and availability.[...]
“RMS has become the vehicle for collusive pricing,” the letter charges, because it relied upon opinions from insurers to switch to the medium term five-year view of risk.
CFA and CJE allege that insurance companies sought this move to justify higher rates and this desire put RMS in a competitive bind.
I don’t know about you folks, but I know that my company’s finance guys, the folks who negotiate reinsurance purchases, and marketing staff (underwriters, for my area of interest) aren’t looking forward to the upcoming model revisions that will apply rate pressure, make reinsurance more expensive, and increase capital requirements to maintain ‘A’ ratings.
I think that there is room for debating the subject of whether decades-long cyclicality should be considered in the insurance world. The consumer advocates do imply a good point — the answer shouldn’t be “whichever is higher”.
Of course, the consumer advocates perhaps need to be reminded that the answer should also not be “whichever is lower’.
It would be nice if the consumer advocates could get their act together and perhaps actually propose something constructive, rather than just pointing at news releases and screaming.
