Hearing of the Oversight and Investigations Subcommittee of the House Committee on Financial Services - Credit-Based Insurance Scores: Are They Fair?
REP. WATT: Credit-based insurance scores are numerical summaries of the credit histories of consumers. The scores are calculated using information contained in a consumer report, information such as past delinquencies, consumer debt ratios and the length of credit. The use of credit-based insurance scores has increased rapidly since the 1990's, and today credit-based insurance scores are widely used.
While common sense tells you that speeding tickets, driving under the influence of drugs and alcohol or automobile accidents should increase automobile insurance premiums, most Americans would probably be surprised to learn that late payments on credit cards can dramatically increase the premiums they pay for automobile insurance. In other words, one's credit history, not one's driving history, is likely to be determinative of the cost of one's automobile insurance. That might be the equivalent to having your driving history determine whether you get a bank loan or determine the interest rate you will pay on your bank loan. The question we need to address is whether this is fair.
Today's hearing is entitled, "Credit-Based Insurance Scores: Are They Fair?" Our objective is to shed light on the growing but often hidden use of credit information in the pricing and underwriting of insurance and to start analyzing, discussing and determining whether that is fair or whether it even makes sense.
A number of consumer and civil rights groups and some states say that it's not fair. They argue that these scores are used to raise premiums, deny coverage for new customers, and deny renewals of existing insurance policies, even in the absence of common sense risk factors, such as moving violations or accidents. They say that the use of credit-cased insurance scores disproportionately hurts young people and minorities.
Some states have already enacted laws or adopted regulations that either ban or restrict the use of credit-based insurance scores. For example, two of our witnesses today represent states -- Hawaii and Washington -- that limit or ban the consideration of credit-based insurance scores in writing automobile insurance. We look forward to their testimony, and I think you will find it interesting.
I've reviewed, for example, the legislative history for the Washington State law and find this interesting quote in their legislative history. "There have been hundreds of complaints, they say, filed in the OIC regarding insurance companies' use of credit scoring for underwriting and rate setting purposes."
"For example, one woman who paid her premiums and never had an accident was told that her premiums went up because her credit rating was bad due to a period of unemployment. A woman who had her insurance premium rates increased by 46 percent even though she paid all her premiums on time discovered that her credit score was low due to a bankruptcy filed by her ex husband. A couple was denied access to reasonable rates because they paid all their bills in cash and therefore had no credit history. There are many reasons for a low credit score that do not take into account individual circumstances or credit worthiness." That's from the legislative history of the Washington statute.
The first federal study on credit-based insurance scores was recently released by the Federal Trade Commission. The FTC was directed under Section 215 of the FACT Act to study whether the use of credit-based insurance scores, quote, "could result in negative or differential treatment of protected classes under the Equal Credit Opportunity Act, and whether such underwriting systems could achieve comparable results through the use of factors with less negative impact," close quote. The FTC study grew out of a compromise between the prospect of an outright federal ban on the use of credit-based insurance scoring, on the one hand, and doing nothing on the other hand.
And I would note that neither one of the two -- both of the members who were responsible for the study on this committee -- the chairman who kind of orchestrated the study as opposed to an outright ban -- and Representative Gutierrez, who I hope will show up here at some point during the course of this hearing because he's on the subcommittee.
The first FTC report focused exclusively on automobile insurance, and while it concluded that credit-based insurance scores are, quote, "effective predictors of risk," close quote, it also found that in three out of four lines of automobile insurance there is, quote, "some", close quote, proxy effect based upon race. While the FTC didn't get to this latter finding until page 69 of the report, I believe that any finding of a proxy effect, however small, should be cause for concern in this day and time.
Several concerns have been raised about the reliability and validity of the FTC's report. One FTC commissioner dissented from the report, noting disagreement with the methodology used to generate the underlying data used in the report because it relied solely on data the insurance industry voluntarily submitted and on publicly available data. The dissent suggested that the FTC could have served insurance companies with Section 6(b) orders to obtain a, quote, "more accurate and complete data set, which would have provided a strong foundation for staff's complex economic analyses," close quote.
Even with perceived shortcomings of the data, the FTC report still concluded that there was some proxy effect from the use of credit-based insurance scores in three out of four lines of automobile insurance. As the dissenting commissioner noted, the study, quote, "still found that credit-based insurance scores have a small effect as a proxy for membership in racial and ethnic groups. Given the incompleteness of the data, it is unclear whether the actual proxy effect might be greater," close quote. Another commissioner's concurring statement to the report concedes that the, quote, "results in today's report are no cause for celebration," close quote, referring to the differences in credit-based insurance scores across racial and ethnic groups.
In short, the FTC's report, the one on automobile insurance that we're considering today, may raise more questions than it answers, especially about whether the use of credit-cased insurance scoring disproportionately impacts minorities.
The FTC is preparing a second report on the impact of using credit-based insurance scoring on homeowners insurance.
Given the serious concerns raised about the validity of the data for the automobile insurance report and the critical importance of the second report on homeowner's insurance, Chairman Frank, Representative Gutierrez and I have requested the FTC to consider using its more extensive authority for the homeowners study to obtain a large and statistically valid data set from insurers. The FTC has advised us that this could take two to three years longer, and one of the things I'll be asking about today is whether this is likely to get a more reliable conclusion or whether it would just take two to three years more to get another study that probably would be perceived as just as unreliable.
Due to the uncertain reliability of using credit-based insurance scores in setting insurance rates, we certainly must proceed with care. This hearing is the first step, certainly not the last, in the process of raising the important questions that need to be asked and in educating ourselves, other members of Congress, and the public about the critical importance of this issue.
In the final analysis, I think it should be clear that neither the FTC report nor today's hearing should deter the states from their traditional role in regulating insurance. The fact that Hawaii and Illinois, or Hawaii and Washington, or Hawaii and North Carolina differ in their legislation and in their regulation of credit-based insurance scores is not necessarily a bad thing. States have historically regulated and controlled insurance and have historically been the so-called "legislative and regulatory laboratories for innovation." State insurance regulators are the best equipped to regulate insurance credit scoring and should continue to do so, certainly until we have a better understanding of the facts and the arguments for and against whether credit-based insurance scoring is fair.
I will now recognize the Honorable Ranking Member, my colleague, Mr. Miller for an opening statement.
BREAK IN TRANSCRIPT
REP. WATT: Thank you so much. Let me just thank all of the witnesses for laying out this issue and kind of setting the framework. Regardless of how you approach it, we need to have this discussion, and it is an extremely important discussion. And I don't think I could have -- we could have had a better panel of witnesses to kind of put out the issues and start our evaluation and discussion.
I'm going to recognize each of the members of the subcommittee for questions in five-minute blocks. And I will recognize myself for five minutes initially. And I kind of want to get right to the bottom of this.
I was struck by Commissioner Rosch's testimony that we need to educate minorities more about how to have better credit or how to get their credit scores up. And I guess my question is, if I got my credit score up, would that make me a better driver?
MR. ROSCH: I think the answer to that is that the commission's report takes absolutely no position whatever with respect to that, Mr. Chairman.
REP. WATT: So if there's not a correlation between my driving, which is what automobile insurance is about, I guess. I don't -- is it? Is that what automobile insurance is about?
MR. ROSCH: Well, it's about that, but it's also about things such as the frequency of claims. I mean, I guess it can be said that there's certainly a -- (crosstalk) --
REP. WATT: But is the frequency of claims related to driving history?
MR. ROSCH: Very definitely it's related to it, but there's not a, as we would put it, a proxy effect involved there. Let me make it clear that what we're talking about here in our report are averages. We're not talking about your particular rates -- (crosstalk) --
REP. WATT: I understand that, but it strikes me as being an -- I mean, I don't think you'll find a stronger supporter of the need for credit education and improving credit histories and credit scores. I think the problem I'm having is what in the world does that have to do with the insurance rates that I pay? Maybe I ought to ask the question this way. Is there some statistical -- has there been a determination that African-American drivers are worse drivers than white drivers?
MR. ROSCH: No, there has not.
REP. WATT: Okay.
MR. ROSCH: There has not, Mr. Chairman.
REP. WATT: All right. I'm just trying to get this on the record so that we make sure that -- that's the base that I'm operating from, and if that is not the case, I don't know how -- what the justification is for basing -- if you know that disproportionately African-Americans and Hispanics have lower credit scores, and you know that there's no correlation between race and safe driving, is there something else I need to know?
MR. ROSCH: Well, let me make it clear what the commission's report said and what it did not say, if I may, Mr. Chairman.
REP. WATT: Yes, sir.
MR. ROSCH: This really cuts right to the heart of your question.
REP. WATT: That's why I'm trying to get to the heart of it.
MR. ROSCH: Okay. The commission's report took definitive positions only with respect to those matters which we felt were completely established beyond per adventure by the statistical analyses that were done. We interpreted your mandate that way. And those statistical analyses showed that there was unquestionably a relationship, a correlation, between credit scores on the one hand and the frequency of claims on the other hand.
REP. WATT: Okay, well --
MR. ROSCH: No. Just to finish up -- (crosstalk) -- Mr. Chairman.
REP. WATT: All right. Go ahead.
MR. ROSCH: We took no position on why that existed, because we did not have a statistical basis for taking a position.
REP. WATT: Okay. That's fair. Let me ask two other questions quickly, because my red light's going to come on, and I try to apply the same rules to myself that -- even more vigorously than I apply them to you all.
I take it, commissioner, that what you're saying is even if you take two more years or three more years to study this issue, the public may deem what you say as having gone through a more methodical process. I guess the question I'm asking is would the end justify the means? Would we have anything better at the end of that two or three years than we have now, based on the way you did this study?
MR. ROSCH: The answer, I think in all fairness, Mr. Chairman, is I don't know. We set forth, in our letter to you and to Chairman Gutierrez and to Chairman Frank what we intended to do in terms of compulsory process the next time around. We also intend to get from consumer groups, like that represented by Mr. Birnbaum, from La Raza, from the insurance commissioners, their input about --
REP. WATT: I understand all that, but I guess one of the concerns I have is that we may be pressing you into a process that you don't think is going to yield a better result. And it might look better to the public that you went through that process.
But if the result is no better and no more reliable in your estimation, I guess I'm beginning to have second thoughts about whether we ought to be pushing you to do a study using a process that you can't verify to me is going to have a better, more reliable, not better, because we're just trying to get to the bottom of this -- more reliable conclusion to it.
MR. ROSCH: Well, I think that's a perfectly legitimate question, Mr. Chairman. That is a decision for this committee and this Congress to make. I cannot sit here and tell you right now that the conclusions, the basic conclusions, particularly that I've described in my opening statement with respect to homeowners' insurance, is going to be one bit different as a result of the use of compulsory process or the input that we receive in the future from these groups.
REP. WATT: Not worried about whether it's different or not, I'm worried about whether it would be more reliable. We're not looking to program the result, but we ought to be able to say, at the end of the day, that the study is reliable. And the question I asked you didn't have to do with whether you were going to change the result or not. The result might be exactly the same. But I don't want to do two or three years waiting for a study that's not going to be perceived as being any more reliable than the study you've already done in a much, much, much shorter and less expensive period of time.
MR. ROSCH: I think that's a perfectly legitimate issue. All I'm trying to say, Mr. Chairman, is that as matters now stand, a majority of our commission feels that the data that we got, the data that we used for the current study, was reliable.
However, we are going to use compulsory process this next time to get the same kinds of data. The problem is that I can't sit here and tell you right now that the quality of that data is going to be any more reliable than that which we got the last time. I'm sorry I misunderstood your prior question.
REP. WATT: All right. Let me ask, just because we are in this line of questioning and I want to be clear, so would you, in the absence of the letter that Chairman Frank, Representative Gutierrez, and I wrote to you, would you have used 6(b) process, or would you have used the same process?
MR. ROSCH: Frankly, I think that I would have voted to use 6(b) process, and I can tell you why I would have voted to use 6(b) process is that even apart from this committee's letter to the chairman, questions have been raised about our lack of compulsory process the last time around.
As I said in my opening statement, we at the Commission feel that having the public have confidence in the way that we prepare our reports is exceedingly important to us. Whether or not that's worth the extra time it will take is a matter for your -- (crosstalk) --
REP. WATT: My time is expired. I'll come back to this on the second round. The ranking member is recognized.
BREAK IN TRANSCRIPT
REP. WATT: One of the great things I've found about being a member of Congress that I didn't find about the practice of law was that when I -- when I was practicing law, I never asked a question I didn't know the answer to already, because you had to live with the consequences. I -- I'm not interested in programming the outcome of the responses that I get, so this is just about making public policy now and getting honest assessments.
I noticed, Mr. Rosch, that you've been very careful to talk about the frequency of claims, and maybe that's a term of art that translates into some other things. I'm interested in knowing whether the frequency of claims has a correlation between dollar amounts, payment of claims, the amounts paid; is that all included in frequency of claims, or is it just the -- what is included in frequency of claims?
MR. ROSCH: Frequency of claims does translate into higher premium -- I'm sorry, total claims paid out.
REP. WATT: Okay, so --
MR. ROSCH: And that does --
REP. WATT: So it is inclusive of -- of more than just the rate at which claims are filed?
MR. ROSCH: That's correct, but what I wanted --
REP. WATT: That's fine. That's not a trick question, I just --
MR. ROSCH: No, no. I understand.
REP. WATT: Just trying to -- trying to understand.
You said that your analysis was limited to policy data. What do you say in response to Mr. Birnbaum's concern that one of the shortcomings of -- of the analysis was that it didn't deal with denials, which probably, possibly, would -- would be disproportionately even greater racially disparate.
MR. ROSCH: That, I think, is speculation at this point, Mr. Chairman.
Let me make --
REP. WATT: But does -- I guess the question -- maybe I asked the question the wrong way. Does your -- does the data you used have the denials in it?
MR. ROSCH: It does not have application data --
REP. WATT: Okay.
MR. ROSCH: -- and let me tell you why, if I may.
REP. WATT: Well --
MR. ROSCH: Let me tell you why, because it's --
REP. WATT: Okay.
MR. ROSCH: -- it really bears directly on whether or not the -- our study was reliable.
The first reason why is because insurance companies, by and large, do not keep application data. And what that meant back in 2001 --
REP. WATT: I actually would be more interested in finding out whether you think that would be a relevant -- I'm satisfied that it doesn't include -- your information doesn't include denials, and I'm satisfied that there are probably reasons why that is the case, very good reasons. Would -- would denial information be an important factor to take into account?
MR. ROSCH: If we could get it, Mr. Chairman --
REP. WATT: Yeah. Okay --
MR. ROSCH: -- but the McCarran-Ferguson Act is -- currently is as much a constraint on the commission as it is on the Congress. This is -- (inaudible).
REP. WATT: Okay. I understand, Mr. Rosch. I'm not being hard on the FTC.
MR. ROSCH: Oh, no. I understand --
REP. WATT: These questions are not -- not aimed at discrediting what the FTC has done. I'm just trying to get to the bottom of this, really. I don't understand how you can take something that appears to me to be unrelated, just -- just common sense, to me. And even if there were a correlation, I'm not sure I would be convinced that it would be appropriate to use something that correlated with race, that had a disproportionate racial content. We outlawed it in the life insurance context. There's certainly a whole wealth of information, body of information that black people live shorter lives than -- than white people, and we said, you just can't -- I mean, you can't -- you can't do that in setting insurance rates, because you've got to do it on a gross basis.
So even if there is a correlation, I'm not sure I buy the notion that we ought be doing this.
MR. ROSCH: Please -- please don't read our report as suggesting that -- that we disagree with that one iota.
REP. WATT: Okay. All right, and I've heard you say that over and over again. That's why I'm -- I'm just trying to get -- I guess I have that bias coming in, whether your study is reliable or not. And at some level, I guess, that's why I'm questioning whether we even need another study, because it seems to me that -- that if there is a proxy effect, even though statistically what you say is that it's a reliable predictor of risk, I'm not sure if it's a fair predictor of what insurance -- how insurance rates ought to be set, I guess is the concern I have.
What do you -- what do you have to say about that, Mr. Kreidler? You all obviously decided, regardless of what the circumstances were, that this didn't -- well, let me ask a different question.
What are the -- what are the insurance companies in Hawaii and Washington state using, if they're not using credit-based scoring?
MR. KREIDLER: Mr. Chairman --
REP. WATT: And what impact does that have on your rate-setting, which might be a better -- better way to get to the bottom of this?
MR. KREIDLER: Well, as you pointed out, you know, Hawaii wound up effectively banning it for auto, whereas we're in a position where we've limited what you can use. The net effect is that I think we've done a better job of limiting the adverse impacts of credit scoring.
Nobody has said it's not a predictor; what we're saying is, is it fair? I mean, if you -- I think your example of life insurance, we dealt with it, but there's also the issue related to lenders and red- lining. Nobody said that if you red-lined out the inner city of -- that somehow you weren't making loans from -- probably more prudently.
The question was was it a surrogate for race, and I think that's the question right here. Is -- is the use of credit scoring a surrogate for race? And if it is, then it should be banned. And we've answered it for life insurance, we've answered it for -- for redlining on lending. I think the FTC came back and showed that yes, there is a disparate impact. And you know, there are some real limitations on the study that they did; not their fault, but because of the data. And I look at it and I say, you know, if this question is as clear as it appears right now, you should take the same approach relative to what you did for life insurance and what you did for relining on lending.
REP. WATT: Commissioner Schmidt, what are you all using, other than credit-based scoring in Hawaii?
MR. SCHMIDT: Yes, we're using driving history, and the accident experience that individuals have. As I noted in my testimony, we not only ban credit bureau rating, but a variety of other factors which have an actuarial basis that show that there are difference in premiums that could be had, based on gender, based on length of driving experience, age, and other factors.
But we do focus on what the actual driving experience and claims history was. During the course of this ban -- Representative Roskam was right that we don't have the -- necessarily the experience of having no credit reports being used and then having credit reports being used to -- to compare it.
But in my testimony, we have banned it for a number of years, and during that course of those years we have had a bad and difficult auto insurance market, based on some of the other laws that we had enacted where we had few companies writing, where we had high premiums, and then we made some reforms that enabled us to drop the premiums from third-highest in the nation to 21st -- one of the biggest drops of any state. And we have a very healthy auto insurance market now, with lower premiums. And through both, we had the ban on credit scoring. So there still are valid criteria for evaluating the insurance rates that companies can use.
REP. WATT: My time has long since expired, and I recognize Mr. Miller.
BREAK IN TRANSCRIPT