Legislature(2003 - 2004)
03/29/2003 09:35 AM House STA
| Audio | Topic |
|---|
* first hearing in first committee of referral
+ teleconferenced
= bill was previously heard/scheduled
+ teleconferenced
= bill was previously heard/scheduled
HB 47-INSURANCE DISCRIMINATION BY CREDIT RATING HB 5-INSURANCE DISCRIMINATION BY CREDIT RATING Number 0040 CHAIR WEYHRAUCH brought before the committee the following two bills: HOUSE BILL NO. 47, "An Act prohibiting discrimination by credit rating or credit scoring in certain insurance rates; and providing for an effective date"; and HOUSE BILL NO. 5, "An Act prohibiting discrimination by credit rating or credit scoring in insurance rates; and providing for an effective date." Number 0350 SAM SORICH, Vice-President, Western Regional Manager, National Association of Independent Insurers (NAII), told the committee NAII has about 100 of its member companies doing business in Alaska; they account for 59 percent of all auto and homeowner's insurance written in the state. He continued as follows: The business of insurance is unlike other businesses, and the pricing of insurance is different than the pricing of a manufactured product. When a store sells a television set, the store know how much ... labor it took to build a TV set, and it knows what the cost of the parts are. When an insurance company sells an insurance policy, we don't know ... the ultimate cost that we will incur. We don't know whether the homeowner and driver will have a loss and, if the loss occurs, what the cost of that loss will be. Now, I guess an easy answer to that is, "Just estimate the cost and charge everybody the same rate." But that would be unfair. ... Good public policy demands that the rate a person pays should be commensurate with ... the person's risk of loss. And people with a low risk of loss should not be forced to pay more than they should to subsidize people with a higher risk of loss. Number 0486 MR. SORICH continued: Insurance companies determine loss by looking at factual data, and the facts show us that a driver with past accidents is more likely to have future accidents than a person with a clean driving record. And the facts show that older homes generally are more likely to have a loss than newer homes. But if we stop there, we wouldn't be doing our best to make sure that the rates we charge are equal to the risk of loss. So, over the years, insurance companies have introduced various characteristics that [try] to get us closer to the goal of linking the risk of loss with the premiums we charge: the age of the driver, the driver's gender, where a car is driven, what type of car, the construction of the home, how close a home is to a fire station. Those are all characteristics that insurance companies use to get closer to the goal of predicting risk. Some insurance companies are now using credit information along with these other characteristics. These companies are convinced that the use of credit gets them closer to achieving a higher level of fairness for their customers. MR. SORICH said Mr. Lo would be talking about the statistical research that shows the strong relationship between credit information and risk of loss. He noted that his own e-mailed statement shows a strong correlation between credit-based insurance scores and risk of loss; also in his statement is a study issued by the University of Texas three weeks ago, a study that the insurance industry had nothing to do with and didn't pay for. He related the conclusion of the study as follows: The research team ... tested whether the credit score for the named insured on a policy was significantly [related] to incurred losses for that policy. It was determined that there was a significant relationship. In general, lower credit scores were associated with larger incurred losses. A regression analysis of the relative loss ratio on credit score was highly significant. This indicates there was less than a 1-in-10,000 chance that the relationship observed between credit score and relative loss ratio could be due to chance alone. MR. SORICH said insurance companies have an obligation to their customers to consider this evidence; forcing insurance companies to disregard this evidence would lead to pricing inequity, which wouldn't be fair. Number 0683 MR. SORICH brought up two issues he'd focus on: the legal basis for insurance companies' use of credit information, and why insurance companies use credit information. He noted that Congress passed the Fair Credit Reporting Act in 1970, which specifically states that insurance companies may use credit reports for underwriting and rating. The federal law does allow states to enact laws that regulate insurers' use of credit information, he said; however, no state has enacted a law that would go as far as HB 5 or HB 47. He emphasized that no law enacted by the states completely prohibits the use of credit information. He noted that there is a question as to whether federal law would preempt a state from absolutely prohibiting the use of credit information because it's inconsistent with federal law. MR. SORICH offered the following five reasons that insurance companies use credit information. First, credit-based insurance scores help insurance companies reach a higher level of objectivity. Subjectivity and bias are removed by the use of insurance scores, which don't consider a person's race, ethnicity, gender, income, home, or religion, for example. Second, insurance companies provide another source of information. Insurance companies are data-driven and only as good as the data they use to make decisions. He said insurance companies use motor vehicle records and past claims history, for example. The availability of credit reports provides insurance companies with another source of information that allows them to make fairer underwriting and rating decisions. MR. SORICH said third, the availability of insurance scores helps insurance companies to become more efficient; rating and underwriting decisions can be made more quickly, which is good for consumers who are shopping for insurance and helps insurance companies keep their costs down. Fourth, the use of credit- based insurance scores allows insurance companies to reach a higher level of fairness. Mr. Sorich said there is a strong relationship between credit information and risk of loss; certain aspects in a person's credit report are predictive of whether that person will have a loss under his/her homeowner's or car insurance policy. If insurance companies are forced to ignore that, he warned, many consumers who would be paying lower premiums because they have good credit history - and most consumers have good credit history - will be forced to pay higher rates than they should be paying. MR. SORICH said fifth, insurance scores help insurance companies to make insurance more available. The availability of this information can convince a company to write a policy for somebody with one claim on a policy, for example; the company would have the assurance of knowing that the person has a good score and is less likely to have a loss. Number 0920 MR. SORICH surmised that the committee will hear criticisms of insurers' practices as they relate to use of credit information; some will be valid. Mr. Sorich suggested looking at the model Act developed by the National Conference of Insurance Legislators (NCOIL), which is a group of state legislators, not insurance companies. He noted that NCOIL's model Act, adopted in November 2002, addresses many of the criticisms leveled at insurers' practices as they relate to insurance scores. MR. SORICH opined that the absolute prohibition proposed in HB 5 and HB 47 should be rejected. He said insurers' use of insurance scores today is providing benefits to thousands of consumers; prohibiting the use of credit information would deprive those consumers of the benefits they're getting today. Number 1059 EDDIE LO, Insurance Manager, Fair Isaac, offered a visual presentation accompanied by handouts. He explained that Fair Isaac, as a pioneer and modeler that has served over 350 insurers in the United States, has put together premium and loss information from insurers and credit information from credit bureaus or consumer reporting agencies on an individual policy basis. In response to a question by Chair Weyhrauch, he clarified that Fair Isaac is not a clearinghouse for insurance companies to use, but is a modeler and software developer. MR. LO explained that Fair Isaac looks at data, builds a model, and then puts forward the model in partnership with three major credit bureaus. Those bureaus pull credit data and run Fair Isaac's model, as well as work with insurers on a daily basis. The insurers provide a name and pull the credit report so that the Fair Isaac score can be calculated by its partners and any questions about the credit report can be answered. Fair Isaac is there in the background, but can answer how the model was built and how credit information was used. MR. LO explained that there are two predictors. The first is the credit score, primarily developed to predict the likelihood on loan repayment for banking and financial services; that isn't what is being discussed today. The second is the insurance score; for that, still using credit information, there is a "whole family of predictors" for personal, auto, and homeowner insurance. That is what is being talked about now. Number 1278 MR. LO discussed charts on page five of the handout. He referred to a chart that shows the number of adverse public records, including foreclosures, judgments, and liens, for example, relating to losses that happened in the past. It shows 96 percent of the people have no adverse record. For the remaining 4 percent, an underwriter can rely on credit and see that expected losses are 54 percent higher than for the rest of the group; thus for those 4 percent there's a very clear distinction based on credit information. MR. LO referred to the next chart, which shows the months since the most recent adverse public record. He said, "Within the last four years, the loss ratio relative to the rest of the group that does not have such adverse public record ... is 68 percent." He explained that loss ratio is the ratio of the losses in relation to the premium collected; when there are more losses, the loss ratio will be higher. He noted that 96 percent of people have zero losses; however, if a person does have "one reason" within the last four years, the loss ratio is very clearly [looked] at from the past, using actual losses and data. Mr. Lo said after four years, a person's loss ratio will improve; for example, it could drop down to 23 percent higher than average, which is still worse than average but is an observable difference. He said adverse public record is a strong predictor, based on data and observation. Number 1416 MR. LO referred to another chart showing trade lines that are 60 or more days delinquent in the last [24 months]. He said it has been shown that the more delinquent a person is, the worse the actual losses have been and thus are expected to be in the future. He noted that 89 percent of people have no such delinquencies in their accounts. In this example, he said, there is a clear trend that if a person has one [delinquency], the expected loss ratio is 29 percent higher; if a person has two or more, it is 80 percent. Mr. Lo highlighted a handout [The Trans Union credit report] that has a sample credit report. MR. LO referred to another chart that shows a graph regarding collections, which are shown in the [Trans Union credit report]. He explained that when an account is seriously in arrears, it is turned over to a collection agency that then puts forward a record on the credit report. Mr. Lo said 97 percent of people have no collections; however, for the 3 percent who do, their loss ratio is a significant 69 percent higher. Number 1488 MR. LO referred to a chart that shows the number of trade lines opened in the last 12 months. He noted that 60 percent of people have not opened up new accounts in the last year; however, for the percentage who have, there is an indication that actual losses were higher and thus future losses will be higher as well. MR. LO said [the charts] show that individual credit characteristics can predict future losses, and can help an underwriter to use credit information hand in hand with traditional underwriting practices. He emphasized that credit [information] should not be used on a sole basis to make underwriting decisions. Noting that the report can show [scores] ranging from 70 percent worse to 50 percent better, Mr. Lo characterized that range as a very helpful hint to the underwriters. He said credit information is also helpful regarding auto insurance, for example. He concluded by saying credit information is predictive of future losses and can be verified independently. Number 1662 SARAH McNAIR-GROVE, Property/Casualty Actuary, Division of Insurance, Department of Community & Economic Development (DCED), informed members that approximately a year ago the division was asked to undertake a review of the insurance industry's use of consumer credit history for rating and underwriting purposes in "personal lines." Based on time and resources, the division did two things. First, it sent out a survey to all insurers in Alaska that wrote personal lines: personal, auto, and homeowner's [insurance]; the survey would provide a broad overview of how those insurers were using credit information. Second, it selected three individual insurance companies to be studied by a team of analysts that would gather more information than was garnered by the survey; those studies, called "market conduct" examinations, are not yet complete. Thus the handout before the committee is just based upon the results of the survey. MS. McNAIR-GROVE noted that one of the most significant issues is whether the use of credit information is unfairly discriminatory. She said providing a definitive answer based upon insurance company data is difficult, if not impossible, because insurers don't collect the type of demographic data needed to answer the question; they don't collect information of income, ethnicity, divorce status, or, in some [personal] lines, age and sex data. Therefore, the division tried to come up with a proxy that could be used; the best proxy it found was zip code data. She said although there are a lot of pros and cons regarding zip code data, it was what was available to the division. MS. McNAIR-GROVE explained that because insurers use credit history in different ways, the division had to find some way to aggregate all the insurers' data in order to make it public. However, the division didn't want to make public the marketing practices of individual insurance companies, and had made an agreement with the companies to that effect. The division also had to find a proxy for what it could use as credit history in order to combine "all these different ways of using it." The decision was made to look at data by market or by tier - terms she said she'd explain shortly. MS. McNAIR-GROVE discussed strengths and weaknesses of the survey approach. She listed the strengths as follows: it allows the division to obtain data from the entire market; provides the division a way by which the data can be combined and made public; and provides a limited means by which the division can begin to address how credit impacts individual policyholders. The weaknesses she noted as follows: it doesn't look at credit history directly; doesn't use the issues that people raise, for example, income and race; and doesn't include other rating factors that might impact some of the results. Number 1862 MS. McNAIR-GROVE turned to some of the results that the division found. First, she defined a credit score as a number that is calculated from a mathematical algorithm or a computer model that's based upon credit information taken from a consumer's credit report. A credit score was first used in Alaska in 1994 for underwriting purposes and in 1998 for rating policies, she noted. She explained that underwriting is the process by which an insurer decides whether or not to offer an individual a policy. Rating is the process by which the insurer decides to offer a policy and then determines what to charge for it. MS. McNAIR-GROVE explained that the division looked at [U.S.] census data and asked, "With these zip codes, what is the income and what is the ethnic composition of those zip codes?" The best available proxy for credit information was "market tier." The division had to make some assumptions, based upon the information that the insurers provided and their own characterization of the type of business they did, whether it was a preferred, standard, or nonstandard business. MS. McNAIR-GROVE defined preferred business as consumers who present the lowest risk to an insurer; it generally includes policyholders with good credit history. Standard business is consumers who present an average risk to an insurer; it generally includes policyholders with average credit history. Nonstandard business is consumers who provide the highest level of risk to the insurer; it generally includes policyholders with poor credit history. She emphasized that market determination is not made entirely by a consumer's credit history; most insurers also use loss history or prior insurance, for example, and a consumer might be in the nonstandard market for reasons other than credit history. Number 1962 MS. McNAIR-GROVE said with the information gleaned from the census and zip code data [shown on graphs in a handout labeled "Insurance Credit Scoring in Alaska"], the division found the following: zip codes that are "predominately Caucasian," with the highest income, tend to have the highest percentages of preferred policyholders and the lowest percentages of nonstandard policyholders. Zip codes with larger percentages of non-Caucasians and lower incomes tend to have smaller percentages of preferred business and larger percentages of nonstandard business. The question that arises when looking at this information is whether this is caused by credit history. MS. McNAIR-GROVE said, in an effort to answer that question, each insurer in the survey was also asked to provide data for the year prior to when the credit history was used so that the division would have a means of comparison. Because insurers began using credit history at different points in time, the comparisons aren't "100 percent good," but she offered her belief that there is some validity to the information. She noted that there were some data limitations, because some of the years in which insurers first starting using [credit history] were fairly [long ago] and it was difficult for those insurers to retrieve the data; other insurers didn't segment their business by preferred, standard, and nonstandard before they used credit [as a predictor]. MS. McNAIR-GROVE reiterated the information regarding the findings based upon zip codes, ethnicity, and income levels. Notwithstanding that, she said, there are differences, the largest appearing to be in the nonstandard market, where a smaller percentage of businesses are classified as nonstandard after insurers have used credit information. MS. McNAIR-GROVE said a person can't jump to the conclusion that credit information isn't a good thing for consumers, based on that information. She explained that all of the other factors are not static. When insurers included credit information in their underwriting or rating practices, they also changed a lot of other factors. For example, they may have changed the number of claims it took to be in the preferred market versus a standard one. Credit, she clarified, is not the only thing that changed over this period of time. She said there is also some evidence that policyholders aren't even getting into the system. She noted that the assigned risk market has been growing significantly since 1999, which was about the time that credit became a predominant tool in the Alaskan market. MS. McNAIR-GROVE said, in spite of these limitations, the data appears to indicate that the use of a consumer's credit is causing some shifts in the marketplace. She said, "In the aggregate, consumers that reside in high-income, high-percentage Caucasian zip codes are less impacted by the use of credit. The limited data does suggest that unequal effects exist on consumers with varying income and ethnic characteristics." Number 2120 MS. McNAIR-GROVE addressed questions posed to [the division] when it was asked to undertake the credit-scoring exam; those questions began on page 11 of the report. The first question read, "Is the correlation between credit history and loss potential sufficient support for the industry to be able to use a consumer's credit history or should the industry be required to also demonstrate causality?" Ms. McNair-Grove answered that correlation alone isn't sufficient for an insurance credit score to be used in an insurer's underwriting and rating practices; that is why, when an insurer proposes to use credit information in its rating plans, the division spends a significant amount of time reviewing those rating plans and the underlying assumptions. The division looks at the statistical correlations and at how credit information is integrated into the entire rating plan, in an effort to understand its effect. MS. McNAIR-GROVE said the division has asked insurers to justify that [the use of credit information] doesn't have a disparate impact on rural versus urban consumers. The statute by which the division reviews insurers' rating plans is very broad. The standards are that a rate must not be inadequate, excessive, or unfairly discriminatory. In addition, a statute requires that when insurers segment the risks - which they are allowed to do - the way in which they segment those risks [must] demonstrate a probable effect upon losses or expenses. MS. McNAIR-GROVE turned to the issue of causality. She noted that the American Academy of Actuaries has a standard of practice called "concerning risk classification." One of the things suggested in the standard of practice is that causality might be appropriate if it's used in a less rigorous sense than a direct cause-and-effect relationship. It also says that risk characteristics should be neither obscure nor irrelevant to the protection provided by the insurance. Mentioning a report from a case called Hartford Accident and Indemnity Co. v. Insurance Commissioner, which referred to causality, she read, "The longer a vehicle is on the road, for example, the more likely it is that the vehicle may be involved in a random traffic accident; thus daily and annual total mileage may be viewed a causal rating factor." Number 2260 MS. McNAIR-GROVE turned to the second question, "Are victims of identity theft further victimized by credit scoring?" She said if that identity theft causes a less favorable credit score [than the consumer would have received otherwise], then the answer is yes, the consumer is adversely impacted by that identity theft. One solution would be to not use the information until it has been corrected and the identity theft problem has [been resolved]. MS. McNAIR-GROVE addressed the third question, "Does it make sense for a consumer to be able to qualify for a home loan but not be able to qualify for homeowner's insurance coverage?" She proffered that it seems counterintuitive that that should happen. However, a financial institution and an insurance company are looking at two different things; the former is looking to see if the consumer will repay the loan, while the latter is looking to see if there will be a loss. MS. McNAIR-GROVE discussed the fourth question, "Why do otherwise similarly situated consumers sometimes pay dramatically different premiums?" She said Alaska has a statute that says rates shouldn't be unfairly discriminatory and that [consumers] with like risk characteristics should be treated similarly; however, Alaska law doesn't require that all insurers charge the same rates. There may be significant differences between one insurer and the next. Differences can also occur based upon the type of business that the insurer chooses to write. For example, she noted, a preferred market generally will have lower rates than a nonstandard market. MS. McNAIR-GROVE addressed the fifth question, "If consumers and regulators do not know the rules of the credit scoring game, how can the interests of Alaskans be protected?" She opined, "You protect yourself if you don't know what's going on." She said some insurers do better than others in making a real effort to learn how credit scoring works and why it works, and to answer why they want to use it. MS. McNAIR-GROVE turned to the sixth question, "Are there Fair Credit Reporting Act conflicts?" She said the Act requires insurers to notify consumers if they have been adversely impacted when action is based upon the consumer's history. She noted that the Federal Trade Commission (FTC) - the agency that oversees the Fair Credit Reporting Act - had spoken to the National Association of Insurance Commissioners and said the following: Simply because an insurer is giving a discount to an insured, based upon the insured's credit history, does not mean the insurer is not taking adverse action against the consumer. If that discount does not result in the consumer receiving the best possible rate available from the insurer, the insurer may [still be] taking adverse action if the consumer would have received the best discount had the consumer's credit history been more favorable. MS. McNAIR-GROVE added that frequently insurers don't give proper notice when they are supposed to; hence there are potential Fair Credit Act violations. TAPE 03-32, SIDE B Number 2392 MS. McNAIR-GROVE pointed out that answers to the remaining questions can be found in the report. Number 2380 LINDA HALL, Director, Division of Insurance, Department of Community & Economic Development (DCED), highlighted the importance of studying what is occurring in general in the Alaska market before addressing specific issues such as credit history. Turning to pages 40 and 41 of the previously mentioned report from the division, she pointed out that although there are 58 [companies] listed as writing automobile insurance and 30 writing homeowner's insurance, many are multiples of the same name; therefore, the impact is dramatically reduced [when the names are listed just one per company]. MS. HALL referred to a handout entitled "2001 Alaska Property And Casualty Market Share." She explained that this information was taken directly out of the division's annual report. She said it concerns her that while there may be 58 companies active in the marketplace, the level of activity is dramatically different. Pointing to [the first page of the handout], she noted that in the homeowner's market, approximately 65 percent of the business is written by two companies. The [second company listed] holds about 29 percent of the market share, while the next lowest drops to approximately 6 percent. Eight of the top twenty companies have less than 1 percent of the market share, she said. MS. HALL noted that the same is true on the next page of the report in terms of private passenger autos, for which approximately 42 percent of the market share is written by two carriers, while ten of the top twenty companies write less than 5 percent market share per company, and seven write less than 2 percent. Ms. Hall emphasized that there is a limited market, that this issue needs to be looked at, and that the numbers of companies listed can be misleading. Number 2291 MS. HALL referred to a letter addressed to Chair Weyhrauch [dated March 25, 2003, in committee packets]. She noted that the division has a four-person consumer services section that handles phone calls. She said the division has only received five formal complaints regarding the use of credit scoring, which are detailed in the letter. She offered the following perspective: "In 2001, we received a total of 492 written complaints; in 2000, a total of 456." CHAIR WEYHRAUCH asked that Ms. Hall again address the letter when she is called back before the committee. Number 2227 OLIVIA POWELL, Staff to Representative Mike Chenault, Alaska State Legislature, testified on behalf of Representative Chenault, sponsor of HB 47. She said the bill addresses the more and more commonplace practice of insurance companies' invading the privacy of individuals' personal credit scores to determine insurance premium rates where no statistical or reasonable correlation can be made between a person's credit history and his/her likelihood of needing to file a claim with an insurance provider. In a time when personal bankruptcies, excessive debt loads, and a troubled economy are causing negative credit scoring among more Americans than any other time in history, insurance companies are exploiting this situation to boost their falling profit margins, she said. She continued: Insurance companies will tell you credit scores are one of the factors that assist an insurance company in determining whether to provide or renew coverage to an applicant and in deciding what premium an applicant should pay. Insurers use credit scores because they believe there is a correlation between financial responsibility and risk of loss. We believe there is no such correlation and that, at times, the practice is discriminatory. MS. POWELL read from a portion of [the "Summary of Conclusions"] from the Division of Insurance's report as follows: Based on the limited data received and evaluated so far, the use of insurance credit scoring in Alaska appears to have different effects on different groups of Alaskan insurance consumers. The survey data indicates that rural Alaska policyholders are more likely to be placed in the nonstandard markets than are urban policyholders. The survey data also suggests that there is a trend for older consumers to move from preferred to standard and even nonstandard with increasing age. MS. POWELL added that the data also suggests "unequal effects exist on consumers with varying income and ethnic characteristics." She said, "The only correlation we see is one of discrimination and bias." She continued as follows: Is it fair that [a] fisherman in rural Alaska pay more for his car insurance because it was a bad fishing year? And is it fair that an elderly Alaskan who doesn't have a large need for credit pay more for his homeowner's insurance? We don't feel that it is right. And we are not alone in thinking this. Number 2134 MS. POWELL continued: Each year, more states are passing legislation regulating and restricting how credit scores may be used to [establish] rating plans. Currently, there are 18 states with laws restricting the use of credit scoring. For example, ... Hawaii Statute 432.10C-207 reads, "No insurer shall base any standard or rating plan, in whole or in part, directly or indirectly, upon a person's race, creed, ethic extraction, age, sex, length of driving experience, credit bureau rating, marital status, or physical handicap." Hawaii obviously felt that the insurance companies were using the credit bureau rating in an unfair way, and they did the right thing by controlling the use of it. Last year alone, over 70 pieces of legislation were introduced nationwide to control the use of credit scoring. Another example is Kansas. Studies there found examples of how credit scoring doesn't always work. According to the Wichita Business Journal, the Kansas Insurance Department did an online survey asking people about their insurance rates. Out of [520] people who participated in the survey, 454 said their insurance rates had increased in the past 12 months, and [267] of those were told by their insurance companies that credit scoring was a factor in that increase. Credit scores can only be effective if they are accurate. Inaccuracies in credit reports are unfortunately more common than one may think. In some studies, as many as one out of every [five] credit reports has some inaccurate data. If this is true, how can credit scores possibly [be] an accurate gauge? And the answer is, they're not. MS. POWELL said, "You've all seen the reports and heard the stories of how people are qualifying for home loans but can't get ... homeowner's insurance." Saying she wouldn't offer further statistics, she closed as follows: "We know that we need to put guidelines on the insurance companies so they can't charge anyone whatever they feel like. And moving House Bill 47 will do that." Number 2056 REPRESENTATIVE HARRY CRAWFORD, Alaska State Legislature, sponsor of HB 5, expressed frustration that the bill has been mischaracterized. He clarified that [the intent] is not to ban the use of credit history or credit information in the setting of insurance rates. Rather, it would ban the use of the credit score, which is totally different. REPRESENTATIVE CRAWFORD listed the determining factors used to derive a credit score along with the general percentages given to each factor: payment history, 35 percent; amount of outstanding debt, 30 percent; length of credit history, 15 percent; recent new applications or opened accounts, 10 percent; and mix of credit and types of accounts and loans, 10 percent. Number 1979 REPRESENTATIVE CRAWFORD said what [the industry] has been able to do with the credit score that differs from a person's personal credit history relates to the length of credit history. For example, he said, it has been proven that those between the ages of 16-25, especially males, have more accidents; [the industry] combines that group with people with a history of missing a payment or two, for example, and thereby ropes in a much larger group. He cited examples. Saying the intent of [HB 5] is to "keep this to a person's personal credit history" Saying credit history can be a good predictor, he added, "What we want to do is ban these secret algorithms that the industry uses." Number 1899 BIRNY BIRNBAUM, Executive Director, The Center for Economic Justice, referred to the following handouts in the committee packet: his testimony; a study ["Insurers' Use of Credit Scoring for Homeowners Insurance in Ohio - A Report to the Ohio Civil Rights Commission," dated January 2003]; charts [the U.S. Census regarding assets, liabilities, and debts of households and families]; some preliminary comments on a University of Texas study ["Preliminary Comments on the BBR Study"]; and ["Comments of Birny Birnbaum on Behalf of the Center for Economic Justice Before the National Conference of Insurance Legislators [NCOIL]", dated November 21, 2002]. MR. BIRNBAUM referred to page 8 of his written testimony, where he'd provided a summary of his education and experience, and said the main point relates to the definition of "fairness." He said the industry's definition of fairness is "the existence of a correlation between a rating factor and risk of loss." If that correlation is there, he posited, then its use is fair. However, by this logic, it would be fair to charge consumers more for auto or homeowner's insurance because they've been laid off, lost their health insurance, or gotten a divorce. He opined that it's important to distinguish between what society or the public thinks is fair and what the insurance industry thinks is fair. Furthermore, he said, he thinks the insurance industry's definition of fairness is quite limited. MR. BIRNBAUM stated his belief that it is both reasonable and necessary to prohibit insurers' use of credit scoring in underwriting and rating personal lines, as well as in determining payment plan eligibility, for the following reasons: it is inherently unfair; it undermines the basic insurance mechanism; it undermines regulatory oversight of rates; and it discriminates against consumers because of income, for example. He opined that [the use of credit scoring] is inherently unfair because it penalizes victims; the vast majority of people who file for bankruptcy have had an economic or medical catastrophe or a divorce, for example, and it's not merely because they can't control their credit-card spending. MR. BIRNBAUM offered his belief that credit scoring penalizes consumers because of decisions made by lenders. From 1990 to 1998, the number of credit card solicitations increased from one billion to four billion, and credit card debt increased fourfold. He said it's clearly a result of lenders' decisions to make credit more available. He cited an example and asked, "Why should consumers be penalized because of decisions that lenders make to enter into more risky lending?" He said the problems are significant and admitted by the industry itself when it supports the NCOIL bill, which contains numerous prohibitions and proscriptions on credit. He continued: So, one would think that there would be some great benefit to consumers to require all this legislative and regulatory oversight, and yet, ultimately, the only rationale offered by the industry is the correlation between credit and risk of loss. But clearly, the near distance of a correlation is not sufficient justification for use of a factor. MR. BIRNBAUM posed another question to think about: "What other risk factor punishes consumers for acting in a legal, rational, and positive manner?" He asked why consumers should be charged a higher auto or homeowner's rate for shopping around for a loan or insurance; for getting a zero-interest credit card to lower their debt payments; for paying cash instead of credit; or for getting a credit card for a first-time 10-percent discount at a department or electronics store, for example. MR. BIRNBAUM remarked, "The big lie here is that the industry wants you to believe that ... a good credit report equates to a good score. That's simply not the case." He said the credit score is based just as much on the presence of so-called positive attributes as it is on the presence of negative attributes. Referring to Mr. Lo's testimony about the small number of consumers who have bankruptcies and delinquencies, he said many times those consumers pay higher rates not just because they've had a bankruptcy or delinquency, but also because they don't have the attributes that the [industry] thinks are good. MR. BIRNBAUM opined that many or all of the industry's claims are unsubstantiated. Responding to the claim that [the industry] will write more business, he referred to page nine of his written testimony, where there are statements by two national agents' associations that have come out against credit scoring. "If credit scoring allowed agents to write more business," he said, "I doubt that they would be against the use of this." As to the claim that most consumers would benefit, he said his research says this isn't true; even if it were true today, there's no guarantee it will be true in the future; and even if most consumers would benefit from something, that doesn't justify the use of an unfair practice. Mr. Birnbaum said the bottom line is that there's no free lunch; credit scoring doesn't reduce claim costs, but shifts premiums from one group of consumers to another. He disagreed with the premise that a company only uses credit [scoring] to offer discounts. Number 1506 DEE HUBBARD, testifying on her own behalf, thanked Representative Crawford for "sticking with this problem that I brought to him back in 2001." She said the issue is whether [credit scoring] is good for insurance companies or consumers. She offered her feeling that if this is so good for consumers, it would have been public knowledge a long time ago. Therefore, she said there isn't much good in it for consumers, and questioned the effects at some later point. She said her family pays cash in order to avoid raising its credit card limit, for example. Saying this issue is all about big money, she gave some examples of what certain chief executive officers (CEOs) are paid and what the companies' one-year targets are in the stock market. CHAIR WEYHRAUCH asked Ms. Hubbard if she had ever learned that a company's use of credit scoring had adversely affected her. MS. HUBBARD said no. She revealed that her insurance [agent] told her in 1999, "Oh, your rate came in really good, you must have good credit." She explained that this is what prompted her to begin asking questions. Number 1295 HOWARD DORSEY testified that his relationship to this bill is on a business level. He explained that before 1999, he was working from 9 to 5 and getting a paycheck, and auto insurance was essentially the only insurance he carried. He said he'd noticed that his auto insurance [rate] decreased as he aged, which made sense because he was more experienced. However, in 1999, when he started his own company, he needed to buy additional insurance for company liability, commercial auto, and general liability and bonding. Over the last four years, the rates have continually increased, even though no claims have been made either on behalf of the company or because of an auto accident. He said his clean driving record dates back 15 years. MR. DORSEY said, regarding credit scoring as his company grows, at times he may be delinquent by 30 to 90 days on a bill, or may have a bill that slips into the trash and ends up in collections. And new companies don't always have sufficient cash flow. He said when he looks for an insurance policy at the end of the term, often, over the last three years, the first quote he receives is the lowest one; after that, the rates are higher, all for the same type of policy for the same term. What has changed is that a credit inquiry has taken place on his credit scoring. He said he went through that in 1996 when he bought his house. The mortgage company told him that while the process was in motion and until it closed, he shouldn't apply for a credit card because every time there was an inquiry, it would show on his credit report and wouldn't look favorable to those that would be underwriting his loan or insurance. MR. DORSEY remarked, "We all know that statistics can be twisted to however we'd like them to be. They offer no justification of the statistics. And when one asks for a copy of the statistics, one is refused a copy, because that is confidential information. It would give ... unfair ... information to competitors." Noting that he has opened two new accounts since last year, he questioned why that should make him a higher risk, because it was due to a business change. Mr. Dorsey thanked Representative Crawford and [Mr. Birnbaum] for their comments. Number 1080 MARIE DARLIN, AARP Capital City Task Force, said AARP has been working with everyone involved with this legislation for the last two years because of the numerous comments received from its members saying their insurance [rates] were increasing, they didn't know why, and they couldn't get any answers from insurance companies when they asked what was meant by credit scoring. Ms. Darlin said AARP has found this issue to be a concern in many states. MS. DARLIN said AARP's main concern is that age seems to play a part. She related her experience that when her insurance bill arrived last July, it was approximately $90 higher than the year before. However, her situation hadn't changed except that she had a newer car. She said she'd researched and found another insurance company, which saved her perhaps $15. MS. DARLIN said she feels this issue needs to be looked at from the viewpoint of how it affects elderly people who are still driving and have no record against them. She said she thinks credit scoring should be studied to find out how it is really done. She said she doesn't feel it's beneficial to anybody as far as providing better coverage or better insurance, for example. She questioned how [insurance companies] can continue to do the credit scoring without providing consumers with the outcome and without explaining what it means. In response to a question by Chair Weyhrauch, she confirmed that her testimony relates more to the legislation that would affect people, rather than the one that would affect businesses; however, she said AARP is hearing more from seniors and retirees who have businesses and are having problems with their insurance. She concluded by expressing concern about both bills. Number 0826 MARK NIEHAUS, General Manager, Progressive Insurance, told the committee that Progressive Insurance is the largest writer of [private passenger] auto insurance through independent agents in Alaska; has been using credit [scoring] for ten years [in the Lower 48], and six years in Alaska; and has filed its credit rating program with the Division of Insurance. He offered his belief that the division does a thorough job in examining the actuarial justification and looking at the disparate impact issue, and did so when reviewing and ultimately approving Progressive Insurance's filings. He said he thinks a good process is currently in place to ensure that the data is there to support what companies are trying to do and that consumers are treated fairly. MR. NIEHAUS paraphrased a portion of his written testimony that read as follows [with some technical changes]: We have filed our current credit scoring methodology and detailed actuarial support with the Alaska Division of Insurance. Our model is public information; there is no "black box." Progressive does not use credit information to refuse to insure a consumer, [or to] non-renew or cancel an existing customer. Credit information that is disputed by the consumer with the credit-reporting agency is not considered in our insurance scoring algorithm. All identified medical and business/commercial debt and liens are excluded from all Progressive credit scoring methods. An applicant or insured who experiences an adverse action as a result of the use of credit information is advised how he or she can obtain a free copy of their credit report. We provide a list of the reasons for an adverse action due to the use of credit to consumers upon request. Progressive is committed to sharing information about how we use credit with regulators, the media, and consumers. We want consumers to understand how credit affects their insurance premiums and we want to help them in developing a plan to improve their credit scores. We have pioneered the use of a new Credit Assistance team for our customers and agents in Michigan and Texas, and we are in the process of rolling that out to more states. The Credit Assistance team can be reached through a toll-free number and it provides: personalized reports to applicants describing how their score on each of the variables considered in the credit-scoring algorithm compares to the average; [and] reasonable credit exceptions based upon prior credit history for persons whose credit information is unduly influenced by extraordinary life events, [for example], catastrophic injury, death of a spouse, [or] business loss Number 0650 MR. NIEHAUS referred to the Division of Insurance's credit scoring report and paraphrased another selection from his written testimony, which read as follows [with some technical changes]: The study found that a disproportionately smaller number of policyholders in lower-income zip codes with higher minority populations qualified for preferred markets, regardless of whether or not credit was used to help determine final market. Furthermore, when credit was used, not fewer but more of these policyholders actually qualified for the preferred market. ... With regard to access to preferred markets by age, the study found that the proportion of policyholders in preferred markets actually peaks for Alaskans aged 61 to 70 years old. MR. NIEHAUS referred to a chart in his written testimony which shows that based on over 23,000 policies quoted [in Alaska during the first half of 2002], the average credit score improves with age. He reported, in regard to the rural-versus- urban issue, that Progressive Insurance separated its data by population density, and it shows that [while the average credit score] changed little, rural areas had slightly better average scores than urban areas. MR. NIEHAUS said Progressive Insurance is opposed to [HB 47 and HB 5]. If they pass, the company would be forced to raise rates on approximately two-thirds of its 17,000 policyholders. Seniors would be especially disadvantaged. Good drivers would be forced to subsidize bad drivers. An unlevel playing field would exist, because direct writers who use mail as a way of attracting customers would be able to pre-screen using credit, since that is specifically allowed under the Fair Credit Reporting Act, whereas companies such as Progressive Insurance that market through independent agents would be unable to use that tool; therefore, agents would be disadvantaged. Number 0578 CHAIR WEYHRAUCH asked if any legal analysis or opinion indicates credit scoring doesn't run afoul of the Fair Credit Reporting Act. MR. NIEHAUS answered that Progressive Insurance has provided that opinion to Representative Crawford and that he'd be happy to provide it to the committee. Number 0375 MR. NIEHAUS continued to explain why Progressive Insurance is opposed to the proposed legislation. If one or both bills were to pass, Alaska would be the only state to ban credit in recent times, which would be damaging both to consumers and to the market. He said the Alaskan auto insurance market is in very bad shape; the NAIC [National Association of Insurance Commissioners] released data on 2001 results that ranked states on the basis of profit margin as a percentage of premium earned, and Alaska was rated 47th out 50 states and the District of Columbia. He said Alaska is one of two states in which insurers have lost money in three of the last four years. MR. NIEHAUS said Progressive Insurance is committed to working with legislators and regulators to find common ground on this issue, and supports a reasonable regulation on the use of credit. He noted that there are two potential avenues in this regard. He mentioned negotiations and discussions last year that resulted in a compromise committee substitute (CS) and that would have addressed many of the concerns; to his understanding, however, it was derailed. He said, "That's something that we would support as being advanced as a possible alternative to an outright ban." In response to a request by Chair Weyhrauch, he said he'd provide a copy of that CS. He added that NCOIL, which is composed of legislators and regulators and gets input from insurance companies and consumers, spent a lot of time developing a model Act [included in the committee packet] that he opined could be used as a basis for a bill to address areas of specific concerns. Number 0238 REPRESENTATIVE SEATON, referring to the previously mentioned model Act and CS, asked Mr. Niehaus whether he offers more support for one of those two alternatives. MR. NIEHAUS responded that the CS from last year addressed the issues that were specifically of concern to Alaskans, whereas the model Act is a broader, nationwide approach. He added, "We could work with either version." Number 0177 REPRESENTATIVE CRAWFORD remarked that Progressive Insurance's usage of credit scoring appears to be a little better than the rest. He asked Mr. Niehaus whether Progressive Insurance's use of credit scoring is representative of the industry. Referring to Mr. Niehaus's previous mention of "the medically coded information," he revealed that two years ago he had an accident, snapped his Achilles' tendon in two, and had to have an operation in Virginia; at that time, he was fully covered by two different insurance companies. After two years, he said, [the insurance company] is continuing to not pay some of the bills, which have been turned over to collections; thus no medical code shows up on the credit report. He asked, "So, how do you get to exclude anything that's been turned over to collections when there's no code there for it?" He continued: Also, you mentioned that no other state has a law this restrictive, but Hawaii has an outright ban on the use of credit scoring, and it's also the most profitable state in the nation, according to the information that you gave me a couple of days ago. And they also have lower insurance rates than Alaska. And I was just wondering, how could they do that without the use of credit scoring? You're saying that to bring Alaska up to more like the national average, but Hawaii makes it without the use of it somehow. [Due to technical difficulties during the tape change, the answer provided by Mr. Niehaus was not recorded.] TAPE 03-33, SIDE A Number 0001 REPRESENTATIVE GRUENBERG referred to Hawaii's Statute 431:10C- 0207, which read as follows: Discriminatory practices prohibited. No insurer shall base any standard or rating plan, in whole or in part, directly or indirectly, upon a person's race, creed, ethnic extraction, age, sex, length of driving experience, credit bureau rating, marital status, or physical handicap. REPRESENTATIVE GRUENBERG asked Mr. Niehaus what cost-savings measures Hawaii had enacted. Number 0065 MR. NIEHAUS recalled that Hawaii had implemented medical fee schedules, for example. Noting that its system is different from Alaska's, he explained, "Their minimum limits are a lot lower. I believe it's $20,000 per person, $40,000 per accident. Hawaii's minimum limits are 50-100." He also noted that [Hawaii] put in limitations on chiropractic visits, for example. REPRESENTATIVE GRUENBERG asked Mr. Niehaus if the prohibitions listed in the previously read Hawaii statute - with the exception of credit bureau rating - are already prohibited under either federal or Alaska law. MR. NIEHAUS replied that Alaska doesn't prohibit the use of age, sex, length of driving experience, or marital status as elements of rating. Number 0221 REPRESENTATIVE GRUENBERG noted that Ms. McNair-Grove had mentioned that the use of credit scoring is a proxy for some factors. He asked, "If it's a proxy for something that's illegal, then why shouldn't we make that illegal too? It's like taking the hood off the Ku Klux Klan." MR. NIEHAUS recalled that Ms. McNair-Grove had testified that the division looks at those issues as part of its review. He said, "We don't gather data on all of those things that were mentioned earlier, in terms of income, in terms of race, ethnicity, religion," for example. He added, "We gave the data to the division by zip code." In response to a question by Representative Gruenberg, he surmised that the zip codes were 5- digit ones, rather than 9-digit. REPRESENTATIVE GRUENBERG explained that in Anchorage, for example, one 5-digit zip code encompasses a very large part of the city and includes wealthy areas as well as poor areas. He suggested that breaking the areas down into 9-digit zip codes, or perhaps [U.S.] census tracts, may provide more information. Number 0303 JOHN GEORGE, Lobbyist for National Association of Independent Insurers (NAII), told the committee he'd served as the director of insurance from 1984-1988. Many problems back then still exist, including that there were only a couple of major companies writing insurance. One of the biggest tasks was developing more insurance companies, he recalled. He emphasized the importance of encouraging more companies to write insurance [for the sake of competition.] He urged the committee to ask Ms. McNair-Grove how the division actually reviews a rate filing, what's in it, and what profit margin it allows. Disagreeing that insurance companies are doing this to boost profits, he said it's a redistribution of profits and an equity issue. He added, "'The right people pay the right premium' is really what this is about." Number 0465 MR. GEORGE continued as follows: A typical rate filing, if you were starting from scratch - an they often put in a filing that modifies a prior one - but if you started from scratch you'd have a file that was three inches thick, probably. And those are public information; you could go over and pull Allstate or Progressive's filing. Nobody ever does, other than the other insurance companies. And I actually have a contract with a couple of outfits that hire me to go over and photocopy those to send to them so that they can see what the other guy is doing. The information that we've heard is secret and all that: it's the recipe. If you were making the best chocolate chip cookies in the world and you had the recipe and it was public information, gee, Nabisco would like that, and they'd start making your cookies. ... And that's really the thing. But those are available to the Division of Insurance to look at. The question is, "Can they keep them proprietary?" And on a "market conducts" examination, they got them all. Look at them - there they are. The question is, "Should they be available for the competing companies?" And if you're going to encourage companies to come in and write, you might say, "Yeah, give them the other guys and they'll compete." But what we're really looking at is a very diverse market - niche markets. One company may prefer to write military or "SR22" business - people that have five drunk driving tickets. You can charge a lot of money. And if you can charge them the right amount of money, you can profit from writing what you consider really bad business. Some would prefer to write only the very cream of the crop. That's their niche, and they really don't want to write the drunk driver, but they charge significantly less. So, they're very competitive. If you look at the end of the Division of Insurance report, you'll see that some companies are using credit for underwriting only. Some are using it for underwriting and rating. Some companies aren't using it at all. So what you find is that there's a lot of different places to shop, and if you don't like someone using your credit information, there's companies that don't. Admittedly, they're not writing a lot of business. But it's a wide market, and you ought to be able to find someone [whose] writing ... works for you. Number 0621 MR. GEORGE said the law allows for fair discrimination. For example, an individual 16-year-old driver may not have a poor driving record and therefore may be discriminated against, but, as a class, 16-year-olds will have poor driving records. Mr. George explained that by using credit or insurance scoring, refinement can be made within a class. One person in a class may be acknowledged for having a better record than someone else in the same class, and therefore could receive a better rate. MR. GEORGE recalled that the Division of Insurance testified last year that [credit scoring] is a strong predictor of loss, and he said he thinks everyone agrees with that. He referred to the previous testimony of Mr. Niehaus that the majority of policyholders in his company benefit [from credit reporting]. He recalled that the Division of Insurance report indicates perhaps seniors would be hurt by [credit] scoring; however, he noted that the [State of] Washington report, which used what he called real people and data rather than census data and zip codes, found that seniors actually did better. He said Mr. Niehaus had testified that seniors do better with Progressive Insurance on their insurance scoring. MR. GEORGE recalled that it had also been previously mentioned that in every zip code that the Division of Insurance report tested - whether urban or rural - the percentage of people in the preferred market increased after [credit] scoring. Highlighting the reverse argument that there wasn't enough data to show this was due to [credit] scoring, he countered that by saying, "They always say, 'We just don't have enough data; we can't say ... for sure what caused that.'" Number 0712 MR. GEORGE said he thinks it is apparent that the insurance industry wants to find things that work; therefore, if an indicator doesn't work, it would be thrown out. They do that in order to have people pay what they should pay. If they failed to do that, he added, it could be said [the insurance companies] were actually discriminating against people by not giving them an appropriate rate by not using credit scoring. Number 0775 REPRESENTATIVE SEATON referred to the previous comparisons with Hawaii. He asked Mr. George to presume that Hawaii has made a distinction that no factors can be used to discriminate between people; rather, a flat rate must be offered to everyone. He asked, "If the insurance companies are going to come out with the same profit margin, no matter what we let them do, why does it matter to the insurance companies themselves whether [they] use this tool or not?" MR. GEORGE replied that, unfortunately, insurance is not like a public utility that is guaranteed a profit. An insurance company puts in a rate and hopes to get it, but in the last two years the insurance companies have had substantial losses, he said. If the insurance companies said they would like to make a 10 percent profit, the division would probably reject that. He asked [Ms. McNair-Grove] to confirm that the amount the insurance companies "put in" is about 5 percent. MS. McNAIR-GROVE replied, "More or less. I can explain that further." Number 0915 MR. GEORGE said it's a matter of competition, and the insurance companies compete fiercely. He compared it to the long-distance companies that come up with all sorts of ways to get people's business. Insurance companies with a niche can get business from somebody else because they have the best rate for that niche. Competition is healthy and provides an appropriate rate to people, he opined. He concluded, "The public really understands that they want to pay their rate, and not the average rate." Number 0933 REPRESENTATIVE CRAWFORD highlighted estimates that up to 75 percent of all credit reports contain erroneous information. He asked Mr. George how he could say the right rate is being charged the right person, if the information being used is based on erroneous information. MR. GEORGE answered that the material corrections that are made make up a miniscule percentage of all corrections, because the report may spell someone's name wrong or get an address wrong, for example. On the other hand, some studies show that up to 40 percent of motor vehicle records are inaccurate. He remarked, "That's even just the ones you get caught on. How many people have actually gone 60 miles an hour in a 55 [mile-per-hour zone] and got away with it, day after day? ... If you just happened to be the guy that got caught, you got a ticket." But credit, he said, is sort of inescapable. Therefore, he concluded that credit is much more credible than maybe someone's driving record. He recommended that Representative Crawford's question also be asked of Mr. Sorich and Mr. Lo. CHAIR WEYHRAUCH invited some of the previous testifiers to return to the witness table to complete their testimony. Number 1074 MR. LO referred to the issue of inaccuracy. For the purpose of discussion, he offered the assumption that there are errors in credit reports. He continued as follows: That ... credit data has all those errors in there. And, in my earlier testimony, I showed you how the data with the error in them on the credit side and the premium and losses were matched together so that we can study the credit versus the ... future loss phenomena. All the credit errors have not been removed. Fair Isaac has no means to remove any errors. And all of those are included in the data. And the results that you have seen are extremely predictive of future losses. So, if the error exists, it did not lower the "predictiveness." So that's point number one [that] I want to clarify. Point number two is: there are errors in there. And there are very stringent Federal Credit Reporting Act dispute resolutions. I'm not the expert on that subject. But there are lawyers, and there [is a] credit consumer reporting agency association that ... can speak to you and answer your questions. They [have] very stringent requirements: 30 days - complaint filed, resolved. Otherwise, that record, that particular trade line, is deleted. So I just wanted to at least leave you with that understanding. The errors are there; nobody denies that. It does not distort the predictiveness, and there are significant regulations already in place addressing the errors. MR. LO referred to a previous point made by Mr. George regarding secret recipes. He said Fair Isaac wants to compete through its insurers. He reminded the committee that Fair Isaac is a modeler only and cannot write insurance. He said there are many insurers that want to come to Alaska and use Fair Isaac tools as an efficiency tool and a predictive tool; however, currently they are prohibited "pretty much" from using Fair Isaac, because Fair Isaac models [would] have to be disclosed to the public. Once disclosed, Fair Isaac would lose its 30 to 40 years of research and investment in producing [models for] predictive results. Therefore, Fair Isaac needs to protect its investment by asking for trade secret protections. He noted that it has done that in at least 12 other states under protection already. He continued: This is not Fair Isaac trying to hide the formula from the consumer. This is our effort to helping [the] insurance department to understand how the model is built, so that [it] can be satisfied that this would not lead to excessive rates, nonfair discrimination, and inadequate rates. This is trying to help the insurance department to form the proper regulations so that [it] can regulate to the best of [its] ability. That's [what] we want to provide; however, we do not want to confuse [consumers]. We do not want them to have to understand what is a credit report, what is a formula, what it means to, basically, calculate his or her own score, in a sense that it will raise more questions to the insurance department, if not to you ... from your constituents. So, we want to address that through education. MR. LO noted that there are a lot of educational efforts currently being made. He referred to the one-page handout entitled "Answers to Your Questions About Insurance Bureau Scores" [available in the committee packet]. He said NAII and all the major trade associations have literature, including the National Association of Insurance Commissioners, which has just released a brochure regarding consumer education on this issue. He summarized that Fair Isaac needs protection, but not to hide from consumers. Fair Isaac doesn't want to generate more questions, but would like the public's questions addressed through education. Number 1300 REPRESENTATIVE GRUENBERG requested copies of the laws of the other 12 states that presently are under protection. He surmised that under those laws, the information is not public, but the division of insurance has access to the Fair Isaac information. MR. LO said, "Correct." Number 1324 REPRESENTATIVE CRAWFORD informed members that he'd requested that his own credit scores be done by three different major credit reporting agencies. He commented, "My credit scores are all over the map; they vary by 100 points." He asked Mr. Lo how it is possible to say that [credit scoring] is a good predictor, when the results of the three companies show that they cannot decide among themselves whether he is a good credit risk. He added that [the scores] don't have anything to do with his driving record or his rental properties, because he hasn't had any claims on his driving record and out of a number of rental properties over the last 25 years, he said he thinks he has had two claims on his homeowner's insurance. He added, "I have a good credit history, but I have a couple of poor credit scores." MR. LO responded that Representative Crawford should be confused about those credit scores because the information [explaining the discrepancy between them] was probably not provided to him or made easily understandable. He confirmed that there are three major credit bureaus, each with its own credit data base. Fair Isaac has built specific models for each. The models have not been made to produce the same score results from each agency. The scores all may be within 100 points of each other and they all would be "within the means." He explained that each company's [results] will be relative within its individual model; therefore, there's no point in comparing model to model, score to score. What is an above-average or below-average score in one model won't be the same in the next. Number 1484 REPRESENTATIVE CRAWFORD opined that his credit history should rank him an "ultra-preferred," but because of "some of these other factors and other ways of weighting these factors," he has been ranked "standard." MR. LO offered the example of standard and nonstandard [business] in one model. Standard starts at [a score of] 650 and higher, whereas nonstandard starts at 650 and lower. In another model, the standard can start at 700. He mentioned different markets and niches. He continued as follows: So, the fact that you are set up by one company in standard and set up by another company in nonstandard, that's no surprise, if that's how each company looks at the business. However, if you are treated as standard in one company and, using the same model, you are treated as a nonstandard another time, that is something you need to investigate ..., because that shows that there's a change in your credit history; that's reflecting the change. But it's not due to the difference between the [models]. It's all, within one model itself, very consistent. Number 1530 CHAIR WEYHRAUCH invited the testifiers from the Division of Insurance back to the witness table to answer questions. He mentioned an overview of credit scoring and market-conduct examinations; he asked when its completion is anticipated. MS. HALL explained that once the report is written, it is given to the particular insurance carrier, which has 30 days to respond to that. During that process, it is still a confidential document. She stated, "Once they have that, they also have an opportunity for [a] hearing, should they request one. So, it's hard for us to tell what period of time they may take on their end. And our reports are not quite finished, is my understanding." Number 1660 CHAIR WEYHRAUCH asked if the division has taken a formal position on either HB 5 or HB 47. MS. HALL answered no. Number 1675 REPRESENTATIVE GRUENBERG noted that on page 11 of the previously cited report by the Division of Insurance, a distinction is drawn between causality and correlation. He distributed an article from the "Daily News Opinion" section of the Anchorage Daily News, dated March 28, 2003, written by George Will, a columnist for The Washington Post, entitled, "Dan Moynihan Saw Beyond The Horizon Of His Time." He read a paragraph of the article, which stated: The Senate's Sisyphus, Moynihan was forever pushing uphill a boulder of inconvenient data. A social scientist trained to distinguish correlation from causation, and a wit, Moynihan puckishly said a crucial determinant of the quality of American schools is proximity to the Canadian border. The barb in his jest was this: High cognitive outputs correlate not with high per-pupil expenditures but with a high percentage of two-parent families. For that, there was the rough geographical correlation that caused Moynihan to suggest that states trying to improve their students' test scores should move closer to Canada. REPRESENTATIVE GRUENBERG said he thinks this points out the possible fallacy in the difference between correlation and causality. He asked Ms. McNair-Grove if she would reiterate the difference between the two words and how that relates to the present issue. Number 1710 MS. McNAIR-GROVE agreed that there is a difference between correlation and causation. She noted that some statistical textbooks have examples of "silly correlations." She said: What we have tried to do in our review of this is get a grip on, "Is this a silly correlation between credit history and loss experience, or is there really something to it?" And the statistical evidence that we have seen has led us to approve these, based on the statutory requirements that we have for approving rate filings. The statute does not require that there be a cause-and-effect relationship; it says they can be demonstrated to have a probable effect upon losses or expenses. REPRESENTATIVE GRUENBERG referred to the top of page 12 of the division's report. He read as follows: As exemplified in the NAIC report, "the longer a vehicle is on the road, for example, the more likely it is that the vehicle may be involved in a random traffic accident; thus, daily and annual total mileage may be viewed a causal rating factor." REPRESENTATIVE GRUENBERG noted that Marilyn Voss-Savant, who is a bright person, has had a number of examples of this in her column. The chance of any random accident occurring is exactly the same regardless of the age of the vehicle, he said. The question is whether they're on the road or not. He said, "It is, each individual time segment is totally complete and unto itself. So is this not a totally false logical example?" MS. McNAIR-GROVE said she wasn't sure how to respond, but that the question posed to the division was, "Is correlation enough?" The point that the division is trying to make, she noted, is that when somebody comes to the division and says there's a correlation between "this factor and loss experience," that doesn't automatically get it approved by the division. REPRESENTATIVE GRUENBERG asked whether the division would like the capability through statute to find out the basis of Fair Isaac's modeling system while preserving that company's proprietary interest. Number 1849 MS. HALL answered that it's one point of consideration she would make. If credit scoring is to be used, she would like the Fair Isaac or any scoring model made available to the division, along with some appropriate confidentiality protections of that. That not only would allow the division to evaluate the model that's being used, but also would allow it to have the details so that when it conducts market conduct examinations, it can test to be sure that an insurance company is actually using the things that company put in its model that it filed with the division. REPRESENTATIVE GRUENBERG requested that Mr. Lo provide that information to DCED. MR. LO agreed to that. MR. SORICH noted that a similar provision is in the NCOIL model Act. Number 1960 REPRESENTATIVE CRAWFORD asked how the use of credit scoring in underwriting would affect the consumer's actual insurance bill. MS. McNAIR-GROVE responded, "If an insurer uses two different companies for their business - one company may be a preferred company and one company may be a standard company - they will have different rates in those two companies. So, depending upon which company the insured is placed with, yes, it does affect your rate." She concurred that in some cases, the rate could be affected considerably. Number 2075 ROBERT CEDERHOLM, Independent Insurance Agent, said he supports the ability of companies to use different criteria for underwriting. He said he believes [credit] scoring is used as just a tool in one of many forms of underwriting. He commented that so many companies have come and gone in Alaska because they did not make a profit. He opined that [insurance] companies want to do what's best for the insured. He noted that the more people an insurance company insures, the less chance that company has of "having any kind of a catastrophe." He added, "That's the point of insurance." MR. CEDERHOLM said insurance companies try to keep the rates low for the insured; if they don't, people have the choice of going elsewhere. He stated, "I think what's missed in this whole thing is that people have choices." He added that if rates go up, then people need to shop around, because it's always in their best interest to do so. He said two or three insurance companies control the largest percentage of the business, which is why so many companies over the years have left the state. He continued: Because they leave the state, it keeps those two or three companies in control of the largest percentage. And when they're in control of the largest percentage, they're spreading the risk amongst themselves, and so they're not the ones that you ever fear ... are going to leave the state. It's the ones that only have a little tiny bit of this that are going to leave. And what happens is, because they have so many clients, they can keep their rates lower, so it feeds upon itself. MR. CEDERHOLM said rates will go up, either because of credit or just due to losses. When companies apply to the Division of Insurance to get rates increased, if the information they give the division doesn't prove they aren't [making a profit], then they are not allowed to have a rate increase. He said the public seems to think insurance companies [change rates on a whim]. Rates are based upon competition, he said. As long as the cost of doing business increases, the charges will increase, whether it's caused by credit scoring or anything else. Mr. Cederholm explained that the reason no one seems to know what [model] any company uses is that it is secret. Furthermore, it has to be secret so that the insurance company has a tool that no one else knows [the details of]; thus it can keep its rates low in order to have the highest possible percentage of clients and, therefore, will not have to raise its rates. He added that people seem to look at the short-term [picture] and not the long-term one. He opined that keeping companies from having flexibility will result in their leaving the state, and the few remaining companies will raise their rates. Number 2137 REPRESENTATIVE CRAWFORD said he agrees with Mr. Cederholm regarding the increase of insurance rates, and "8 percent is about what it's going to take to have companies remain profitable here." However, he said there are examples that, because of credit scoring, some people's rates have increased 100 percent because "they got placed in another group." He said, "Just because I was shopping for lower interest rates on my mortgages for my rental properties and had a lot of hits, my insurance rates went up 25 percent one year and 40 percent the next." He asked if that was fair that his insurance rate was affected because he was doing the American thing by shopping around for the best price. MR. CEDERHOLM replied that credit scoring is a new tool that, like all new tools, needs fine-tuning. He said that is what's currently being done. He suggested the rise in rates that Representative Crawford experienced after his medical situation may have been due to a glitch in the system. He also suggested that Representative Crawford call another company that doesn't use [credit scoring], because it would be in his best interest to shop around. MR. CEDERHOLM noted that as an independent agent, he writes for 10 different companies. He said every day he has customers come in complaining about their current company, and he shops them out to another one of the 10 companies that he represents. If they are still not happy, he said, they can go to another independent agent. Number 2238 REPRESENTATIVE CRAWFORD responded that he uses independent agents also, and that shopping around for a lower rate is exactly what [negatively impacted] his credit score. He said, "Instead of fine-tuning it, or making it ... more like a laser, what you're doing is making ... a huge lasso around so many people that they're getting roped in to a group that they shouldn't be a part of." Number 2279 ARTHUR PARKS, Underwriter Administrator, State Farm Insurance Company ("State Farm") for the Pacific Northwest Dome, testified that State Farm's underwriting model is different from the model previously talked about, in that State Farm developed its own product, using its loss data. State Farm scored three million of its own policyholders and then tracked the loss experience with them. The company identified the credit characteristics that were most predictive and combined those with prior loss information to develop an underwriting score. State Farm only uses this model at the time of initial application, for new business only; it is not used to re-underwrite people, renew them, or raise their rates at renewal. In addition, he noted, their model is not used as the sole reason to deny someone coverage, and it often qualifies people for State Farm's preferred companies. He offered an example. MR. PARKS told the committee that whenever a credit score does keep people from getting the company's very best rate, State Farm provides notification of "that line in the Fair Credit Reporting Act." He explained that the notification tells the people that they were adversely impacted by information contained in the consumer report, and where they can get a copy of that report. In addition, State Farm provides those people with up to four reasons - related either to credit or loss history - that explains what impacted their scores. He listed some of the more common reasons given: prior losses, the number of revolving accounts, the number of late payments, and collections not related to medical bills or utilities. MR. PARKS referred to the previous discussion regarding causation. He said credit components don't cause people to have future losses, but they are a predictor. He said, "What we've done is we've tried to capture a highly predictive element that is extremely complete to help better predict the future losses that someone might have." TAPE 03-33, SIDE B Number 2390 MR. PARKS referred to a 1999 study by the American Insurance Association that broke income levels into nine groups, from those people making below $15,000 annually to those making more than $125,000. What the study showed, he explained, is that there is no significant difference in the credit score that came back "for those individuals." Mr. Parks said [the use of credit scoring] isn't necessarily about decreasing the loss rate; it's just about trying to award equitable distribution of how premiums are charged for the losses that will be experienced. He concluded by reiterating that State Farm continuously validates its own underwriting model. He added that as the company gains additional experience with the model, it will be adjusted as appropriate. Number 2302 MICHAEL LESSMEIER, Attorney at Law, Lessmeier & Winters, Lobbyist for State Farm Insurance Company, testified that the issue before the committee is really one of fairness. He said repeated studies show that this tool works. He added, "If it works, and if it allows us to price our product appropriately, commensurate with the risk that is presented, we should be allowed to use it." He referred to a paragraph [page 29] of a review done by the American Academy of Actuaries for the NAIC, dated November 15, 2002, which read as follows: Causality is not a requirement for any element in a risk classification system. For example, drivers with past accidents and driving violations have been shown to have higher rates of accidents in the future, and therefore driving record is a useful and commonly accepted element of risk classification systems for automobile insurance. However, histories of past accidents and violations do not cause driver[s] to have more accidents. The rating practice that does exist is based on the fact that, as a group, drivers who have been accident-prone in the past are likely to be accident-prone in the future. MR. LESSMEIER opined that the review is excellent. He noted that in response to a recent study by the University of Texas, as well as the Alaska and [State of] Washington studies, the NAIC asked the American Academy of Actuaries to do a follow-up study to address some of the questions asked. He said there is some hope that this study will be completed in the next couple of months. MR. LESSMEIER discussed [last year's proposed committee substitute (CS) for SB 320, sponsored by Senator John Cowdery, Version 22-LS1462\T, Ford, 4/30/02]. He said it addresses a number of concerns raised at today's hearing. For example, it has a confidentiality provision such that, upon request, an insurer shall file its model with the Division of Insurance, and it provides appropriate protections for that; it requires an insurer that takes an adverse action based on credit to give notice to the insured in writing, stating the significant factors of the credit history or credit score that resulted in the adverse action; and it provides that if a disputed credit history is used and later found inaccurate, there would be a retroactive fix. Furthermore, it lists a number of things that an insurer cannot consider in terms of using credit; each was discussed in great detail with much debate, he said, and six were ultimately agreed upon. Mr. Lessmeier urged the committee to use this as a platform because it has involved a significant amount of work and doesn't totally ban the use of credit. MR. LESSMEIER referred to Appendix A in the previously noted report by the Division of Insurance, which shows that the division sent out solicitations to 91 companies. By the time the "response time" came out, there were responses from 27 companies that no longer wrote business in Alaska. He opined that [credit scoring] is a useful tool that's good for policyholders. He encouraged the committee not to ban [credit scoring] and to consider last year's proposed CS for SB 320. Number 2136 REPRESENTATIVE GRUENBERG noted that the provision he'd previously spoken about with Mr. Lo is found in the [proposed CS for SB 320] on page 4, lines 16-22, and a similar provision is found in the NCOIL version on page 5, subsection (9)(b). He offered his estimation that [the proposed CS for SB 320] is a "much more comprehensive treatment of the subject of the Fair Isaac issue." Number 2099 MR. LESSMEIER noted that credit [as a basis] has been used in Alaska for approximately three years. He referred to the previously cited statistics regarding the percentage of the market held by [State Farm]. Mr. Lessmeier said during that three-year period [State Farm] has had five complaints about the use of credit. He said, "This has not been an issue that has presented a significant number of problems for our policyholders and, in fact, that is verified by the lack of number of complaints with the Division of Insurance - not just the lack of ... complaints, but the lack of number of complaints that are found to be valid." He added that he doesn't know that there has ever been a single complaint regarding the use of credit in Alaska found to be valid. Number 2049 REPRESENTATIVE SEATON said one thing he has found disturbing is the number of inquiries made about somebody's credit history. He said, "It seems to be one of these factors that there could be many reasons for." He asked Mr. Lessmeier if that is a factor that he finds necessary to use in [an insurance company's] determinations, or whether it could be eliminated, which would eliminate a lot of opposition from people who "obviously know that there could be many reasons for getting credit checks, and if that's going to influence your automobile insurance, then that seems, on the face of it, unfair." He added, "The other factors I could understand. This one I can't quite." MR. LESSMEIER answered that [last year's proposed CS for SB 320] addressed that issue with a provision that, although it didn't totally eliminate [credit inquiries], did limit the number of them. He added that there was "a little bit of a qualifier on that." Number 1982 ELIZABETH MOCERI, Regional Counsel, Allstate Insurance Company ("Allstate"), told the committee Allstate is the second largest insurer in Alaska. She said Allstate is concerned about the insurance market in Alaska and about the availability of insurers, because it believes in a healthy, competitive insurance market, as well as healthy markets for its own agencies. Ms. Moceri said Allstate has exclusive agents who only sell Allstate and who are waiting for the company to use credit rating so that they can sell more insurance. She explained, "For Allstate, it's our growth strategy. We use credit and ratings so we can sell more insurance at a more competitive price." She noted that Allstate agencies are small businesses. Allstate also supports independent agencies in Southeast Alaska, she added. She mentioned letters from Allstate agents who are concerned about the market. MS. MOCERI pointed out that Allstate only uses credit in underwriting; if a consumer comes to Allstate, the company runs the credit and either accepts or rejects the consumer. She said, "When we use it in rating, when you come to us, we have a price for you. And we have a rate for you, even if you don't have a credit history." She explained that this enables Allstate agents to write more business. She said Allstate is writing more business in the states in which it has introduced credit in rating. Noting that Allstate conducted its own study with its own insured, she said, "We took a look at their credit, we took [a look] at their loss, and there's a connection. It's predictive and it works. And if it didn't work, we wouldn't use it." Number 1888 MS. MOCERI noted that Hawaii is one of the states that she works with. In Hawaii, she reminded the committee, age, sex, and length of driving experience cannot be used [in rating for insurance]; for example, a 16-year-old driver pays the same as a 40-year-old woman who has a lower loss potential. She said there are huge subsidies built into the risk; [Hawaii's] are the thirteenth highest in the nation in terms of overall rate. "A lot of people are paying for other people," she remarked. She said Allstate has to turn away people who've had tickets, which makes it difficult for the company to grow. MS. MOCERI said, "They did pass tort reform in 1999. And right now we're sort of seeing some legislative creep-back." She commented that from an insurer's point of view, it is expensive to do business in Alaska, which has unique laws relating to attorney fees. She said it is unique to Alaska that people can buy up to $1 million in uninsured motorist [insurance], while only having $100,000 limits for someone else. She emphasized that [Allstate] pays a lot of money in claims in Alaska, which goes into the cost of "the money you're paying out." Number 1783 REPRESENTATIVE GRUENBERG referred to Ms. Moceri's previous statement regarding the 40-year-old woman who pays for the 16- year-old driver, and asked, "Isn't that the nature of insurance to spread the risk?" MS. MOCERI responded that [Allstate] looks at groups and is able to refine credit to be predictive for each group. Using credit goes beyond the use of age, gender, and marital status, and allows companies to take in people that they would normally not have been able to in the past. The ability to offer insurance is needed for a healthy, competitive market, she opined. She suggested, "You could have complaints here of people saying, 'I can't find insurance.'" She said the Division of Insurance has indicated that its assigned risk pool - people who cannot find insurance - has increased. REPRESENTATIVE GRUENBERG commented that he has heard a lot from people in the insurance industry regarding the problems caused by the "uninsured motorist" law in Alaska. He suggested that at some point perhaps the legislature ought to revisit that issue if it is causing many problems for people. Number 1695 MR. NIEHAUS commented that the committee has not heard from the approximately 12,000 policyholders who would suffer if credit were eliminated; therefore, he urged the committee not to throw the baby out with the bathwater. He said a lot of work had gone into making a CS, and that the Division of Insurance has done a great job in making sure that "insureds" file the data properly and that it justifies all of their actions. He urged the committee to consider the CS [for last year's SB 320]. Number 1659 MR. SORICH, regarding the previously mentioned consumer inquiries and inquiries showing up on the credit report and being reflected the insurance score, noted that the Fair Isaac model does not include nonconsumer-initiated inquiries. Also, he noted that under that model, inquiries that the consumer makes for mortgages or car loans within a 30-day period are condensed into one inquiry. Noting that this provision is also in the NCOIL model Act, he said he thinks it is a positive provision in the model. [HB 47 and HB 5 were held over.]
| Document Name | Date/Time | Subjects |
|---|