TO THE POINT
Agency workflows will feel the pinch
By William Horton, CPCU
In the years since the advent of comparative rating systems, writing personal auto insurance has been a relatively stable process. In recent years, the growing acceptance of the SEMCI concept has made the process more efficient, causing data to flow easily between agency management systems, rating software, and company mainframe computers.
However, storm clouds are gathering in what had been mostly blue sky. Two fronts are driving the clouds, one being the growing trend of using credit scores in underwriting auto insurance, and the other being the proliferation of insurance company Web sites for agents. There's no doubt that both will have a major impact on the way we do business, and if left unchecked this gathering storm could be disruptive to our current workflows and choices of companies.
In Connecticut, where our agency is located, the vast majority of the companies we represent are awaiting approval of new credit-based auto programs. Some of these are fairly simple programs, but several are complex. It's reported that one major carrier will have 96 different rate levels based on credit and underwriting. It appears that the trend will be toward complex programs like this one. The goal is to be able to write virtually any auto risk for the "correct price" based on credit and underwriting factors.
How will the companies be able to rate and underwrite these new programs? Apparently there will be a "black box" at the company Web site, and all the agent has to do is provide the client information, and the firm price quote will be generated. The company system will order the credit report, the CLUE and the MVR to develop an accurate rate. Then, if the agent wants the policy, it can be ordered via the Web site.
This would be a fine system if I were a direct writer. But I'm not. My agency represents 17 companies that write auto insurance. Keying in underwriting information 17 times to get 17 quotes won't help my bottom line.
What about the comparative raters? While most companies say they will continue to provide their rates to the vendors, that won't be enough to ensure a seamless workflow.
The heart of the matter
The major obstacle to the current workflow is the credit report itself. Agents have long since gotten used to ordering CLUE reports and MVRs to get an accurate rate. And since the reports are universal, ordering them once assures accurate underwriting with all the companies. Not so when ordering a credit report.
Currently, all the companies we represent are using different credit models. Some reports say "acceptable" or "not acceptable." Others give a numeric score, and most give a score such as "superior," "good" or "fair." So what's good for one company isn't necessarily good for another. This disparity is frustrating for agents, especially given the fact that most reports come from the same credit database. Using myself as a guinea pig, I was surprised at the range of prices I qualified for, given different interpretations of the same data.
The increasing complexity of the rating process makes it harder to get an accurate quote from a comparative rating product. And even if you can get a rate, you have to order a separate credit report for each company--which often entails using additional software programs. So what's an agent to do? Are there any alternatives?
Some improvements will come by way of changes in the current technology and workflow. Rating and management systems need to link to the Internet to get accurate information and give accurate quotes, while enabling agencies to avoid multiple data entry.
Inroads are being made. Real-time processing between agencies and companies is taking place. IVANS is licensing the real-time technology (formerly referred to as WARP) developed by Applied Systems and is distributing it as Transformation Station. As of August 2001, Transformation Station has several in-production companies and some 40 "live" agencies.
Another important development is AMS TowerStreet's SmartWorks product. Conceptually a home run, it blends agency- and Internet-based multi-company rating with underwriting information and data validation. As of August 2001, SmartWorks, which supports personal lines insurance, is available in four states.
Another solution would be to decrease the number of companies we do business with, or to shift business to companies that don't use the credit score/Internet business model, which would allow us to keep our present workflow. In an ideal world, though, agents shouldn't have to make this sort of choice. Companies should keep the agent's "ease of doing business" in mind--and avoid use of a complex credit model.
Enough blame to go around
Part of this impending crisis is our own doing because we agents didn't provide adequate input to our vendors and carriers and demand that they develop useful technology solutions. We agents need to voice our concerns collectively with the following goals in mind:
- Remind insurance companies that we are not exclusive agents. We need the companies to keep our ease of doing business in mind. Companies also need to cooperate with our rating and automation vendors to help smooth out our choppy technology stream. And we need to encourage standardization in the use of the credit models to make their usage more like MVRs and CLUE reports.
- Remind rating vendors that if we can't get comparative rates, we're no longer independent agents. Rating vendors need to step up to the plate and get us products that work. The technology is there. Agents need to buy rating vendors' new products promptly and work with those vendors in perfecting the products.
- Remind automation vendors that they must cooperate with the rating vendors and the companies to develop innovative products. Our databases don't do us any good if they are islands unto themselves. The Internet is about sharing of information, and we need to be able to share, also.
By working together, we're more likely to have the products and systems in place that will make all of us big winners in the age of the Internet. *
The author
William Horton, a 17-year veteran of the insurance business, is the manager of personal lines for Stone Insurance Agency in Meriden, Connecticut.
Liability problems may loom for agents
By Kevin Hennosy
Twenty years ago a statistics professor at a large Midwestern university lectured to his students on the concept of correlation. To illustrate his point the professor showed figures for the population of the state of Ohio and the population of New York City. For several decades, these two figures were roughly equal and grew at the same rate. The professor described the relationship between these two figures as a statistical correlation. He then asked rhetorically, does this mean that the state of Ohio controls population growth in New York, or vice versa? After a few seconds of silence he answered his own question, "These numbers demonstrate correlation, not cause--don't confuse the two. It can get very expensive if you do."
For the past decade, in response to aggressive marketing by credit bureaus, insurance professionals have become aware of a correlation between select elements of credit information and certain rates of property/casualty claims. Over the past three to five years, growing numbers of insurers have modified underwriting and pricing methods to incorporate credit information.
When this business trend began, producers and their trade associations generally supported the use of what can be called credit-based underwriting and pricing (CBUP). Insurers presented the concept to producers as the greatest thing to hit the insurance industry since Blaise Paschal published the Theory of Probability. Numerous producer-training sessions included sections that demonstrated the correlation between credit and claims.
As insurers implemented CBUP programs, producers began to hear complaints from customers who were denied coverage or charged higher rates. And even those consumers who benefited found the concept hard to grasp, although they didn't complain about lower rates. In short, people do not understand why having good credit makes the tree in their front yard less likely to fall on their house.
Then came the accusations of wrongdoing. Organizations that advocate for minority groups began to level charges of "disparate impact upon protected classes" and accused insurers, as well as producers, of redlining.
At about the same time, some insurers directed producers to directly contract with credit bureaus for credit information, while the credit bureaus inserted "hold-harmless clauses" in those contracts with producers that said, "If you get sued for decisions made by using information we provide, you are on your own!"
Policymakers, like producers, were originally unwilling to challenge the correlation that insurers so clearly presented in testimony and reports. Insurers and credit bureaus provided officials with data of the insurers' choosing and under confidentiality agreements that established the correlation between credit and claims. In state after state, the use of CBUP was approved.
According to a July 31, 2001, Conning & Company study, 58% of 92 insurers surveyed use credit factors in the underwriting and rating process. These insurers tend to be large companies that concentrate risk through segmenting techniques in "an aggressive price penetration strategy."
In a public statement announcing the report, Conning & Company Assistant Vice President Clarence Smith said: "Insurers have been sold on the competitive advantage of utilizing credit data in underwriting and pricing."
As the use of CBUP has grown, so has the skepticism of policymakers and producer groups. If you want to start a fight around the halls of government, just bring up the topic of insurers using credit information in underwriting. The practice has become a hotly contested issue in state legislatures and insurance departments across the country.
What is it?
Just what is CBUP? The answer to that question varies to some extent from insurer to insurer; however, at a meeting of the Independent Insurance Agents of Texas (IIAT), a panel discussion that addressed the practice provided a general description of CBUP procedures.
Insurers contract with one of several credit bureaus to conduct CBUP activities. These bureaus include Fair, Isaac & Co.; Trans Union; and Equifax. The bureaus provide credit history data on individuals and compile data elements from insurers' policyholder data.
According to the IIAT panel discussion, Fair, Isaac & Co. uses data from approximately 12 insurance companies that the agency believes represent the industry as a whole. Fair, Isaac draws data elements that include premium and incurred losses for a selected time period, which the agency compares to policyholders' credit reports for the same time period.
The agency then compiles similar data from four time periods into a database that holds elements of 1.4 million policies, $1.5 billion in earned premium and $900 million in incurred losses. Fair, Isaac then compares these data to two sections of credit reports--credit account trade line (balance, delinquency, payments due) and additional credit inquiries. The credit score results from this analysis.
While proponents of the use of credit scores often describe them as a uniform and scientific measure, the companies that use credit scores do so in parochial ways. Each company decides how to interpret the scores in order to predict future losses. Differences in companies' books of business and business aims shape the use of credit scores in underwriting and pricing.
Some companies never take possession of credit scores by requiring or encouraging producers to obtain them directly from third-party credit agencies. This approach places liability for underwriting and pricing decisions squarely with the producer, especially in cases where credit agencies require producers
to agree to a hold-harmless arrangement before granting access to their credit score information.
Why?
Why do insurers use credit information in underwriting and pricing?
The record of the IIAT panel discussion contains the following explanation: "The theory is that a customer with a poor credit history is more likely to present a claim and could also be a more likely moral risk." A review of numerous pieces of testimony filed with state and federal officials by insurer advocates echoes this assertion.
The National Association of Independent Insurers (NAII), a lobbying group for property/casualty companies, distributes a pamphlet with the title "What's Credit Got To Do With It?". The NAII pamphlet says: "Several independent studies have proven a strong connection between credit history and the likelihood of an individual filing a claim."
The pamphlet begins with an acknowledgement that "a number of state legislators and consumer groups have expressed concern about [CBUP's] impact on consumers." The NAII reassures the reader that "major insurance companies, and organizations such as Fair, Isaac and Company and Tillinghast-Towers Perrin have studied the issue and found data to support using credit history as a predictor of claim filing."
Critics
Numerous stakeholders in the insurance business have questioned the premise that the NAII and other groups put forth with such certainty.
Consumer advocates have punished the use of CBUP in testimony and the news media by associating the practice with junk science, bigotry and anti-competitive behavior. Insurers deny and dismiss these charges as anecdotal and tainted.
Any objective observer must agree to the anecdotal nature of what consumer advocates bring to the debate; however, since insurers withhold data and methods necessary for scientific analysis, it seems unfair for insurers to degrade anecdotal analysis offered by critics.
Consumers
First and foremost, consumer advocates question whether credit information provides a measure of insurable risk. Consumer groups hammer home to policymakers that no independent study has been conducted to prove that such a measure exists. Furthermore, consumer advocates remind policymakers that the credit history data has earned a notorious reputation for inaccuracy.
Insurer advocates counter that credit history data can be seen as more accurate than motor vehicle data; however, there are problems with the insurers' position. The study that insurers most often point to compiled and compared complaint data where consumers asked for changes to credit history and motor vehicle records.
Researchers found that more consumer-initiated corrections were made to motor vehicle data. On this basis, insurer advocates have called credit data "more accurate" and have ignored the fact that most consumers have no idea what their credit history is, when it is accessed or how to change it. As a result, correction-based assessments of accuracy serve as a poor measure.
Consumer advocates charge that credit information provides a means for insurers to redline. Credit scores can predict race through means that have nothing to do with risk of financial loss. For example, one element that leads to lower credit scores in minority communities is the lower rate of checking account ownership, which results in no small part from bankers' unwillingness to open branches in minority neighborhoods.
Persons who do not own a checking account will have a lower credit score even if they pay their bills on time using cash or money orders. Such persons show greater initiative and pay higher costs to pay their bills, but they will rank lower on a credit score than the person who writes a check or pays online.
The result of CBUP, consumer advocates charge, will be that the money-order payer, who will more often be a racial or ethnic minority and/or a female, will be less likely to be offered insurance or will pay higher rates than the check-writer, who is more likely to be white.
Consumer advocates also decry the use of CBUP because the practice forces consumers to divulge private and sensitive information that may not serve to measure risk in order to apply for insurance. Credit information, which is highly marketable, falls under the relatively lax financial privacy protection provisions of the Gramm-Leach-Bliley Act. Leading consumer advocates point out that insurers may collect credit information for resale, or cross-marketing, value alone.
Another concern from a consumer perspective relates to whether CBUP practices may violate prohibitions in federal antitrust law, even in the light of the McCarran-Ferguson Act. By using CBUP, insurers can establish and enforce categories of purchasers that restrict consumer choice. That's because the defining elements of these categories were established by pooling data elements from 12 of 1,500 property/casualty insurers.
Producers
Producer advocates have also expressed uneasiness with CBUP. Beyond expressing concern over how their prospects and customers fare under the practice, insurance producers understand that credit-based decision-making facilitates direct marketing campaigns over the Internet and through boiler-room phone solicitations. Insurers' use of CBUP might be viewed as another volley in "the war on the agent." They also express concern about their being exposed to consumer dissatisfaction and legal liability for adverse decisions.
Public officials
In recent years, as the use of CBUP has expanded among property/casualty insurers, state officials have scrambled to establish laws and regulations governing the practice. The provisions of these policy initiatives differ tremendously in scope and approach.
The use of CBUP has received approval from most insurance regulators, although more than 20 states have adopted limitations on the practice. On the whole, state officials have been behind the curve on this issue, and the National Association of Insurance Commissioners (NAIC) has been "Absent With Out Leave."
In the October 1995, issue of the NAIC Research Quarterly, two Washington State Insurance Department staff members asserted that insurers were using questionable science to justify the use of credit history in underwriting and pricing: "The literature certainly indicates a strong need for comprehensive research by a disinterested and qualified third party. Much of the research to date on the accuracy of information and the use of credit reports in insurance has been conducted or financed by entities with an economic interest in the outcome. Additionally, other studies are questioned in the literature reviewed as using unscientific methodology."
Politics
In 1996, the NAIC promised a public policy study of the use of CBUP. The NAIC did not conduct a scientific study of the use of the connection between credit and claims, but it did compile a survey of competing viewpoints--at least initially. The working group charged with developing the report released a draft for public comment in September 1996.
The draft drew considerable political opposition from advocates for insurance companies and credit agencies. Company trade associations charged that the report contained inaccurate and misleading and anecdotal charges, although the company charges were every bit as anecdotal as were the offending sections. Company advocates also complained about a list of policy options in the report, especially one that suggested putting restrictions on the use of CBUP.
During this time period, numerous property/casualty companies were placing pressure on the NAIC by withholding financial database filing fees. Database fee income then represented 40% of NAIC revenue, and the boycott-like activity contributed materially to NAIC operating deficits. In response to this pressure, the NAIC membership and staff leadership looked for ways to make property/casualty insurers happy. In January 1997, the NAIC released a watered down report that expunged many negative comments about CBUP.
The sanitized version of the NAIC report has been useful to company advocates' lobbying efforts. In particular, the insurer advocates have used the NAIC report to counter the charges of disparate treatment of protected classes lodged by fair housing advocates. The NAII states: "After much study, the [NAIC] could not find any studies in the insurance field that demonstrate that the use of credit history in underwriting an insurance risk has had a disproportionate impact on protected classes." Once again, the NAIC did not make an independent assessment of this issue; it merely compiled information presented by competing lobbying groups.
Credibility
In the absence of any clear standard established by the NAIC, individual states have issued numerous reports on the issue of CBUP. At times, industry advocates have lifted portions from these reports and presented them as proof that CBUP works. This practice
has caused regulators much consternation. For example, the NAII pamphlet contains the following statement: "The Arizona and Virginia Insurance Departments also have reaffirmed these findings (studies by Fair, Isaacs, insurers and others) through comprehensive studies of their own on the issue."
When Rough Notes contacted Arizona and Virginia regulators, they jumped at the chance to comment on the NAII's characterization of the departments' work. Arizona Insurance Director Chuck Cohen said: "The pamphlet definitely makes a significant overstatement as to the nature and conclusions of the Arizona 'study.'"
Director Cohen continued: "The statement is a reference to a 10/16/95 letter to then Director Chris Herstam from David Cox, an independent consulting actuary retained by the department to review materials related to credit underwriting practices by Allstate Insurance Co." That letter, Director Cohen pointed out, was written by Cox in response to a subpoena. Cox did not conduct an independent assessment of the issue or opine on other studies cited by the NAII.
"In short, as the current Arizona Director of Insurance, I certainly do not agree that this issue has been comprehensively studied or resolved in Arizona," said Director Cohen. He also expressed his belief that the NAIC should reactivate its study of the issue and develop a model law.
Virginia regulators also expressed disappointment with how their report to the legislature was characterized. The report found a correlation between credit and claims but found no causal relationship. In addition, while the report offered no evidence of redlining under Virginia law, the insurance department "continues to monitor" how CBUP impacts policyholders in its jurisdiction.
Representatives of both departments called the NAII characterization of their reports "overstated." Virginia regulators stressed that the data reviewed for their study were prepared and provided by the insurers and should not be used to extrapolate national conditions.
Black box
To this date, no scientific study exists that can answer the question of whether the use of CBUP "provides an accurate predictor of loss." Companies may indeed hold this proof, but insurer advocates have not been willing to share this information in public session.
The result has been that insurers have looked like Flimflam men who offer the secrets of life in a black box that they refuse to open for inspection. What if the black box is empty? If insurance companies wish to settle this controversy, then the causal relationship between credit score and financial loss must undergo scientific testing.
To receive the credibility
that scientific testing brings, independent researchers (with independent funding) must use transparent and replicable test data and methods. This means no research reports from lobbying houses, credit bureaus, insurance consulting firms, or academics whose chairs and schools receive funding from the industry, and their activities cannot be shielded behind claims of trade secrets or confidentiality.
Even without scientific proof, a great number of industry insiders remain firm in their belief in the validity of CBUP. As Clarence Smith of Conning & Company observed, insurers have been "sold" on the idea of CBUP.
From a related perspective, some observers within the industry believe that insurers will use CBUP out of fear that competitors who do "might have something." On condition of anonymity, a compliance officer for a large property/casualty insurer said, "We don't know if credit scores tell us anything, but it worked for Progressive so we are going to use it." *
The author
Kevin Hennosy, an insurance writer specializing in the history and politics of insurance regulation, covers the proceedings of the NAIC (National Association of Insurance Commiss-ioners) for Rough Notes readers. Hennosy began his career with Nationwide Insurance Companies and then served as public affairs manager for the NAIC. He has written extensively on insurance regulation and testified before the NAIC as a consumer advocate. He is currently writing a history of insurance and its regulation in the United States.
(Crisis ... )
(Secrets ... )
(Secrets ... )
(Crisis ... )
(Secrets ... )