2001: Insurance Scoring In Personal Automobile Insurance - Breaking the Silence

Price : $1,250.00

Many insurers contend that the use of credit data is one of the most powerful underwriting and pricing tools available. More insurers are beginning to use credit data in underwriting and/or pricing; others are expanding their use of credit data. How are credit data being used? Are consumers likely to benefit from insurers' use of credit data? Will there be consumer backlash? One sure thing: consumers are concerned with how insurers are using credit data. Bottom line: the use of credit data is risky business. (113 pages)

Table of Contents

1. Introduction



2. Summary and Conclusions



3. Personal Auto Underwriting and Rating 101—A Primer
   • Underwriting and Risk Classification
     • The Personal Automobile Underwriting Process
     • The Principles of Personal Automobile Risk Classification
     • Risk Classification—Theoretically Sound Yet Subject to Criticism
     • Risk Classification and Rating Methodologies Continue to Evolve
   • Insurers Encounter Many Obstacles in Pricing
     • Regulatory Restrictions
     • Socially Unacceptable Criteria
     • Data Integrity
   • Increasingly Insurers Rely on Data from Outside Parties to Underwrite Risks



4. The Expansion of Credit Data into Personal Auto Insurance
   • A Brief History
     • Fair Isaac—The Pioneer
     • Credit Reporting Agencies as Information Providers
     • The Fair Credit Reporting Act
     • The Credit Report Dispute Process
   • A Synopsis of Credit Scoring Models
     • Four Scoring Models—Their Strengths and Weaknesses
     • Custom Models—Important Considerations
   • The Application of Credit Data into Insurers’ Business Processes
     • To What Extent Are Insurers Using Credit Data? Why Do Some Insurers Still Not Use Credit Data?
     • When Did Insurers Begin to Use Credit Data?
     • How Are Insurers Using Credit Data?
     • How Frequently Do Insurers Use Credit Data? What Are the Criteria Dictating Use?
   • Strategic Approaches to Exploit the Benefits of Insurance Scoring
   • Conclusion



5. The Insurance Scoring Debate—A Fresh Look from Both Perspectives
   • Proponents of Insurance Scoring—Their Claims
     • Proponent Claim #1—Correlation Between Credit Data and Loss Ratio
     • Proponent Claim #2—Credit Data Allow for Underwriting Exceptions
     • Proponent Claim #3—Credit Data Are Nondiscriminatory
     • Proponent Claim #4—Credit Data Reduce Underwriting Subjectivity
     • Proponent Claim #5—Most Consumers Have Good Credit
   • Critics of Insurance Scoring—Their Claims
     • Critic Claim #1—Credit Data Overlap with Data Already Being Used
     • Critic Claim #2—Credit Data Are Unfairly Discriminatory
     • Critic Claim #3—Credit Data Are Not Accurate
     • Critic Claim #4—Insurance Scoring Is Socially Unacceptable
   • Conclusion



6. A Long Rocky and Uncertain Road Ahead
   • The Major Stakeholders and Their Perspectives
     • Consumers
     • Regulators
     • Distributors
   • Financial Services Convergence—Opportunities and Challenges
     • Bank/Insurers Affiliations
     • Mergers and Acquisitions

Appendices

• A.Conning & Company Ratings on Companies Covered in This Study



• B.Glossary



• C.Supplemental Data

Introduction

More and more often we hear the term “insurance scoring.” It seems that the terms “insurance scoring” and “credit scoring” are used interchangeably. What is insurance scoring? Conning defines this term as a score derived from credit report data that has been determined to be highly predictive of future claim costs. Insurance scoring or insurance scoring models as discussed in this study are not synonymous with credit scoring as used in the loan underwriting process in the banking industry to predict loan defaults. Fair Isaac and Company Inc. (referred to throughout this study as “Fair Isaac”) is the pioneer in credit scoring and insurance bureau scores. Insurance and credit scoring models have proliferated throughout the financial services industry during the 1990s. According to E-Loan an online mortgage broker and lender nearly 80% of all mortgage-lending decisions now use a credit score as the primary determinant of an individual’s credit risk. In fact 92% of respondents to Conning’s survey of the 100 largest personal automobile insurance writers indicate that they use insurance scoring models today.

Insurers strongly support the use of credit data in the underwriting and pricing of personal automobile insurance – pointing to evidence that the use of these data enhances their ability to predict future claim costs. In addition they claim that data-driven underwriting and pricing result in more efficient accurate and consistent underwriting and pricing.

Nonetheless insurance scoring has become one of the most contested topics in the history of personal automobile insurance underwriting and pricing. The controversy surrounding the use of credit data escalated in the late 1990s and has not lost momentum in the new millennium. At the root of the controversy is credit data predictability. For example is an insurance score a predictor of loss ratio performance or is it the other data (i.e. age marital status location or a combination of these rating variables) that are incorporated already in insurers’ underwriting and pricing methodologies? Is the insurance score a proxy for failure to maintain assets appropriately and manage one’s finances or is it a surrogate for identifying people unemployed disabled divorced single with children or not collecting child support payments? Questions such as these have caused insurance scoring to be controversial.

Early in its research Conning identified common claims by proponents and critics of insurance scoring. Both sides present cogent arguments and supporting data to substantiate their claims. Insurers are using scoring models to refine risk classifications enhance pricing and marketing methodologies and develop new retention models. Conning learned that smaller insurers are using it predominantly in their underwriting processes while larger insurers appear to be focusing on underwriting pricing and more sophisticated market segmentation techniques. The latter may offer a tremendous competitive advantage.

In conducting research for this study Conning identified four scoring models that it believes insurers are using today. These models have evolved over time to become more complex hard to duplicate and difficult for consumers and regulators to understand.

Consumers’ credit behavior (i.e. consumption or lack of consumption of credit) is likely to depend on lifestyle financial knowledge and condition experience of a catastrophic event (e.g. loss of job medical illness divorce) cultural beliefs and attitude. Each of these factors has the potential to increase or decrease one’s insurance score a score that may affect their underwriting acceptability and/or the price they will pay for automobile insurance. Conning estimates that the use of insurance scoring models will affect the underwriting and/or pricing decision for as many as two million households that apply for automobile insurance in 2001.

The increased use of these models comes at a time of concern among consumers about the accuracy of credit data and how credit data are used. The Federal Trade Commission also is concerned about an emerging trend in identity theft. Most recently the Treasury Department reported that the number of identity theft occurrences reported by banks in 2000 has more than doubled over 1999. Further deterioration of consumer household credit quality continues and increased concern about privacy is driving a plethora of legislation that could result in more challenges to insurers’ use of credit data.

Nevertheless insurers continue to use credit data. In fact they are expanding their use of credit data and more are beginning to use the data. Yet some may not even be aware of all the potential risks. Insurers point out that their use of credit data benefits consumers because most have a good credit rating and underwriting decisions now are more objective and efficient. Thus the underwriting decision process has moved from a subjective judgment to one that is intensely data-driven with less bias and animus. Critics contend however that the underwriting process has become less humane.

This study was undertaken with several goals in mind. First Conning sought to verify that the use of credit data is an extension of insurers’ efforts to refine their underwriting and risk classification processes and make them more efficient accurate and consistent. The use of credit data does not signal a major change in insurers’ underwriting and pricing objectives. Second the study examines the positions of both proponents and critics on the use of insurance scoring. Conning sought to research both sides understand the essence of their positions and not become immersed in all the complex mathematical formulae. Third Conning sought to learn how the industry is using credit data today. This study summarizes the results of Conning’s survey. Finally this study examines the implications of insurers’ increased and expanded use of credit data in the underwriting and pricing processes.

Conning conducted both primary and secondary research as part of this study. First we surveyed the 100 largest personal automobile insurers and conducted follow-up telephone interviews with a number of them. Companies accounting for 43% of the 1999 personal automobile premium volume responded to the survey. Second Conning conducted an extensive interview with Fair Isaac. Finally Conning conducted telephone interviews and solicited feedback from both independent and exclusive agents to obtain their perspectives on insurers’ use of credit data.

The basic structure of this study is as follows.

• Chapter 2 contains a summary of Conning’s key findings and conclusions.



• Chapter 3 examines the personal automobile insurance underwriting and classification processes. In this chapter Conning explains how the use of credit data in risk selection is consistent with insurers’ underwriting objectives. This includes an examination of insurers’ underwriting and risk classification objectives and the impetus behind the emergence of third-party data.



• In Chapter 4 Conning discusses Fair Isaac and the origin of insurance scoring. We examine the Fair Credit Reporting Act and the duties imposed on credit reporting agencies and insurers. We also examine the different approaches insurers are taking in their use of credit data. The information presented is based on survey responses.



• Chapter 5 examines the claims being made by both proponents and critics of insurance scoring. The chapter relies heavily on secondary research due to the proprietary nature of insurers’ scoring models.



• In Chapter 6 we examine the implications of insurers’ use and expansion of credit characteristics into underwriting and pricing. Later we examine the emerging trends that are likely to determine insurers’ future success or failure in their use of insurance scoring models.