Fighting Insurance Fraud - Survey of Insurer Anti-Fraud Efforts
Introduction and Executive Summary
This report presents findings from a survey conducted by the Insurance Research Council (IRC) and Insurance Services Office, Inc. (ISO), to learn what insurance companies are doing to fight fraud. IRC and ISO distributed a survey to all property-casualty insurers in the United States. The 353 responses represent small, medium, and large companies and 73 percent of the property-casualty insurance market. Findings show how insurers perceive the problem of fraud and the corporate resources and strategies their companies are using to fight it.
Said to be the second largest economic crime in America after income tax evasion, insurance fraud is both pervasive and expensive.[1]1. Coalition Against Insurance Fraud, "Coalition Fact Site." World Wide Web: http://www.insurancefraud.org/ Newsletter/MarApr97/facts.htm, September 2001. Defined as "any deliberate deception committed against an insurer or producer for the purpose of unwarranted financial gain,"[2]2. Insurance Institute of America, "Insurance Fraud and Anti-Fraud Activity," Focus Series, 2000, p. 1. insurance fraud drives up costs for insurers and premiums for policyholders. Fraud may be committed by applicants for insurance, policyholders, third-party claimants, and professionals who provide services to claimants, agents, and company employees. During the past decade, insurers, law enforcement officials, state and local governments, and industry groups have marshaled their resources to identify and combat fraud, trying to keep pace with what some call "the fraud amoeba," or the new and old scams that threaten the insurance business.[3]3. National Insurance Crime Bureau, International Association of Special Investigation Units, and Coalition Against Insurance Fraud, "National Insurance Fraud Forum," October 2000. World Wide Web: http://www.insurancefraud.org, October 2001; R. Panko, "Getting a Jump on Crime," Bests Review, October 1999, p. 76.
This report is the third in a series from the IRC describing what insurers are doing to fight fraud.[4]4. Insurance Research Council, Fighting Fraud in the Insurance Industry, 1992, 1997 (2nd ed.). The previous reports found that a large segment of the property-casualty insurance industry had adopted systematic programs and had increased funding to detect and deter insurance fraud. This report shows that insurance companies of all sizes continue to be active players in a fraud-fighting network that includes industry groups as well as federal and state law enforcement agencies.
Since the 1997 report, new tools have strengthened the ability of insurers to fight fraud. The Internet has enabled insurers to share information about suspicious claims with each other and has raised the prospect of sharing with law enforcement and other financial institutions. State and federal laws have strengthened the ability of insurers to identify suspect claims, to report successful investigations of suspect claims to law enforcement, and to have prosecutors win convictions. Advances in computer technology have led to the development of enormous claim databases, which facilitate identification of fraud networks and rings, and sophisticated data-mining programs, which help to identify patterns within those files. Industry groups have developed training programs to assist insurers in detecting and investigating fraud and information campaigns to raise public awareness about the threat and cost of insurance fraud.
The current study builds on the IRC's previous studies by doubling the size and expanding the composition of the sample. The current sample of 353 includes many small and medium insurers that did not participate in the earlier studies. The IRC's 1997 report contained findings from 150 insurers, of whom 65 percent had market share of 0.1 percent or higher, 24 percent had market share of between 0.1 and 0.01 percent, and 11 percent had market share of 0.01 percent or lower. The 2000 sample has a very different distribution: 20 percent have market share of 0.1 percent or higher, 41 percent have between 0.1 and 0.01 percent, and 39 percent have 0.01 percent or lower (see Appendix 1). In addition, the current report covers practices, such as use of databases and statistical modeling, that were not available or not applicable to most insurers when the prior reports were written. For these reasons, with only one exception, findings from the three reports are not formally compared.
The report opens with a review of the problem of fraud in the insurance industry from the insurers' perspective: estimates of its seriousness, the effectiveness of company fraud-fighting efforts, and the number and type of claims most likely to be affected (Chapter 2). The report continues with a discussion of corporate anti-fraud programs (Chapter 3) and concludes with a review of methods companies use to detect and investigate suspect and fraudulent claims (Chapter 4). Throughout, differences in practices and perceptions between small, medium, and large insurers are analyzed when they are significant. Comments that respondents wrote on the survey are highlighted in the text to illustrate individual perspectives on survey questions.
Methodology
The insurer survey (see Appendix 3) was mailed to all insurers listed in the A.M. Best Company property-casualty file for 1999. That file provided contacts for 1,042 insurers (683 individual companies and 359 insurer groups). Throughout this report, the terms company and insurer are used interchangeably to describe single- and multiple-company insurers. The survey was mailed to insurers in September 2000. The surveys were most often addressed to company claims directors or claims executives; however, in approximately 30 percent of mailed surveys, the CEO or presi-dent was used as the contact. The survey instructed recipients to forward the surveys to persons responsible for fraud control at their companies. A second mailing of the survey was conducted in October 2000. A total of 390 respondents returned the survey. If a respondent from a lead company indicated he or she was responding on behalf of all insurers in a group, only the lead respondent's answers were used in the analysis. After duplicate respondents from the same company were eliminated, the analytic data set included 353 insurers representing 73 percent of the property-casualty insurance market for 1999.
Whereas previous IRC studies on insurance fraud targeted insurers with top market shares nationally and in specific states, this study distributed surveys to all insurers registered with A.M. Best in 1999. As a result, the present study contains a greater proportion of smaller insurers than do previous IRC reports (see Appendix 1). For the purpose of analysis, the data set was divided into three groups by company size, as measured by premium volume. The three groups are as follows:
- Large insurers (n=50). These insurers are the top fifty respondents in the study ranked by direct written premium volume. Each of the insurers in this group had at least $550 million in direct written premiums in 1999. These fifty participating insurers also represent 64 percent of the property-casualty market in 1999. (Market share is calculated by dividing a company's direct written premiums into total direct written premiums for the property-casualty market in 1999.) The large insurers in this study also represent 72 percent of the auto insurance market and 57 percent of the workers compensation market in 1999.
- Medium insurers (n=163). This category is composed of insurers with at least $31 million but less than $550 million in direct written premiums in 1999. These 163 participating insurers represent 8 percent of the property-casualty market in 1999.
- Small insurers (n=140). The group of small insurers is defined as those with less than $31 million in direct written premiums in 1999. These 140 participating insurers represent 0.5 percent of the property-casualty market in 1999.
The 353 respondents to the survey identified their positions within their companies as follows: 40 percent indicated they were a home office claims manager or vice president; 23 percent that they were in SIU management; 20 percent, a CEO or other senior management position; 5 percent, claims staff; and the remainder either SIU staff, a regional or state claims manager or vice president, a legal officer, or "other/missing" (3 percent each).
Key Findings
Key findings of this study are the following:
- Insurers consider fraud "a serious problem" but their companies' anti-fraud efforts only "moderately effective." Sixty-eight percent say their companies' anti-fraud programs address claims fraud "thoroughly," 19 percent say they address premium fraud "thoroughly," and 25 percent say they address application fraud "thoroughly." Slightly more than one-third (37 percent) think the amount of fraud their companies have experienced has increased over the past three years. Forty-two percent think that 21 percent or more of total claims contain "soft" fraud, but only 6 percent think that 21 percent or more of claims contain "hard" fraud.[5]5. "Soft" fraud was defined on the survey as exaggeration of an otherwise legitimate claim, and "hard" fraud was defined as deliberate attempts to stage a type of loss, such as accident, theft, or arson. They agree that fraud is most prevalent in the private passenger auto and workers compensation lines of business.
- Eighty-two percent of the 353 insurers responding to the survey say they have an anti-fraud program at their companies. One hundred percent of the large insurers, 91 percent of the medium insurers, and 64 percent of the small insurers have an anti-fraud program. Comments written on the surveys indicate that those without anti-fraud programs specialize in lines where fraud is secondary or rare. Other data indicate that those without formal anti-fraud programs are active in fighting fraud using standard underwriting and claim procedures. Sixty-three percent of the companies say that the state or states in which their companies do business require an anti-fraud plan. However, only 13 percent of insurers doing business in these states (n=213) consider state requirements and guidelines "very useful."
- Because fewer than one-third of respondents answered questions about their companies' expenditures, estimates of industry-wide spending on anti-fraud efforts are not reliable. The response rate suggests that insurers are unable to isolate anti-fraud expenditures in their budgets or unwilling to share what figures they have with other insurers and the general public.
- To detect fraudulent claims, at least four out of five insurers use fraud awareness training (87 percent), manual red flags or indicator cards (81 percent), and external database searches (80 percent). No more than one in four use some of the newer methods, such as e-mail, automated red flags or indicator cards, mathematical or analytical techniques, or geographic data mapping. However, higher percentages report using external databases, with the most popular, ISO ClaimSearch®, being used by 74 percent.
- Among companies that conduct training in fraud recognition, most use home office claims and special investigative unit (SIU) personnel. Only 48 percent say their companies have tried to educate their policyholders or the public on how insurance fraud affects them. Large insurers are more likely than medium or small insurers to be active in public education about insurance fraud.
- Insurers agree that strong legal and industry support is "very important" to fighting fraud. Insurers favor tough enforcement of existing law by federal, state, and local law enforcement, combined with support from state insurance fraud bureaus and industry associations and organizations.
- Two hundred of the 353 companies in the sample report having an SIU. All of the large insurers, over two-thirds (69 percent) of the medium insurers, and just over one-quarter (27 percent) of the small insurers have an SIU. This representation projects to 71 percent of the total market having an SIU, a decline from the projected 76 percent of the total market that had an SIU in 1996, possibly explained by differences between the study samples, specifically a 2000 sample that better represents the industry as a whole.
|
