Driving under the influence of alcohol (DUI) has become an issue of national concern, at both state and federal levels. Increased awareness of the danger of drinking and driving was reflected in a substantial amount of legislative activity during the 1980s--mainly to raise criminal sanctions for drunk driving--and in renewed efforts at law enforcement (Jacobs, 1989). At the same time, there was a consumer revolt against the rising price of autobobile insurance, especially in certain jurisdictions such as California, where Proposition 103--which called for sweeping changes in auto insurance regulation--was enacted by voters in 1988.
Coexisting with applicable criminal law is the drunk driver's liability for property damage and personal injuries resulting from accidents under tort law, which is designed to deter careless behavior by making individuals financially responsible for the consequences of failing to take due care. Individuals purchase insurance to cover such losses when they occur. Such insurance may blunt the potential effect of tort liability, especially if premiums are not set to accurately reflect insureds' expected losses (Abraham, 1986, pp. 45-49; Boyer and Dionne, 1989; Shavell, 1979).
But insurance may not reflect insureds' expected loss for several reasons. First, precautions are costly for the insurer to observe and typically can be observed only after an accident. Second, these ex post observations of care are often inaccurate. Public records of drivers' key performance measures are often not well maintained. Errors, especially omissions, in public accident records have been documented.(1) Third, inferences about a person's driving behavior based on traffic violations may be even more problematic. The relationship between an adverse event, such as a traffic conviction, and subsequent driving performance is not understood. Fourth, some adverse events are rare, and by using a rare event for insurance classification, the insured may be placed in a risk category with a much higher premium based on that one relatively rare mishap (Palfrey and Spratt, 1985). A pricing system based totally on prior adverse events, especially rare ones, would place a substantial premium penalty on each event (Lemaire, 1985). Basically good drivers who make a very occasional lapse may be unduly penalized. But the alternative approach of relying largely on demographic factors for risk classification purposes has been viewed as inequitable, because the factors are beyond a policyholder's control and cannot be said to have caused an adverse record.(2)
To obtain information on underwriting and premium-setting practices of automobile liability insurers, we conducted surveys in 1991 of the 18 largest automobile insurers in the United States, seven companies specializing in nonstandard automobile risk, and all state insurance departments--the latter to document pertinent statutes and regulations. With these data and data from secondary sources, including state assigned-risk plans or other risk-pooling arrangements and the Behavioral Risk Factor Survey, this study addresses several questions: What are the premium penalties imposed by insurers for DUI charges and other chargeable accidents? What is the mix between underwriting and premium-setting practices? Why does observed variation in premiums exist, especially in surcharges imposed on drivers with traffic law violations and chargeable accidents? What are the potential causal factors reflecting both state and company policies? What effect have insurance surcharges on drunk drivers had on the propensity to drive under the influence of alcohol? The article concludes with a discussion of the policy implications of our findings.
How Driving Records Affect Premiums
The automobile insurance industry comprises three different types of insurers: standard insurers, nonstandard insurers, and the involuntary market. The nonstandard market specializes in insuring risks that are rejected by standard insurers, but at appreciably higher premiums (U. …