This document has been archived.

Skip To Content Skip To Left Navigation
NSF Logo Search GraphicGuide To Programs GraphicImage Library GraphicSite Map GraphicHelp GraphicPrivacy Policy Graphic
OLPA Header Graphic

NSF Press Release


NSF PR 01-106 - December 19, 2001

Media contact:

 William Harms

 (703) 292-8070

Program contact:

 Daniel Newlon

 (703) 292-7276

Drunken Driving Costs and Risk Measured More Accurately by Economists

Drunk drivers are at least 13 times more likely to cause a fatal crash than sober drivers, according to a new study by Steven Levitt, Professor of Economics at the University of Chicago and Jack Porter, Professor of Economics at Harvard University.

Using an innovative approach to studying drinking and driving, Levitt and Porter were also able to determine which law enforcement strategies are most likely to reduce accidents caused by drunken driving. The research was supported by the National Science Foundation (NSF).

During the holiday season, which is one of the most dangerous times for people to be on the road because of increased incidents of drunken driving, police frequently launch random roadblocks to apprehend drivers. Levitt's research shows that those roadblocks are less effective than increased general surveillance.

"Our results suggest that policies focused on stopping erratic drivers with greater frequency might be more successful," write Levitt and Porter in "How Dangerous are Drinking Drivers?" in the current issue of the Journal of Political Economy. A pilot program using dedicated patrols in Stockton, Cal., reduced involvement of drunk driving crashes by 10-15 percent, the authors point out.

The study provides a more accurate measurement of the risks and costs of drunken driving than was available in previous studies based on data gathered at roadblocks.

To reach a more universal understanding of the impact of drinking on driving, Levitt and Porter studied fatal two-car crashes. By comparing the number of two-car crashes involving two drinking drivers, one drinking driver or no drinking drivers, they are able to apply mathematical formulas to determine the percentage of people estimated to be driving drunk.

They looked at records from 1983 to 1993 in the Fatality Analysis Reporting System administered by the National Highway Transportation Safety Administration and calculated the percentage of fatal accidents in which police said a driver had been drinking as well as those in which the driver had been legally drunk (0.10 percent alcohol in the blood, the most common definition at the time.)

They found that drivers who had been drinking were seven times more likely to cause a fatal crash than sober drivers and those who were legally drunk were 13 times more likely.

"The peak hours for drinking and driving are between 1 a.m. and 3 a.m. when as many as 25 percent of drivers are estimated to have been drinking," Levitt said. During those time periods, about 60 percent of the fatal crashes are caused by drivers who have been drinking, the research shows.

Overall, alcohol is a factor in 30 percent of fatal crashes, which cause 40,000 deaths each year and are the leading cause of death for Americans aged six to 27.

They also used Federal Highway Administration data to help estimate the cost to society generated by loss of innocent lives due to drinking and driving. For each arrest, the cost to society would be $8,000 based on standard economic estimates for the value of a person's life. That cost shared among all drinking drivers, including those not involved in fatal crashes, would be 16 cents a mile for drinking drivers and 30 cents per mile for those who are legally drunk.

Public policies to limit drinking and driving focus frequently on adding taxes to the cost of alcohol and providing criminal penalties for driving while intoxicated. Enforcing drunk driving laws is a more effective means of reducing fatalities than increasing taxes, the research shows. The results of enforcement are uneven among drinkers, however, the scholars said.

"Interestingly, higher beer taxes and tougher punishments for first-time offenders are generally associated with greater danger posed by drinking drivers on average" because the people who choose not to drive are less drunk than the remaining pool of drivers who have been drinking, the scholars found.

The most effective way to decrease overall fatalities is punish "a relatively small fraction of hard-core drunk drivers," the paper points out. Although the punishments do not reduce the numbers of drunken drivers, those who are drunk drive with more care.




National Science Foundation
Office of Legislative and Public Affairs
4201 Wilson Boulevard
Arlington, Virginia 22230, USA
Tel: 703-292-8070
FIRS: 800-877-8339 | TDD: 703-292-5090

NSF Logo Graphic