Total U.S. Highway Deaths Fell in 2003

The Center for Auto Safety is the nation’s premier independent, member driven, non-profit consumer advocacy organization dedicated to improving vehicle safety, quality, and fuel economy on behalf of all drivers, passengers, and pedestrians.

Safety Groups Question Revision by Transportation Department

By Greg Schneider
Washington Post Staff Writer
Wednesday, August 11, 2004

Bush administration officials said yesterday that fewer people died on U.S. roadways in 2003 than the government had previously reported and took credit for an unexpected improvement in highway safety.

Safety advocates accused the administration of sensationalizing the numbers, and questioned how the statistics could have changed so significantly from estimates earlier this year.

A total of 42,643 people died in traffic accidents last year, down from 43,005 the year before and marking the first drop in fatalities in six years, Transportation Secretary Norman Y. Mineta said in a news conference. He noted that because Americans also drove more than ever, the rate of fatalities per 100 million miles traveled dropped to an all-time low of 1.48.

"With the Bush administration’s unprecedented focus on safety, we have reduced the number of trips that have ended in tragedy," Mineta said. "Americans are more likely to reach their destinations safely."

The administration’s positive message was a sharp reversal of the gloom that surrounded an earlier estimate of fatality numbers for 2003, which had shown deaths going up instead of down and reaching the highest total in more than a decade. The previous estimate, released by the National Highway Traffic Safety Administration in April, was based on incomplete data and didn’t take into account an unexpected improvement late last year, said Jeffrey W. Runge, head of the agency.

Safety advocates accused the administration of presenting the new numbers in a sensationalistic way. "We just think it smacks of election-year politics," said Joan Claybrook, head of Public Citizen and NHTSA chief during the Carter administration.

Runge has cited the earlier estimate many times over the past few months to emphasize the need for safer driver behavior. But yesterday he focused on the low rate of deaths instead of the total number, and proclaimed that "now the nation is moving in the right direction."

The revision shaved 577 deaths off the total for 2003 and added 190 to the total for 2002. The government always revises its fatality estimate after more fully analyzing crash data reported by every state, but in recent years has made only minor changes.

Safety advocates said they were perplexed by the unusually steep revision. In fact, from 1996 to 2002, the agency added an average of 143 fatalities to each year’s estimate, according to Clarence Ditlow of the Center for Auto Safety. To reduce the current number so significantly "flies in the face of earlier years," Ditlow said in an e-mailed analysis. "The revisionary death toll published by NHTSA today is a vain effort to conceal a dismal vehicle safety record for the past four years."

Runge said the agency’s statisticians were troubled by this year’s wide margin of error and are still studying it. The most likely explanation, he said, is that safety belt use increased sharply following the Bush administration’s unveiling last summer of the "Click It or Ticket" promotional campaign.

Overall safety belt use rose to 79 percent among U.S. motorists, according to NHTSA — a 4 percent jump and the highest level ever recorded, Runge said. At the same time, a smaller percentage of fatalities involved unbelted motorists — 56 percent in 2003 compared with 59 percent the year before.

"It looks like many of the improvements we saw were actually enjoyed after that heavy increase in safety belt use," he said. The agency’s projections had not taken that change into account.

The numbers also benefited from an unexpected reduction in alcohol-related fatalities, which decreased to 17,013 in 2003 from 17,524 in 2002, based on the revised statistics. Runge and Mineta said the improvement was possible because all states have now adopted a 0.08 percent blood-alcohol level as the standard for prosecuting drunken driving. They also noted that while SUV registrations increased by more than 11 percent last year, the number of people killed by SUV rollovers increased by "only 7 percent."

"That’s fewer rollover deaths than would have been predicted by the increase in registrations," Runge said. He added that he does not have the data yet to determine whether the difference was caused by safety belt use.

Claybrook said it was wrong to claim progress when SUV rollover deaths continued to increase even as passenger car rollover deaths declined. She said the upbeat presentation of the statistics minimized their seriousness, and questioned the attempt to claim credit for a decline in the rate of fatalities per miles traveled. "That’s been going down for more than 30 years," she said.