Fatality Reduction by Automatic Occupant Protection in the United States

Fatality risk of occupants in cars equipped with air bags plus manual belts (at 1993 use rates) is 23 percent lower than in "baseline" cars with manual belts at 1983 use rates. In similar comparisons, the fatality reductions for the four types of automatic belts range from 11 to 19 percent. All reductions are statistically significant. In the 1993 model-year mix of cars with air bags or automatic belts, at 1993 belt use rates, the average fatality risk is 20 percent lower than for baseline, manual-belt cars at 1983 use rates.

[This report's findings on air bags have been superseded by a 1996 evaluation titled Fatality Reduction by Air Bags: Analyses of Accident Data through Early 1996 .]

The individual estimates of actual fatality reduction are averaged to derive current "best" estimates, with approximate confidence bounds, for five types of automatic occupant protection, relative to cars with manual belts at 1983 use rates. For example, the two estimates of fatality reduction for driver air bags (21 and 24.4 percent) average out to 23 percent.

Also, the overall effectiveness of the occupant protection program for model year 1993 cars is estimated by taking a weighted average of the five individual estimates, based on the distribution of FARS of model year 1993 cars through June 1993 (which was 59 percent air bags, 13 percent motorized 2 point belts without disconnect, 4 percent motorized 2 point belts with disconnect, 19 percent nonmotorized 3 point belts and 5 percent nonmotorized 2 point belts). The weighted average represents the difference between the actual fatality rate in model year 1993 cars, at 1993 belt use, and the fatality rate that would occur in those cars if they were equipped only with manual belts, at 1983 use rates.

It is clear that the occupant protection program - the automatic protection requirement of FMVSS 208, in combination with a nationwide effort to increase belt use - has reduced fatality risk. The average fatality reduction for model year 1993 cars at 1993 belt use rates is 20 percent relative to the baseline of manual-belt cars at 1983 use rates, with confidence bounds of 15 to 25 percent. Furthermore, the average fatality reduction has improved since 1991, when it was 16 percent. A market shift from automatic belts to cars with air bags and manual belts, and increased use of these belts, are contributing to the improvement.

Each of the five individual types of automatic protection has a positive and statistically significant "best estimate" of fatality reduction relative to the baseline, as evidenced by the positive lower confidence bounds. It is too early for a definitive rank-ordering of the systems. Air bags and motorized 2 point belts (without disconnect) have the highest "best estimates," but the overlap in the confidence bounds for the various types is still substantial.

As stated above, all of the estimates represent actual effectiveness of automatic systems, as currently used. Continued increases in the use of 3-point belts supplied with air bags, as well as other manual or automatic belt systems would boost "actual" effectiveness. As the market share for air bags increases toward 100 percent, the weighted average will rise and approach the estimate for air bags.