The long-term overall effect of ABS on fatal crash involvements is close to zero in both cars and LTVs. The observed effect in cars is a 1-percent reduction (90% confidence bounds range from a 2% increase to a 4% reduction). The observed effect in LTVs is a 1-percent increase (confidence bounds: -6% to 4%).
But the overall effect of ABS on all crash involvements, including nonfatal involvements is beneficial and statistically significant in both cars and LTVs. The observed reductions are 6 percent in cars (confidence bounds: 4% to 8%) and 8 percent in LTVs (confidence bounds: 3% to 11%).
The overall reduction of nonfatal-injury crashes is approximately the same as the reduction of all crashes.
Fatal run-off-road crashes increase with ABS by a statistically significant 9 percent in cars (confidence bounds: 3% to 15% increase). The long-term effect is substantially smaller than in the early years of ABS (28% increase), but it is still a significant increase. The observed effect in LTVs is a non-significant 6-percent increase (confidence bounds: -16% to 3%).
On wet, snowy, or icy roads, where ABS is most likely to activate, the increase in fatal run-off-road crashes is a statistically significant 34 percent in passenger cars (confidence bounds: 20% to 50% increase). On these roads, all three types of fatal run-off-road crashes increase significantly for cars and so do fatal rollovers of LTVs.