by
Charles Lave, PhD, Chairman
Patrick Elias, B.A.
Dept of Economics
University of California, Irvine
Performed under a grant from
AAA Foundation for Traffic Safety
1730 M Street, N.W., Suite 401
Washington, DC 20036
September 1992
The new 65 mph limit would allow state highway patrols to shift resources from speed enforcement on the interstates to other safety activities and other highways -- a shift many highway patrol chiefs have argued for. If the chiefs were correct, the reallocation of patrol resources should lead to a reduction in statewide fatality rates. In addition, the new speed limit might attract drivers away from other, more dangerous roads; again, leading to a change in the statewide fatality rate.
To test these ideas we began by examining aggregate data. We combined the states into two groups: a test group, the states that raised their speed limits; and a control group, the states that did not. We calculated the change in overall fatality rates that occurred after the new speed limit: the test group improved 3.5% more than the control group.
We also found evidence that highway patrols had, in fact, shifted resources and drivers had shifted roads in the manner predicted by our theory.
We then turned to a regression analysis of the data from the individual 65 mph states. We used the basic model developed by Garber and Graham that holds constant the effects of long term trend, driving exposure, seat belt laws, and economic factors; and fitted this to monthly time series data on a state-by-state basis. We found that using our dependent variable, the statewide fatality rate, instead of the rural interstate fatality rate, reversed the Garber/Graham results: the average decline in state-by-state fatality rates following the new speed limit was 3.4 percent.
Did the 3.4 percent decline in the 65 mph states represent the result of the new speed limit, or was it possibly the effect of some downward break in the long term trend? To check this possibility, we ran the same regression model on the 55 mph states, giving them fake-65 mph dummy variables starting in June 1987. The estimated coefficients of the fake-65 mph dummies contradicted this alternative explanation -- there was no evidence for a general shift in the long term fatality trend.
To obtain more efficient estimates of the speed limit effects, we ran the analysis on the combined sample of the 65 mph states, restricting them to a common estimate of the 65 mph dummy variable. This indicated that the new speed limit had produced a 5.1% decline in the fatality rate. To test the robustness of these results, we performed a Chow test by removing seven representative states from the sample: the resulting regression estimates were statistically similar to those of the full sample.
Taken as a whole, these different analyses lead to the conclusion that overall statewide fatality rates fell by 3.4 to 5.1 percent in the states that adopted the 65 mph limit.
Why did the new speed limit lower the fatality rate? 1) Drivers may have switched to safer roads; 2) highway patrols may have shifted resources to activities with more safety payoff; and 3) the speed variance among cars may have declined -- it might decline on the interstates as law abiding drivers caught up with the speeders, and it might have declined on other highways as their speeders switched to the interstates. The evidence indicates that events 1 and 2 did occur; we have no evidence for event 3. Future research ought to be directed toward disentangling the relative contribution of these factors.