Speeding to conclusions

November 06, 1997|By George F. Will

WASHINGTON -- Statistics and probabilities can be puzzling, even paradoxical, as the man knew who wrote:

Very, very, very few

People die at ninety-two.

I suppose that I shall be

Safer still at ninety-three.

Another example: In the movie ''Fathers' Day,'' Billy Crystal wants Robin Williams to facilitate a deception by pretending to cry. Mr. Crystal suggests that Williams imagine that he is a tragic hero like Lou Gehrig. Mr. Williams asks, ''Who's that?''

Mr. Crystal, dumbfounded, says, ''Everybody knows Lou Gehrig -- the baseball player, he died of Lou Gehrig's disease.''

Mr. Williams, flabbergasted that someone named Lou Gehrig died of something named Lou Gehrig's disease, exclaims: ''Wow! What are the odds of that?''

Which is prologue to today's subject: What were the odds in 1987 that the increases in states' speed limits would result in decreasing statewide fatality rates? Herewith a story about the vagaries of statistical analyses and social policies.

Two scholars who can explain the counterintuitive results of increased speed limits are Charles Lave and Patrick Elias of the department of economics at the University of California at Irvine. The story they tell begins with the 1973 Yom Kippur War, the oil embargo, and the federal law coercing states to enforce 55 mph limits. Federal highway funding would be reduced for states not meeting compliance requirements, which included speed monitoring programs and reports of the proportion of drivers violating the new limit.

The primary reason for the 55 mph limit was fuel conservation, and when the energy crisis passed, Americans grew restless. In 1987, states were allowed to raise their limits on rural interstates and 40 adopted 65 mph limits. Opponents stressed not conservation but safety, predicting carnage.

Concerning developments immediately after 1987, the economists note that construing the evidence is a more complex task than some analysts realize. Their conclusion is that, up to a point, higher speed limits save lives.

After 1987, the higher speed limit reduced statewide fatality rates by 3.4 percent to 5.1 percent, compared to the rates in states that did not raise limits. True, the actual number of fatalities continued to increase after 1987, but the volume of traffic increased even faster. The critical measurement is fatalities per vehicle mile traveled in the entire state. Some studies found that raising speed limits on particular highways increased, or did not decrease, fatalities on those particular highways. However, such studies failed to gauge the ways in which all of a state's highways comprise a single interdependent system.

Messr. Lave and Elias say the 55 mph limit caused the misallocation of traffic and of police resources. The federal government had demanded strict compliance with the 55 mph limit, which forced state highway patrols to concentrate on speed enforcement on the interstate highways, which have the densest concentration of high-speed traffic-- but are also the safest highways.

And state police patrol resources were then decreasingly available for such safety programs as truck inspections. Furthermore, many drivers who wanted to speed switched to less traveled, less patrolled but less safe roads.

The 55 mph limit might have decreased fatalities on some roads by increasing patrolling and decreasing traffic volume from what it would have been without that limit. However, the effect on a state's total highways system was apt to be a net subtraction from safety.

Raising speed limits lured some drivers back to safer, more heavily patrolled roads, and allowed highway patrols to shift resources back to the programs they thought most effective. And it decreased a real killer -- speed variance among vehicles.

Many collisions occur when cars are overtaking and passing one another. Speed variance among drivers increases when speed limits are set so low that there is a high rate of noncompliance.

When in November 1995 Congress empowered states to set such limits as they chose, fatalities did not increase the 10 percent to 14 percent predicted by the Cassandras who foresaw 4,400-6,000 extra deaths per year. Neither did fatalities increase even the 3 percent that would have been expected, extrapolating from recent trends. Instead, Mr. Lave concludes that fatalities have declined slightly (0.14 percent) even as total vehicle miles increased 2 percent.

This is a cautionary tale about the complexity of discerning reality in a welter of statistics. It also is an encouraging tale. Sometimes the unintended consequences of a policy are good.

George F. Will is a syndicated columnist.

Pub Date: 11/06/97

Baltimore Sun Articles
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.