Speeding, Coordination, and the 55 MPH Limit

Do laws coordinate or restrain? A number of recent papers discuss the optimality of the 55 mph national maximum speed limit (NMSL), and evaluate the tradeoff of timelost vs. lives-saved resulting from the lowered speed (James Jondrow et al., 1983; Dana Kamerud, 1983; Thomas Forester et al., 1984). These papers all implicitly accept the conventional wisdom-speed kills, slower is safer. This conventional wisdom leads to laws designed as limits on behavior, whereas . . . the crucial element is often coordination. People need to do the right things at the right time in relation to what others are doing (Thomas Schelling, 1978, p. 121). There are indeed some traffic laws that establish conventions of expected conduct: we ask that motorists drive to the right, not because driving on the left is evil, but because it is important that the direction of flow be commonly agreed upon. Likewise, traffic lights are best viewed as a coordinating device: allowing free flow to alternating lanes of traffic to reduce the confusion and loss of time in unsignalized intersections. For peculiar historical reasons, speed laws evolved as limits on driver behavior, rather than as signaling devices meant to coordinate it. Guided by the limit-rationale, police concentrate on those drivers who exceed the legal speed, and tend to ignore those drivers who disrupt coordination by traveling much slower than the norm. This paper tests these differing views of the law by examining the current effects of the 55 mph NMSL-should it be viewed as a coordinating mechanism or a limiting mechanism? I measure the effects of limit-defying behavior (speeding), and absence of coordination (speed variance) on the fatality rate. Based on analysis of 1981 and 1982 state cross-section data, I find that there is no statistically discernable relationship between the fatality rate and average speed, though there is a strong relationship to speed variance. When most cars are traveling at about the same speed, whether it is a high speed or a low one, the fatality rate will be low-presumably because the probability of collision will be low. Variance kills, not speed.