Arnould and Grabowski

They use existing estimates of restraint effectiveness in crashes and a fixed distribution of traffic accidents as do Arnould and Grabowski. For safety belt use, however, they use a different technique. They asked a panel of 5 specialists in occupant-crash protection what they thought the percentage of occupants belted in crashes for fleets of cars with detachable passive belts would be. The composite median usage was 17 percent with a composite lower boimd (.1 fractile) of 10 percent and a composite upper level (.9 fractile) of 35 percent. The panel of specialists was asked the same question for discourage-defeat passive belts which are not easily detached. The composite median usage for the more coercive belt system was 43 percent with a lower bound of 15 percent and an upper bound of 60 percent. 

In October 1981 NHTSA published the final regulatory impact analysis upon which the rescission of mandatory passive (automatic) occupant restraints was based. The basic approach in anal is of the safety nefits and dollar costs is similar to that of earlier studies and some of the same data is used to make estimates. The text of the 1981 NHTSA document and the benefit-cost anal is found in Appendix A of the report permit comparison of 

Annual traffic fatalities prevented range in steady state from 2,050 if belt use is 23 percent to 8,750 if belt use is 60 percent. The higher estimate is quite close to that of Arnould and Grabowski for NHTSA s estimate of effectiveness. The valuation of the safety benefits is based on a foregone earnings approach instead of the willingness to pay, and the values of risk reduction are lower than in other studies.  

The cost estimates for passive belts are based primarily on manufacturers figures which range from 50 to 150 dollars. NHTSA estimates that the 

---

The first independent attempt to estimate conventionally the benefits and costs of mandatory passive restraints was made by Richard Amould and Henry Grabowski. For the reduction in fatalities and injuries in crashes they use two sets of estimates. One set is based on a 1981 field team study of rural traffic accidents which shows, for example that fatalities are reduced by 34 percent by air bags and lap belts together, 32 percent by lap and shoulder belts together, 28 percent by passive belts, and 25 percent by air bags alone. The field study estimates are only appronmately one half of the NHTSA lab study estimates which is the other set. Arnould and Grabowski assume that 60 to 70 percent of occupants with passive belts would use them and that 0 to 20 percent of occupants with air bags would also use lap belts. Occupant protection is assumed to have no affect on chances of accidents. The 1 5 distribution of traffic accident injuries and the estimates of restraint effectiveness in crashes are used to calculate the fatalities and injuries prevented. 

In steady-state equilibrium the annual total social costs of passive belts are estimated to be from 0.5 to 1.0 billion dollars per year and the same type of costs of air bags are estimated to be from 2.5 to 6.5 billion dollars per year depending on which cost estimates are used. Comparable benefits are estimated to be from 2.1 to 5.5 billion dollars per year for passive belts and 3.6 to 6.4 billion dollars per year for air bags. Costs and benefits are expressed in 1975 dollars. For passive belts the estimated net benefits are positive for each case examined. For air bags, the net benefits are smaller and always negative for the better costs estimates. Arnould and Grabowski conclude that passive belts are a much more cost effective approach to occupant crash protection than air bags. 

Air bags and passive belts, 1975doUars, Arnould and Grabowski (1981)... 

Richard J. Arnould and Henry Grabowski. Auto Safety Regulation An Analysis of Market Failure The Bell Journal of Economics 12 (Spring 1981) 27-48. 

---