The number of monthly motor vehicle fatalities in the United States, January 1975 - September 1995, was modeled as a function of vehicle miles traveled, number of unemployed persons, number of employed persons, number of persons not in the labor force, and the number of new car sales. Miles traveled and the number of unemployed persons were highly reliable in predicting monthly fatalities. In particular, the model accounted for short-term increases in fatality trends beginning in 1984, 1986, and 1993. The long-term trend toward decreases in motor vehicle fatalities does not seem to have changed.
Abstract