Since 1975, the fuel economy of passenger cars and light trucks has been regulated by the corporate average fuel economy (CAFE) standards, established during the energy crises of the 1970s. Calls to increase fuel economy are usually met by a fierce debate on the effectiveness of the CAFE standards and their impact on highway safety. A seminal study of the link between CAFE and traffic fatalities was published by R. W. Crandall and J. D. Graham in 1989. They linked higher fuel economy levels to decreases in vehicle weight and correlated the decline in new car weight with about a 20% increase in occupant fatalities. The time series available to them, 1947–1981, includes only the first 4 years of fuel economy regulation, but any statistical relationship estimated over such a short period is questionable. This paper reexamines the relationship between U.S. light-duty vehicle fuel economy and highway fatalities from 1966 to 2002. Cointegration analysis reveals that the stationary linear relationships between the average fuel economy of passenger cars and light trucks and highway fatalities are negative: higher miles per gallon is significantly correlated with fewer fatalities. Log–log models are not stable and tend to produce statistically insignificant (negative) relationships between fuel economy and traffic fatalities. These results do not definitively establish a negative relationship between light-duty vehicle fuel economy and highway fatalities; instead they demonstrate that national aggregate statistics cannot support the assertion that increased fuel economy has led to increased traffic fatalities.