I exploit cross-province and time-series variation in Canadian mandatory se
at belt legislation to empirically test the offset hypothesis. The results
of this study offer modest evidence of the existence of offsetting behavior
by drivers. Specifically, increased use of seat belts by drivers after the
enactment of seat belt legislation should have led to a 29 percent decreas
e in driver fatalities. However, econometric estimates indicate that the in
troduction of seat belt legislation is significantly correlated with only a
21 percent decline in driver fatalities. Furthermore, the use of legislati
ve and nonlegislative factors to control for the effects of independent det
erminants of traffic fatalities is critical, as the omission of these effec
ts implies that seat belt laws are associated with roughly a 32 percent dec
rease in driver deaths, thus implying an absence of significant offsetting
behavior.