Despite the ubiquitous nature of these policies, there is disagreement in the literature regarding the direction and size of the impact that speed limits have on traffic-related fatalities. We argue in this paper that the mixed results in previous work may arise because these studies have missed an important component of the implementation of speed limit laws. More explicitly, they have failed to adequately control for the deterrent effect of enforcement and sanctions. We develop the argument that the observed impacts of speed limits will be overly large when the certainty and severity of punishment are not accounted for. We test this assertion in a cross-sectional time series analysis of state-level traffic fatalities from the years 1990–2006 and find that lower speeds do save a significant number of lives. Interestingly, we find that the impact is significantly overestimated for 65-mph limits and significantly underestimated for 70-mph limits when enforcement, penalties, and the interaction of the two are excluded. The results also suggest that fines have a rather modest impact on fatalities unless states employ a sufficient number of troopers to enforce posted limits. In addition to clarifying previous findings related to speed limit policy, therefore, the findings contribute to the general application of deterrence theory by empirically confirming that the importance of sanction severity is dependent on the perceived certainty of punishment.