Quantifying the Causal Effect of FMCSA Enforcement Interventions on Truck Crash Reduction: A Quasi-Experimental Approach Using Carrier-Level Safety Data

Abstract

The Federal Motor Carrier Safety Administration (FMCSA) and partners in the states are supposed to minimize collisions and other serious safety offenses by effective enforcement interventions, such as compliance reviews, audits, and planned inspections not to mention enforcement interventions. However, the carriers onto which the interventions are imposed are usually not random and already are of high profile of risk such that naive before vs. after comparisons is misleading information. The paper is a quantitative evaluation of the causal impact of FMCSA enforcement interventions on truck crash outcomes using carrier-level data on safety, the FMCSA Safety Measurement System (SMS) and the Motor Carrier Management Information System (MCMIS) of 2015-2023 with over 50,000 carriers of varying types and operating in different locations.
Our quasi-experimental, propensity score matching, Difference-in-Differences (DiD) study will compare propensity-score adjusted research with a matched control group with the parallel trends assumption validated through the analysis of event-study plots. Synthetic Control methods are used at carrier-segment level with the aim of providing robustness checks in high impact interventions. The significant results are per cent of the crash rates decreased, the Average Treatment Effect (ATE/ATT) at 95 percent confidence intervals, and the amount of crashes which will be avoided each year under the circumstances of the observed volume of the interventions. Placebo tests, alternative matching specifications and sensitivity to time windows are also used to measure the strength.
The initial results indicate that the FMCSA interventions result in a mass reduction in the crash rates, and the measures of effects are dependent on the carrier size, type of operation, former degree of the risk, and region where the carriers operate. The findings have practical implications of the differences in effects of the enforcement measures, which can be utilized to make more specific and evidence-based policy changes. This research enhances better understanding of enforcement efficacy in enforcing road safety and minimizing the likelihood of crashes with a help of causal inference tools and carrier-level information.