A logit analysis of vehicle emissions using inspection and maintenance testing data

Abstract Many states use vehicle inspection and maintenance (I/M) programs to identify high polluting vehicles and ensure that they operate in accordance with standards. While I/M programs are generally regarded as a valuable means to curb urban air pollution, they have been often criticized for their cost-ineffectiveness. One criticism has been centered on the blanket approach that requires all vehicles within the program boundaries to participate regardless of their emission conditions. This paper explores the basis for a selective sampling of vehicles most likely to be pollution violators. Using I/M testing data from Portland, Oregon, it estimates logit equations for the likelihood of carbon monoxide and hydrocarbon emission violations given a set of vehicle characteristics. The results indicate that vehicle age, engine size, and odometer reading all play a significant role in determining the probability of emission test failure.