Analysis of Pedestrian and Bicycle Casualties with Regional Panel Data

Abstract

An analysis is presented of pedestrian and bicycle casualties by using cross-sectional time series data for the regions of Great Britain. A fixed-effect negative binomial model is used that accounts for heterogeneity in the data and the distributional properties of count data. Various factors associated with those killed and seriously injured as well as with slight injuries are examined. These include the average age of vehicles in the region, the road length of various road classes, vehicle ownership in the region, per capita income, per capita expenditure on alcohol, age cohorts, and various proxies for medical technology improvements. Various specifications of the models are estimated. Generally, it is found that more serious pedestrian injuries are associated with lower-income areas, increases in percent of local roads, increased per capita expenditure on alcohol, and total population. Statistical effects are more difficult to detect in models with serious injuries for bicyclists, but alcohol expenditure is strongly associated with increased injuries. This work has implications for transport policy aimed at increasing the modal share of pedestrians and bicyclists. Further research is needed to clearly understand some of the trends found in the analysis, especially the effect of changes in medical care and technology on total injuries.