FORECASTING HIGHWAY CASUALTIES: THE BRITISH ROAD SAFETY ACT AND A SENSE OF DEJA VU

This paper reexamines the effect of the introduction of the British Road Safety Act of 1967. We construct a dynamic model relating monthly road casualties to road traffic, rainfall, and alcohol consumption, standardizing for the seasonality in the data. An intervention variable captures the effect of the Road Safety Act. The findings confirm Ross's earlier conclusion that the Road Safety Act significantly reduces casualties. However, we find that the Road Safety Act only accounts for 2.7 percent of the variance in road casualties, while miles-driven and rainfall account for 48.8 percent, and alcohol consumption explains 4.2 percent. Our model forecasts accurately for 24 months beyond December, 1972, the last month used for estimation.