The effect of long term non-invasive pavement deterioration on accident injury-severity rates: A seemingly unrelated and multivariate equations approach

- A method is developed to measure the effect of pavement deterioration on safety.
- SURE and multivariate tobit models are estimated.
- Pavement condition and accident injury-severity rates are forecasted over time.
- Pavement forecasts are used in the injury-severity rate prediction mechanism.
- It is shown that safety should be considered as a criterion in pavement management.

This paper seeks to measure the effect of long term non-invasive pavement deterioration on accident injury-severity rates, and demonstrate the potential of considering safety as one of the criteria in the pavement management decision making process. Using data from Indiana, a system of seemingly unrelated regression equations (SURE) is estimated to predict pavement deterioration curves over a 30-year projection period based on three commonly used pavement performance indicators. The annual predictors of the pavement roughness, rutting depth, and pavement condition rating are then used in a multivariate tobit equations model of vehicle accident injury-severity rates. The results provide the expected change of the no injury, injury, and fatality rates, due to the non-invasive pavement deterioration, and are compared to a budget-unrestricted scenario under which rehabilitation occurs routinely. Even though the aim of the paper is not to provide an optimal pavement management program, the findings suggest that safety should be considered as one of the decision making criteria.