Development of discrete models of infrastructure deterioration with panel data

Statistical models of infrastructure facility deterioration are typically estimated using panel (i.e. pooled cross section and time series) data of in-service facilities. For example, bi-annual ratings of bridges have been used to develop discrete models of component deterioration by a number of researchers. These models have not accounted for the presence of unobserved heterogeneity in the panel data, which may lead to biased coefficient estimates. In this paper, we adopt a random effects specification to control for unobserved heterogeneity in a probit model of bridge deck deterioration. The proposed model yields improved results in comparison with a simple probit model. An implication of the present study is that unobserved heterogeneity needs to be accounted for in developing probabilistic infrastructure deterioration models.