Abstract
State-level bridge expenditure forecasts are used by highway agencies for financial need forecasts, long-term planning and budgeting, asset valuation and infrastructure user cost allocation. It has been observed that the levels of normalised bridge expenditure vary widely across the states and across time periods, and therefore, it may be inappropriate to use an average value of normalised expenditure for agencies business processes. Further, there is a need to investigate the underlying causes of this variation so that bridge managers can better estimate their expected costs. This paper shows how aggregate ( jurisdiction-level) bridge data can be used to develop annual maintenance/repair expenditure (Amex) models. In the case study of the paper, data from US states were used, and the Amex was expressed as a function of bridge condition, bridge inventory size, traffic loading and climate severity. The paper quantifies the magnitude of the increase in Amex per unit area due to changes in factor levels that include traffic loading, annual temperature, precipitation and the conditions ratings of the superstructure and substructure. The results indicate that there are significant scale economies in state-level bridge expenditures – for every additional 1000 m
2
of bridge deck inventory, the bridge Amex decreases by $0·0916/m
2
.