As the proportion of highway needs that can be supported by fuel taxes declines, the attractiveness of toll funding rises. Public agencies and local authorities are examining their options, and some states have broadly increased the range of local toll-related options available. These entities need an analytical tool that will allow them to make realistic "first-cut screening" estimates of potential revenue so they can make prudent determinations about the potential viability of a toll-supported option. Most of the factors bearing on revenue prediction are highly variable - corridor demand, diversion rate, and acceptable tolls - and increase in variability with the length of the project. However, it is imprudent to treat these factors as point estimates for simplicity of analysis because such estimates can easily mask either a truly viable project or clearly bad project. This paper develops the analytical framework for a more robust revenue estimate. This analysis leads to a spreadsheet analysis tool to provide a preliminary determination of the viability of a toll road or toll-supported project. The tool uses inputs typically available to local agencies and output measures representing various aspects of viability, including economic measures such as net present value or benefit/cost ratio. This analysis treats inputs and outputs as random variables, uses sensitivity analyses to identify critical variables, and develops probability distributions that improve the accuracy, reliability, and decision-making value of the toll study. The results provide to local agencies a "first-cut screening tool," to be supplemented by investment-grade traffic and revenue studies if the findings so warrant. The paper emanates from a research project sponsored by the Texas Department of Transportation.
Abstract