LEARNING FROM EX-ANTE EX-POST COST-BENEFIT COMPARISONS - THE COQUIHALLA HIGHWAY EXAMPLE

The purpose of cost-benefit analysis (CBA) is to help public sector decision-making. The ''help'' varies according to when it is performed. CBA can be performed ex ante (EA), ex post (EP), or in the interim-in medias res (IMR) of a project. We propose a fourth class of CBA-one that compares EA with EP or with IMR CBA on the same project. In fact, this type of comparison has not been conducted in the literature. We suggest that without such research it is impossible to evaluate the practical value of CBA as a decision-making tool. This article demonstrates the value of such comparisons, and contrasts them with other classes of CBA. Specifically: (1) it compares the advantages of comparison studies with other classes of CBA; (2) it categorizes four major types of error in CBA studies-omission errors, forecasting errors, measurement errors, and valuation errors-and models the impact of these errors on actual and estimated net benefits over time; (3) it examines the causes of the four different types of error; and (4) it compares three different classes of CBA on the same highway project: one clearly EA, one 18 months later (an IMR study) and one 7 years later (which we treat as an EP study). There are major differences in the estimates of net benefits. Contrary to what might have been expected, the largest source of difference was not due to errors in forecasts, nor differences in evaluation of intangible benefits, but from major differences in declared and actual construction costs of the project. That is, the largest errors arose from what most analysts would have thought were the most reliable figures entered into the CBA. We conclude that comparison studies are potentially the most useful for learning about the accuracy and efficacy of cost-benefit analysis to decision-makers and evaluators.