Optimism Bias in Multilateral Development Banks’ Infrastructure Appraisals

The appropriateness of using Cost-Benefit Analysis as a decision-making criterion in transport infrastructure planning has been widely questioned due to systematic deviations from appraisal estimates—primarily in investment costs, delivery times, and demand forecasts—and a persistent optimism bias among evaluators that leads to projects with negative social returns, commonly referred to as white elephants. However, there remains relatively little empirical evidence on the extent to which social profitability is affected by these deviations, largely because ex-post evaluations are seldom conducted. In this study, we construct a dataset of infrastructure projects evaluated and financed by Multilateral Development Banks (MDBs) and find that deviations remain modest relative to other studies and evaluators. Most indicators suggest that project profitability is unbiased and that only a relatively small proportion of projects yield negative social returns. We argue that this outcome may be attributed to the MDBs’ internal appraisal policies, particularly regarding project governance (with the banks acting as both evaluators and financiers, and applicant countries serving as project sponsors and executioners) and the mandatory implementation of ex-post evaluations, suggesting that the optimism bias phenomenon is more related to strategic misrepresentation than to a cognitive bias.