This article extends discussions of potential biases that can exist in applying cost–benefit analysis. While there is extensive evidence that capture can result in stakeholder manipulation of inputs, there are also claims that the analysis is inherently theoretically bias in favour of over acceptance. The article shows that, contrary to these latter claims, treating projects in isolation is unlikely to produce such bias; indeed, it is as likely as not to lead to suboptimally low acceptance rates. The reason for excessive acceptance of projects therefore is largely due to institutional capture of the analysis for either self-interest or natural human over-optimism.