Mott MacDonald's Optimism Bias Report
The link is to a treasury web page with various PFI reports on it.
Mott MacDonald researched "optimism bias". I think that it is not entirely fair to call it "optimism bias". Generally people don't know all the details of a large project when they start it.
However, Mott MacDonald found that there were various reasons why projects both traditional and PFI/PPP ended up having an "optimism bias".
Meanwhile back at the ranch only Public Sector Comparators have an "optimism bias" penalty added to them. Noone is likely to complain because they want the projects to proceed and they know if they complain they won't get their projects.
This has only changed because PbR means that people really are looking at the real costs (eg what they have to pay out rather than fiddle factors).
Such arcane issues of accounting really do matter because when they are wrong (which is generally the case at the moment) wards get closed, new hospitals are not built and patients don't get treated.