Abstract
Researchers have already proven significant systematic cost over-runs measured from the decision to build estimate even in the most developed countries of the world. The reference class forecasting was introduced to counter this bias. However, the precise workings or cost over-run determinants are however still under-researched. We selected the case of the National Highway Construction Program in Slovenia and projects, completed between 1995 and 2007. The purpose was to determine the precise cost performance through time and analyze the cost estimating mechanism to see whether and how it influences the cost performance. The details of the cost estimating practice have been studied and compared with the existing practice in the USA and elsewhere. A representative sample of 36 projects could be constructed, valued at USD 2.7 billion (2006 prices) with a total length of some 235km. Almost half of the cost performance variance could be explained by the cost estimation mechanism (cost-based estimation with historic bid database). In different variants, this approach appears to be dominant in the most developed countries of the World due to its relative simplicity. We find, that due to the behavior of bidders in the tendering process, the same mechanism also implies that some systematic cost over-run will likely occur, even if no other causes were present. This adds a new perspective to the suggestions of Flyvbjerg and others, who suggested that the dominant (but not exclusive) cause for the persistence of systematic cost is strategic misinterpretation. The findings suggest that project sponsors should perform a supporting analysis on how price changes on the construction market feed in to the cost performance to further support the reference class forecasting approach. The findings also suggests, that in case of strong cost performance shifts, longer periods need to be included in the analysis, before one can conclude, that the improvement or worsening is not temporary and is not the inherent result of the cost estimating mechanism.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have