Statistical analysis of optimism bias

A summary of results for the Full Period and stage 3 is displayed in Table 1. It should be noted that the Full Period represents all the interaction between governments/instrumentalities and bidders/contracting parties, and is heavily influenced by the former. Stage 3, by contrast, is impacted primarily by the performance of the contracting party: whether a Traditional provider, or PPP consortium.

In Table 1 we find that for the Full Period, i.e. the period from Original Approval to Final Actual, the cost over-runs experienced by Traditional projects due to a combination of scope changes and contractor efficiency were 35.3 percent. In the case of PPP projects the cost over-run was less than a third of this, at 11.6 percent. However, the most telling findings are provided by the results for Stage 3, which depend largely on the performance of the respective providers. Here we find that for the Traditional projects in our sample, an expected cost of $4.53 billion at contract signing was over-run to the value of $672.5 million, representing a 14.8 percent increase in cost. By contrast, the $4.95 billion in contracted PPP projects had on average over-run their budgets by only $576 million, or 1.2 percent, which is not statistically different from zero. In summary:

• Optimism Cost Bias -Traditional projects are found to be subject to significant optimism bias during the contracted stage, with $672.5 million of extra costs over a contracting base of $4.53 billion, while at the contracting stage, PPPs cannot be said to be subject to optimism bias.

 Table 1 Total cost of traditional & PPP projects ($m)

_______________________________________
Source: ACG/University of Melbourne

► Table 2 Traditional & PPP projects (%): value weighted time over(under)run

_________________________________________
Source: ACG/University of Melbourne

In Table 2, we present the summary results investigating optimism time bias as the percentage time over/under-run weighted by project size. While smaller Traditional projects were found to perform relatively well in terms of timing bias, we found that larger project size had a strong negative impact on the timeliness of delivery. Therefore, when we weighted optimism time bias by project value, we found that Traditional procurement strongly under-performed again. While the weighted time over-run for the Full Period was roughly twice as long for Traditional projects

(25.6 percent) as for PPPs (13.2 percent), for Stage 3 (Contractual Commitment to Final Actual) the Traditional projects were on average 23.5% behind time, while PPPs were 3.4 percent ahead of time. In summary:

•  Optimism Time Bias - On a value-weighted basis, Traditional projects are found to be subject to significant optimism bias, with a 23.5 percent time over-run during the contracted stage (Stage 3), while at this stage, PPPs cannot be said to be subject to optimism bias (and on average completed projects slightly ahead of time).