Case Study
The Risk Driver Impact Approach to Estimation of Project Cost Risks
(The content of this paper is the opinion of Evin Stump, Sr. Systems Engineer at Galorath Incorporated) Perhaps the most commonly used approach to quantitative estimation of project cost risks is to estimate a “nominal” cost for each work element of the project’s work breakdown structure, subjectively assign cost risk distributions to those nominal values, then run a Monte Carlo simulation to obtain a risk distribution of the cost sum. When this is done, the Monte Carlo tool used must have the capability to allow the user to assign correlations between the various work element costs, else serious errors can result. In projects, work elements generally are not designed for statistical independence, nor is it likely they could be in any practical sense, even if we wanted to do so. If there are significant correlations and they are not well specified, the result is a serious underestimation of the risks.
