Estimating Required Contingency Funds for Construction Projects Using Multple Linear Regression Thesis

Cost overruns are a critical problem for construction projects. The common practice for dealing with cost overruns is the assignment of an arbitrary flat percentage of the construction budget as a contingency fund. This research seeks to identify significant factors that may influence, or serve as indicators of, potential cost overruns. The study uses data on 243 construction projects over a full range of project types and scopes gathered from an existing United States Air Force construction database. The author uses multiple linear regression to analyze the data and compares the proposed model to the common practice of assigning contingency funds. The multiple linear regression model provides better predictions of actual cost overruns experienced. Based on the performance metric used, the model sufficiently captures 44% of actual cost overruns versus current practices capturing only 20% The proposed model developed in this study only uses data that would be available prior to the award of a construction contract. This allows the model to serve as a planning tool throughout the concept and design phases. The model includes project characteristics, design performance metrics, and contract award process influences. This research supports prior findings of a relationship between design funding and design performance as well as the influence of the contract award process on cost overruns. While the proposed model captures 44% of actual cost overruns, its application reduces average contingency budgeting error from -11.2% to only -0.3% over the entire test