A probabilistic evaluation of highway project costs

Abhishek Bhargava, Purdue University

Abstract

Researchers and engineers worldwide continue to ponder over the reliability of construction cost forecasts made at the early stages of project development. The accuracy of such estimates influences the selection of projects for implementation. After a project has been incorporated in the statewide transportation improvement program, more detailed estimates of its cost are prepared at the remaining stages of project development, namely, design estimate, engineer's estimate, and bid estimate. This dissertation presents a methodology for identifying contracts that are likely to experience cost overrun at different stages of project development process. This was done on the basis of contract specific factors. Also, the dissertation analyzed escalation patterns of cost estimates across project development stages. The methodology used risk-based econometric models to determine probability of an escalation pattern and cost overrun. An application of the methodology was demonstrated using data from Indiana. Monte Carlo simulations were conducted to analyze the variability in the probability of an escalation pattern and the variability in the probability of cost overrun. Using the models, a software tool was developed to identify contracts that are more likely to experience cost overrun in Indiana. The methodology can, however, be used by other highway agencies. For the case study, the results showed that for expansion (capacity addition) and bridge contracts, the probability of cost overrun increased with increase in contract size. Large pavement contracts (greater than $5 million), on the other hand, were more likely to experience a cost underrun as compared to small pavement contracts. Pavement contracts in urban areas were more likely to experience cost overrun compared to their rural counterparts. For expansion and bridge contract cost overrun, no significant difference was found between urban and rural areas. Expansion and pavement contracts on Non-NHS highways were more likely to experience cost overrun, compared to such contracts on other transportation systems. Bridge contracts on Interstates were more likely to experience cost overrun compared to other transportation systems. Furthermore, it was found that as the time duration between project proposal and design completion increased, the tendency of a contract to experience cost overrun increased. Longer time spans between project proposal and letting were also found to increase the probability of cost overrun. In addition, the risk of occurrence of a particular escalation pattern influenced the probability of cost overrun.

Degree

Ph.D.

Advisors

Sinha, Purdue University.

Subject Area

Civil engineering

COinS