Transportation project bundling policy impacts on construction time delay: some empirical evidence

It has been hypothesized that project bundling policies can significantly reduce the likelihood and magnitude of time delays, and that such an effect varies across project types and bundle sizes. This paper attempts to address this question using binary logit and accelerated-failure time specifications to quantify the effect of project bundling policies on time delay likelihood and severity, respectively. In the context of this study, a bundling policy is expressed in terms of the number of projects bundled to form a single contract. The study data consisted of contract information from construction projects in six (6) work categories that involved thirty-six (36) different project types. The results suggest that the time delay likelihood and severity are significantly influenced by project bundling policy, and that this relationship varies significantly across the different project types: with regard to Traffic projects, an increase in bundle size always decreases the likelihood of project delay and the delay severity; regarding Small-structure projects, an increase in bundle size always increases the likelihood of project delay and the delay severity; and, regarding Bridge and Road (pavement) projects, bundling size was found to have a dichotomous effect – increasing the bundle size has an increased or decreased impact on the time delay propensity or severity depending on the bundle size. Further, a bundle consisting of projects with high similarity or are spatially proximate is found to be generally associated with a lower time delay risk, consistently for across the project types and most significant in the case of Road projects. The study results can provide highway transportation agencies with knowledge not only on which project types could be bundled and their appropriate bundle sizes to reduce the likelihood or magnitude of time delay, but also on those project types and bundle sizes where bundling would have no such effect. As bundling practices become increasingly prevalent globally, transportation agencies can replicate this paper’s methodology to develop more realistic estimates of the durations of their bundled projects.