Regression modelling of risk impacts on construction cost flow forecast

Significant risk factors inherent in construction cost flow forecast were identified in this study. The aim of this paper is to develop regression models to assess the impacts of the identified risks on the baseline forecast at the in‐progress stage of construction.

Two stages were involved in data collection. The first was a structured questionnaire survey administered on 370 UK contractors to identify significant risk factors inherent in cost flow forecast. The second stage was the collection of forecast and actual cost flow data from 55 case study projects. Variations between these pair of data sets were measured at 30 per cent, 50 per cent, 70 per cent and 100 per cent completion periods. Respondents were then requested to score on a Likert type scale, the extent of occurrence of the significant risk factors in the case study projects. This pair of data sets were used in regression modelling.

Significant risk factors were identified from the questionnaire survey analysis as: changes to initial design, variation to works, production target slippage, delay in agreeing variation/dayworks and delay in settling claims among others. Using the identified significant risk factors and the periodic variability measurements, multiple linear regression models were developed. The models were promising in that they helped to establish the fact that the phenomenon under consideration could be modelled. They also provided some insights in explaining the observed variability between the baseline cost flow forecast and actual cost flow based on risk impacts.

The developed models showed a promising level of accuracy but also indicated that the phenomenon under consideration is not strictly linear and may need to explore some other form of modelling.

The developed models provide invaluable information to the construction contractors regarding the likely impacts of significant risk variables on cost flow baseline forecast at different stages of construction so that a pro active risk response can be put in place.

This study makes an original contribution of providing a modelling insight into the phenomenon of how risks inherent in construction could impact the baseline cost flow forecast at different stages of construction. The information is invaluable in making pro active risk response.