MODELING CONTRACTOR���S BID MARKUP ESTIMATION FOR PUBLIC BUILDING CONSTRUCTION PROJECTS IN GAMO ZONE, SOUTH ETHIOPIA REGIONAL STATE

Abstract

A markup is an important factor applied by the estimator to the total cost of a particular job activity or bid to cover overhead and profit. A study on cost estimating practices in the construction sector in Ethiopia has highlighted the need to change the current practice of estimating bid markup, especially in the construction of public buildings. Furthermore, this inability to estimate bid markup correctly so many contractors leave the construction industry, aside from the loss of money and having construction quality issues and project completion delays in the majority public building projects. This study aims to identify and analyze factors that influence the size of bid markup for public building construction projects and develop a model to support local contractors' decisions when estimating bid markup size. The study utilized an integrated review of various literatures and a questionnaire survey as data collection methods. 28 factors influencing bid markup size were identified through literature review and key influencing factors were ranked using analysis. The study developed a multiple linear regression (MLR) equation containing nine factors based on stepwise regression technique, with a correlation coefficient (R) of 0.819 and an adjusted coefficient of determination (R2) of 0.67. The fuzzy neural network (FNN) modeling method was also used, showing an R2 value of 83.2, RMSE of 1.681, and MAPE of 1.366, showing good results. Statistical performance measures indicate that the FNN modeling approach is more effective in predicting bid markup than the MLR approach, although both models are considered viable tools for predicting bid markup. Both approaches can therefore be considered satisfactory predictors of bid markup and can serve as a useful starting point for estimators working on estimating tasks for public construction projects.