An econometric evaluation framework for investment in construction safety

A growing body of empirical research has been focused on the relationship between investment and return on safety in the construction industry. However, these previous studies mainly described the accidents occurrence in terms of actual numbers, which were based on the hypothesis that when there is an investment, there will be an effect on construction safety performance and hence the number of accidents will be reduced. In reality, investment on construction safety does not guarantee in reducing the actual numbers of accident occurring in a construction project. The occurrence of accidents sometimes is out of control and indiscipline and hence remains random. Hence it presents a gap between the existing theory and the actual practice which means there is a need to develop theoretical framework that can closely model the accidents' randomness. To fill the gap existed in the previous studies, this paper presents an econometric analysis framework for evaluating the effectiveness of safety investment in construction project. The fundamental assumption is that the investment on safety can only reduce the probability of accident and the number of accidents occurring will behave as a random number in a range with some distributions, or, statistically speaking, shift the distribution of accidents. Firstly, the distribution of construction accidents will be determined by an econometric model. Then the differences in cost with and without safety risk reduction and mitigation programme will be calculated as the benefits of the investment. Individual simulations have no certainty on whether there is benefit, but by doing sufficient numbers of simulations (e.g. 10,000 or more) the distribution of return on investment (ROI) will be acquired. Detail steps for implementing the framework is provided, the results of the econometric evaluation will be made as the expected ROI and its confidence interval, due to the random nature of ROI. The framework developed in this paper will be the base and instruction for future work including using actual project data to test and validate the framework and proposing optimized safety investment models.