ABSTRACT
This chapter presents the weaknesses in ‘deterministic’ DEA measurement particularly on the upward bias (Ozcan 2014; Simar et al. 2014), which ignore the uncertainty issues in the context of noise in data, small sample (sampling error) and probability. These weaknesses are dealt with by using the extension of the classical DEA approach known as Bootstrap DEA (BDEA), the Sensitivity Analysis and Chance-constrained DEA (CCDEA). This chapter addresses the importance of the assumptions of uncertainty issues in uncertain climate and shows how risk management efficiency results will change. These three approaches are formulated to suit the purpose of this study to achieve accurate, robust and reliable results, and the objective to provide comparative discussion on the traditional DEA results, implications and conclusion.
