Efficiency Assessment and Emission Governance for Haze Based on an Improved Network Data Envelopment Analysis Method

Accurate evaluation of emission governance efficiency can build foundation to develop haze control strategy towards sustainable development. By features of the haze, we view the haze formation stage as the first sub-process and the haze control stage as the second sub-process. This paper proposes an additive aggregation network data envelopment analysis (DEA) model with undesirable intermediate measures and undesirable outputs, which have not been thoroughly studied in previous literature. Our methodological contribution is: the newly developed network DEA model is nonlinear and cannot be converted into a linear program, and an improved second-order cone programming approach is developed to solve this problem. Using the data of haze control in China, our new findings are: (1) The variation in the overall efficiency results from the use of different weights of preference over two sub-processes. Although the efficiency of the haze formation changes with very small values in some provinces, the efficiency of the haze control changes with relatively large values. (2) Decision makers can adjust their preferences on the information of the haze formation and haze control stages to achieve the goal of reducing haze.