Shadow Prices and Abatement Cost of Soil Erosion in Shaanxi Province, China : Convex Expectile Regression Approach

Using panel data of 83 counties observed from 2000 to 2015, the marginal abatement cost (MAC) of soil erosion is accurately evaluated by directional distance function with convex expectile regression. Besides, we further explore how the external variables affect shadow prices and the choice of abatement solutions. Incorporating the investment of the SLCP into the production process, we could choose the most cost-efficient abatement solution from both the input side (i.e., increasing the SLCP investment) and output side (i.e., downscaling the primary industry or downscaling the non-primary industries). Our main findings are the following: First, after taking into account the inefficiency and noise, this novel data-driven approach could alleviate the overestimation bias and obtain more acceptable abatement costs, which plays a critical role in making scientific allowance standards. Second, the temporal-spatial distribution of abatement costs shows that the abatement potential from the SLCP has shrunk gradually, which implies that there might exist cost-inefficiency in some counties if the SLCP continues. Last, but not least, we recognize that natural conditions, like wind speed and vegetation quality, could influence the input-side abatement cost and output-side abatement cost differently. These results have profound policy implications, asking for more cost-efficient strategies in the future