Measuring green total factor productivity of China's agricultural sector: A three-stage SBM-DEA model with non-point source pollution and CO2 emissions

Agricultural sector is the basic industry supporting the construction and development of national economy. With the continuous deterioration of agricultural pollution, measuring and understanding agricultural total factor productivity (AGTFP) is the important premise to achieve agricultural green development and clean production. Regarding carbon emissions and agricultural non-point source pollution (ANSP) as undesired outputs, this paper applies a three-stage Data Envelopment Analysis (DEA) method combined with the Slack-Based Measure (SBM) model to eliminate the influences of environmental factors and random errors and explore the real AGTFP of 30 provinces in China from 2000 to 2017.On this basis, the spatial distribution and dynamic changes of AGTFP before and after adjustment are further discussed to seek the underlying reasons. The empirical results demonstrate that AGTFP is lower when carbon emissions and ANSP are both considered, and it decreases from east to west in StageⅠbut the Northeast region surpasses the East to achieve the highest productivity after removing the interferences. Moreover, China's AGTFP has been restricted by the external environment, and superior external environment disguises the poor management efficiency of Beijing and Shanghai. Therefore, some policy implications to balance agricultural environment and economic development are proposed. • Non-point source pollution and CO 2 are integrated to measure agricultural green efficiency. • A three-stage SBM-DEA approach is employed to calculate the real agricultural green total factor productivity. • Different environmental factors have different effects on the slack of agricultural inputs. • China's agricultural green total factor productivity is restricted by the external environment.