Measurement and analysis of agricultural green total factor productivity based on farmers’ perspectives

Improving agricultural green total factor productivity (AGTFP) and hastening agricultural green transformation are unavoidable choices for comprehensively building a strong socialist, modernized country. Based on a comparative analysis of micro-measurement methods, this study analyzed the status of AGTFP at the farmer household level based on the technically optimized Malmquist-Luenberger index. The kernel density estimation method and the Dagum Gini coefficient method were further used to reveal the dynamic evolution of AGTFP and its regional differences in the micro-sample. The main findings are as follows: 1) From the measurement results, the mean value of AGTFP in the microfield in 2014, 2016 and 2018 was 1.0030, with a good overall development trend. The mean value of AGTFP of farmers in 2016 was 1.0099, and agricultural green development had a good growth trend. The mean values of technical efficiency change and technical progress change were 1.0165 and 0.9928, respectively, indicating that the improvement in farmers’ green agricultural technical efficiency was the main driving factor while the change in technical progress was relatively slow. In 2018, the mean value of AGTFP by farmers was 0.9960, which showed a decreasing trend. The corresponding mean values of technical efficiency change and technical progress change were 0.9765 and 1.0200, respectively, indicating that the technical efficiency improvement of green agriculture did not achieve a sustainable spillover effect and that the innovation function of technical progress change played a role in the improvement. 2) In terms of contributing factors, the use of subjective environmental assessment scores or objective provincial-level environmental pollution data as proxies for non-desired outputs among farmers with higher levels of AGTFP, agricultural green technological progress, and agricultural green technological efficiency was found to be more effective. For farmers with high levels of AGTFP, both green technological advances and green technological efficiency in agriculture were drivers of green growth, and the contribution of the latter was greater than that of the former. 3) From the perspective of a dynamic evolution pattern, in terms of AGTFP, the concentration in 2016 and 2018 was high, showing distinct clustering; however, the divergence phenomenon was not obvious, and the number of farmers with a high level of green development in 2018 was much higher than that in 2016; in terms of the agricultural technical efficiency of farmers, there was no bifurcation in 2016 and 2018. The number of low-level farmers in 2018 was higher than that in 2016, indicating that there was a regression phenomenon, and the difference between the agricultural technical efficiency of high- and low-level farmers was obvious. In terms of agricultural green technical progress of farmers, the overall trend was increasing, the number of low-level farmers in 2016 was lower, and the number of high-level farmers was relatively higher, while in 2018, the number of high- and low-level farmers remained the same, and a spatial clustering effect was evident. In 2018, the number of farmers with low levels of agricultural green technology progress decreased “precipitously.” On the premise that the number of farmers remained unchanged, this part of the low-level farmers moved to the middle- and high-level groups, forming the dynamic transfer effect of “internal push and external pull.” 4) From the perspective of regional disparity, the overall gap in AGTFP in the sample period was decreasing, with a decline of 22.32%. From the source decomposition, the hyper-variance density was the main cause of the overall regional disparity in AGTFP. From the contribution rate, the contribution rate of hyper-variance density was much higher than the contribution rate of intra- and inter-regional disparity, indicating that the cross-over problem between different regions was the main cause of the overall disparity in AGTFP at the farmer level. Further, from the intra-regional disparity, the disparity of AGTFP at the household level decreased within the eastern and western regions; from the inter-regional disparity, the disparity between the eastern and western, eastern and central, and central and western regions decreased continuously during the sample period, and the synergy was the highest, but this gap was susceptible to environmental factors.