Agricultural mechanization and land productivity in Rural China: A multinomial endogenous switching regression analysis

Abstract

This study investigates the determinants of adoption of different mechanization strategies (i.e. no-mechanized farming, semi-mechanized farming, and fully mechanized farming) and their impacts on land productivity. A multinomial endogenous switching regression model is used to estimate farm household data derived from the 2016 China Labor-force Dynamics Survey. The empirical results show that farmers’ decision to adopt semi-mechanized farming is positively affected by household size, access to credit, farm size, irrigation rate, subsidies, and machinery service; their decision to adopt fully mechanized farming is primarily determined by age of household heads, farm size, land use certificate, subsidies, and machinery service. The estimates of the treatment effects show that the adoption of both semi-mechanized and fully mechanized farming exerts positive impacts on land productivity, and the larger impact is associated with the adoption of fully mechanized farming. The disaggregated analyses by gender, farm size, and geographic locations reveal that land productivity effects of agricultural mechanization are heterogonous. In particular, we show that female-headed households adopting semi-mechanized farming obtain higher land productivity than their male-headed counterparts do by adopting the same mechanization strategy; the farm size–land productivity relationship is positive for semi-mechanized farming adopters but negative for fully mechanized farming adopters; both semi- and fully mechanized farming adopters living in central China obtain the highest land productivity relative to their counterparts residing in the western and eastern China.