Alternative methods and models for analyzing the impact of mechanization -- a case study of selected rice growing household in North China

Production effects of mechanization were evaluated using production function analysis and decomposition techniques. First, a t-test was employed to examine if differences exist in yield, input utilization, and farmer's characteristics between mechanized and non-mechanized farms. The results of the test indicated that yield was higher on mechanized farms. In addition, mechanized farms used more fertilizer and chemical inputs and less labor than non-mechanized farms. Mechanized farmers had more education, while non-mechanized farmers had more farming experience. Secondly, Codd-Douglas production for both mechanized and non-mechanized farm groups were estimated. Using a dummy variable for mechanization, the results showed that farm mechanization increased returns-to-scale by reallocating the contribution of inputs use (i.e., the change in output of inputs). Thirdly, the yield effect of mechanization was investigated using decomposition analysis employing a production function framework. The model disaggregated total yield difference between mechanized and non-mechanized farms into the effects of neutral technological change, non-neutral technological change, and the interaction of technological change and changes in inputs use. The results showed that the major source of yield differences between the two farm types was brought about by non-neutral technical change, i.e., a shift in the slope coefficients of the production function, which means differences in resource allocation for the two farm types. The results indicate that mechanization increased rice yield significantly. Lastly, the farmer's decision to buy a tractor, use a rented tractor, or not use a tractor was investigated by employing the Qualitative Response model. The results indicate that the farmer's asset condition was the most important factor which influenced the decision.