A Technique To Estimate Input Productivity From Farm Data

Unfortunately, procedures are not available for handling variations induced by unquantifiable difference in location, soil, weather or management, particularly if the data source is farm survey information from relatively small samples for a single production season. Estimation problems occur regardless of whether classical, profit or trans-log approaches are used. Procedures suggested by Hoch and Hoch and Mundlak for handling these disturbances in classical production functions require a priori knowledge to devise a weighting system or observations over time to provide estimates of weights. Profit functions as proposed by Lau and Yotopoulos require data to be of such nature that a production function can be specified either in the normal form or that the relationship between profit and input quantities can be specified and estimated. The price of the product is also required to be either a function of quality or of selling costs or that some common or average price is utilized.