Wheat Production Under Alternative Production Functions

In an earlier study [1, pp.407-415] it was assumed that a linear production function would be the best to depict the relationship between wheat production and explanatory variables in the Punjab. After a careful scrutiny, four of the explanatory variables, viz., area under Mexican Wheat varieties, area under local wheat, fertilizer application and rainfall, were employed to explain variations in wheat production. As this linear relationship gave a good fit, it was claimed that the equations, derived from 1967-68 to 1971-72 output and input data, can be useful in forecasting wheat production in the Punjab for future years well in advance of actual wheat harvests and, in fact, a forecast of wheat production in the Punjab for 1973-74 was published [2, pp. 106-112], based on the findings of that model. Whether the linear production function related the dependent and independent variables in the best possible manner was not tested. The results of our analysis in the present paper indicated that the re-specification of the relationship will be of little use.1 In addition, in an earlier model the rainfall relevant to wheat production was defined as the total rainfall during the months of November-January of each wheat season. Publication of the article containing that model elicited suggestions that use of seven-month rainfall from July to January (instead of three-month rainfall used in the model) might improve the results of the model. Accordingly, it was decided to test that model using