Sankhyā - The Indian Journal of Statistics

Next for every relation the effect variable should refer to a distinct subject or group of subjects in the economic sphere the purpose of the relation being to describe how the group reacts to changes in the explanatory variables. [...] Third and last the relations of the system should be autonomous in the sense that the behaving units should be able to shift their reaction pattern without that the shift need necessarily change the other relations of the system. [...] And little wonder for it lies it the nature of formal analysis that the econometric-mathematical wing of the research front is exposed to the danger of being carried away by the mathematical apparatus forgetting about the economic significance of the deductions. [...] At the one extreme we have approaches where the analysis is mainly theoretical the empirical evidence entering only on the setting of the problems and in confronting the theoretical results with factual experience; at the other extreme the approaches are mainly empirical with only the first traces of a logical theory. [...] We note that the extension to evolutive series is bought at the price of some reduction of the genrality for in our final theorem the non-correlation properties of the residual refer to unconditional expectations on the basis of the joint probability distribution of the effect variable and_ the explanatory variables ADDENDUM Cowles Commission Monograph No.