Statistical Issues in Price Research

S MR. FOX has suggested, there is great diversity among the procedures currently used or advocated in the analysis of demand. The diversity exists even though all the workers may concur in thinking of the nature of our economic system on the lines of the Walrasian model, consisting of a very large number of simultaneous equations connecting all economic variables. The diversity of procedures appears as soon as the workers try to use their common model for prediction. Unfortunately, the model is only vaguely specified, and is extremely complex. Anyone who wants to use it in clarifying social problems must accept the necessity for simplification. Wherever there is simplification, there are sources of error in prediction. Each worker will choose his own simplification, attempting to avoid the errors that he thinks can be easily avoided, and those that he thinks have serious effects on the appropriateness of social action. Under such conditions, it is not to be expected that there will be any one procedure that ought to be considered the correct one. At least for the present, there seems to be little hope of avoiding a resort to opinion. As price research continues, there may be a consensus among workers to the effect that a certain type of error can be handled conveniently by adopting a specified technique; in such a case, the refinement can be expected to become part of the working knowledge of most researchers. In another case, there may be no feasible way of determining the relative seriousness of a group of errors; in such a case, workers may have to rely on intuition, or on a comparison of computation difficulties. The uncertainty that surrounds our work can be seen by considering the relation between "shock" and "error" models. Assume that one of the equations of a complete economic model is the following