The Decentralization of Governmental Control Planning
We have already argued (Chapter XIV) that, in the absence of an indicative plan, a governmental control plan can be formulated only if some governmental econometric model of the working of the economy is constructed in order to know what will be the effect upon the economy of a given pattern of governmental controls. The construction of such an econometric model presents exactly the same problems of decentralization as those discussed in Chapter XIII in connection with the formation of an indicative plan. A single econometric model which accounted for every possible variable in the economy could not possibly be handled. A number of decentralized, partial models, might be constructed by the different departments and agencies of government. The outputs of one such model (to take a very simple example, an estimate by the Ministry of Agriculture of the total wages which would be paid to agricultural wage earners if certain conditions, including the level of governmental subsidies to agriculture, were given) could be used as the inputs of another model (for example, an estimate by the Treasury of the demand for consumption goods by all wage earners). These partial models could be revised until they were consistent with each other, the whole process of revision being helped by a highly simplified but comprehensive framework model of the broad aggregates in the economy. This whole process has been described (pp. 205-6) as a possible procedure between the different sections of a single central statistical office. Clearly it could operate in a very similar manner between all the various departments and agencies of government with some central statistical office ultimately responsible only for the aggregated, framework model.