Knowledge that Transforms

They asserted that Gt is a function of Rt but that R* is in no way a function of Gt (see p. 51). From this identity it of course follows that the cycles in R* will generate cycles in R unless offset by G. Investigating the period April 4, 1951-May 31, 1967 as a whole, B-S found 1) monthly cycles in R* and G, 2) no monthly peak in R, and 3) no cross covariance between G and R.' Based on these results they concluded that the effect of Federal Reserve open market operations since the Accord has been practically to eliminate the very strong monthly element cycling in member bank reserves, . (p. 59). In addition, they found a monthly cycle in currency outside all and significant coherence between this series and both R* and G. On the basis of these data they suggested an interpretation of open market operations as being undertaken, or at least having the effect of existing primarily, for the purpose of offsetting a monthly cycle in currency outside all banks (p. 60). But B-S committed important methodological and analytical errors which require us to set aside their results and conclusions and search for correct ones.

J. M. Montias is primarily concerned with the effectiveness of the iteration, the only completely decentralized planning procedure described in my model. He has two main points. First, he shows that when production in every sector is centrally planned, the addition of a iteration to the planning procedure may actually increase supplydemand imbalances. Secondly, he argues that given the existence of capacity limitations, the retrospective iteration is likely either to generate gross outputs in excess of these limits or to generate an inferior bill of final demands in the course of the plan execution. Both of Montias's points have very important implications for the interpretation of the model, and I would like to comment on them.