On the Adequacy or Inadequacy of Keynesian Balance-of-Payments Theory: A Reply
回应了三位学者对作者1978年文章的评论，重点讨论凯恩斯模型在固定汇率下的一致性和完备性问题，指出批评者关于冲销操作的假设不符合现实。
This paper is written in reply to the contributions by Willem H. Buiter and Jonathan Eaton, Alan Deardorff, and Norman C. Miller which appeared in the September 1981 issue of this Review as comments on my 1978 article. I propose to deal mainly with the issues raised by Buiter and Eaton, Deardorff, and Miller in his Section IV, as the point of departure of the first two sets of authors and Miller in the section noted is essentially the same. Miller, in his Sections II and III, however, takes a different view, arguing that the Keynesian models I criticized were meant to be nonequilibrium constructions in important respects. If this is the case, my remarks would have been inappropriate, since they were made specifically in the context of models in which all markets are supposed to clear. I personally do not think that Miller's basic premise in these two sections is correct, but readers will of course make their own judgements. The papers on the whole contend rightly that it is my Propositions 2 and 3, concerning the inconsistency and incompleteness of the Keynesian models under fixed-exchange rates, which are the crucial ones. The other two propositions may be easily accommodated by introducing wealth into the demand functions of the models (as is clear from the choice of words in those two statements). Proposition 2 asserted that models which require the demand for money to equal the domestic supply in each country force the balance of payments to vanish, which contradicts the values given by their defining equations. The commentators argue that this is incorrect because I did not properly consider the exchange-market-intervening activities of the authorities. Proposition 3 stated that neither bond market clears in Keynesian models because Walras' Law permits the exclusion of only one market-equilibrium equation. Again, there is unanimity of approach, although not of results, in the three papers. Each of them attempts to utilize a balance-of-payments equation to make the system determinate. My reply to these positions is the following. The authors' argument against Proposition 2 assumes that the monetary authorities are able to sterilize exactly and simultaneously all balance-of-payments flows. This is, of course, patently false empirically. It requires the ability of the monetary authorities to predict precisely what the aggregate equilibrium currency flows will be if it steps in to totally offset them. This involves knowledge of a completely different order than that conventionally assumed in general equilibrium models. It is as if the Walrasian auctioneer were first, as usual, to call out his prices to all agents. But then, in an aside, each exchange authority is informed of the total sum of its residents' demands and supplies of foreign exchange, so that, in the same marketing period, it can formulate a response to exactly counterbalance their plans. This is not the way the world works. Exchange authorities, lacking omniscience, sterilize currency flows after they have taken place, and the possibility that the market period may be reasonably short does not alter this fact. The difference, moreover, is not one of theoretical hairsplitting, since the two alternative sterilization assumptions do give different results. In addition, if the simultaneous sterilization interpretation is made for the Keynesian models, they can only deal with the case in *London School of Economics. I would like to thank Howard Petith and Kyoung Mihn for helpful comments on a previous draft of this paper.