Knowledge that Transforms

The object of this paper is, unfortunately, destructive. It is to argue that almost all of the models in the Keynesian balance-ofpayments literature suffer from internal contradictions and deficiencies which make them unsuitable for balance-of-payments theory. Doubling the Keynesian closedeconomy equations does not provide an appropriate description of the international economy. More careful attention needs to be given to portraying the money markets than is usually the case, and assets must be included in the various demand functions in any acceptable analysis. Contributions employing the monetary approach to balance-of-payments theory are largely free of the criticisms made in this paper; see, for example, those by Frank Hahn, Murray C. Kemp (1962, 1964, 1970), Takashi Negishi, Rudiger Dornbusch (1973a,b), Pentti J. K. Kouri and Michael G. Porter, Michael Mussa, Jacob A. Frenkel and Carlos A. Rodriguez, Richard K. Anderson and Akira Takayama, and the author(1970, 1972, 1975, 1976, 1977). The fact that theorists using a different approach have written papers which do not fall into the difficulties discussed below does not, however, necessarily imply either that the profession at large is, or the particular theorists involved were, aware of these problems in the Keynesian literature. We have, for instance, the following quotation from a well-known article in the monetarist literature by Dornbusch.

An investment model is suggested as an improvement over intuition in setting government policies aimed at providing an optimal social infrastructure for boom towns. Using the decision environment and economic characteristics of a Rocky Mountain state boom town, the model shows that low interest rates and early front-end investment produce the greatest stability, while delayed investment contributes to instability. Policy implications derive from the fact that ad valorem property taxes for investment are generally collectible only after construction is completed. Of equal importance to the timing and amount of investment funds is the source of repayment funds. These could be broadened to include wage or use taxes during the construction period.

I enjoyed the story of the plumbing fixtures cartel being drained of its assets. But if I need a plumber's friend, David Mills and Kenneth Elzinga have failed me. With the possible exception of their second paragraph,' nothing in their comment convinces me to eliminate the offending material. I will try to explain why in paragraphs 1) and 2) below. Then I will comment briefly on the antitrust implications in paragraph 3). 1) Mills and Elzinga maintain that my resolution of the deterrence problem is either ineffectual or unnecessary. It is ineffectual in the absence of detection and unnecessary in its presence. They are right about the first. No deterrence is possible without detection. But who would think otherwise, or that I had claimed otherwise? As for the second, they appear to believe that deterrence follows immediately from detection. This belief is obviously mistaken-as our crowded jails prove. More direct proof is furnished by the experience of the International Air Transport Association (IA TA). For detection, this cartel depends on our Civil Aeronautics Board and Department of Justice and its own compliance department (consisting of some fifty investigators) to inspect tickets, receipts, and accounting records at offices of the 'members and their approved travel agents (see IATA Review). For deterrence, it relies on the fines determined by due process before its Breaches Commission or the federal courts. In 1974, the Breaches Commission levied fines of $1.9 million (see Aviation Week); in fiscal 1975, the Civil Aeronautics Board obtained judgments totaling $556,594 (see its Reports to Congress); in September of 1975 the Justice Department obtained fines totaling $655,000 (see Aviation Week). These fines measure IA TA's success at detection and its failure at deterrence. In the same way that lax enforcement of the criminal laws leads to jails full of prisoners, inadequate punishment of cartel breaches increases their expected payoff and stimulates both the breaches and the fines in which they result. The IA TA's penalties have been too small or too uncertain to be regarded as anything more than a normal cost of doing business. Deterrence does not follow from detection.2 2) Mills and Elzinga object to the minimum-variance criterion for choosing among several joint maximizing points. I will be glad to consider an alternative if Mills and Elzinga will offer one. Instead of suggesting a definite alternative or, even better, the principles which govern its choice, they declare in effect that it could be anything.3 If focal points really do depend on analogy, accident, casuistry, and the other things in the list quoted from Thomas Schelling, they are analytically useless: being consistent with everything they explain nothing. 3) Mills' and Elzinga's remarks about antitrust implications disquiet me. They themselves do not recommend that we go about prosecuting oligopolists for having stable market shares, but their remarks alert me to the danger that my analysis will *Federal Reserve Bank of Dallas. IThere they object to my mild chastisement of standard theory for its fascination with the prisoners' dilemma. But since they advance the objection in an incidental manner I will disregard it, pointing out, however, that standard theory consists of the unwritten as well as the written word. For a more detailed criticism of the prisoners' dilemma as a model of oligopoly, see my (1976) paper. 2For more on these problems, see my (l977a,b) papers. 3Concerning the two examples they give (the gasoline marketers and the firms separated by the MasonDixon line), it is impossible to say whether the assumed arrangements indicate the unique joint maximum or a choice from among many.

In a rapidly inflating domestic economy, explanation of the inflationary process must of necessity focus on the determinants of the money supply. By analogy, in a closely integrated economy under fixed exchange rates the behavior of the sum of individual countries' money stocks-the ,world money should play an important role in determining the behavior of the world price (an index of national price levels). This is recognized in analytical models of the international monetarist variety where, under the assumption that all goods are traded (or more generally that relative prices are not affected in the long run by monetary disturbances), the price level adjusts to equate the demand for money with the supply. Strictly fixed exchange rates imply that national money stocks can be treated as components of a Hicksian composite commodity, the money stock, since exchange-stabilization operations prevent variations in the relative values of national currencies. Closely integrated capital markets insure that the money stock is redistributed rapidly from country to country in response to payments disequilibria of monetary origin, thus ensuring a tendency towards rapid return to balance-of-payments equilibrium (which can, of course be frustrated by systematic attempts at neutralization of reserve flows). Closely integrated goods markets insure that the price levels of various countries move in harmony abstracting of course, from divergent trends in productivity and/or tastes that may cause changes in relative prices, including both the terms of trade and the ratio of the price of nontraded to traded goods. In such a world, one can view the stock of money as determining the price of a composite commodity, the components of which are national output levels. Though far too simple for many purposes, this Humean or Ricardian view of the economy is instructive in periods dominated by disturbances of monetary origin. For this type of analysis to be complete, however, the question of what determines the supply of money in the must be answered. This paper seeks to answer this question within the confines of a conceptually (though not necessarily algebraically) very simple model. The is assumed to be divided into two parts, Europe and the United States. money stocks consist of commercial bank liabilities only, and the money stock is defined as the sum of the money balances held by the public of each country.' Various institutional arrangements are considered, including a gold standard, a dollar standard, and the Euro-dollar system. The model provides a first answer to such questions as: does it make any difference to inflation whether monetary expansion originates in one region or the other; what asymmetries does a dollar standard introduce into the international monetary system; what determines the size of the Euro-dollar market and in what sense, if any, is its growth inflationary? Two key assumptions are used to answer these questions, namely, 1) that reserve * Professor of international economics, Graduate Institute of International Studies, Geneva, and visiting professor of economics, Harvard University. The original version of this paper was prepared for the Money Study Group's Oxford Seminar in honor of James Meade, September 25-27, 1974. It includes material developed in connection with a research project on National Economic Policy and the International Monetary System at the Graduate Institute of International Studies under a grant from the Ford Foundation. ITo sum national money stocks, they must of course be expressed in terms of the same currency, existing fixed exchange rates providing the required conversion factor.