THE METHODS EMPLOYED in analyzing the effects of economic changes on a country’s balance of payments have undergone drastic revision in the course of the last 50 years. The evolution of economic ideas and analytical methods generally follows, although sometimes with a considerable lag, the emergence of economic problems requiring solution. Balance of payments analysis, too, has been influenced directly by the changing character of international economic problems; in addition, however, it has also been affected by changing methodological fashions in the mainstream of economic thought.

THE METHODS EMPLOYED in analyzing the effects of economic changes on a country’s balance of payments have undergone drastic revision in the course of the last 50 years. The evolution of economic ideas and analytical methods generally follows, although sometimes with a considerable lag, the emergence of economic problems requiring solution. Balance of payments analysis, too, has been influenced directly by the changing character of international economic problems; in addition, however, it has also been affected by changing methodological fashions in the mainstream of economic thought.

In the period following World War I, when the problems of resource allocation occupied the center of the stage, the tools of value theory—demand and supply schedules and their elasticities—were applied to the then new problem of exchange devaluation.1 After World War II, economic theory was dominated by the memory of the Great Depression of the 1930s and by Keynes’s method of analyzing it. In particular, partial equilibrium analysis of the labor market and of the determination of wage rates had come to be regarded as faulty and misleading, since a change in the wage rate—the most important price in the economy—would have macroeconomic repercussions leading to shifts in the labor demand and supply schedules. Applying this methodological point to balance of payments analysis, it was easy to see that the “elasticities approach” to analyzing the effects of a change in the exchange rate—for many economies, another very important price—was subject to a similar criticism. A devaluation tends to affect not only the relative prices of traded and domestic goods but also aggregate income and expenditure. The second effect induces shifts in the demand and supply schedules and thus invalidates the basic supposition of the simple “elasticities approach,” namely, unchanging demand and supply schedules.

This difficulty was avoided by following a method that was much more in keeping with the Keynesian framework of macroeconomic analysis. This approach, which developed largely on the basis of research in the Fund conducted under the guidance of E.M. Bernstein, views the balance of payments on current account as the difference between national income and national expenditure (or absorption).2 The effect of any economic change—for instance, a devaluation—on the current balance is, therefore, best assessed by ascertaining its effects on output (income) and on absorption, and by subtracting the latter from the former. This “income-absorption approach” has been criticized on two grounds: first, like the elasticities approach, it does not deal with the balance of payments as a whole but only with the current account; second, while it lends itself readily to an analysis of the effects on the current account of changes directly affecting income and absorption—for instance, of an increase in government expenditure—it is much less suited for an assessment of changes affecting, in the first instance, exchange rates and prices.

As public preoccupation with the insufficiency of aggregate demand and with unemployment gave way during the postwar period to concern about inflation, the Keynesian analytical tools were supplemented, and in some instances replaced, by the simple, if somewhat old-fashioned, instruments of monetary analysis. While there is still controversy about the role of monetarism in solving problems of inflation and unemployment, the monetary approach—all the proponents of which are not necessarily “monetarists” in the narrower sense—has come to occupy a central place in the analysis of balance of payments problems. Several features of this approach, besides the growing popularity of monetarism in general, help to explain its recent ascendancy. Its supporters could argue that “the balance of payments is essentially a monetary phenomenon”; 3 that it is relatively easy to apply, especially if certain plausible simplifications are accepted; and that its strength lies precisely where the other approaches falter, namely, in the analysis of the overall balance of payments-of current and capital transactions taken together. It may be worth elaborating briefly on these aspects of the monetary approach.

Some Characteristics of the Monetary Approach

What is ordinarily called “the balance of payments surplus or deficit,” or the “overall balance of payments,” refers to the net balance of certain financing items in the double-entry bookkeeping system of the external accounts that reflects the net monetary impact of the other transactions recorded in the balance of payments statement. Since it is often the objective of balance of payments analysis to ascertain the effect of some economic change—the imposition of a tax or tariff, a rise in the rate of inflation, a change in the exchange rate, an increase in the price of an important raw material, etc.—on the overall balance, it is plausible to approach the problem by analyzing the “money account” rather than the numerous accounts recording transactions of various goods, services, and capital items. The change in the money account—in most instances simply the change in international gross reserves—is directly linked to monetary balance in the national economy by the condition that the change in external reserves must equal the difference between the change in the demand for money and the change in the supply of money of domestic origin.

It has often been pointed out that each of the three approaches could in principle produce the right answers if it were correctly applied, that is to say, if proper allowance were made for all the repercussions throughout the economy of the change whose effect is being analyzed.4

It has, however, proved difficult in practice to set forth a suitable framework for use with either the elasticities approach or the incomeabsorption approach within which the requisite information would be marshaled in a comprehensive and consistent manner. For applied research and background work for policy discussion on balance of payments problems, the monetary approach suggested itself, therefore, as apparently simpler and more manageable than the other approaches. It is based on the postulates of a stable demand function for money and of a stable process through which the money supply is being generated.5 The demand for money, it is argued, depends on a relatively small number of economic factors, and the effects of economic changes on the demand for money are therefore easy to assess because they can operate only through one or several of these few factors. A similar argument can be made with respect to the determination of the supply of money. By focusing directly on the relevant monetary aggregates, this approach eliminates the intractable problems associated with the estimation of numerous elasticities of international transactions and of the parameters describing their interdependence, which are inherent in other approaches.

The apparent simplicity of the monetary approach to the balance of payments is, however, somewhat deceptive. Even though for many purposes the demand for money can be conveniently expressed as a function of a small number of variables, it is still just as much the resultant of all the influences that come to bear on the economy as are national income and national expenditure. Again, domestic credit creation, which is often taken as being determined exogenously, may in fact be systematically influenced by factors determining the demand for money or by some of the events whose monetary effects are being examined. These considerations do not invalidate the monetary approach; they merely draw attention to the possibility that it will be seen, on further examination, to be not quite so superior in terms of simplicity of application as had first been thought.

The elasticities approach and the income-absorption approach are best applied to the analysis of changes in the merchandise trade balance and the balance of net exports of goods and services, respectively. If more comprehensive balance of payments concepts are to be analyzed, for example, the basic balance or the overall balance, the analysis has to be supplemented so as to cover the additional balance of payments categories (for instance, capital movements), usually in a manner that does not fully conform to the approach in question. The monetary approach, by contrast, leads directly to the determination of the overall balance of payments as the difference between the change in the demand for money and the change in the net domestic assets of the banking system (domestic credit creation). If the effect of some economic event on other balance of payments components, say, on the current account balance, is to be ascertained, the simple model based on the demand for and supply of money, and the factors directly influencing them, can be extended by the inclusion of the requisite additional relationships. In this connection it is well to remember that the choice of the monetary approach to the analysis of the balance of payments does not confine the analyst to a small and simple model. As Johnson has urged in his recent significant extension of the monetary approach,

the general thrust of the new approach ought not to be identified with, and assessed according to the plausibility of, the particular simple models that its proponents have employed to derive some of the central conclusions that differ from those implied by the conventional models, or to express those conclusions as logical consequences of general equilibrium models constructed with full use of the mathematical and economic expertise now required of scientific model construction.6

Development of Monetary Balance of Payments Models at the Fund

While the monetary approach to balance of payments analysis in a sense dates back to David Hume and the classical specie-flow mechanism, two developments contributed to the renewed interest in this approach in the period following World War II. One important factor was the renaissance of academic interest in monetary problems, which was spearheaded by members of the University of Chicago. For the last quarter century, Professor Milton Friedman has been the main contributor to the rehabilitation of the quantity theory of money; and during the last decade, Professors Robert Mundell and Harry Johnson have intro-duced the monetary approach to the balance of payments in academic circles. The other factor, which preceded the emergence of academic interest in this topic, was the development of a direct concern with monetary policy questions that were encountered by central bankers and by other national and international officials. As is clearly brought out in the paper by J.J. Polak (Ch. 2 of this volume), written in 1957, acceptance of the Keynesian analytical framework by the economics profession had left a gap between problems that could easily be solved with the help of Keynesian tools and those frequently encountered by officials concerned with monetary and balance of payments questions. In the construction of a theoretical basis for solving such problems and in the design of procedures for quantitative analysis in this area, much of the innovative work was done by the staff of the International Monetary Fund. The studies contained in this volume represent a collection, albeit not a complete one, of contributions to this field by Fund staff members written between 1957 and 1974.

The initial impetus toward this research came from the staff’s work on problems of less developed countries. There were, perhaps, four reasons for this. First, in the 1950s, many less developed countries lacked the detailed national income and product accounts necessary for an analysis of national income and balance of payments determination along Keynesian, or income-absorption, lines; nor was it feasible to apply the elasticities approach in an adequate manner. However, monetary statistics were usually obtainable as a result of the central bank’s exercise of its supervisory authority over the banking system. Similarly, balance of payments data were available as a by-product of the customs administration and from banking sources. In view of the availability of these two sets of data in a large number of countries for which other statistical information was scarce, the thought naturally presented itself to develop a framework of analysis that could take full advantage of this data base.

Second, the nature of the Fund’s work on balance of payments problems of member countries made it desirable to have available a framework for quantitative analysis that was sufficiently manageable (in the days before long-distance access to computers) to be serviceable during staff missions to foreign capitals. The monetary approach permits a meaningful approximate analysis of the relevant aggregates with the help of models that are small enough to be calculated with pencil and paper.

The third reason is more fundamental. Less developed countries typically have a simpler financial structure than do more developed countries. In the absence of well-developed asset markets and financial instruments, there are relatively few alternatives to either holding funds in monetary form or spending them on domestic or foreign goods and services or on foreign financial instruments. In these circumstances, the implication for the external balance of a difference between the amount of money newly supplied through domestic credit creation and the additional amount that residents wish to hold is more obtrusive than it is in countries with a more complex financial structure.

Finally, a monetary framework for analyzing the balance of payments effects of economic policy was particularly appropriate for many developing countries, particularly in Latin America, in which control over domestic credit was in fact relied on as a major instrument—perhaps the most important one—of demand management and balance of payments control.

The Fund’s approach to monetary management—or, as it came to be called, to financial programming—for the purpose of achieving balance of payments equilibrium evolved during the 1950s, initially in staff work on Latin American member countries. It emerged from the need to discuss with the authorities of a member requesting financial assistance from the Fund the adequacy of the policy program proposed by them and the quantitative conditions (“credit ceilings”) under which the member would continue to have access to the Fund’s resources made available in a stand-by arrangement.

This approach rested on Professor Triffin’s analysis in terms of “money of external and internal origin”—concepts that correspond to net foreign assets and net domestic assets of the banking system, respectively—and the relation of the former to the overall balance of payments. 7 The central element of this approach was the estimation of the prospective demand for money on the basis of forecasts of real gross domestic product (GDP), an assumption about future price inflation, and any other relevant information. By controlling domestic credit creation during the period under review so as to equal the estimated change in the demand for money, the authorities could keep the external accounts in balance and the change in international reserves to zero. If an external surplus was to be achieved, perhaps in order to permit repayment of indebtedness, domestic credit creation would to that extent have to be kept below the forecast change in the demand for money; and if a deficit could be temporarily tolerated, domestic credit creation could be allowed to exceed the anticipated change in the demand for money.

There was a theoretical difficulty in the initial formulation of this scheme: the growth of output and the change in the price level had to be assumed to be known without prior knowledge of the magnitude of domestic credit creation. But this shortcoming can be—and has been—surmounted by iterative calculation carried to the point at which sufficient consistency is obtained between the estimated changes in output and prices, on the one hand, and the calculated value of domestic credit creation, on the other hand. A source of greater concern was the difficulty of forecasting the national price level. In many instances, this difficulty was resolved by assuming that the national price level would move in conformity with the world price level, which could be forecast more easily. This means that the “law of one price,” a prominent postulate in much of current balance of payments theorizing, was used as a standard element of this approach. In other instances, where the importance of prices of nontraded goods and services made reliance on the law of one price obviously inappropriate, the national price level was forecast on the basis of an appraisal of the government’s announced wage policy and its probable implications for the general price level. Although price forecasting was always a troublesome aspect of the application of this procedure, forecasting the demand for money on the basis of a given nominal GDP was generally a more important source of error.

Selections Contained in This Volume

This general procedure, of which there were a number of variants, was taken as a point of departure for a small model designed for monetary analysis of income formation and payments problems, which was constructed by J J. Polak in 1957 (Ch. 2 of this volume). This marked the beginning of formal research on monetary balance of payments models in the Fund. The first seven papers in this volume (Chs. 2-8) illustrate the further development of this research and its application by Polak himself and by other staff members. The last three papers (Chs. 9-11) are contributions by staff members whose university training had exposed them to the new academic approach to monetary balance of payments analysis before coming to the Fund. With these papers, written in 1973 and 1974, the two roots are joined, therefore, and a proper cutoff point for this volume is reached.

In his 1957 article, Polak sets out his own rehabilitation of the quantity theory of money, which differs in important technical detail from that of Milton Friedman and his disciples. In particular, he uses the assumption of constant income velocity of circulation—which differs however among countries—for a novel analytical configuration in which the unit period of the model for each country is chosen so as to make the average income velocity equal to 1.0 and the predicted change in income from period to period equal to the change in money. Despite its appearance, this relation between income and money is not an identity but a behavior relation, which may be looked upon as a demand function for money or, alternatively, as a function determining current income from the income of the preceding period and the change in the money supply.

Polak uses an import function as a second behavior equation and assumes exports and capital movements, as well as domestic credit creation, to be determined exogenously. The current values of income, imports, and money are, therefore, the result of the historical course of domestic credit creation and the exogenous elements in the balance of payments. Compared with the approach based on Triffin’s analysis and developed in the Fund’s country work, Polak’s method has the advantage that it does not require forecasting of nominal income, which is endogenously determined (as it is in some of the more recent contributions to the academic literature on the monetary approach to the balance of payments). This advantage is bought, however, at the cost of having to forecast exports, net capital movements, and domestic credit creation. Polak’s findings cover the qualitative results with respect to the long-run effects of credit policy and changes in exports and capital movements on income and the balance of payments that have been obtained from small monetary models of open economies in the more recent literature. In addition, however, they also extend to short-run effects and dynamic adjustment paths.

The basic theoretical article in 1957 was followed in 1960 by an empirical study by J.J. Polak and Lorette Boissonneault (Ch. 3), which applies the model to 39 countries. Several analytical refinements are introduced, but the main achievement of this study is without doubt the successful application of the theoretical model to a wide variety of countries. The study shows that the model can predict imports reasonably well from monetary data, so that—in conjunction with an appropriate forecast of exports and net capital flows—it is possible to determine the effects of alternative domestic credit policies as well as other autonomous or policy-induced changes on the balance of payments.

In 1961, J. Marcus Fleming and Lorette Boissonneault (Ch. 4) followed up on the earlier Polak papers, focusing specifically on the relationship between money and imports. In their investigations covering 36 countries, they find a good correlation between actual imports and imports predicted on the basis of the model. The relationship appears to be better in small and less developed countries than in large and industrialized nations. This paper also uncovers what has sometimes been referred to as the “excessive lag in the Polak model,” namely, the tendency for predicted imports to lag behind actual imports, and provides a full analysis of its possible causes.

In the next selection, S.J. Prais recasts the Polak model, which relies on difference equations, in differential equation form (Ch. 5). Besides adding mathematical elegance, the formulation of the model in terms of continuous time allows precise specification of the relation between the stock and flow variables in the system. Prais specifies a domestic expenditure function emphasizing the role of deviations of actual from desired money holdings as the link between the real and monetary sectors of the economy. This particular specification, which was implicit in the Polak model, has come to be widely used in the recent literature.8 As a result, this paper has been one of the most frequently cited of the papers contained in this collection.

Rudolf R. Rhomberg (Ch. 6) also focuses attention on this relation between money and expenditure and argues that the short-run predictive properties of the Polak model may be improved by using the Prais expenditure function. He then estimates the entire structure of the Polak model by multiple regression techniques, in contrast to previous empirical investigations by Polak-Boissonneault and Fleming-Boissonneault, in which the parameters were estimated as average ratios over the sample period. Moreover, by estimating the structural equations rather than the reduced form of the Polak model, Rhomberg is able to increase its predictive performance.

The paper by Victor Argy (Ch. 7) formulates the entire model in real terms, while its predecessors were cast in nominal terms. Moreover, Argy considers the interest rate explicitly, which was treated only peripherally in the other models. He shows how two policy targets (real income and the balance of payments) may be achieved by selecting proper values for the two policy instruments in his model (government expenditure and the money supply), and he investigates the implications of alternatively using money, or total domestic credit, or central bank credit as the policy variable. A passive monetary policy, which holds the interest rate constant, is also considered.

The following paper by J.J. Polak and Victor Argy (Ch. 8) broadens the original Polak paper by incorporating capital markets, so as to make it more relevant for monetary analysis of advanced industrial countries. Polak and Argy consider the effectiveness of credit ceilings versus monetary ceilings on the balance of payments in the face of unanticipated domestic and foreign disturbances. They conclude that a credit ceiling is always a superior stabilizer as far as the balance of payments target is concerned. As regards the target of stabilizing output, however, it depends on the kind of disturbance threatening stability whether a credit ceiling or a monetary ceiling is to be preferred.

A similar question is considered from a different angle by Manuel Guitian (Ch. 9). He emphasizes that in an open economy only domestic credit creation can be controlled by the monetary authorities, while the money supply is also influenced by factors outside their control. He concludes—reinforcing the point made by Polak and Argy—that domestic credit is the appropriate policy instrument for the control of the balance of payments.

The paper by Mohsin Khan (Ch. 10) applies the monetary approach to ten less developed countries. Khan pioneers in formulating the model in a continuous-time disequilibrium framework. This allows him to estimate the time pattern of adjustment to the final equilibrium values via a system of linear differential equations. His simulations show that the model is able to explain the behavior of the balance of payments (and income) in a satisfactory manner for a wide variety of countries.

Bijan Aghevli and Mohsin Khan (Ch. 11) conduct a cross-sectional test involving 39 less developed countries. They concentrate on the long-run implications of the monetary approach for changes in international reserves. The authors conclude that the monetary approach performs well when applied to cross-sectional data. Given the wide variety of countries included in the cross-sectional study and the restrictive assumptions made—such as perfect capital mobility, full employment, and fixed exchange rates—they consider the satisfactory results obtained to be evidence of the robust nature of the monetary theory of the balance of payments.

Concluding Observations

A few general remarks about certain similarities and differences between the earlier Fund work and the approach developed in the academic literature of the last decade may be offered in conclusion.

Both approaches consider the balance of payments to be an essentially monetary phenomenon and stress the importance of the demand for money and of the money supply process in an open economy. Proper specification of the money supply process is perhaps the most important single element distinguishing the monetary approach from earlier analyses. In Keynesian models, it was customary to consider the money supply as determined by the monetary authorities and thus as exogenous to the model. This practice was retained even after foreign trade in goods and services was incorporated in the model, making it explicitly a model of an open economy, and was defended by reference to the possibility of offsetting the effects of changes in international reserves on the money supply.9 While the framework of the monetary approach permits analysis of the consequences of insulation of the money supply from the influence of the foreign balance, the adherents of this approach emphasize that sterilization is impracticable in the longer run—certainly for small less developed countries with relatively undeveloped financial systems, but even for larger developed countries with fully articulated financial markets and institutions.

Specification of the money supply as endogenous and subject to being influenced by surpluses and deficits in the overall balance of payments is responsible for certain results, common to Fund and academic models alike, that were considered novel by students of traditional Keynesian models. For instance, as long as there is a balance of payments surplus or deficit, the money supply is changing and equilibrium has not yet been reached.10 Or, a lasting increase in the rate of domestic credit creation will, in the long run, bring about a steady loss of reserves until the country’s reserves are exhausted. Or again, a devaluation of the exchange rate will raise the level of a country’s international reserves but will not affect the long-run equilibrium level of its overall balance of payments. These and similar results can be obtained with the standard Keynesian model, provided only that the money supply is allowed to be affected by the foreign balance. (See Argy, Ch. 7.)

Both the Fund approach and the more recent academic literature deal with long-run consequences of economic changes for the balance of payments. Both approaches lead to the conclusion that a change that does not interfere with the monetary balance of payments adjustment process—for instance, a lasting autonomous change in the level of exports or capital inflows or a once and for all change in the exchange rate—will have no lasting influence on the balance of payments but will, under a regime of pegged exchange rates, alter the level of. reserves. A change that does interfere with the adjustment process—such as a lasting change in the rate of domestic credit creation, which permanently alters the relation between the demand for money and that portion of it which is supplied from domestic sources—would have a lasting effect on the balance of payments and in the long run an indefinitely large effect on the level of reserves.

The Polak model and the other models used at the Fund are designed as dynamic models and differ in this respect from most of the academic monetary balance of payments models. The Fund models can be used not only for deriving the long-term results just mentioned but also for assessing short-term effects of policy changes and other exogenous events, and for estimating time paths of balance of payments adjustment. This is particularly important in the context of applying the analysis to current policy problems, where the short-run effects of policy measures and other exogenous changes, even though they may be known to be transitory, could well be of greater interest than the ultimate effects after long-run equilibrium has been re-established.11

It is clear that the construction of dynamic models requires decisions with respect to the specification of certain behavior relationships and to the time lags involved that can be left unarticulated in long-term equilibrium models. Since many alternative specifications are possible, there can be numerous short-term dynamic models corresponding to a single long-term equilibrium model. Results concerning short-run effects or effects on particular accounts of the balance of payments that can be obtained with these specifications, however useful they may be, are not unique and are in this sense less generally valid than are the basic conclusions drawn from the long-run equilibrium properties of these models.

The final point of comparison concerns the application of the monetary approach to the determination of changes in exchange rates. This is now a prominent area of research by monetary economists using tools of analysis that are related to, or evolved from, the basic monetary model of the balance of payments. Just as this model can be used under conditions of pegged exchange rates to assess the effects of various economic changes on reserve movements, so can it be used in a world of floating exchange rates to ascertain the effects of such changes, including intervention policies, on the level or rate of change of the exchange rate. Although the analysis of exchange rate changes within the monetary approach to the balance of payments has at times been part of the Fund’s work on financial programs of member countries, published work at the Fund in the 1950s and 1960s dealt only incidentally with this topic.


This type of analysis goes back to F. Y. Edgeworth and Alfred Marshall, and indeed to John Stuart Mill. The economist credited with the first statement of the elasticity conditions for stability of “the foreign exchange market”—or, more accurately, the conditions under which a devaluation will improve the trade balance—was C. F. Bickerdike, writing in 1920 (“The Instability of Foreign Exchange,” Economic Journal, Vol. 30 (March 1920), pp. 118-22). See Lloyd A. Metzler, “The Theory of International Trade,” Ch. 6 in A Survey of Contemporary Economics, Vol. 1, ed. by Howard S. Ellis (Philadelphia, 1948), pp. 210-54, especially pp. 225-32.


These ideas were discussed in internal Fund documents and in the Fund’s Annual Reports, for example, in an unpublished research paper by J.J. Polak, “Depreciation to Meet a Situation of Overinvestment” (September 10, 1948), and the Annual Report, 1950, pp. 23-24; they became more widely known through a published research paper by Sidney S. Alexander, “Effects of a Devaluation on a Trade Balance,” Staff Papers, Vol. 2 (April 1952), pp. 263-78.


Jacob A. Frenkel and Harry G. Johnson, “The Monetary Approach to the Balance of Payments: Essential Concepts and Historical Origins,” in The Monetary Approach to the Balance of Payments, ed. by Jacob A. Frenkel and Harry G. Johnson (University of Toronto Press, 1976), p. 21.


Several reconciliations of the elasticity and absorption approaches have been attempted. One that is perhaps most deserving of citation in the present context is S. C. Tsiang, “The Role of Money in Trade Balance Stability: Synthesis of the Elasticity and Absorption Approaches,” American Economic Review, Vol. 51 (December 1961), pp. 912-36.


See Michael Mussa, “Tariffs and the Balance of Payments: A Monetary Approach,” in The Monetary Approach to the Balance of Payments (cited in footnote 3), pp. 187-221.


Harry G. Johnson, “The Monetary Approach to Balance-of-Payments Theory: A Diagrammatic Analysis,” Manchester School of Economic and Social Studies, Vol. 43 (September 1975), p. 221.


See, for example, Robert Triffin, “Esbozo General de un Analisis de las Series Estadisticas Monetarias y Bancarias de America Latina sobre Bases Uniformes y Comparables,” in Memoria: Primera Reunion de Tecnicos sobre Problemas de Banca Central del Continente Americano, Bank of Mexico (Mexico City, 1946), pp. 410-30.


See, for example, Rudiger Dornbusch, “Devaluation, Money, and Non-Traded Goods,” in The Monetary Approach to the Balance of Payments (cited in footnote 3), pp. 168-86.


An exception is J.J. Polak and William H. White, “The Effect of Income Expansion on the Quantity of Money,” Staff Papers, Vol. 4 (August 1955), pp. 398-433, where the money supply is allowed to be affected by the foreign balance in an otherwise traditional Keynesian model.


Nor would sterilization be consistent with full equilibrium, since it implies changes in the banking system’s holding of certain financial assets.


This was also pointed out in Gottfried Haberler’s review of The Monetary Approach to the Balance of Payments (cited in footnote 3), in the Journal of Economic Literature, Vol. 14 (December 1976), pp. 1324-28.