Monetary Dynamics and the Mundell-Fleming Priority Question: Evidence from the Adaptations in

Despite the importance of this amendment to the original paper, it contains two mistakes, which tend to undermine its validity. The first error is as follows: In citing CJEPS61, Mundell reverses the words in its title. Whereas the correct title is listed in the references in the present paper, Mundell cites his own work as “Employment Policy and Flexible Exchange Rates.” ¹⁷ This error disappears in the International Economics adaptation as the references in it are identified by chapter number. The chapter in question has the title stated identically to that of the original publication.
More importantly, the second mistake concerns the meaning of the term “fiscal policy.” The Long Footnote states that the definition of fiscal policy in CMP is “…an increase in government spending financed by government bond issues with no change in the money supply” (Mundell [1963e, 481]).¹⁸ Since the key point of note in the fixed exchange rate portion of CMP is precisely the endogeneity of the money supply and the alteration in its value as government expenditure increases, this italicized negation seems bizarre. Surely it is a typo, although one that appears in identical fashion in every adapted version of CMP.
IV. The Changed Definition of Monetary Policy
The first paragraph of the Long Footnote claims that the perfection of capital mobility necessitates a change of definition of a given monetary policy. In Mundell’s IMF lecture he states, “In my earliest works on the model I identified monetary policy with interest rate policy…Later, however, when I made the assumption of perfect capital mobility, monetary policy had to be redefined and was correctly treated as an open market operation”(Mundell [2001, 224])^19,20 The tone in this footnote is that this is the first time that the idea has crossed the author’s mind. But this is far from true. Mundell had used the interest rate as an indicator of monetary policy many times before and at least once after (MD, CJEPS61, IMFSP, BNL).²¹ In contrast he had used the money supply (or domestic credit) at least once previously, in his article in Kyklos, although somewhat inconsistently.²²
Furthermore the evidence is compelling that he had considered before the dilemma that is presented when a central bank is setting the interest rate in a perfect capital mobility situation. Namely, he had faced it in writing his MD (Mundell [1960]). In it he states (237) that when it comes to capital mobility “Two extreme cases can be identified: the foreign balance schedule is flat if capital is perfectly mobile and vertical if capital is completely immobile. Figures IV and V [in both of which the foreign balance schedule is shown as being horizontal] are drawn on the assumption that capital is almost completely mobile.” Surely the reason that Mundell uses that ungainly expression, “almost completely mobile,” is that he is aware that there is an apparent contradiction between the notion of perfect capital mobility and the claim that central banks are setting interest rates as their monetary policy implementation. In the International Economics adaptation he uses “virtually perfect capital mobility,” or “extreme case of perfect mobility.” (160)
Before dealing with the validity of this change in definition, let us see instead how it relates to his description of his prior work. In the first paragraph of the Long Footnote in CMP he points out that in CJEPS61 the central bank is maintaining a constant rate of interest. In this case, he claims “…the money supply is allowed to expand in proportion to the increase in income…” (Mundell [1963e, 481])²³ But, in fact, this conclusion can not found in CJEPS61. This overly-specific claim is just one mention of a Fleming [1962, 374, 379] conclusion which Mundell asserts at least five times in CMP.²⁴ While this claim may be correct under certain circumstances, the specific assumptions which are needed to validate it are not made in CMP, although they are so made in the well-defined model in Fleming.
The middle paragraph of the Long Footnote deals explicitly with the definition of a given monetary policy. In fact, many of the observations of this paragraph apply to a situation of capital mobility of any degree, but Mundell phrases his claims in a way which makes it sound as though these points arise only in the perfect capital mobility case. This is simply not true.²⁵
He claims that the reason he changed definitions for the present paper is that “monetary policy cannot in any meaningful sense be defined as an alteration in the interest rate when capital is perfectly mobile, because the authorities cannot change the market rate of interest.” This sentence makes two mistakes. First, it is easy to define monetary policy in terms of interest rates, even when capital mobility is perfectly mobile. One way is to distinguish between the target rate of interest and the actual rate of interest (as is done today by the Bank of Canada in implementing monetary policy. Although they are very close, both values are reported on the Bank’s website (2005)). Another way to have a consistent definition is to think in terms of a dynamic model, as in MD. At a point in time the central bank can set the level of the interest rate, however the degree of capital mobility determines the speed at which that level must adjust. (Operationally this may not be very different from the distinction made immediately above.) (Mundell notes later in CMP that ”…many of our actual observations about the economic world are observations of disequilibrium positions…” (Mundell [1963e, 484]) so certainly for empirical applications of this model a change in interest rates is perhaps possible, if only transiently, even in a perfect capital mobility setting.)
The second mistake in this sentence is that it suggests that this problem arises only in the perfect capital mobility case. This is incorrect. As Kyklos, which deals with a general degree of capital mobility, emphasizes, the central bank can not arbitrarily set the value of the interest rate, because neutralization of the balance of payments can not be carried out indefinitely ([1961a, 163-66])). Presumably CJEPS61 is a short-run analysis, and in this way it is consistent with Kyklos in that time frame. But as time proceeds the money supply adjusts to the needs of the domestic goods market and to restore balance to international payments. CMP (Mundell [1963e, 480-81]), with perfect capital mobility, and Kyklos with a general degree of capital mobility, both state this point unambiguously. Therefore clearly it does not depend on the degree of capital mobility.
The same critique can be leveled at the next sentence in the Long Footnote: “Nor can monetary policy be defined, under conditions of perfect capital mobility, as an increase in the money supply, since the central bank has no power over the money supply either (except in transitory positions of disequilibrium) when the exchange rate is fixed.” (Mundell [1963e, 482]) As Kyklos (160-66) shows this claim applies not only to the perfect capital mobility case, but to a general level of capital mobility, in the Mundell-Fleming specification of the capital account.
To summarize our observations concerning these two sections, we can say that the Long Footnote clearly asserted propriety over results which had previously been reported by Hume, Polak, and Fleming. Further, in order to maintain continuity with previous work, Mundell points out that the diagrammatic tools and some of the definitions had been employed his earlier publications. What differed about the CMP was the definition of a given monetary policy. This change of definition was forced upon the author by his sudden realization of the nature of the environment towards which the world economy was moving.²⁶ The implicit claim is that the results reported in CMP could have been derived by a conscientious readers of the earlier works.

The third paragraph in the Long Footnote (in all adapted versions of CMP except that in International Economics) deals directly with MD. Since that paper is, of the important contributions to International Economics, the one which has the largest amount of material directly altered in its body, we deal with this portion of the Long Footnote in a separate section. But to understand fully those alterations we need first to delineate the constraints that applied to the adaptations undertaken in 1967.

V. The Situation in 1967
In 1967 Mundell was under contract to produce for Macmillan, Inc. a volume of collected papers which was eventually titled International Economics. Presumably, the goal for the editing which went into the manuscript was to make his research output appear to be consistent and progressive when laid out in a single-volume format.²⁷ The central focus of the output would be CMP, but inconsistencies with his earlier or concurrent work would be downplayed, while the element of continuity would be emphasized. Three further considerations were relevant to these adaptations. First it would be desirable to show continuity to work that had been done before Fleming [1962] in order to establish priority over that work. Second, the work had to appear to be an internally coherent, substantial body, with a focus on open economy macromodelling. Finally the work had to be presented as Mundell’s alone and therefore acknowledgements to previous work by others would be modified to make it appear that this body of work sprang fully formed out of Mundell’s analysis. Of course, acknowledging the influence of contemporaries such as Fleming, Krueger, or Sohmen was out of the question.²⁸ Vanity footnotes which had served to get these papers in print could be elided, so long as that was not detrimental.²⁹
In the very same year³⁰ CMP was to be included in the book of readings edited by Caves and Johnson [1968] for the American Economic Association. The editors imposed strict guidelines which permitted only the most limited modifications. Unless in 1968 two markedly different republications of CMP were to be countenanced, therefore, the form of CMP in International Economics was essentially determined by what was in the original publication. But Mundell did make two substantial changes in CMP in editing it for International Economics where it appears as (Mundell [1968h]). First, he included CJEPS64 (with the first three pages elided) as an appendix, something which Caves and Johnson had explicitly refused to do.³¹ In addition, he removed the third paragraph of the Long Footnote. Minor modifications of other footnotes were also undertaken.
Therefore in order to forge tighter connections among his papers, Mundell had to revise the earlier ones. This he did, with substantial modifications made to MD. These changes include the elimination of four of the nine figures, the rewriting of a whole (five-page) section, and the elision of the last paragraph in the appendix.³² We provide the details about these changes in the next section of this paper. The adaptations of other papers were less extensive. In particular, footnotes in all papers, but crucially specific ones in Kyklos and CJEPS61, were modified or deleted. In these papers, as well, footnotes were added in order to draw tighter connections between them and either IMFSP or CMP.³³ But one should note that there has been a change of title with the adaptation of IMFSP which has caused a number of well-known economists to mis-cite this paper.³⁴
MD was crucial to the success of the macro portion of International Economics because it was the first paper Mundell had written in the area. Furthermore, it had sustained the greatest amount of criticism from subsequent papers which he had published. And it clearly predated Fleming’s work on the topic of the monetary-fiscal mix. To present an error-free version of this paper with the early date attached would then support Mundell’s claim of priority over contributions by other authors in the field. MD was obviously different from most other papers which he had authored, in that it took a short-run dynamical approach to processes of adjustment in open economies. It was therefore complementary and not necessarily inconsistent with the comparative static framework which had brought CMP such acclaim.
VI. The Adaptation of MD
Mundell’s “Monetary Dynamics…” [1960] (MD) paper has endured a huge swing in its standing both within the profession and in its own author’s mind. Immediately after its publication, Johnson wrote a letter to Mundell “…saying something to the effect that it carried the subject to a different level…”³⁵ (Mundell [2001, 221]) In contrast, Dornbusch in his discussion of Mundell’s Nobel Prize has only one word to say about this paper: “formidable.”³⁶ He provides no further consideration of it, despite his extensive coverage of many other works, including most notably CMP, Barter Theory (Mundell [1962b]) and Optimal Currency Areas (Mundell [1961b]).
Persson writing for the Nobel Committee says of this article that it presents a model in which “…differences in the speed of adjustment on different markets…” ([2000,xvi]) have consequences for the proper assignment of stabilization policy to various targets, summarized by the Principle of Effective Market Classification. While this principle is just barely presented in MD, as we note below, in fact speeds of adjustment have nothing to do with it.
In MD the speeds of adjustment are parametric and not necessarily higher in one market than another. Furthermore the process of adjustment is strongly influenced by central bank behavior, summarized by a reaction function, rather than by market equilibration processes, therefore the “overshooting” model of Dornbusch [1976] has nothing to do with the analysis in this paper. On the contrary, the money-capital market equilibrium condition is not introduced in the whole paper, and the money supply does not appear as a variable anywhere in the extensive appendix.³⁷
As portrayed, the central bank has the option of choosing the market, in response to whose disequilibrium, it is adjusting the value of its setting of the interest rate over which it has control. Should it respond to disequilibrium in the foreign exchange market (imbalances in international payments), or to disequilibrium in the goods markets? The private sector, by the process of elimination, is then relegated to providing adjustment in the other market, the one which the central bank is ignoring. But in all cases the speeds of adjustment are given parametrically. The main focus is on the state of excess demands in various markets, with particular emphasis on the size of disequilibrium in the balance of payments, and the way in which this depends upon the degree of capital mobility in the Mundell-Fleming specification of the capital account.
While the profession at large after a brief attraction has spurned the article, Mundell has moved in the opposite direction. His own early references to MD dismissed it because it is “…a first approximation…” to work published only a year later. (Mundell [1961a, 158]).³⁸ Further complaints are that it had done little more than “…examined a few…dynamic implications…” and that it left “…much more work…to be done on the theory of flexible exchange rates.” (Mundell [1961c, 516]) By the time the Long Footnote appeared these complaints had crystallized. Repeating earlier claims of shortcomings, CMP states that MD merely focused on “…purely dynamical aspects…” of the comparison between alternative exchange rate regimes, and, more seriously, the paper “..suffers from precisely the defects…[which CMP]…tried to avoid…” (Mundell [1963e, 482])³⁹
These comments, of course, were written before the adaptations to MD which were undertaken in preparation for its publication in International Economics. These adaptations afforded the author the opportunity to remove the defects which were pointed to by the Long Footnote in CMP, and of course this is precisely what he did. In the process four figures are removed from the original form of MD, including the two diagrams with horizontal FF loci. Also elided is the discussion of those cases representing the situation in which capital is “almost completely mobile.” Instead the chapter talks about “virtually perfect capital mobility,” and even deals with “the extreme case of perfect mobility.” But now the reader is no longer confronted with a contradiction between what a figure appears to be implying (horizontal FF is the extreme case in which capital mobility is perfect) and the way it is described in the text (almost, but not quite, complete mobility).
Once the adaptations were done, the improvements justified the removal of the sniping comments in the papers which had been written in1961. Also the entire third paragraph of the Long Footnote in CMP was elided.⁴⁰ Ironically, before the adaptations every important macro Keynesian paper which Mundell had written between 1961 and 1963 referred to MD. But in their adapted versions only one of the papers (IMFSP) makes such a reference, and then only because that paper employs similar dynamic tools to those used in MD. Understandably all the other papers, arguing along comparative static lines, failed to make any reference to it.⁴¹
Clearly this was a useful strategy in editing the manuscripts for the volume. The novelty factor that these dismissals highlighted, which encouraged publication of later papers by denigrating earlier work, was no longer a useful tactic when the papers all appear together as they do in International Economics. The attempt to forge these papers into a coherent unified body of work required the editing to downplay any inconsistencies between later and earlier work.⁴² The emphasis in the Long Footnote in the International Economics adaptation is now entirely on how any apparent inconsistencies are in fact illusory. No defects remain once the adaptation is carried out and CMP appears in the form of Chapter 18 in International Economics (Mundell [1968h]).
With the dismissive remarks removed, all these disparagements are forgotten in the assessment which Mundell makes with the benefit of forty years of hindsight. Now Mundell [2001] is mystified as to why this paper is not rightfully recognized as being an article central to the development of the open economy macromodel. “It took me some time before I realized that some economists did not count the model…[in MD]…as part of the Mundell-Fleming model.”⁴³ [2001, 219] He insists in contrast that “…in some respects this first in the series was the most important and set the methodology for the others⁴⁴…I needed a coherent and plausible international macroeconomic model that was consistent with a full-employment economy…this formulation…fits the world of today better than the variable output versions⁴⁵…The model found a new application for economic dynamics…” [2001, 220]
Questions arise as to what caused the profession to turn its attention away from this article, and what caused Mundell to change his assessment of this paper in the opposite direction. The answers involve the changing nature of the analytics of open economy macromodels. The claim of early Mundell had been that “…an explicit dynamic model is…essential to even a minimal understanding of the meaning of the adjustment mechanism.” (Mundell [1960, 228]) But the late Mundell (Kyklos, CJEPS61, CMP, CJEPS64) had phrased its analysis entirely in comparative static terms. Such analysis was far easier to understand, and this was particularly true when put in the form a checkerboard square. Clearly dynamics, especially when phrased in disequilibrium terms, were being superseded by statics. CMP is the apotheosis of the comparative static analysis, and the first adaptation of it contains the checkerboard square explicitly in a tabular form (Mundell [1963c, 18]).

In addition, in order to try to forge the link between MD and CMP, and thereby to establish the basis for the assertion that MD is part of the Mundell-Fleming model, it would be useful to show that the MD framework is capable of handling comparative static analysis.

While the early Mundell dismissed MD with the statement that it dealt with “purely dynamic aspects” of the comparison of exchange rate regimes, the late Mundell found much to praise in the comparative static capabilities of that framework.⁴⁶ The late Mundell claims that “The comparative statics of the model could show the effects of expenditure changes on interest rates and the relative prices,” (Mundell [2001, 221]) but in fact the early Mundell did not actually carry out such an experiment.
For those readers who can not muster the ingenuity to tackle this exercise for themselves, the late Mundell does it for them. “One implication of the model was that a domestic boom (shift up and right of the XX curve) would raise interest rates, attract capital inflows, appreciate the real exchange rate, and worsen the balance of trade, a conclusion that would hold under either fixed or flexible exchange rates. This was very relevant to an understanding of the economy of Canada…in the 1950s,…the Reagan boom in the early 1980s and the German unification boom in the context of the exchange rate mechanism crisis in the early 1990s. Under the old Keynesian model, which typically assumed capital immobility, it was generally assumed that domestic expansion would weaken the currency.” (Mundell [2001, 221]) ⁴⁷
There are a number of problems with this discussion. From the Canadian perspective, most analysts see the high value for the Canadian dollar during the early 1950s as being due originally to an export surge as world demand for Canadian primary commodities increased.⁴⁸ The correct portrayal of this situation therefore requires that both the XX and the FF loci shift right in response to this shock. As a consequence the trade account initially improved, being tempered somewhat by the high value of the Canadian dollar, whose movement seemed to coincide with the change in the terms of trade.
In the middle of the decade the current account worsened, in response to an investment surge, which took the form of heavy increases in imports of iron and steel. Once again the common view is that “non-price factors” are at work. It is true that the current account worsened, but this is not due to movements in the terms of trade. Nor did the interest rate move markedly in the direction that the MD model suggests it should have. The Canadian dollar fell in value at that time, along with the worsening of the Current Account.
The most noteworthy feature of the decade, however, occurred at its end. It is generally agreed that it was the misguided tight monetary policies of James Coyne as the Governor of the Bank of Canada which brought an abrupt end to what had until that point been a fairly successful experiment with flexible exchange rates. In light of these various episodes, we can state clearly that the 1950s was not a period which would be called in its entirety one of domestic boom.⁴⁹