Monetary Dynamics and the Mundell-Fleming Priority Question: Evidence from the Adaptations in
Yüklə 155 Kb.
|
| This Canadian example recalls, as well, the discussion in CJEPS61 whose argument is that fiscal policy would have gotten the economy out of the recession faster under flexible rates than it would have under fixed rates.50 This is quite contrary to what Fleming [1962, 372] and Rhomberg [1964] were saying at the time. By that point Mundell’s contribution to the ineffectiveness debate concerning goods market interventions was limited to a discussion of commercial policy’s impotence.51 It was precisely because of the Canadian recession, starting in 1957, that the discussion in CJEPS61 had been currently topical. Of course, one of the key conclusions of CMP concerns the inefficacy of fiscal policy under flexible exchange rates. In the later paper Mundell recants his previous position on this matter, and embraces the Fleming and Rhomberg view, but without acknowledging their position nor that they had held this view long before he switched over to their side.
Perhaps the most remarkable aspect of the analysis of the comparative statics above is that it equates the movements of real and nominal exchange rates. It is true that these examples are ones where nominal exchange rates moved, in some cases drastically. But is it obvious that real exchange rates moved as well? And, whether they did or not, does the model in MD help us to understand this effect? The answer appears to be that it does not, since nominal exchange rate values seem to be indeterminate in the model. For example, the model is designed so that the same value of the real exchange rate is attained even if the nominal value of the exchange rate is kept constant (as under a fixed exchange rate regime).) Thus, the MD framework would suggest that Canada’s experience in the late 1950s and early 1960s would have occurred even if the authorities had been pursuing a fixed exchange rate regime. Such a view is inconsistent with the entire literature written on this subject.52 It is real exchange rate values alone whose movements can be analyzed in the MD framework. The second purpose of the adaptations of MD was to firm up the equivalence of the comparative statics of the model under alternative exchange rate regimes. The key point of the paper was to examine Friedman’s assertion that “If internal prices were as flexible as exchange rates, it would make little economic difference whether adjustments were brought about by changes in exchange rates or by equivalent changes in internal prices.” The conclusion of Mundell’s paper follows this quotation immediately: “…while this view, under certain circumstances, may be valid in statics, it is entirely erroneous in dynamics.” (Mundell [1960, 227-28]) It was noted in the original version of the paper that there arise “…monetary differences between the two systems based on consideration of fixed debts.” (Mundell [1960, 228]) The adaptation undertakes to eliminate even this distinction between the systems.53 The way this point is presented in the International Economics adaptation is by noting “…that equal percentage changes in the price level and exchange rate have different monetary effects inasmuch as a price-level increase reduces, while an exchange-rate appreciation increases, the real value of cash balances.” In order to eliminate this discrepancy, the adaptation offers a simple solution: “An exact identification between the two is possible, however, if an exchange-rate appreciation is accompanied by an equal percentage reduction in nominal cash balances, or, alternatively, if an increase in the price level is accompanied by an equal percentage increase in nominal cash balances.”54 (Mundell [1968h, 157]) While attempting to make more explicit what the central bank is accomplishing, this sentence in fact has just the opposite effect. However, while the tools of the central bank are not described in specific terms, its actions are. The basic building block of the MD paper is: “The rate of interest is assumed to be determined by the monetary policy of the central bank—which means that the latter must always supply funds to the public (through, say, open market operations) to make any given interest rate compatible with equilibrium in the capital market…” [1960, 229] Whether interest rates can be pegged at the same time that the central bank is undertaking the “exact identification” operations noted above is an open question. But there is a further task which the monetary authorities must deal with: “…if the monetary authorities stabilize the exchange rate they must be prepared to buy and sell foreign exchange reserves at a fixed price, and if they stabilize the price level they must buy and sell goods and services at a fixed price…” [1960, 233] It seems to us that in adding on the burden of the “exact identification” operations, the central bank is overwhelmed with tasks. Do the authorities have a sufficient number of tools to carry out all the chores that they are being asked to cover?55 The third major change in the International Economics adaptation concerns the “Principle of Effective Market Classification.” This principle is mentioned in only three of Mundell’s original publications and with somewhat varying interpretations. In MD and IMFSP the setting is explicitly dynamic. In BNL the setting is entirely static, but the adaptation of that paper which appears as Chapter 14 in International Economics (Mundell [1968d]) is so substantial that it is virtually unrecognizable. In the later added sections of that adaptation, which we enumerate below, the Principle is reiterated in dynamic terms. Interestingly, in the original MD article the expression “effective market classification” does not appear until the very last paragraph of the body of the paper [1960, 250]. Indeed, the general observations upon which that principle is based are not introduced until the concluding section of the paper. This seems to treat the Principle as an afterthought, rather than being a key idea which could tie together a number of contributions. We have been unable to find a succinct statement of the Principle in the corpus of Mundell’s research.56 In order to make this Principle a more substantial contribution, the International Economics adaptation of MD includes a discussion of it in the very center of the paper [1968c, 163], in the section already completely modified to eliminate the perfect capital mobility defects which CMP had noted. Although that discussion does not add anything new to the concept, indeed, in many ways is merely repeating what appears in the unchanged conclusion of the paper, its earlier, more definitive declaration at this point in the paper reinforces the purported substantial contribution which some economists think this Principle makes.
One of the longest papers which Mundell has published in a professional journal appeared in September 1963 in the Banca Nazionale del Lavoro Quarterly Review. At twenty-four pages, this paper, entitled “On the Selection of a Program of Economic Policy with an Application to the Current Situation in the United States,” was essentially a contemporary of the CMP. Its model and purpose were quite different however and so it needed to be kept under wraps when the process of editing for International Economics was undertaken. Extensive adaptation of this paper would not create much attention as it had been published in an Italian journal that is not every day reading material for members of the profession. The paper starts by providing a general consideration of the choices which policy-makers face when they carry out their official duties. From this point a graphical device is used to exemplify these general observations. The model is further specified by applying particular values, that are said to be relevant to the US and European situations, in order to analyze how the US authorities might escape from the dilemma in which they found themselves at the time, in 1963, with a balance of payments deficit and high levels of unemployment. We can give a sense of the nature of the analysis, by citing some of its conclusions. It found that the US balance of payments deficit could have been reduced by $2 billion and output could have been increased by $30 billion by increasing government expenditure by $12 billion and raising (long-term) interest rates by 1 percentage point. (Mundell [1963d, 280]). There was one aspect of the paper which made it a very attractive candidate for inclusion in International Economics: it dealt with a specific application of the Principle of Effective Market Classification, and in a way which made it more plausible for dealing with the US situation. Namely, it included feedback effects from foreign countries in the markets for goods and services, so it did not picture the US as a small open economy. Nonetheless, on the financial side it continued with the assumption that the US could set its rate of interest independently of the exogenously fixed foreign rates of interest. The resulting adaptation was a complex affair (Mundell [1968d]), one that left even its author rather baffled. In personal correspondence, Mundell says that he wished that he had included this paper in the volume.57 Apparently he had forgotten that fully a third of the paper does in fact appear there, albeit in a disguised form. Almost every one of the other papers which are adapted for the volume has that fact noted in the first footnote in each chapter. But this chapter waits until the fourth page until it reveals the source of the material. Even so this footnote is misplaced in that it indicates that only the preceding material derives from BNL. In fact, the first eight pages of the chapter repeat BNL material very closely.58 But further, the title of both the chapter (“The Nature of Policy Choices”) [1968d, 200] and the reference to BNL are completely different from the original paper. While titling the chapter should be whatever the author wishes, the title of the original paper is a given. Mundell misstates it by referring to it as “The Nature of Policy Choice.” [1968d, 204] After this portion, which follows the original article closely, the rest of the chapter is quite different. The sections here include “Mathematical Aspects of the Theory of Policy,” “Anatomy of Policy Failure,” “Incomplete Information and Policy Dynamics,” and “Alternative Mathematical Formulations.” The key difference from the original is that in these added sections there is no identification of what the specific policy tools under analysis might be. In particular, the interest rate is left out of the discussion entirely (although there is a passing reference to the fact that a particular diagram may be thought of as a general case for which IMFSP (Chapter 16, in the adapted form in which it appears in the International Economics volume) would serve as an illustration). [1968d, 208] Leaving out discussion of the interest rate was essential to the process of adaptation, because by 1963 Mundell was arguing in CMP that the use of interest rates as a gauge of monetary policy was inappropriate, even unacceptable. That policy instead was to be measured by the size of domestic credit expansion, rather than by money supply changes or interest rate levels, along the lines of what we have noted at length above. To have included this portion of the BNL in its adapted version in International Economics would have been directly to contradict CMP. This contradiction can be exemplified in a stark fashion by a comparison of the original of BNL with CJEPS64 (which appears in an adapted form in International Economics as the appendix to Chapter 18), both of which are two-country models of the world economy. For BNL the interest rate in the US is an exogenous variable, set by the monetary authorities. The foreign interest rate too is given, independent of the values in the US, and there is no analysis of what these stances imply for money supplies. In contrast, CJEPS64 assumes that interest rates must have the same value in the two countries. Thus it includes the feedback effects of not only interactions in the goods markets, but also those in the financial markets. The focus here is on the endogenous determinations of: the level of world interest rates; and the distribution of international reserves. Clearly it would have been hard to reconcile the view which this paper takes with that which is presented in BNL. The elision of two-thirds of BNL and its replacement with general theorizing enabled the author to avoid a head-to-head comparison. In his Mundell-Fleming Lecture Mundell says that the original BNL “…article was, I think, the first fully-developed global empirical model of the world economy in a Keynesian framework, a precursor of the forecasting models used by professional forecasting companies like Otto Eckstein’s Data Resources and Laurence Klein’s WEFA.” [2001, 219] As we now show, this statement constitutes an example of very generous assessment of one’s own work. Let us look briefly at the data sources in it, and at a few of its limitations. Herewith is the only discussion of the way in which the parameter values in Mundell’s BNL paper have been “estimated.” There are no data sources that are identified explicitly. As Mundell [1963d, 284] put it: In developing the numerical values of one or more of the parameters I have benefited from (without necessarily accepting their results literally) published or unpublished work of T.C. Liu, D. Meiselman, W. White, F. de Leeuw, R. Rhomberg, von Hohenbalken and Tintner, D. Edwards, A. Okun, J. Vanek, P. Kenan [sic], and P. Bell, in addition to more well-known econometric models. However, it is as well to emphasize again the tentative and exploratory nature of these estimates and my hope is that they will serve to simulate further research on the subject. In addition to noting the weaknesses of the parameter estimates it is also necessary to point out possible refinements to the model itself—though the high price of increasing complexity must certainly be paid. It is sufficient to enumerate some of the more important limitations:
There follow eight more limitations of this sort. Of course, none of this discussion finds its way into the adaptation of this material which appears in International Economics.
We have seen how, for the important chapters in International Economics (Chapters 8, 11, 15-18), the most extensive changes were those that were undertaken of MD. Furthermore that article is the lynchpin upon which a priority claim could be staked, since incontrovertibly it was written before and independently of Fleming [1962]. But to make such a claim one needs to first establish that MD has grounds for being a part of the Mundell-Fleming canon. To see whether there are such grounds we start with first principles, but quickly apply the insights which we have uncovered in this paper. There is no disagreement about what constitutes Fleming’s contribution to the Mundell-Fleming model. Clearly it is the one paper published in IMF Staff Papers in 1962. Similarly, there seems to be little disagreement about what is the locus classicus of Mundell’s contribution to the model: CJEPS63 (which is a specific example from a whole set of versions of a single framework. These versions are sufficiently similar that we have identified them all here with a single designation, CMP). The interesting question is whether any other papers need to be included in the rubric, and if so, which ones are they. To this, Mundell has an answer, but one that we find unpersuasive: “…I supposed it [the expression ‘Mundell-Fleming model’] included all my papers on international macroeconomics…” (Mundell [2001, 219]) With further consideration even Mundell would acknowledge that this is probably a bit of an exaggeration. Barter Theory was certainly written before CMP was presented for the first time to an audience outside of the Fund. 59 Nonetheless, his recollections (Mundell [2001, 225]) make clear that this paper, being one which fell into the category of “return to the classics,” is not based on the Keynesian model, and therefore is not eligible for inclusion in the Mundell-Fleming rubric.60,61 Similarly, taking Mundell at his word, if CJEPS61 is “…the least relevant article…” (Mundell [2001, 223]) for the issues discussed in Fleming [1962], then there are ample grounds for not including it in this category. As we have argued in this paper, since this is one of the first of the checkerboard square papers, it certainly has lots of affinities with CMP. Therefore, there are two, not necessarily contradictory, reasons for excluding this paper. For Mundell the motivation is that he wishes to hide how different are its results, the last ones he reported before reading the Fleming paper, from the results laid out in CMP immediately after the Fleming paper was published. For us, the two papers are very repetitive. If the author feels that the later paper is the superior one, then we will accede to his view. In either case, these arguments boil down to reasons for not including CJEPS61 in the Mundell-Fleming model. As to IMFSP, we respect McCallum’s [1996, 145] citation of it as being part of the model. Nonetheless, this seems to us to confuse the issue, since it brings in a very different facet of the modeling exercise, and perhaps that is why McCallum made his citation. Namely this paper is written entirely in dynamic terms. There is no attempt in it to do comparative statics. In our view, the legacy of the Mundell-Fleming model resides in its textbook simplicity and the ineffectiveness results which are its distinctive feature. On this view, too, we justify our excluding it: IMFSP deals only with fixed exchange rates. Since the unique element of the Mundell-Fleming story is the comparison between the comparative static results under the two polar exchange rate regimes, this paper does not satisfy a number of different criteria. For us the differences are too great and we are unwilling to include this paper in this category. A similar argument would automatically exclude both Kyklos and MD. Nonetheless it is worth going through the further differences which make them implausible candidates for inclusion in the Mundell-Fleming rubric. Kyklos should be excluded for the same reasons that IMFSP is. It deals just with fixed exchange rates, and virtually its entire focus is on dynamic adjustment mechanisms. There is little interest in the comparative static results, the analysis is focused on whether the economy will be permitted to attain equilibrium, or whether the sterilization policy of the central bank will thwart the adjustment process. This focus on dynamics is neatly captured by the title of the paper “The International Disequilibrium System.” Remarkably the conclusion of this paper is that the presence of capital mobility makes little difference to the adjustment process as we have known it since Hume. This is exactly the opposite of the claim in CMP. That leaves just MD to consider. It seems to us that the argument for excluding this paper from the Mundell-Fleming system is stronger than for any other contribution mentioned above. This conclusion is consistent with popular opinion on the matter. 62 The differences are staggering: MD uses a definition of monetary policy that by the time of International Economics the author “no longer liked or accepted.” (Mundell and Swoboda [1969, 262]). As a result, the author has indicated that comparisons with CMP are essentially ruled out. Furthermore, this rejected definition of monetary policy is at the very heart of MD. Whether the exchange rate is fixed or it is floating, the authorities in every case are gauging their monetary policy stance by the level and rate of change of the value of the domestic interest rate. Furthermore, given the structure of the model in MD it is hard to see how one could conduct an exercise analyzing an expansionary monetary policy. This is a substantial deficiency given the importance of such questions in the Mundell-Fleming canon. MD casts its entire analysis in dynamic terms. The comparative statics, which the recollections carry out retrospectively and are the heart of Mundell-Fleming, are rebuffed as being uninteresting by MD itself even in its adapted form. MD sees the nature of the exchange rate regime as being essentially irrelevant for the equilibrium to which the economy settles down. Mundell-Fleming asserts precisely the opposite, that one can not determine what the consequences of a policy initiative will be until he specifies the exchange rate regime which the authorities are pursuing. MD assumes sufficient price flexibility that the economy is always in a full employment situation. Mundell-Fleming does just the opposite, looking at how financial policies can have an influence on the level of output and employment in the economy. MD models the central bank as being quite sophisticated, able to follow a reaction-function type of rule in responding to imbalances in various markets. In strong contrast, Mundell-Fleming sees policy actions as being exogenous events. MD may have been “formidable” to Dornbusch, but we find it to be virtually impenetrable. We are not able to convince ourselves even that the rule for the fixed exchange rate regime in fact keeps the exchange rate constant, or the rule for the flexible exchange rate regime keeps the money supply (or the price level, in an inflation-targeting regime) constant. In contrast, Fleming’s model is straight-forward since he has provided us with all of his equations. And Mundell’s diagrams in CMP are easy to use as well. The tractability of these frameworks is therefore wildly differently as between MD and the Mundell-Fleming model. MD explicitly assumes that the degree of capital mobility is less than perfect. In contrast, CMP assumes from the very beginning that capital mobility is absolutely perfect. It sees as the key defect of MD precisely the fact that it can not handle this extreme case, which seems to be a stereotype towards which the world economy is moving. It is this defect in MD that inspired the change in definition concerning monetary policy, which led the author to develop the correct version of the checkerboard squares table. Obviously, there is no way that MD should be considered part of the Mundell-Fleming model.63
We have shown in this paper that the analysis of the Mundell-Fleming priority question boils down to a focus on the two key papers: Fleming (1962) and Mundell’s CMP.64 The question then becomes, Can the argument be sustained that Mundell’s CMP was developed prior to and “completely independent” of Fleming (1962)?65 The final answer to this question is perhaps obvious, but the details which we have turned up in arriving at this conclusion make a fascinating story, one which we intend to lay out in our next paper. References: Yüklə 155 Kb. Dostları ilə paylaş:
|