This article investigates how Milton Friedman's views on exchange rates in emerging nations evolved through time. Early on he supported flexible exchange rates for most less developed countries. He endorsed several forms of flexibility, including auctions and the “crawling peg.” Starting in the early 1970s, he favored a system characterized by fixed rates and no central banks. From that point on he argued that while flexible rates were the preferred option for advanced nations, they were the second-best solution for (most) developing countries. This article helps elucidate Friedman's views on the use of pegged exchange rates during stabilization programs (Chile and Israel) and on exchange rates in socialist countries (China and Yugoslavia).
In testimony from 1973 in front of the Joint Economic Committee in Congress, Milton Friedman (1973c: 114) said, “While I have long been in favor of a system of floating exchange rates for the major countries, I have never argued that that is necessarily also the best system for the developing countries” (emphasis added). What were, then, his views on the less developed countries (LDCs)? Did these views evolve as a result of discussions with policymakers during his travels to India, Israel, Brazil, Chile, and South Africa? Was he influenced by his Chicago colleagues T. W. Schultz and Arnold C. Harberger, who worked on economic development? Did he think that, in spite of poor institutions and shallow markets, some developing nations would benefit from floating rates? And if not, what regime did he favor? Was his thinking on the subject rigid, or did he take a flexible, case-by-case approach?
In this article I investigate the evolution of Milton Friedman's views on exchange rates and monetary regimes in poor and middle-income nations. I show that Friedman's perspective on poorer nations corresponds to what Stanley Fischer (2001) has called the “bipolar view” of exchange rate regimes or what others have called the “two corners solution”: countries should opt for either very hard pegs—currency boards or straight “dollarization”—or flexible exchange rates; they should avoid middle-ground arrangements, such as Bretton Woods–style adjustable pegs. Interestingly, in his influential paper, Fischer (2001) does not mention Friedman as an early proponent of the bipolar view. I also show that for Friedman, “flexible exchange rates” was a rather broad category and included arrangements different from cleanly floating exchange rates.1 In some cases he endorsed auctions, and in some instances he supported variations of the “crawling peg,” the system of mini devaluations implemented in a number of Latin American countries, including Brazil.2 He thought of these arrangements as transitional regimes, as systems to be implemented in countries that were about to move to one of his preferred “two corners.” I further maintain that Friedman's views evolved through time: during the 1950s and 1960s his preferred arrangement for most (but not all) developing countries was some form of flexibility. Starting in the early 1970s, however, he favored a “unified currency,” a system characterized by a (very) hard peg and no central bank, or “dollarization.” In other words, his ranking of preferred regimes changed from “flexibility first” to “unified currency first.” Importantly, and contrary to what his critics have said, Friedman was not dogmatic, and he accepted exceptions to these rankings. This pragmatism was reflected in his advice to Chinese leaders in 1988. What never changed, however, was his critical view of the fixed-but-adjustable Bretton Woods regime.
From a history of thought perspective, there are several reasons why it is important to understand Friedman's policy position on exchange rates and monetary policy in developing countries:
Although thousands of pages have been written about Friedman's views on monetary rules and currencies, there has been very little work on his perspective on the emerging economies. Hanke 2008 is the only work that directly addresses Friedman's views on exchange rates in the less developed world and points out that Friedman followed a pragmatic and flexible approach on these matters.3
A thorough study of Friedman's views on currency regimes will help us understand why, according to the International Monetary Fund, in early 2022 almost every advanced country had flexible rates, but only two developing nations had a “free floating” regime.4 Indeed, most emerging countries still have some variation of pegged-but-adjustable exchange rates that Friedman criticized time and again. What explains, then, this asymmetry in terms of his influence on policy? Why have advanced nations adopted a Friedmanite view, while emerging markets have insisted on having the system that Friedman criticized repeatedly and severely?
An analysis of Friedman's writings (including his correspondence) will help elucidate his views on the use of pegged exchange rates during stabilization efforts. This controversial issue involves a comparison between Chile and Israel, two countries that Friedman visited and wrote about.5 Chile pegged its currency to the US dollar in 1979, while Israel did it in 1985. In both cases, the goal was to reduce inflation to mid-single digits. However, while Israel succeeded, Chile failed. In 1994, Friedman published an article analyzing why, in his view, the outcomes were so different in these two nations. He also exchanged letters with a number of associates trying to explain Chile's currency collapse of 1982.
This article's contribution, then, is to extend the literature on Friedman's thinking on monetary and currency regimes in several directions. First, it expands the work by Dellas and Tavlas (2018), Lothian and Tavlas (2018), and Nelson (2020a)—just to mention some important recent works on Friedman's monetary and exchange rate views—to the case of developing countries.6 Second, this article furthers Hanke's (2008) analysis on Friedman's views regarding exchange rate regimes in emerging countries, including the important question of whether fixed exchange rates provide a constraint on monetary policy in countries that have a central bank. Some of the issues tackled in this article, and not addressed by Hanke (2008), include Friedman's views on forex auctions and crawling pegs, along with his policy recommendations in countries as different as India, South Africa, Yugoslavia, Chile, Brazil, and China. And third, in this article I expand the work by Edwards and Montes (2020) and dig deeper into Friedman's views on the reliance on exchange rate anchors. More specifically, I discuss in detail Friedman's reaction to the collapse of Chile's experiment with pegged exchange rates in 1982, something that was not addressed in Edwards and Montes 2020 or Hanke 2008.7
Before proceeding, a word on sources. Friedman's output is immense and includes lectures, speeches, roundtables, symposia, debates, and interviews. The Milton Friedman Archives at the Hoover Institution (MFAHI) contains 218 works classified under the “exchange rates” tag; 46 of those are on “flexible exchange rates.” But there are many works cataloged under different tags that deal with currencies and monetary regimes in different countries, and there are many additional works and letters that have not yet made their way to the archive. One such item is a pamphlet in Spanish with Friedman's lectures in Chile in 1975, the year he met with General Pinochet and argued that a “shock treatment” was required to end inflation (Friedman 1975). Although I have made an effort to be as comprehensive as possible, it is likely that I have missed some of Friedman's works on the exchange rates and monetary regimes in developing countries. I believe, however, that I have analyzed the most important pieces—including key correspondence—and that I have been able to reconstruct Friedman's thinking on the subject.
Friedman's Early Views and the Influence of Lloyd W. Mints and Henry C. Simons
In his influential 1948 article “A Monetary and Fiscal Framework for Economic Stability,” Friedman addressed, for the first time, the exchange rate regimes question in the context of a monetary rule. He wrote that
under the [monetary] proposal, the aggregate quantity of money is automatically determined by the requirements of domestic stability. It follows that changes in the quantity of money cannot be used—as they are in the fully operative gold standard—to achieve equilibrium in international trade. . . . The international arrangement that seems the logical counterpart of the proposed [monetary] framework is flexible exchange rates, freely determined in the foreign exchange markets, preferably entirely by private dealings. (Friedman 1948b: 252; emphasis added)
In a footnote, Friedman pointed out that it was possible to present his argument in a different way: “Flexible exchange rates can be defended directly. Indeed, it would be equally appropriate to present the proposed domestic [monetary and fiscal] framework as a means of implementing flexible exchange rates.” He would make this point many times during the years to come. There was a strict connection between monetary policy and the exchange rate regime; it was not possible for a country to simultaneously have free capital mobility, active monetary policy (including a monetary rule), and rigid exchange rates. This proposition would come to be known as “the impossibility of the Holy Trinity.”
In April 1948, Friedman participated in a roundtable on Canada's payments problems. During the discussion, transmitted on radio by NBC, he argued that Canada would benefit from a market-determined exchange rate. In his memoirs, Friedman points out that until that time Canadian officials had not thought of adopting floating exchange rates, as they eventually did—partially prompted by Friedman's suggestions—in September 1950. In the same memoirs, Friedman recalls that in 1950, when he was working on the Schuman Plan, he presented a memorandum to the German authorities, suggesting that exchange controls “would be rendered utterly unnecessary if the simple step were taken of letting the exchange rate go free” (Friedman and Friedman 1998: 189; emphasis added; see also Friedman 1950).8
“The Case for Flexible Exchange Rates” was written in 1950 as a memorandum for the US Economic Cooperation Administration and subsequently published, in 1953, in Essays in Positive Economics.9 Friedman begins with a simple argument: countries face recurrent shocks that alter their balance of payments position. There are four ways of solving the imbalances created by these shocks: (a) changes in exchange rates; (b) changes in internal prices and income; (c) controls, including tariffs and quotas; and (d) use of monetary reserves. Friedman argues that while a flexible exchange rate regime would result in smooth and gradual adjustment, infrequent large devaluations would be highly destabilizing. He writes, “The system of occasional changes in temporarily rigid exchange rates seems to me the worst of two worlds: it provides neither the stability of expectations that the genuinely rigid and stable exchange rate could provide in a world of unrestricted trade . . . nor the continuous sensitivity of a flexible exchange rate” (Friedman 1953: 164).
Friedman then addresses changes in internal prices as an adjustment mechanism. He writes that since wages and prices are rigid downward, “an incipient deficit that is countered by a policy of permitting or forcing prices to decline is likely to produce unemployment rather than, or in addition to, wage decreases” (Friedman 1953: 165). Regarding the last two options, he points out that direct controls in the form of import tariffs, licenses, and quotas have significant efficiency costs and that the use of monetary reserves would be ineffective if the shocks are large and persistent. In the rest of the essay, Friedman addresses a number of objections commonly raised against flexible exchange rates, and he delineates practical ways of implementing a new international financial system characterized by market-based exchange rates.
The closest Friedman got to addressing the case of LDCs in the 1953 essay is in a section titled “The Sterling Area,” in which he deals with currency unions and trade. Although at the time the sterling area included a number of poor countries in Africa and Asia, Friedman's analysis is oriented toward Great Britain. For instance, there is no mention that the poorer members of the area are commodity exporters and that their terms of trade tend to be highly volatile.10 Friedman notes that a possible arrangement is a “mixed system of fixed exchange rates within the sterling area and freely flexible rates within sterling and other countries” (Friedman 1953: 193). He adds that there should not be trade restrictions within the currency union. He also notes that in an ideal monetary union, both monetary and fiscal policies would be subject to a single political authority.
Many of the arguments presented by Friedman in his 1950 memorandum had been made a few years earlier by his Chicago teacher, and later colleague, Lloyd W. Mints.11 For example, in a review of a 1944 League of Nations report, which was mostly written by Ragnar Nurkse (1944), a strong supporter of pegged exchange rates, Mints (1945: 193–94) wrote,
It is doubtful that fluctuating exchanges, under conditions of internal monetary stability . . . would be disequilibrating. . . . And it is beside the point to contend that exchange fluctuations “involve constant shifts of labor and other resources between production for the home market and production for export.” . . . [Under fixed rates] the adjustment must come by way of a change in domestic prices, including wage rates, whereas with free exchanges the necessary adjustments can be obtained largely by means of changes in the prices of international goods. The important consideration is that the latter prices are more flexible than wage rates.
In his book Monetary Policy for a Competitive Society (1950), Mints included two chapters on exchange rates, in which he argued that flexible rates could provide a solution to the international adjustment problem. In chapter 5, titled “Fixed vs Flexible Exchange Rates,” Mints criticizes purchasing power parity—he calls the doctrine “a fallacy” (Mints 1950: 97)—and the nascent Bretton Woods system. In the preface to the book, Mints writes, “I am greatly indebted to Professor Milton Friedman, who read the penultimate draft of the manuscript. In consequence of his many suggestions several chapters have been rewritten” (vii).
It is well known that Henry Simons influenced Friedman on monetary policy in general and on monetary rules in particular (see Friedman 1967). What is less known, however, is that Simons also had an impact on Friedman's views on exchange rates. Simons was more guarded than his colleague Lloyd Mints in supporting flexible rates, but in several of his writings he criticized the gold standard and argued in favor of what he called “independent national currencies.” In a little-known 1934 essay titled “Currency Systems and Commercial Policy,” Simons (1934: 346) wrote that the gold standard was “conducive in depressions to policies of extreme protectionism and economic isolation.” In his view, this problem would not exist if there were “a system of independent currencies . . . [that] would involve . . . change[s] in the exchange rates.” He pointed out that if monetary policy were geared at stabilizing the price level, flexible exchange rates would not be volatile and that “adequate future markets for foreign exchange would surely develop” (347).
In a 1943 paper on the future of the global economy after World War II, Simons argued that free trade was the most important requirement for a durable peace. He then pointed out that the adoption of flexible rates by the main countries would facilitate a world based on free trade. He wrote, “My own predilection is for essentially independent currencies (or currency blocks), each stabilized in terms of an inclusive domestic price index, and all traded freely (i.e., without intervention by central bank treasuries, or stabilization funds) on well-organized exchange markets (forward and spot). It is hard, however, to point an easy or promising course toward such a monetary world” (Simons 1943: 149; emphasis added).
Neither of the papers referenced above was included in the volume edited posthumously by Aaron Director. However, in several of the essays that were included in the book Simons addresses exchange rates, and in all of them he points out that his preferred regime is independent national currencies that float with respect to each other (Simons 1948: 63, 81, 83, 85).
The analysis in “The Case for Flexible Exchange Rates” clarifies the reasons why Friedman favored flexible exchange rates: they permitted the implementation of a monetary rule, they facilitated free trade, and they allowed getting around (nominal) price and wage rigidities. In the years that followed, Friedman reiterated and refined his arguments on the superiority of flexible exchange rates for advanced countries, including his view that speculation in currency markets was stabilizing, a point also made by Henry Simons (1934).12
India: The 1955 Memorandum and the 1963 Visit
In 1955, two years after he published “The Case for Flexible Exchange Rates,” Milton Friedman traveled to India to advise the Nehru government. The trip was part of the US program for technical assistance to poor nations. His companion in this endeavor was Neil Jacoby, a professor at UCLA. At the time, the Indian government was preparing its Second Five-Year Plan under the leadership of Professor P. C. Mahalabonis. In the mid-1950s, one of the most hotly debated issues surrounding India's development program was how to allocate foreign exchange. The view among Indian economists was that foreign exchange had to be rationed and allocated to those “strategic” industries that contributed to import substitution and industrialization. The view that foreign exchange scarcity was a constraint to growth was formalized a few years later by Chenery and Strout (1966) in their “two gaps model.” (As I note below, Friedman believed that there was only one gap, and that that was the “fiscal resources gap.”)
In early November 1955, Friedman presented to India's finance minister C. D. Deshmukh an eleven-page memorandum titled “Some Initial Comments on Current Problems of Economic Development in India.” The memo was never released officially by the Indian government, and it found its way to the public domain only some forty years later, when it was presented at a conference on development economics in Hawaii (Friedman  1992). Section 6 of the memorandum deals with the “foreign exchange problem” and is extremely critical of India's system of exchange controls. Friedman made a number of simple and, for that time, highly controversial policy recommendations that clashed with the dominant views of the 1950s and 1960s; see below for a contrast between Friedman's and Gunnar Myrdal's views on India. A summary of Friedman's ( 1992) main arguments is as follows (emphases added in the quotations):13
There are only two ways to deal with India's large and persistent external imbalances: “First, to inflate or deflate internally in response to a putative surplus or deficit in the balance of payments; second, to permit the exchange rate to fluctuate . . . [a method] that has been adopted by Canada with such conspicuous success.” In discussing the disinflation channel, Friedman referred to a point made in his 1953 essay, in which he wrote that “if internal prices were as flexible as exchange rates, it would make little difference whether adjustments were brought about by changes in exchange rates or by equivalent changes in internal prices.” But he noted that domestic prices and wages tend to be rigid downward. Consequently, he pointed out that “an incipient deficit that is countered by a policy of permitting or forcing prices to decline is likely to produce unemployment” (Friedman 1953: 165).14
An alternative to straight floating, and one that India could prefer, was to “auction off whatever amount of foreign exchange it is decided to be released, permitting the purchasers to use it for anything they wish and in any currency area they wish.” He noted that an auction system had the advantage of generating some additional government revenue. On the negative side, foreign exchange auctions would only eliminate “distortion in the pattern of imports . . . [but would] not produce the appropriate adjustment of exports to imports.”
Under a floating system, the rupee could eventually move toward convertibility. “In a world of inconvertible currencies, a country that offers convertibility, albeit at a fluctuating rate, has a special attraction for investors and traders.”
In addition to dealing with the “foreign exchange problem,” the 1955 India memorandum addressed other macroeconomic issues, including inflationary pressures, investment, and public sector imbalances. Friedman argued that at the end of the road there was a close link between the overall deficit and forex “scarcity.” He pointed out that “there are not two problems, a total resource gap and a foreign exchange gap, but only one, a total resource gap.”
The authorities had good reasons to keep Friedman's report out of the public eye: they had no intention of following his advice. The Second Five-Year Plan continued to rely on exchange and imports controls and a pegged exchange rate.15 In 1955, the exchange rate was 4.77 rupees per dollar, a price that remained in place until 1966, several years after Friedman's second visit, in 1963. After the 1966 devaluation, the official value of the dollar became 7.55 rupees per US dollar.
Milton Friedman returned to India in 1963. This visit was part of a yearlong world tournée that included a grueling schedule of lectures, presentations, and meetings. During the trip, Friedman spent time in countries whose economic systems he wanted to analyze in some detail: Israel, Yugoslavia, Turkey, Greece, India, Hong Kong, and Japan. Many of them faced serious external imbalances and devaluation pressures. In his talks and presentations, Friedman was repeatedly critical of the Bretton Woods regime and presented the merits of flexible rates.16
In 1963, Friedman spent a little over two months in India, where he gave a number of lectures on government deficits, inflation, growth, exchange controls, and the balance of payments. The most important event was a two-part lecture delivered in Mumbai (Bombay). In contrast to his 1955 visit, this time there was no secrecy, and the lectures were promptly published by the Council for Economic Education. A few years later, Friedman included them as chapter 1 of his 1968 volume, Dollars and Deficits (see Friedman  1968).
The second of the Mumbai lectures dealt, almost exclusively, with problems related to the balance of payments, exchange controls, and the exchange rate system. Friedman made several points that he would repeat again and again, and in country after country. Once more, he strongly criticized the Bretton Woods system of pegged-but-adjustable exchange rates. He pointed out that in an inflationary context, it was impossible for a pegged exchange rate to be continuously near its long-run equilibrium. Further, given India's inflationary pressures, it was not possible to use stepwise devaluations to solve the external imbalance. He said, “The temptation will be to change its [the rupee's] value from its present level. . . . And then try to hold it at the new fixed level. That would be another mistake. . . . A far better solution, as already noted, would be to allow exchange rates to go free and find their own levels whatever they may be” (Friedman  1968: 58–59; emphasis added). In this talk, Friedman discussed, again, the possibility of implementing an auction system. Since his 1955 memo had been withheld by the authorities, for most of the Mumbai audience this was a new and audacious proposal.
The published version of the 1963 lectures includes the Q and A sessions. Not surprisingly, most of the questions were centered on his strong criticism of India's exchange rate regime and his suggestion of repealing controls and adopting an auction system. When asked if devaluing the rupee was the only way of dealing with the growing external imbalances, Friedman went back to the costs of disinflation in a pegged exchange rate regime, a point that, as noted, he had emphasized in the 1953 essay. In Mumbai, he said, “In principle there is another way. You could drive down prices in India by 30 or 40%. Deflation would be an alternative. What matters is the relation between prices in India and prices in the rest of the world, account being taken of the exchange rate. . . . As a practical matter, I think that no government will in fact try to force a deflation of 30 or 40%” (Friedman  1968: 64–65; emphasis added).
At the end of this second visit to India, Friedman was deeply disappointed. In his memoirs, he wrote that “the intellectual climate of opinion about economic policy was almost wholly adverse to any changes in the direction that seemed to me to be required. There was a deadening uniformity of opinion. . . . I encountered again and again the same stereotyped responses. . . . It was as if they were repeating a catechism” (Friedman and Friedman 1998: 315).
In 1968, five years after Friedman's second visit to India, Gunnar Myrdal published his three-volume oeuvre Asian Drama: An Inquiry into the Poverty of Nations (Myrdal 1968). The book has more than two thousand pages and covers many aspects of the development challenges faced by a group of Asian countries. The first two volumes include detailed discussions on agriculture, industrialization, public sector financing, international trade, foreign aid, and inflation, among others. Volume 3 deals with demographic issues and contains sixteen technical appendixes. Milton Friedman and his views on India are not mentioned in any of the volumes. Although Friedman's 1955 memo had not been released to the public, my conjecture is that Myrdal and his team had access to it and decided to ignore it. But, this is just a conjecture that has not been confirmed. What is indisputable, however, is that Myrdal had access to Friedman's 1963 lectures, published in India the same year he delivered them as “Inflation: Causes and Consequences.” The short pamphlet circulated widely, and because of Friedman's controversial policy advice, including his arguments in favor of flexible exchange rates and forex auctions, it was profusely discussed in the press and policy circles.17 Of course, not citing an author who espouses different views is common practice among academics. In that regard, the absence of references to Friedman and his recommendations for India is not surprising. What is surprising, however, is the almost complete absence of an exchange rate–related discussion in the main text of Asian Drama. The terms devaluation or exchange rates do not appear in volume 1 (chaps. 1–14), which includes a forty-eight-page chapter on India (chap. 7) and a lengthy chapter titled “Foreign Trade and Capital Flows” (chap. 13). Volume 2 (chaps. 15–28) has one reference to “devaluation,” and that is in a footnote (Myrdal 1968, 2:919n2). Most of the discussion on exchange rates is relegated to appendix 8, “A Note on Positive Operational Controls.” The first section of this appendix is titled “The Foreign Exchange Front,” and it opens with a discussion of “reasons for or against devaluation” (Myrdal 1968, 2:2078). Throughout his analysis—which is less technical than other appendixes and does not include any equations—Myrdal is extremely skeptical with respect to the role of exchange rate adjustments in LDCs. This was not new for him. In An International Economy: Problems and Prospects, he wrote that the historical evidence demonstrated that “devaluation of the currency . . . is neither a wholesome nor an efficient means of curing a structural disequilibrium in trade and payments” (Myrdal 1956: 94). Myrdal's views were consistent with the then-popular import substitution strategy and “elasticities pessimism” perspective. What he writes on exchange rates in appendix 8 of Asian Drama is the opposite of Friedman's views on the subject:
Since devaluation cannot—either in the short or the long run—be expected to stimulate a very considerable increase in export volume, it cannot free countries like India and Pakistan from the necessity of preventing or severely limiting imports other than those of essential consumption goods and development goods. Devaluation is not an alternative to import controls. . . . It should be frankly recognized that the concept [devaluation] is not applicable to these countries. (Myrdal 1968, 2:2081; emphasis in the original)
A comparison of Friedman's and Myrdal's views and advice for India provides one of the best illustrations of how divergent views on economic development were in the 1950s and 1960s. On the one hand, there was the “markets approach,” represented by Friedman, his colleagues Ted Schultz and Al Harberger, and Peter Bauer, among others; and on the other hand, there was the “planning approach,” represented by scholars such as Myrdal, A. W. Lewis, and Albert Hirschman. At the time, the planning approach was clearly dominant among development economists; see Edwards 2015.18
“Unified Currency” Regimes
After his 1963 visit to India, and in spite of his increasingly busy schedule as an adviser to politicians in the United States, Friedman continued to travel and lecture in both advanced and emerging markets. During the 1960s and early 1970s he visited, among other places, Japan, Korea, Iran, Spain, Yugoslavia, Brazil, Australia, New Zealand, Zimbabwe, and South Africa. In every country he gave talks and interviews and met with policymakers, academics, and representatives of the private sector. On almost every occasion, the discussion would turn to exchange rates, imports and capital controls, inflation, and macroeconomic stability.
In September 1965, during a symposium organized by the American Enterprise Institute, Friedman stated that in spite of his crusade for flexible rates, he thought that there were some exceptions to the flexibility rule: “I am not, and never have been, in favor of floating rates under all circumstances for every country, whatever may be its circumstances. On the contrary, I think it would be most unwise at the present moment, for example, for Hong Kong to give up its present fixed rates and that is true also for other countries” (Friedman 1966: 108–9). The use of the term fixed rates is important, as Friedman made a clear distinction between “fixed exchange rates,” which he sometimes called “unified currencies,” and “pegged exchange rates.”19 He underlined this difference in a paper he presented in 1965 at the Fifteenth General Meeting of the Mont Pèlerin Society in Stresa, Italy (Friedman 1968b). “Unified currencies” are irrevocably fixed, cannot be altered through administrative or political decisions, and include “dollarization”; they exist in jurisdictions where, generally, there are no central banks. “Pegged exchange rates,” in contrast, may be changed by the monetary or political authorities, including by the central bank; the Bretton Woods system is an example of pegged rates.20 To illustrate what he meant by “unified currencies,” Friedman gave the example of the United States: there is a unified currency between Illinois and New York; he also mentioned Hong Kong and Singapore as examples of unified currencies.
Regarding central banks and exchange rates, he also wrote that “a central bank or its equivalent is a necessary condition for balance of payments problems. In modern conditions, one can go further; with rare exceptions, it is both a necessary and sufficient condition for balance of payments problems” (Friedman 1968b: 270). He then referred to Malaysia, a country that had just created a central bank and transitioned from a unified currency to a pegged regime. He predicted that “within a few years . . . [Malaysia] will join the other ‘developing nations’—to use the most recent euphemism—by experiencing an exchange crisis” (270).21
Israel and the Horowitz Lectures
In 1972, Friedman delivered the Horowitz Lectures in Israel, where he came back to the connection between central banks and exchange rate crises in LDCs. He argued that given the emerging countries' proclivity to use money creation and the inflation tax to finance unproductive government projects, a system with a hard peg and no central bank—what he called a unified currency system—was preferred to one where there was a central bank and a currency whose value fluctuated according to market forces. This represented a departure from what he had said in India in 1955 and 1963. During the rest of his life, Friedman maintained the view that a unified currency regime was the best option for most emerging countries. The word most in the preceding sentence is important, as Friedman was extremely pragmatic and flexible and recognized that often politics stood in the way of first-best solutions. As will be seen below, this pragmatic view was reflected in his views regarding Yugoslavia and China. It was also reflected, although in a more controversial way, in his analysis of Chile.
In the first Horowitz Lecture, delivered in Jerusalem, Friedman surveyed the historical evidence on monetarism and focused on the connection between money creation and inflation in the United States and the United Kingdom.22 In the second lecture, in Tel Aviv, Friedman dealt with the developing countries, with special emphasis on Israel, a country that was then considered to be middle income. He noted that because they tended to rely on the inflation tax, poorer countries had, on average, significantly higher rates of inflation than advanced nations. The inflation tax, he pointed out, was “the only tax that can be levied without explicit legislative enactment or executive announcement” (Friedman 1973b: 48). Worse yet, the inflation tax was usually implemented in conjunction with an array of controls that “discourage private investment, often lead to a flight of private capital, and produce economic waste and inefficiency” (44). The situation described by Friedman is known in the literature as “fiscal dominance.”
By this time (1972), Friedman had already delivered his American Economic Association presidential address on the Phillips curve and the absence of a long-run inflation-unemployment trade-off. Robert Gordon (2011) has argued that the reason the “natural rate” idea was developed in Chicago, and not on the East Coast or in Europe, was because of Chicago economists' involvement with developing countries. For instance, Arnold Harberger, Friedman's longtime colleague and occasional travel companion, had been working on inflationary pressures in Latin America for many years, and he often discussed the most pressing issues with Friedman. Also, a number of students in the department of economics had written (or were writing) doctoral dissertations on rapid inflation and failed stabilization attempts; see, for example, Diz 1966 on Argentina and Luders 1968 on Chile.
Halfway through his second Horowitz Lecture, Friedman emphasized the point made at the 1965 Mont Pèlerin Society meeting in Italy: a unified currency was not the same as a pegged exchange rate. Under a unified currency system, money creation can happen only through an inflow of international reserves; in a pegged regime, in contrast, the monetary authority can “use the printing press or the central bank's bookkeeper's pen” (Friedman 1973b: 46).
An important question was whether a unified currency regime could be replicated in countries with central banks, as opposed to currency boards or a dollarized regime. The answer, said Friedman, was that, in principle, it was possible. However, the evidence suggested that it was politically difficult. A simple peg did not impose a true constraint on central banks.23 Good and able central bankers were either kicked out or “kicked up” when they questioned the reliance on the inflation tax. This, he pointed out, had been the case in Thailand in the late 1960s. Friedman's conclusion was simple and very controversial: “I conclude that the only way to refrain from using inflation as a method of taxation is to avoid having a central bank. . . . A unified currency assures a maximum degree of integration of the country in question with the greater world” (Friedman 1973b: 47; emphasis added). During the Q and A session, the moderator of the second Horowitz Lecture asked whether there was an optimal sequencing in the adoption of an appropriate monetary/exchange rate regime in a developing country such as Israel. He said that he understood that Friedman thought that “a unified currency . . . was a first step and a flexible rate a second step” (64; emphasis added).
Friedman's response was long, pragmatic, and all-encompassing. He argued that the answer had to consider the overall reality of each developing country, including political pressures and the institutional structure. When the big picture was taken into account, he affirmed, the unified currency regime was his preferred monetary arrangement for most poorer countries; it was the regime to adopt as soon as possible and to maintain over time. He said, “The reason why I regard a floating rate as a second best for such a [developing] country is because it leaves a much larger scope for government intervention” (Friedman 1973b: 64–66; emphasis added).
A year after delivering the Horowitz Lectures, Friedman confirmed, in congressional testimony, that in his view the best strategy for LDCs was to adopt one of two monetary and exchange rate regimes. As quoted at the outset of this article, Friedman stated in his testimony that he favored floating exchange rates for developed countries but was open to other kinds of rates for developing countries. “Indeed,” he then added, “in April of last year I gave a series of lectures in Israel [where] . . . I recommended as probably the optimum policy under current conditions for a developing country that it [irrevocably] peg its exchange rate to its major trading partner rather than have a floating system” (Friedman 1973c; emphasis added).
Hong Kong, Unified Currencies, Currency Boards, and Central Banks
Friedman admired Hong Kong's economic and monetary system. During his 1963 visit, he was interested in learning how the currency board system worked exactly. He consulted bankers, business people, and economists. To his surprise, “none of them had any clear idea of how the system worked, but by the time I finished checking with them, I thought I did” (Friedman and Friedman 1998: 318). In 1983, as a result of uncertainty about its political future, there were major capital outflows that threatened the stability of the Hong Kong dollar.24 In his memoirs, Friedman tells of his role in helping the banker John Greenwood and Hong Kong authorities design the system that linked (irrevocably) the Hong Kong dollar to the US dollar (rather than to the pound sterling, as had been the case until then). He wrote,
In 1983, when there was an exchange crisis in Hong Kong, John [Greenwood] was the architect of the monetary reform that led to the Hong Kong dollar being unified with the U.S. dollar. . . . During the course of the detailed negotiations that led to the final reform, John was on the phone almost nightly conferring with Alan Walters, then in London as an adviser to Margaret Thatcher, and with me in San Francisco. (Friedman and Friedman 1998: 326)
Hanke (2008, 2021) makes the connection between Friedman's views, Hong Kong's successful currency board, and Argentina's failed attempt at pegging the peso to the dollar in 1991. As Hanke (2008: 279) recounts, Friedman was skeptical about the Argentine experiment because, in contrast with Hong Kong, Argentina continued to have a central bank. In a 1994 paper titled “Do We Need Central Banks?,” Friedman (1994b: 46) wrote, “If a country has a central bank, then pegging its currency to another country's currency is going to cause trouble. Because a central bank is going to insist on doing things and once it does things, it would sooner or later establish a situation in which the pegged exchange rate is not held.” As I show below, Friedman also blamed the central bank for the failure of Chile's experiment with a fixed exchange rate stabilization program in the late 1970s and early 1980s.
During the early 1990s, nations that for decades had been in the Soviet orbit began to reform their economies. One of the important areas of reform was the adoption of a new monetary and exchange rate regime. Hanke, Jonung, and Schuler (1992) developed a plan for a currency board in Estonia, and Schuler, Selgin, and Sinkey (1991) worked on a similar program for Lithuania. Although Friedman was not involved in those projects and did not travel to the Baltic nations to consult with the authorities, he was in contact with the authors of the plans, and he supported them. Regarding the Hanke et al. (1992) plan, Friedman explicitly wrote, “A currency board such as that proposed by Hanke, Jonung and Schuler is an excellent system for a country in Estonia's position.”25 Also, Friedman supported Estonia's currency board in an interview in the Swedish newspaper Dagens Industri, on December 17, 1991. Interestingly, neither Hanke et al. (1992) nor Schuler et al. (1991) cite Friedman in the bibliography of their work.
Flexibility in a Second-Best World: The Crawling Peg
In his December 1968 radio program, Friedman talked about the “crawling peg,” a monetary regime then in vogue in several Latin American countries, including Brazil, Chile, and Colombia (Friedman 1968a). Friedman began by explaining that under a number of conditions it is not possible to maintain a pegged exchange rate indefinitely, as the Bretton Woods regime mandated. Instead of setting the exchange rate and attempting to defend it, under a crawling peg regime the authorities change the parity by a small amount very often, sometimes daily. Because of this feature, the system is also known as a “mini devaluations” regime. Friedman said, “Personally, I think this [the crawling peg] is a far better system than the present fixed rate” (emphasis added). This endorsement of the mini devaluations regime was consistent with Friedman's enthusiastic views on indexation in general, a perspective that he developed during the 1970s as the rate of inflation in the United States began to move gradually up (Nelson 2018; Friedman 1974b).26
In the radio presentation Friedman also discussed possible mechanisms for defining the rate of the crawl. He gave a specific example, in which the parity would slide quarterly and be allowed to fluctuate by 1 percent above and below a predetermined central value. Friedman also addressed the possibility of monitoring the evolution of the stock of international reserves to determine the rate of adjustment of the parity. He explained that even though market participants would know ahead of time that the exchange rate would change gradually and continuously, such a system would not necessarily generate speculative flows (Friedman 1968a).
In 1972, Friedman came back to the crawling peg issue during the Horowitz Lectures. In the Q and A period, a participant asked “whether a ‘trotting’ peg as in Brazil would not be a way to limit the harm done by . . . government” (Friedman 1973b: 61). He answered as follows: “The Brazilian system seems to me better than no attempt to change exchange rates but less good than an exchange rate that changes more rapidly. I have never been in Brazil but I have been in Korea for one day, which makes me an expert, and Korea also had a system very similar to Brazil's. They justified it on the grounds that their commercial and financial markets were so poorly developed” (61–62). The quip about Korea reflects Friedman's sense of humor. However, there were some important differences between the Korean and Brazilian systems. While in Korea the main objective of the policy was to maintain the degree of international competitiveness of exports by making sure that the won was somewhat undervalued, in Brazil it was part of a broader set of measures aimed at adapting all economic activities to a high level (three digits) of inflation.27
In 1973, Friedman visited Brazil for ten days; see Boianovsky 2020 for details. Friedman had long been interested in the Brazilian economy and had mentioned its history with inflation in his American Economic Association presidential address (Friedman 1968c).28 He argued that Brazil's experience showed that a stabilization program could result in short-term unemployment. The reason, Friedman pointed out, was the divergence between actual and expected inflation, and not the existence of a stable Phillips curve (Friedman 1968c: 8–9).29
On his return from Brazil, Friedman penned a number of pieces on inflation and indexation. In Friedman 1974a, he wrote, “The thing that fascinated me about the Brazilian experience was how closely what they did paralleled in detail what had been recommended in 1886 by Alfred Marshall—although I'm sure the Brazilian people had never read Alfred Marshall's paper.”
Friedman also supported a crawling peg approach to exchange rates in his 1975 visit to Chile.30 After a talk on stabilization and the inflation tax, he was asked whether the system of mini devaluation followed by Chile at the time fed inflation. He answered as follows: “Mini devaluations . . . don't result in higher real costs [or inflation]. They are simply a response to price increases. . . . If prices in Chile increase by 10% each month, then it is necessary to devalue by 10% in order to maintain a stable real value of foreign currency” (Friedman 1975: 39).
A close reading of Friedman's writings on the crawling peg and indexation—including the works in which he addresses Brazil and Chile—indicates, clearly, that for him a system of mini devaluations was a transitionary regime that would work in countries with rapid rates of inflation. Friedman did not see the crawling peg as a permanent exchange rate and monetary arrangement, nor did he recommend it in countries with low inflation, such as India or China; see the discussion below for details.
Exchange Rates in Socialist Countries: Yugoslavia and China
Friedman, of course, realized that many countries faced serious political constraints in selecting their exchange rate and monetary regimes. In the real world, deciding between a flexible rate and a unified currency (or other arrangements) was not easy. This was particularly so in socialist countries embarking on reform processes aimed at expanding the role of the market and the private sector.31 Yugoslavia was one of such countries. In early 1973, almost two years after the collapse of the Bretton Woods system, Friedman visited Belgrade as a guest of the Institute of Investment Analysis. During a talk at the National Bank of Yugoslavia he discussed exchange rates in an emerging socialist country that did not belong to the Warsaw Pact.32 Friedman began in a familiar way by stating that Yugoslavia had two alternatives: flexible rates or an irrevocable fixed rate. He then described the “unified currency” regime and argued that if Yugoslavia decided to adopt it, the currency of reference should be the German mark.33 He affirmed that in countries with a high degree of political centralization it would be politically difficult to give up the national currency (Friedman 1973a: 15).
Friedman next discussed a “less satisfactory, but still a good idea”: “a freely floating exchange rate.” After providing details on how flexible rates functioned, he pointed out that, at the moment, this was not a realistic option for Yugoslavia. He said, “Yugoslavia has first to develop, as I hope it will, a market in foreign exchange, so that there exists a market evaluation of its domestic currency in terms of foreign currencies.” He ended up by discussing the prospect for exchange rate regimes around the world: “Whatever policies Yugoslavia and other developing countries choose to follow, I believe that they will benefit a great deal from the new [international monetary] arrangements. The new system of floating exchange rates will be consistent with a rapid growth in world trade and with the gradual elimination of tariffs and other barriers to international trade” (Friedman 1973a: 16).
Friedman visited China three times—in 1980, 1988, and 1993.34 By his own account, the most interesting of those visits was in 1988, when he met for two hours with Zhao Ziyang, the secretary general of the Communist Party, to discuss the challenges faced by China. Zhao was the main force behind the market-oriented reforms, and at the time he was under significant political attack from the more conservative wing of the Communist Party. Eight months after Friedman's visit, the Tiananmen Square protests took place, and Zhao, who refused to use force against the demonstrators, was dismissed from his post by the party's elders. Some analysts have argued that one of the causes for his dismissal was the meeting with Friedman; Zhao spent the rest of his life under house arrest.
A few days prior to their meeting, which took place on September 19, 1988, Friedman sent the secretary general a short memorandum that served as the basis for their conversation. After pointing out that he was not an expert on China, Friedman highlighted four areas in which he thought reforms should be deep and move fast. The first area of action was to “end exchange controls, establish a free market in foreign exchange, and permit the foreign exchange to be determined by the market.” The other three critical policy goals were to “end inflation,” “decontrol individual prices and wages as rapidly and as fully as possible,” and “replace centralized government controls of the economy and state enterprises by decentralization and private control” (Friedman 1990: 119–23).
Friedman stated that there were several universal laws in economics, and that one of them was that government controls—including controls in the currency market—did not work. He then mentioned his work in India thirty years earlier and noted that his advice to free the exchange rate was not followed. He added that “continued observation of India in the decades since indicates that much of the subsequent corruption and inefficiency in India, as well as the lack of any significant improvement in the living standards of the ordinary people, derives directly from the continued existence of exchange controls and multiple exchange rates” (Friedman 1990: 121–22).
During their meeting, Zhao produced his own list of three challenges that China needed to solve in the next few years. Eliminating exchange controls and freeing the exchange rate was not one of them. The secretary general's list included “price reform, curbing inflation, . . . [and] implementing a shareholding system for enterprises” (Friedman 1990: 132). After hearing this list, Friedman tried, more than once, to move the conversation back to exchange rates and flexibility, emphasizing that it was a policy that spilled over into many other areas and sectors. Friedman's arguments for eliminating controls in China were very similar to those he had given around the world to presidents, prime ministers, members of Congress, and representatives of the private sector. Interestingly, he did not mention unified currencies, even once. His references to Hong Kong were centered on the fact that there were no exchange controls. My conjecture is that Friedman stayed away from unified currencies for pragmatic political reasons. It was difficult to believe that the Chinese Communists would be willing to give up monetary policy completely and fix the value of the yuan to that of some foreign capitalistic power. As I discuss below, in 1981, when discussing the Chilean experiment with a pegged exchange rate, Friedman acknowledged that implementing a unified currency regime was significantly more difficult, from a political point of view, than what he had originally thought when delivering the Horowitz Lectures.
Toward the end of his meeting with Zhao, Friedman made an important recommendation regarding the sequencing of reforms that highlighted, once again, the importance that he attached to the exchange rate regime. He said, “If there are political obstacles and the two [ending inflation and freeing the exchange rate] cannot be tackled at the same time, then freeing the control over the exchange rate is more important” (Friedman 1990: 140).
Friedman and Exchange Rate–Based Stabilization Programs: Chile and Israel
In 1994, Friedman published an essay in which he analyzed the experiences in Chile and Israel with pegged exchange rates during stabilization episodes. He discussed why in Chile the policy ended up in failure and a major crisis in mid-June 1982, while in Israel it succeeded. Friedman begins by pointing out that there was an element of luck: immediately after Chile pegged the exchange rate with respect to the US dollar, in 1979, external conditions soured. The dollar strengthened in global markets, and the terms of trade turned against Chile. In contrast, when, in 1985, Israel pegged the value of the shekel, external shocks were favorable (a drop in the price of oil and a weaker dollar).35 An important policy difference was that Israel devalued the shekel by 20 percent before pegging it. By doing this, it built a “cushion” for real appreciation to take place during the transition to lower inflation. Chile, instead, pegged the exchange rate at a time (1979) when the peso was already overvalued (Edwards and Edwards  1991). Also, while Israel instituted income policies that included a temporary wage and price freeze, Chile put in place a backward-looking wage indexation system that resulted in automatic increases in real wages. Finally, Israel pegged the exchange rate to the US dollar as a temporary measure to guide expectations; after a few months, the shekel was devalued “at irregular intervals to offset the difference between the roughly 20% inflation in Israel and the lower inflation in its trading partners” (Friedman 1994a: 241). Chile, in contrast, announced that the pegged rate would remain indefinitely, even in light of overvaluation (Edwards and Edwards  1991).
Friedman visited Chile in November 1981, only a few months before the 1982 crisis, to attend the regional meetings of the Mont Pèlerin Society. His paper for the gathering was titled “Monetary System for a Free Society,” and in the last section he discussed exchange rates and monetary policy in Chile.36 In it, we read as follows:
Only Chile has in recent years effectively unified its currency with that of a major developed country. . . . Experience since I gave the lecture in Israel [where he suggested a unified currency for most LDCs] has not led me to alter my views on the economics of the issue [the superiority of the unified currency], though it has led me to become far more modest about judging political feasibility (in the sense of likelihood of adoption). Perhaps the example of Chile, if its policy continues to be as successful as it has been so far, will lead other developing countries to follow suit. (Friedman 1995: 174; emphasis added)
Two aspects of this paragraph are worth discussing. First, Friedman states that Chile had effectively implemented a unified currency. This, however, was not strictly the case. What Chile had done was adopt a pegged exchange rate in the tradition of Bretton Woods, with an additional verbal commitment to maintain the parity and to follow a “passive” monetary policy. However, from a legal perspective, the authorities could, at any time, undertake active monetary policy or give up the peg and devalue the peso. Furthermore, Chile had not eliminated the central bank and replaced it with a currency board, as Friedman had suggested years earlier at the Mont Pèlerin Society meeting in Stresa, Italy, and as he had forcefully advocated in the second Horowitz Lecture. The president of the central bank was Sergio de la Cuadra, one of Friedman's former students and a prominent “Chicago boy.” The previous two governors of the central bank were also Chicago graduates: Alvaro Bardón and Pablo Baraona. It was very unlikely that Chile's military would have agreed to give up the peso and adopt a foreign currency as legal tender. In addition, the existence of backward-looking wage indexation meant that a key adjustment mechanism—disinflation and the reduction of wages—was absent.
Second, Friedman's statement implies that Chile's pegged rate experiment had been successful. The persistence of inflation—it was still 12 percent per year—the large current account deficits financed by short-run speculative capital (14 percent of GDP), and an acute real exchange rate overvaluation call that statement into a question.
During his presentation at the Mont Pèlerin Society meeting, Friedman departed from his prepared remarks (and from the paper) in two ways. First, he talked about the relation between economic and political freedom. Friedman stated that it was important for Chile to move toward democratic rule. However, and not surprisingly given the military's severe censorship, the media did not report his remarks on the subject. Friedman became frustrated by this fact and talked about it extensively in an interview he gave in Peru, immediately after his trip to Chile.37 Second, in his improvised remarks—which were summarized on page 2 of the newspaper La Segunda, on November 20, 1981—he expressed concerns regarding the exchange rate policy, something he had not done in the paper. He said that in the written version his remarks about Chile had been too brief and somewhat elliptic. According to the news story, he said that
at this moment, international institutions don't have a guarantee that Chile will stick to its policy. . . . [Chile] also faces a speculative attack against its currency, triggered by the expectation that Chile may devalue, departing from its original goal. If Chile reaffirms the credibility of its current policy and allows its monetary aggregates to reflect changes in the balance of payments, then in the next crisis speculative forces will help to stabilize the system.
Friedman also noted that at the time the Chilean peso seemed to be overvalued.38 However, in his improvised discussion, Friedman did not delve into the mechanisms through which overvaluation was corrected under a pegged rate; he did not mention that under fixity the only way of addressing major overvaluation was by generating a massive disinflation.39
Seven months after Friedman's visit, Chile could not defend the peg any longer, and on June 14, 1982, it devalued the peso.40 The crisis that followed was one of the deepest ever faced by a Latin American nation. This was the crisis that many of the Chicago boys' detractors would blame on Friedman-inspired policies. (See, e.g., Klein 2010.)
On July 8, 1982, three weeks after the devaluation, Friedman wrote to Peter D. Whitney, the economic counselor at the US embassy in Santiago. He said, “I was surprised at the change [a stepwise modest devaluation followed by preannounced further adjustments] since it seems to me the appropriate alternative to the policy that Chile was following, if an alternative were to be adopted, was a fully floating exchange rate, not a prescheduled series of devaluations.”41 In a letter to journalist José Rodriguez Elizondo, written four months after the devaluation (October 15, 1982), Friedman commented on the policies that Chile's new economic team, led by his former student Rolf Luders, was likely to undertake: “He [Luders] may be, because of the type situation and because of a lack of previous commitments, more flexible. . . . Whether he can succeed in face of the tactics of the military is something else again on which I am not a competent judge” (MFAHI, box 188, folder 13).
In his 1998 memoirs, Friedman writes that it is doubtful “that there is a very good time for a country like Chile that has a central bank to peg its currency. I have consistently taken the position that a country like Chile with a central bank should let its currency float. The alternative is to abolish the central bank and unify its currency with that of its major trading partner” (Friedman and Friedman 1998: 405; emphasis added). However, as noted above, this was not the message that Friedman transmitted during his visits to Chile. To be sure, he noted that the Bretton Woods regime was unstable, but he did not say in public, even when asked about it, that Chile faced two options: either abolishing the central bank or floating.
On August 5, 1997, Friedman wrote a long letter to Robert J. Alexander, a Rutgers professor with whom he had had a number of exchanges regarding Chile and other Latin American countries. The purpose of the letter was to comment on the recently published book by Juan Gabriel Valdés, Pinochet's Economists: The Chicago Boys in Chile (Valdés 1995). Friedman takes issue with Valdés's narrative about the 1982 currency crisis. It is worthwhile quoting Friedman extensively:
Valdés has no understanding of what produced the 1982 depression. What produced it was the departure from the basic Chicago School economic principles that Valdés oversimplifies. [Minister Sergio] De Castro's mistake in pegging the Chilean currency to the U.S. dollar produced the disaster. My view has always been that a country like Chile, if it has a central bank and a separate monetary unit, should allow the exchange rate to float. That was the policy that was followed until 1979 when De Castro made the major mistake of pegging the Chilean currency to the U.S. dollar in the hope that that would impose the discipline necessary to eliminate inflation. In my opinion that was a bad decision under any circumstances, but it turned out to be a disastrously bad decision because of . . . the drastic appreciation of the U.S. dollar. (MFAHI, box 188, folder 10; emphases added)
Friedman addressed Chile once again in 2001, during a debate with Robert Mundell (Friedman and Mundell 2001). Friedman argued that Chile's 1979 “hard peg” policy was “disastrous” as a consequence of the strengthening of the US dollar in 1980–81. He also made this point in an addendum to the paper presented to the 1981 meeting of the Mont Pèlerin Society in Chile, in which he wrote that “the preceding three paragraphs [the ones that mentioned that Chile had unified its currency], correct when written in 1981, no longer are. Chile ended the pegging of its rate to the dollar in 1982, after the sharp appreciation of the US dollar plunged Chile into a disastrous recession” (Friedman 1995: 174).
To summarize, during his 1981 visit, Friedman was not openly critical about the exchange rate policy in Chile. He did not explicitly approve it, but he did not condemn it either. This ambiguity was not consistent with Friedman's usual bluntness and directness. At this point, one can only make a conjecture about what was going through his mind and why he was not more forthcoming. A possible explanation has to do with politics and with the likely effects that his criticism could have had on markets and capital outflows. It is also possible that he met with Minister of Finance Sergio de Castro, who at the time was trying to convince General Pinochet to cut wages by 10 percent, in order to generate the relative price adjustment that the economy required, and that Minister de Castro asked him not to “rock the boat.”42 My conjecture, then, is that given the circumstances, and the role that his former students played in the government, Friedman opted for circumspection and prudence. However, what is clear from his postcrisis writings and correspondence cited above is that he was never fully comfortable with Chile's strategy of pegging the exchange rate while maintaining the central bank.
Friedman's Long-Term Influence on Exchange Rate Regimes
Friedman's views on exchange rate and monetary regimes were, eventually, extremely influential in the advanced nations. According to the IMF, in 2021, all but two advanced countries had a floating regime. In his memoirs, George Stigler (2003: 161) wrote that the decision to abandon the Bretton Woods system and to float the dollar, taken by Richard Nixon in August 1973, was the result of Friedman having persuaded Secretary of the Treasury, and former Chicago colleague, George Shultz of the merits of a flexible exchange rate regime.
Friedman's ideas have been much less influential in developing countries. In 2021, just two of them—Mexico and Russia—had what the IMF calls a “free float.” In addition, only twenty-four developing nations—most of them tiny islands—had a “unified currency regime.”43 In 2022 most emerging countries still had “intermediate regimes” that were not very different from the Bretton Woods system so often criticized by Friedman.
There are a number of possible explanations for the lack of unified currencies among midsize and large developing nations. Perhaps the most important one is the absence of successful modern experiences that serve as examples of best practices. The failure of Argentina's experiment with a currency board between 1991 and 2001 generated great skepticism regarding the merits of superfixed currency systems.
Many of the causes behind the collapse of the Argentine experiment were related to issues raised by Milton Friedman throughout the years; see Hanke 2008, 2021 for detailed discussions. For example, the Central Bank of the Republic of Argentina (BCRA) was not abolished. Moreover, starting in 1995, the BCRA began to relax its operational rules, and money was created with very low backing of hard currency. Worse yet, fiscal policy was procyclical, and fiscal imbalances, mostly driven by provincial profligacy, grew significantly over time. This situation became particularly serious after 1995, when, as a result of contagion coming from Mexico's Zapatistas crisis, there was a sudden stop of capital flows. As in Chile twenty years earlier, wages were not flexible enough to allow for relative price realignment; the importance of price and wage flexibility had, of course, been a recurrent subject in Friedman's writings on exchange rates.44 There were also elements of bad luck, similar to those alluded to by Friedman in his postmortem on Chile. The Argentine peso was fixed to the US dollar at a time when the dollar strengthened significantly in the global markets, world interest rates increased substantially, a succession of crises in emerging markets (Russia, Turkey, Brazil) resulted in an across-the-board reduction in capital flows, and the terms of trade turned seriously against Argentina. Eventually, and when the IMF withdrew its support in December 2001, the one peso–one US dollar peg had to be abandoned. By 2003, Argentina was in shambles, and the reputation of monetary regimes based on unified currencies and currency boards had suffered a severe blow.45 A fact that is usually ignored in discussions about Argentina and its 2001 crisis is that throughout most of the convertibility period—from February 1991 to April 2001—the central bank was presided over by two Friedman students: Roque Fernandez and Pedro Pou.
The reasons behind the very low number of LDCs with free-floating rates are more straightforward and easier to understand. In most emerging economies commodity exports have (highly) volatile prices. Aizenman, Edwards, and Riera-Crichton (2012) have shown that terms-of-trade volatility is much higher in middle-income and poor countries than in advanced nations. In order to reduce the transmission of terms-of-trade instability into the real exchange rate, most LDCs have opted for a combination of managed currencies and changes in international reserves buffer stocks. Using data from 1990 through 2009 for Latin America, Aizenman et al. (2012) showed that this combination of policies indeed improves macroeconomic outcomes.
Interestingly, in Friedman's massive and comprehensive writings, there is very little emphasis on the role of terms-of-trade fluctuations in selecting the exchange rate regime. Arnold Harberger, his longtime colleague at Chicago, has pointed out that when choosing the appropriate monetary system, countries like Panama and Chile, for example, face very different realities. Panama has a very steady flow of foreign exchange earnings stemming from the Panama Canal; this, plus a relatively high degree of wage flexibility, fully justifies adopting a unified currency system. Chile, on the other hand, faces highly volatile terms of trade coming from fluctuations in the international price of copper and other minerals. According to Harberger, Chile would have made a mistake if it adopted a unified currency (Edwards and Harberger 2021). Harry G. Johnson (1969: 16), another of Friedman's Chicago colleagues, made a related point, when he argued that “in a banana republic . . . the currency will be more useful if it is stable in terms of command over foreign goods than if it is stable in terms of command over bananas.”
There are two important issues that, in my view, Friedman did not stress sufficiently in his massive and comprehensive writings on exchange rate regimes. First, and as noted above, he did not emphasize enough the role of terms-of-trade fluctuations in selecting the exchange rate/monetary regime. Second, when advocating for unified currencies without a central bank, Friedman did not discuss in sufficient detail the role of central banks as lenders of last resort. This, of course, was a topic he had worked on many times, and it was central in A Monetary History of the United States, in which he and Anna Schwartz analyzed many banking crises in the United States. However, and surprisingly, there is very little about the topic in his writings on monetary and exchange rate regimes in emerging nations.
In a recent paper, Rockoff (2022) traces the evolution of Friedman's thinking on “bailouts,” from A Monetary History to the United States to the savings and loans crisis, Penn Central, Lockheed, and Chrysler. Very much along the lines of this article's findings, Rockoff concludes that Friedman was not dogmatic; he was flexible and pragmatic and saw the issue through the lens of cost benefit-analysis. Moreover, he thought that it made a difference if the bailout referred to a brick-and-mortar company or a financial institution. According to Rockoff (2022: 64), “For Friedman . . . in the case of financial institutions the benefits of a bailout might well outweigh the costs.”46 He insisted, however, that the bailouts had to protect depositors and creditors, not shareholders or management. What is not clear, however, is how a possible bailout would be undertaken in a country without a central bank and a unified currency. One possibility, which was tried in Argentina during the convertibility period (1991–2001), is to have banks build buffers, or liquidity reserves, held in low-risk securities issued by advanced countries and deposited in international banks. In Argentina, however, and for political reasons, the government ended up relaxing those buffers in April of 2001, paving the way to the megacrisis of December of that year.
Comments by Maury Obstfeld, Doug Irwin, Michael Bordo, Bob Aliber, Ed Nelson, Kurt Schuler, John Lipsky, Lars Jonung, Leon Montes, Fernando Losada, Bruce Caldwell, Steve Hanke, Hugh Rockoff, Eugene White, Roberto Chang, and George Tavlas have been extremely useful. Comments and suggestions by the referees have helped improve the article significantly.
Of course, Friedman was not the only supporter of flexible rates for advanced countries. Other prominent backers of flexibility included James Meade, Harry G. Johnson, Gottfried Haberler, Lloyd Mints, and Fritz Machlup. On Friedman’s views on monetary policy, see Nelson 2020b. On the “corners solution,” see Eichengreen 1994. On exchange rate regimes, see, e.g., Tavlas, Dellas, and Stockman 2008.
See Boianovsky 2020 for a discussion of Friedman’s views on inflation and indexation in Brazil.
On Friedman, exchange rates, and Chicago, see Irwin 2018, 2019; Edwards and Harberger 2021; and Edwards 2020. Hanke (2008: 282) notes that Friedman mistrusted LDCs’ central banks and that this mistrust affected his views on exchange rates in those countries.
Mexico and Russia. Until 2021 Chile was classified as a “free floating” country. After the invasion of Ukraine, Russia imposed severe exchange controls, and her regime moved to anything but “floating.”
In his 1998 memoirs (with his wife, Rose), Friedman dedicated one chapter to each of them: Chile, chap. 24; Israel, chap. 27.
See, e.g., the bibliography in Nelson 2020b for a list of works on Friedman.
Edwards and Montes (2020) concentrate on Friedman’s view on economic and political freedom. The analysis of exchange rates is restricted to what he said in Chile in 1975 and 1981. In that regard, then, the present article may be thought of as both a prequel and a sequel to the much narrower analysis in Edwards and Montes 2020.
In the 1950s and after, Canada had moved to floating, while the United Kingdom considered, twice, adopting a flexible exchange rate. The arguments used by supporters of flexibility mirrored closely those made by Friedman in his 1950 memorandum. The first time was the ROBOT plan of 1951–52, and the second one was in 1952, during the negotiations for the Collective Approach. See Schenk 1991 for details.
The memo version was never released to the public. Its introduction is different from that of the published version. In it, Friedman refers to Keynes’s Tract on Monetary Reform and argues that it is not possible to simultaneously have fixed exchange rates, stable internal prices, and free capital mobility. MFAHI, box 43, folder 13.
As the discussion in the rest of this article will show, Friedman rarely emphasized the role of terms of trade volatility.
On Mints’s views and his influence on Friedman’s monetary views, see Dellas and Tavlas 2021.
In a 1965 American Enterprise Institute symposium, Friedman discussed a paper by James Meade and emphasized the experiences of Canada and Peru in the post–World War II era as instances of stable regimes with exchange rate flexibility. In a 1967 debate with former undersecretary of the Treasury Robert Roosa, Friedman pointed out that flexible rates were very likely to boost international commerce. In a 1969 discussion with MIT’s Charles Kindleberger, Friedman argued that the amount of exchange rate risk under fixed and flexible rates was similar. In a 1973 testimony in front of the Joint Economic Committee in Congress, two years after the “gold window” was closed, Friedman stated that recent events vindicated his long-held views on exchange rates. He said, “The exchange rate of the dollar has . . . declined in an orderly fashion. . . . There has been no sudden crisis, no closing of exchange markets, no changes of rates by 10 percent overnight” (Friedman 1973c: 117; available in the Friedman Archive at the Hoover Institution).
The quotations correspond to the online version of the memorandum; no page numbers are given because there are none in the online version.
See Friedman 1948a for an early discussion on Canada adopting flexible rates.
The Statement published a summary of Friedman’s recommendations based on a leaked copy of the report. In early 1957 Friedman published a brief paper in Encounter, where he summarized his views on India.
He also visited the Soviet Union (a place he disliked intensely), Lebanon, Cambodia, Pakistan, Taiwan, Malaysia, Thailand, Vietnam, Singapore, and the Philippines. See chap. 20 of Friedman and Friedman 1998 for details about this trip.
In 1963, Friedman also published a short paper on exchange rates in India’s promarket magazine Swarajya, titled “Exchange Rate Policy.” He wrote that “the Achilles heel of the Indian economy at the moment is the artificial and unrealistic exchange rate. The official exchange rate is the same today as it was in 1955. In the interim, prices within India have risen some 30 to 40 per cent; whereas prices in the US, UK, Germany have risen far less, at most by 10 per cent” (Friedman 1963).
Myrdal reacted very negatively to the news in 1976 that Friedman had been awarded the Nobel Prize, a prize that he had received in 1974. See Myrdal 1977.
Hanke (2008: table 1) underlines the fact that for Friedman there were three exchange rate regimes: fixed, pegged (Bretton Woods) and floating. Hanke’s discussion on Argentina’s failed “convertibility” experiment using Friedman’s perspective is illuminating.
A slightly revised version of the paper was included as chap. 12 in Dollars and Deficits.
His prediction about a currency crisis in a few years did not become a reality.
Friedman usually published several versions of his lectures, with minor variations. In this case, the material of the Horowitz Lectures was also published, in 1973, in a volume in honor of Moses Abramovitz and as the lead article in the first issue of the Liberian Economic and Management Review.
As noted earlier, and as pointed out by Nelson (2020b: vol. 2, chap. 13), Friedman thought that in the 1960s, the fixed exchange rate between the US dollar and gold, at $35 an ounce, did not act as a true constraint on the Fed.
This quotation corresponds to a blurb provided by Friedman and published on the back cover of Hanke, Jonung, and Schuler 1992.
Boianovsky (2020) provides a fascinating discussion on the connection between Brazil’s inflationary experience and Friedman’s thoughts about the Phillips curve.
Friedman corresponded with Alexander Kafka, Brazil’s legendary executive director at the IMF, as early as 1962. See MFAHI, box 29, folder 1. See also Boianovsky 2020.
On Friedman and the evolution of Phillips curve analyses, see Hoover 2008. On the role of Brazil’s experience in Friedman’s thinking, see Boianovsky 2020. On Harberger and the Latin American students at Chicago, see Edwards and Harberger 2021.
For details on the economic and political aspects of that visit, see Edwards and Montes 2020.
Friedman was also aware of political constraints in other countries, including South Africa. In 1976 he traveled to Cape Town and Johannesburg, where he lectured and met with policymakers and political leaders, including Zulu chief Gatsha Buthelezi. During his lectures and several meetings, he once again criticized the pegged-but-adjustable regime and argued that “the wise policy for a country like South Africa is to allow a free float.” Conscious of political constraints, he did not mention the unified currency option (Friedman 1976: 46).
This was his second visit to Yugoslavia. He had been there in 1963.
During his second visit to Chile, Friedman pointed out that under a unified currency, the currency of reference should be that of the major trading partner. He noted that if Australia decided to fix its exchange rate, it should do it relative to the Japanese yen. See La Segunda1981.
In 1990, the Hong Kong Center for Economic Research published a book titled Friedman in China, which included four lectures delivered in 1980, two lectures from the 1988 visit, a memo to Secretary General Zhao Ziyang, and a transcript of the two-hour conversation between Friedman and Zhao (Friedman 1990).
In Edwards and Montes 2020, Leon Montes and I discussed the political implications of Friedman’s two visits to Chile and his recommendation of a “shock treatment.” Our analysis of the exchange rate issue is quite general and does not go into the details of how Friedman reacted to the currency and banking crisis of 1982.
The paper was published in 1995, with an addendum written after the currency crisis of 1982 (Friedman 1995).
Letter to journalist José Rodríguez Elizondo dated December 18, 1981, MFAHI, box 188, folder 13.
Apparently, the reporter got a tape with Friedman’s improvised remarks. Some of the terms appear in quotation marks and are in English.
As noted, he had made this point as early as in 1950 when he said that a “decline of 10 percent in every internal price in Germany” was equivalent to a 10 percent devaluation of the mark relative to the dollar, and he had repeated it in almost every poor country he had visited during the previous twenty years (Friedman 1953: 164–65).
It is difficult to know to what extent Friedman’s remarks affected market expectations of an imminent crisis and, thus, contributed to the decline of capital flows in the following months. What is known, however, is that the probability of devaluation, as measured by interest rate differentials, almost tripled in the fourth quarter of 1981, relative to the previous quarter (Edwards and Edwards  1991: 68).
MFAHI, box 189, folder 2; emphasis added.
De Castro was not successful. At the time Pinochet’s advisers were in two camps: those like de Castro who favored the pegged exchange rate policy, and a group led by José Pinera who advocated for a devaluation followed by a floating rate. See de Castro 2018 for a first-person account of the rift.
The IMF (2021: 8) considers three main regimes: hard pegs, soft pegs, and floating.
The IMF was also held (partially) responsible for the debacle. Its own evaluation suggested that it did not raise important issues with the authorities in a timely fashion (IMF 2004).
In 2000, Ecuador gave up its currency and adopted the US dollar as legal currency. As a result, inflation has stayed in check. The absence of a central bank, however, did not constrain a populist government from running up debt very significantly in the 2007–17 period. See Edwards 2019 for details.
The page number in Rockoff comes from the typescript.