Milton Friedman and the Road to Monetarism: A Review Essay

Abstract

Milton Friedman was the most influential economist in policy circles since John Maynard Keynes. Friedman almost single-handedly resuscitated the importance of monetary policy to academic and policy thinking while leaving his mark in such areas as the natural rate of unemployment and the long-run Phillips curve, the choice of exchange rate regimes, the destabilizing effects of discretionary policies, the benefits of monetary policy rules, the determinants of consumption, the demand for money, the effects of inflation expectations on nominal interest rates, and the narrative approach to monetary history. The objective of Ed Nelson's two-volume book, Milton Friedman and Economic Debate in the United States, 1932–1972 (Chicago: University of Chicago Press, 2020), is to provide an account of Friedman's views in major monetary policy debates during the period identified in the book's title, although Nelson frequently draws on Friedman's works after 1972. The sweep of the book is astonishing, both in terms of the wealth of material presented and in terms of the author's extensive knowledge of the doctrinal and contemporary economics literature. Nelson has seemingly read everything that Friedman wrote during the latter's engagement with economics—from the time of his graduate studies in the early 1930s to his death in 2006.To provide some perspective of the book's immense compass, consider the following. The two volumes of the book contain a combined fifteen chapters totaling 1,324 pages. The documentation of sources is impeccable: by my count, volume 1 contains 2,016 footnotes (164 pages); volume 2 contains 1,321 footnotes (115 pages). Volume 1 references 234 of Friedman's solo-authored works and 38 of his coauthored works; volume 2 references 232 of Friedman's solo-authored works and 35 of his coauthored works. These references do not include newspaper and magazine articles—including triweekly columns that Friedman wrote for Newsweek from 1966 to 1984—and electronic media items, which are listed separately. The bibliography of references in volume 1 totals ninety-eight pages; that in volume 2 totals ninety-three pages. Nelson interviewed more than 250 people in preparing the book.The end product is a work of the highest scholarship that will stand as the preeminent work on Friedman's monetary economics for present and future generations. The book is not a biography, although all the ingredients that compose a biography are included in various places. Nor is the book a linear account of the development of Friedman's monetary scholarship. The book tells the story of the development of Friedman's monetary thinking through the windows of a selection of debates that engaged Friedman. For scholars of both Friedman and the development of macroeconomics, this book is to be carefully studied, although each chapter can be read separately from the others. The payoff will be a comprehensive understanding of the emergence of Friedman's views and their place in modern literature.Most of the chapters of the book are divided into overlapping sections titled “Events and Activities,” ‘Issues,” and “Personalities.” The “Events and Activities” sections cover some of Friedman's main engagements in economic debates during the particular years covered in the chapter. The “Issues” sections cover major policy or research issues with which Friedman was involved during the years in question—including those mentioned at the beginning of this review, the area of exchange rate regimes being the major exception. The “Personalities” sections focus on specific individuals—twenty individuals in total—with whom Friedman interacted (or to whom Friedman reacted) on the particular policy identified in the “Issues” sections. The reader of the book will encounter some distinguished economists in the “Personalities” sections, including Alvin Hansen, Paul Samuelson, James Tobin, Robert Lucas, and Thomas Sargent. In the midst of his chronological account, Nelson pauses to provide a five-chapter exposition of Friedman's macroeconomic and microeconomic frameworks and his views on policy rules.A substantial body of the book shows how Friedman's monetary economics can be made consistent with present-day macroeconomics. In what follows, I focus mainly on doctrinal issues related to the development of Friedman's monetary thinking.Friedman earned a bachelor of arts degree in 1932, at Rutgers University, where he came under the influence of his teacher, Arthur Burns, with whom he would form a close, lifelong personal friendship. That relationship was temporally ruptured in the 1970s over a dispute about monetary policy after Burns had become chair of the Federal Reserve Board; I discuss the nature of that dispute below. After his graduation from Rutgers, Friedman's professional career path over the remainder of that decade had, in Nelson's view, “convoluted elements” (1:27), reflecting Friedman's quick succession of moves from one institution to another. Friedman undertook graduate studies in economics at Chicago in the 1932–33 academic year, earning a master of arts degree in 1933. He then did graduate work at Columbia University in the 1933–34 academic year, where he was influenced by Harold Hotelling, the mathematical statistician and economic theorist. He returned to Chicago for the 1934–35 academic year as a research assistant to Henry Schultz, a pioneering econometrician. From 1935 to 1937, he was employed by the National Resources Committee in Washington, D.C., constructing estimates of consumer spending. Beginning in 1937, he began an association (that lasted until 1981) with the National Bureau of Economic Research (NBER). Friedman's work at the NBER in the late 1930s entailed assisting Simon Kuznets in the latter's work on professional income: from 1937 until 1940, Friedman worked with Kuznets as the latter's “salaried assistant . . . at the National Bureau and, in effect, Ph.D. student of Kuznets” (1:69). Based on their work together, Friedman and Kuznets coauthored the 1945 book Income from Independent Professional Practice, which Friedman submitted as a doctoral thesis at Columbia. He was awarded a PhD from Columbia in 1946. During the 1940–41 academic year, Friedman held a visiting position at the University of Wisconsin.The picture that emerges is that of a fledging, well-traveled economist who had received first-rate training in applied economics and statistical theory—and had shown exceptional potential in those fields—during a period of time that would see the Keynesian revolution sweep through the economics profession. By the mid-1940s, Friedman had become a theoretical statistician of the first rank, having published several important articles in that area. Together with L. J. Savage, in the late 1940s and early 1950s he applied statistical analysis to the theory of choice under conditions of uncertainty. His work on statistical theory had a big payoff: in 1951, Friedman was awarded the John Bates Clark Medal for, as Nelson put it, having “built bridges between mathematical statistics and economics” and for having “used statistical theory to generalize the analysis of utility to a world of uncertainty” (1:7). By 1951 Friedman had reached what Nelson calls “the heights of the profession” (1:176).Amid Friedman's rise in the 1940s, however, a transformation was taking place, slowly at first, in his research agenda. This transformation would initially have a profound and negative impact on his professional reputation. As Nelson explains, after having been awarded the Clark Medal, Friedman's “standing in the profession was about to crash” (1:176). What was the nature of that crash? Why did it happen? To shed light on these questions, we must take up Nelson's account of Friedman's career path after he completed his one-year appointment at the University of Wisconsin.Friedman began working at the US Treasury in 1941. According to Nelson, “Prior to joining the Treasury, Friedman had largely accepted the theoretical contribution of the General Theory” (1:88–89); and Friedman “came to the US Treasury already having a Keynesian perspective” (1:91). Nelson expresses the view that Friedman's employment at the Treasury reinforced that perspective (1:91–92).The main issue that Friedman and his Treasury colleagues confronted in the early 1940s was to find a way to restrain aggregate demand in light of the large military buildup associated with the United States' entry into World War II, while, at the same time, facilitating the transfer of resources to the defense sector. The particular Keynesian perspective that Friedman and his Treasury coworkers adopted was not that of Keynes's General Theory (1936); instead, it was that of Keynes's 1940 study, How to Pay for the War. That latter study addressed the excess demand conditions of a wartime economy from the perspective of the inflationary gap—that is, the excess of total nominal spending over the level that is consistent with price stability.Nelson points out that, under inflationary-gap analysis, fiscal deficits were “ipso facto a stimulant to aggregate demand” (1:94). What mattered for inflation was the magnitude of the fiscal deficit. Moreover, “an accommodative monetary policy stance was not a key part of the sequence in which deficit spending generated an increase in nominal aggregate demand” (1:94; italics in original). Consequently, “lower deficit spending was crucial to restraining inflationary pressure in a way that monetary restraint was not” (1:94). In the context of the inflationary-gap analysis, Friedman and his Treasury colleagues focused on ways of getting taxes to keep pace with the rises in government spending in order to contain inflation.In May 1942, Friedman submitted a memorandum to the House Ways and Means Committee in which he dealt with ways to control inflation. To achieve that objective, Friedman argued that consumer spending would have to be restricted. The best way to do that, he argued, was via income taxation. The other ways of avoiding inflation mentioned were through price controls and rationing, controls on consumer credit, a reduction in government spending, and selling war bonds to the public. Friedman, however, did not take a position on the effectiveness of price controls; as an employee of the Treasury, he would have been constrained in criticizing price controls. As Nelson points out, Friedman did not view the reaction of monetary policy to deficit spending “as decisive in determining whether deficit spending raised the price level” (1:96).By the mid-1940s, however, Friedman had begun to alter—if only modestly—his views on Keynesian economics. Nelson notes that in a 1944 review of the book Saving, Investment, and National Income, by Oscar Altman (1944), Friedman put distance between himself and Keynesian income-expenditure theory (1:121–22). Friedman concluded that the book provided little evidence to support its main contention—namely, that investment is the main dynamic variable that determines income and employment (1:122). Similarly, in a statement—which Nelson does not cite—published in the Congressional Record on April 16, 1946, Friedman called for the elimination of price controls, imposed by the Office of Price Administration (OPA). He argued, “A major effect of OPA price controls has been to disguise rather than prevent price increases” (Friedman 1946: A2336). He also called on the control of the quantity of money to reduce inflation: We can and must take measures now to control the basic causes of inflation by limiting the supply of cash and bank deposits. This will require that Government collect as much and spend as little as possible; that we put the Federal Reserve System once again in control of the volume of cash and deposits by drastically raising reserve requirements and that we pin down the liquid assets in the hands of the public by a realistic debt policy, even if that means higher interest rates on the Federal debt. (A2336)The following point merits comment. Friedman's written track record in the first half of the 1940s conveys an unmistakable Keynesian perspective. Indeed, the Keynesian influence on his thinking was acknowledged by Friedman in his memoirs. Referring to his work while he was with the Treasury, Friedman stated, “I had completely forgotten how thoroughly Keynesian I then was” (Friedman and Friedman 1998: 113). The point that I want to make is that the argument that Friedman was a Keynesian in the first part of the 1940s should not be pushed too far. In the first half of the 1930s, Friedman had been of student of Frank Knight, Lloyd Mints, and Jacob Viner at Chicago, each of whom was a quantity theorist and an advocate of a very limited role of the government in economic affairs. In the 1950s and after, Friedman acknowledged the influence that those three economists, as well as Henry Simons (see below), had exerted on his monetary economics (see, e.g., Friedman 1956: 3–4). Friedman's Keynesian persuasion in the 1940s had to have confronted—and clashed with—his Chicago training. Consequently, Friedman's early Keynesianism was far from rock-solid, which helps explain his changing views during the 1940s.After joining the economics faculty at the University of Chicago in the fall of 1946, Friedman's views on macroeconomic policies continued to change. Nelson singles out the years 1948–51 as pivotal. He refers to Friedman's “epochal 1948–51 rethinking of monetary matters” (1:40). Similarly, Nelson considers that the years 1948–51 marked “a shake-up in Friedman's thinking, especially with regard to monetary economics” (1:32), and “from 1948 to 1951, he became a monetarist” (1:295). The “shake-up” would initially diminish Friedman's standing in the economics community because it marked Friedman's move away from Keynesian economics, with its deemphasis on the role of money in the economy, and toward the adoption of the quantity theory of money (1:32).The shake-up saw Friedman embrace monetary economics as a primary field of interest. In 1948, he published the paper “A Monetary and Fiscal Framework for Economic Stability” in the American Economic Review, in which he proposed a monetary rule for the first time. Under his particular rule, the stock of money would be increased when there was an increase in the budget deficit—by the amount of the deficit. It would be decreased when there was a surplus in the budget—by the amount of the surplus. The budget would be balanced over the course of the business cycle, or, alternatively, it would lead to a deficit that was sufficient to provide some specified secular increase in the quantity of money at a level of income corresponding to reasonably full employment (Friedman 1948a: 137). As Nelson points out, the proposal required an estimate of the full-employment level of income as part of the process under which the cyclically adjusted budget balance would be determined (1:141).Thus, under the proposal, Friedman focused on the need to control the supply of money, using the government's fiscal position to do so. Nelson does not regard the proposal to be a monetarist one. He points out that “the 1948 article's analysis is one that takes policy-induced changes in the quantity of money as having their effects on the economy via the fiscal-multiplier effect of deficit spending and not [as under Friedman's monetarist framework] via reactions of yields. The multiple-interest-rate channel of monetary policy transmission that Friedman emphasized from the 1950s onwards is absent from this analysis” (1:139). Nevertheless, Nelson views the 1948 paper as a move away from Keynesian analysis because it contained “an acknowledgement on Friedman's part that money-financed deficits had greater repercussions for spending than deficits financed by issuance of longer-term securities” (1:139).Nelson singles out two other features of the 1948 article that would play prominent roles in Friedman's subsequent work on policy rules (1:140–41). First, the article introduced Friedman's famous phrase “long and variable lags” into the economics literature. Second, Friedman noted that economic forecasters had established a poor track record; therefore, it was preferable not to rely on forecasts in the area of policy formation.On the basis of the above discussion, I believe that it is fair to argue that an initial shake-up in Friedman's thinking on macroeconomics occurred between the early 1940s and the publication of his 1948 AER article. What caused this initial shake-up? Part of the answer lies in one of Nelson's “Personalities” sections—namely, the one dealing with Chicago economist Henry Simons (1:57–67).While a student at Chicago in the early 1930s, Friedman did not take a course with Simons. Moreover, the two Chicagoans did not overlap in their teaching positions at Chicago—Simons died unexpectedly a few months before Friedman began teaching at that institution in the fall semester of 1946. Nevertheless, the influence of Simons's views on Friedman's 1948 AER article was pervasive. By 1948, it was evident that Friedman had studied Simons's policy views and had been heavily influenced by those views. As Nelson points out regarding the influence of Simons on Friedman's work, “A debt to Simons undoubtedly existed and was repeatedly acknowledged by Friedman” (1:57). Friedman's 1948 article contains numerous references to Simons's views and cites three of Simons's works. The similarities between Simons's views and those of Friedman, 1948 vintage, are striking.1Nelson shows that, beginning in the 1950s, Friedman would move away from Simons's policy playbook. According to Nelson, “Friedman's 1948 proposal . . . proved to be the high-water mark of the agreement between his and Simons's framework” (1:58). Friedman would drop his 1948 proposals that the issuance of Treasury bills be discontinued and that the existing stock be withdrawn from the market. Simons believed that a monetary rule could best be made effective in a highly regulated financial system. Such a system would have prevented a wide range of financial transactions between borrowers and lenders and acted as a brake on the accumulation of capital. Friedman “had no interest in imposing wide-ranging additional restrictions on the terms of private lending and borrowing” (1:61). By the late 1950s, he deemphasized the importance of 100 percent reserves.What caused Friedman to subsequently reduce the emphasis that he had placed on the 100 percent reserves scheme? Nelson convincingly argues that there were two main reasons. First, although Friedman was concerned about the instabilities in the credit-creation process, in the early 1950s he realized that the possibility of a key instability that had contributed to banking crises in the 1930s and earlier—namely, that associated with changes in the deposit-currency ratio—had been greatly reduced by deposit insurance, introduced in 1934. Second, in the 1950s Friedman became convinced that open-market operations would offset variations in the reserve deposit ratio and remaining variations in the currency deposit ratio and, thus, achieve the goal of maintaining the money stock (1:59).The upshot of the above discussion is that Friedman began his engagement with monetary economics in 1948 highly influenced by the work of Simons, although he would, over time, substantially modify the Simons playbook. That playbook was by and large also advocated by Lloyd Mints, who taught the graduate course on money at Chicago until his retirement in 1953. It is also likely that Friedman had been influenced by Mints; Friedman and Mints interacted extensively in the late 1940s and early 1950s, with the result that there had been considerable cross-fertilization of their respective views (Dellas and Tavlas 2021). In the mid-1940s, Mints's and Simons's views on money—including the advocacy of rules, the use of the government's fiscal position to generate changes in the money supply, 100 percent reserve requirements, and flexible exchange rates—became known in the profession as “the Mints-Simons program” (Hansen 1946: 73). There was a reason why Mints and Simons had been identified with a particular program: each of those policy positions was an outlier within the economics profession. Taken together, the program was viewed as downright eccentric. It was for that reason that, in 1951, Harry Johnson (1951: 382) wrote that Simons and Mints were the leading members of “the Chicago radical school.” Yet in the early 1950s, a third name became identified with that program—that of Milton Friedman. As I discuss below, the person who made that identification was Clark Warburton.Nelson shows that Friedman's monetary economics took on their monetarist character in the early 1950s.4 The distinctive characteristic that marked Friedman's move to monetarism was his embrace of the quantity theory of money. As Nelson argues, “By 1950 he had cast his lot with the quantity theorists” (1:153), a move that greatly diminished his standing in the profession.Essential elements of Friedman's monetarism included the following. (1) By the mid-1950s, Friedman had articulated a comprehensive empirical critique of the Fed's policies in the late 1920s and early 1930s. (2) Friedman came to believe that open-market operations should be the instrument of choice for generating changes in the quantity of money. He abandoned his fiscal-based monetary rule in favor of a 3 to 5 percent money-growth rule. (3) He came to believe that a fractional-reserve banking system is not “inherently” unstable, although, as mentioned, he continued to support the 100 percent reserves proposal. (4) He produced empirical evidence showing that the demand for money is a stable function of a few variables. (5) Contrary to his views in the 1940s, he came to believe that fiscal policy is an ineffective stabilization tool, although he believed in the effectiveness of the automatic fiscal stabilizers throughout the 1950s. (6) He developed the permanent income hypothesis, which is an early example of individuals solving a dynamic model to engage in forward-looking behavior. (7) Beginning in the mid-1950s, in the Chicago Workshop in Money and Banking, which he supervised, Friedman compared the empirical performances of the quantity theory and the Keynesian income-expenditure theory using empirical methods, with the result that the former was shown to be far superior to the latter in predicting nominal income (proxied by nominal consumption expenditure). Nelson discusses each of these changes in Friedman's framework in detail.As Nelson explains, not all of the above-listed changes took place within the three years spanning 1948 to 1951. For example, Friedman's first public espousal of a constant-money-supply growth rule did not take place until 1956. Nevertheless, as Nelson also explains, there was enough of a change in Friedman's views by the early 1950s to account for an epochal rethinking. Nelson writes, “His move to monetarism came in light of his study of empirical evidence, as well as his reconsideration of a large body of literature, including that of monetary economists, like Fisher and Pigou” (1:32). What was the nature of the empirical evidence on which Friedman could draw that led to Friedman's rethinking about the role of money in the economy?A good place to begin concerns fiscal policy. Nelson points out that “by the early-1950s . . . Friedman's rating of the effects of fiscal policy was much diminished. He now saw deficit spending per se as not exerting a great influence on aggregate demand. But he was redoubled in his conviction that monetary policy had strong effects” (1:296). As Nelson explains, “After 1948 Friedman became skeptical about the idea that fiscal policy actually had much of a distinct impact on aggregate demand. The empirical basis for this skepticism began to see print even in the 1950s, with Friedman's 1952 article on wartime monetary relations a particularly notable example” (1:319). The article to which Nelson refers is Friedman's “Price, Income, and Monetary Changes in Three Wartime Periods.” Based on his ongoing work with Schwartz, Friedman presented the paper at the December 1951 meetings of the American Economic Association; it was published in the AER in May 1952. Friedman provided empirical evidence on the determinants of inflation during the Civil War, World War I, and World War II. He found that price behavior in all three episodes was proximately explained by the stock of money per unit of output. It could not be explained by any other variable examined, including measures of fiscal policy.In my view, although the results of the 1952 AER paper contributed to Friedman's emerging monetarism, that project could account for only a part of Friedman's changing views. For example, the results could not explain Friedman's switch to open-market operations conducted by the Fed as the preferred monetary instrument, since the Fed did not exist in the Civil War and, once the Fed was established, it did not start experimenting with open-market operations until after World War I.5What else, then, helped create the “shake-up” in Friedman's monetary thinking in 1951? To shed further light on this question, we must look to the work of Clark Warburton, another of the “personalities” in Nelson's book. It has long been recognized that Warburton's views anticipated Friedman's. Recent research, published after Nelson finalized his text, suggests a causal connection. Lothian and Tavlas (2018) showed that throughout the course of 1951, Warburton and Friedman carried on a lengthy correspondence in which Warburton severely criticized Friedman's initial monetary framework. In Tavlas 2019, I showed that Warburton expanded his criticisms of Friedman's framework into two articles reviewing that framework, published in successive issues of the Journal of Finance in 1952 and 1953. Warburton began his 1952 article with the observation that Friedman's monetary framework bore a striking resemblance to the Simons-Mints platform: “It is similar to proposals for monetary reform developed during the past two decades by Professors Lloyd W. Mints and the late Henry C. Simons” (Warburton 1952: 328). In other words, Friedman's initial position in his engagement with monetary economics was that of the Chicago monetary tradition of the 1930s and 1940s.In his 1951 correspondence with Friedman and in his Journal of Finance articles, Warburton's arguments included the following. (1) A monetary system in which banks' reserves are the liabilities of the central bank is not “inherently unstable,” as had been argued by Simons, Mints, and Friedman. In such a system, the central bank can use open-market operations to offset shifts from deposits to currency and, thus, control the quantity of money by issuing its own liabilities. Simons, Mints, and Friedman had mistakenly analyzed the workings of a commodity-based monetary system—such as the early 1930s gold standard—in which the use of open-market operations is constrained by the central bank's holdings of commodity reserves. (2) The use of the government's fiscal position was far too unwieldy and unreliable as a mechanism with which to produce changes in the money supply. (3) Simons, Mints, and Friedman had failed to distinguish between the inherent characteristics of a monetary system and the way the system had been managed. The “inherent instability” that the Chicagoans had identified was, in fact, attributable to poor central-bank management, particularly in the early 1930s. (4) The Fed was responsible for initiating and exacerbating the Great Depression. (5) Friedman's policy rule was too complicated to be useful in practice. A far better rule was one under which the quantity of money increased by 4 percent a year. Earlier, in the 1940s, Warburton had compared the empirical performances of the quantity theory with an income-flow-based analysis of the Keynesian type, a comparison that presaged Friedman and Meiselman's (1963) work.Nelson discusses the similarities between Warburton's views and Friedman's. With regard to Warburton's work comparing the Keynesian-type income-flow-based analysis with the quantity theory, Nelson writes, “Just as Friedman would argue subsequently, Warburton held that the money/spending relationship had proved more resilient in wartime than the Keynesian consumption function” (1:115).6 Moreover, Nelson points to the many similarities between Warburton's views and those of Friedman: “Few cannot be struck by the extent to which Warburton anticipated Milton Friedman's work on matters relating to monetary policy. On the interpretation of the Great Depression, the advocacy of constant money growth, and several other matters, Warburton . . . was taking a stand [in the 1940s] . . . that Friedman would take up only after 1948” (1:114).Thus, Nelson, like previous writers, credits Warburton as having anticipated Friedman's views on money. On the basis of recent research, I believe that it is necessary to go beyond that characterization and conclude that Warburton influenced Friedman's views precisely at the time (1948 to 1951) that Friedman's views were undergoing an “epochal . . . rethinking on monetary matters” (1:40).As Nelson points out, “Friedman wrote prolifically—and yet he produced nothing that consolidated his views into a single definitive statement” (1:9). Friedman left “no monograph that could be regarded as a compendium of his monetary views.” A central objective of Nelson's book is to present an encompassing picture of Friedman's monetary framework after Friedman became a monetarist. In doing so, Nelson presents a series of vignettes that describe and analyze Friedman's views on key issues relating to money. I provide an overview of several of those issues.Why did Friedman focus on the quantity of money and not on the quantity of credit? Nelson explains that Friedman vie