An Interview with Paul A. Samuelson

Interviewed by William A. Barnett, University of Kansas, December 23, 2003

An interview with Paul A. Samuelson, by William A. Barnett, which appeared in print in the journal, Macroeconomic Dynamics (published by Cambridge University Press), in September 2004 under the title An Interview with Paul A. Samuelson. To our knowledge, this is the first and only interview of Samuelson published in a professional economics journal. In addition, this is the only interview conducted solely by the editor of Macroeconomic Dynamics, William A. Barnett. The interview confers Samuelson’s views on the economics profession from 1929 to the present and an overview of Samuelson’s career as one of the greatest economists of all time.

This interview is also published in Inside the Economist’s Mind: Conversations with Eminent Economists. The book, edited by Paul A. Samuelson and William A. Barnett, is published by Blackwell/Wiley and appeared in 2007. Professor Douglas Gale, New York University:

The interviews in this volume are unique intellectual documents in the history of economic thought, economic policy, and biography. Scholars will value them as primary sources. Readers with only a passing interest in economics will be delighted by their entertaining insights into the minds and lives of these great thinkers. This is one of the most valuable projects in academic economic publishing for a long time, and we should all be grateful to the journal, Macroeconomic Dynamics, for collecting these archival treasures over a number of years.

Paul A. Samuelson (15 May 1915 – 13 December 2009) was the first American to win the Nobel Prize in Economics. The Swedish Royal Academies stated, when awarding the prize, that he “has done more than any other contemporary economist to raise the level of scientific analysis in economic theory”. Economic historian Randall E. Parker calls him the “Father of Modern Economics”, and The New York Times considered him to be the “foremost academic economist of the 20th century”. He was author of the largest-selling economics textbook of all time: Economics: An Introductory Analysis, first published in 1948. It was the second American textbook to explain the principles of Keynesian economics and how to think about economics, and the first one to be successful, and is now in its 19th edition, having sold nearly 4 million copies in 40 languages.

William A. Barnett is Oswald Distinguished Professor of Macroeconomics at the University of Kansas. Barnett has been a leading researcher in macroeconomics and econometrics. He is one of the pioneers in the study of chaos and nonlinearity in socioeconomic contexts, as well as a major figure in the study of the aggregation problem, which lies at the heart of how individual and aggregate data are related. He is editor of the Elsevier monograph series International Symposia in Economic Theory and Econometrics, and editor of the journal Macroeconomic Dynamics. He has published 17 books (as either author or editor) and over 130 articles in professional journals.

‘The U.S. Economy: The Last 50 Years and the Next 50 Years’. Nobel Laureates Franco Modigliani, Paul A. Samuelson, Robert M. Solow (18 September 2000)

Order your copy of Inside the Economist’s Mind: Conversations with Eminent Economists on Amazon now!

Figure 1. Paul A. Samuelson.

It is customary for the Interviewer to begin with an introduction describing the circumstances of the interview and providing an overview of the nature and importance of the work of the interviewee. However, in this case, as Editor of this journal, I feel it would be presumptuous of me to provide my own overview and evaluation of the work of this great man, Paul Samuelson. The scope of his contributions has been so vast (averaging almost one technical paper per month for over 50 years) that it could be particularly difficult to identify those areas of modern economic theory to which he has not made seminal contributions.[1] In addition to his over 550 published papers, his books are legendary. He once said: “Let those who will—write the nation’s laws—if I can write its textbooks.”

Instead of attempting to provide my own overview, I am limiting this introduction to the following direct (slightly edited) quotation of a few paragraphs from the Web site, The History of Economic Thought, which is maintained online by the New School University in New York[2]:

Perhaps more than anyone else, Paul A. Samuelson has personified mainstream economics in the second half of the twentieth century. The writer of the most successful principles textbook ever (1948), Paul Samuelson has been not unjustly considered the incarnation of the economics “establishment”—and as a result, has been both lauded and vilified for virtually everything right and wrong about it.

Samuelson’s most famous piece of work, Foundations of Economic Analysis (1947), is one of the grandest tomes that helped revive Neoclassical economics and launched the era of the mathematization of economics. Samuelson was one of the progenitors of the Paretian revival in microeconomics and the Neo-Keynesian Synthesis in macroeconomics during the post-war period.

The wunderkind of the Harvard generation of 1930s, where he studied under Schumpeter and Leontief, Samuelson had a prodigious grasp of economic theory, which has since become legendary. An unconfirmed anecdote has it that at the end of Samuelson’s dissertation defense, Schumpeter turned to Leontief and asked, “Well, Wassily, have we passed?” Paul Samuelson moved on to M.I.T. where he built one of the century’s most powerful economics departments around himself. He was soon joined by R.M. Solow, who was to become Samuelson’s sometime co-writer and partner-in-crime.

Samuelson’s specific contributions to economics have been far too many to be listed here—being among the most prolific writers in economics. Samuelson’s signature method of economic theory, illustrated in his Foundations (1947), seems to follow two rules which can also be said to characterize much of Neoclassical economics since then: With every economic problem, (1) reduce the number of variables and keep only a minimum set of simple economic relations; and (2) if possible, rewrite it as a constrained optimization problem.

In microeconomics, he is responsible for the theory of revealed preference (1938, 1947). This and his related efforts on the question of utility measurement and integrability (1937, 1950) opened the way for future developments by Debreu, Georgescu-Roegen, and Uzawa. He also introduced the use of comparative statics and dynamics through his “correspondence principle” (1947), which was applied fruitfully in his contributions to the dynamic stability of general equilibrium (1941, 1944). He also developed what are now called “Bergson–Samuelson social welfare functions” (1947, 1950, 1956); and, no less famously, Samuelson is responsible for the harnessing of “public goods” into Neoclassical theory (1954, 1955, 1958).

Samuelson was also instrumental in establishing the modern theory of production. His Foundations (1947) are responsible for the envelope theorem and the full characterization of the cost function. He made important contributions to the theory of technical progress (1972). His work on the theory of capital is well known, if contentious. He demonstrated one of the first remarkable “Non-Substitution” theorems (1951) and, in his famous paper with Solow (1953), initiated the analysis of dynamic Leontief systems. This work was reiterated in his famous 1958 volume on linear programming with Robert Dorfman and Robert Solow, wherein we also find a clear introduction to the “turnpike” conjecture of linear von Neumann systems. Samuelson was also Joan Robinson’s main adversary in the Cambridge Capital Controversy—introducing the “surrogate” production function (1962), and then subsequently (and graciously) relenting (1966).

In international trade theory, he is responsible for the Stolper–Samuelson Theorem and, independently of Lerner, the Factor Price Equalization theorem (1948, 1949, 1953), as well as (finally) resolving the age-old “transfer problem” relating terms of trade and capital flows, as well as the Marxian transformation problem (1971), and other issues in Classical economics (1957, 1978).

In macroeconomics, Samuelson’s multiplier–accelerator macrodynamic model (1939) is justly famous, as is the Solow–Samuelson presentation of the Phillips Curve (1960) to the world. He is also famous for popularizing, along with Allais, the “overlapping generations” model which has since found many applications in macroeconomics and monetary theory. In many ways, his work on speculative prices (1965) effectively anticipates the efficient markets hypothesis in finance theory. His work on diversification (1967) and the “lifetime portfolio” (1969) is also well known.

Paul Samuelson’s many contributions to Neoclassical economic theory were recognized with a Nobel Memorial prize in 1970.

Barnett: As an overture to this interview, can you give us a telescopic summary of 1929 to 2003 trends in macroeconomics?

Samuelson: Yes, but with the understanding that my sweeping simplifications do need, and can be given, documentation.

As the 1920′s came to an end, the term macroeconomics had no need to be invented. In America, as in Europe, money and banking books preached levels and trends in price levels in terms of the Fisher–Marshall MV = PQ. Additionally, particularly in America, business-cycles courses eclectically nominated causes for fluctuations that were as diverse as “sunspots,” “psychological confidence,” “over- and underinvestment” pathologies, and so forth. In college on the Chicago Midway and before 1935 at Harvard, I was drilled in the Wesley Mitchell statistical descriptions and in Gottfried Haberler’s pre-General Theory review of the troops. Read the puerile Harvard book on The Economics of the Recovery Program, written by such stars as Schumpeter, Leontief, and Chamberlin, and you will agree with a reviewer’s headline: Harvard’s first team strikes out.

Keynes’s 1936 General Theory—paralleled by such precursors as Kahn, Kalecki, and J.M. Clark—gradually filled in the vacuum. Also, pillars of the MV = PQ paradigm, such as all of Fisher, Wicksell, and Pigou, died better macroeconomists than they had earlier been—this for varied reasons of economic history.

Wicksell was nonplussed in the early 1920′s when postwar unemployment arose from his nominated policy of returning after 1920 back to pre-1914 currency parities. His long tolerance for Say’s law and neutrality of money (even during the 1865–1900 deflation) eroded away in his last years. For Fisher, his personal financial losses in the 1929–34 Depression modified his beliefs that V and Q /V were quasi constants in the MV = PQ tautology. Debt deflation all around him belied that. Pigou, after a hostile 1936 review of The General Theory (occasioned much by Keynes’s flippancies about Marshall and “the classics”), handsomely acknowledged wisdoms in The General Theory’s approaches in his 1950 Keynes’s General Theory: A Retrospective View.

I belabor this ancient history because what those gods were modifying was much that Milton Friedman was renominating about money around 1950 in encyclopedia articles and empirical history. It is paradoxical that a keen intellect jumped on that old bandwagon just when technical changes in money and money substitutes—liquid markets connected by wire and telephonic liquid “safe money market funds,” which paid interest rates on fixed-price liquid balances that varied between 15% per annum and 1%, depending on price level trends—were realistically replacing the scalar M by a vector of (M₀, M₁, M₂, . . . , M₁₇, a myriad of bonds with tight bid-asked prices, . . . ). We all pity warm-hearted scholars who get stuck on the wrong paths of socialistic hope. That same kind of regrettable choice characterizes anyone who bets doggedly on ESP, or creationism, or. . . . The pity of it increases for one who adopts a simple theory of positivism that exonerates a nominated theory, even if its premises are unrealistic, so long only as it seems to describe with approximate accuracy some facts. Particularly vulnerable is a scholar who tries to test competing theories by submitting them to simplistic linear regressions with no sophisticated calculations of Granger causality, cointegration, collinearities and ill-conditioning, or a dozen other safeguard econometric methodologies. To give one specific example, when Christopher Sims introduces both M and an interest rate in a multiple regression testing whether M drives P, Q /V, or Q in some systematic manner congenial to making a constant rate of growth of money supply, M₁, an optimal guide for policy, then in varied samples the interest rate alone works better without M than M works alone or without the interest rate.

FIGURE 2. New York, February 19, 1961. Seated left to right, participating guests who appeared on the first of The Great Challenge symposia of 1961: Professor Henry A. Kissinger, Director of the Harvard International Seminar; Dr. Paul A. Samuelson, Professor of Economics at MIT and President of the American Economic Association; Professor Arnold J. Toynbee, world historian; Admiral Lewis L. Strauss, former Chairman of the Atomic Energy Commission and former Secretary of Commerce; Adlai E. Stevenson, U.S. Ambassador to the United Nations; and Howard K. Smith, CBS news correspondent in Washington, moderator of the program. The topic: 'The World Strategy of the United States as a Great Power.'

The proof of the pudding is in the eating. There was a widespread myth of the 1970′s, a myth along Tom Kuhn’s (1962) Structure of Scientific Revolutions lines. The Keynesianism, which worked so well in Camelot and brought forth a long epoch of price-level stability with good Q growth and nearly full employment, gave way to a new and quite different macro view after 1966. A new paradigm, monistic monetarism, so the tale narrates, gave a better fit. And therefore King Keynes lost self-esteem and public esteem. The King is dead. Long live King Milton!

Contemplate the true facts. Examine 10 prominent best forecasting models 1950–1980: Wharton, Townsend–Greenspan, Michigan Model, St. Louis Reserve Bank, Citibank Economic Department under Walter Wriston’s choice of Lief Olson, etc. When a specialist in the Federal Reserve system graded models in terms of their accuracy for out-of-sample future performance for a whole vector of target macro variables, never did post-1950 monetarism score well! For a few quarters in the early 1970′s, Shirley Almon distributed lags, involving [M_i(−1), M_i(−2), . . . , M_i(−n)], wandered into some temporary alignment with reality. But then, outfits like that at Citibank, even when they added on Ptolemaic epicycle to epicycle, generated monetarism forecasts that diverged systematically from reality. Data mining by dropping the M_i’s that worked worst still did not attain statistical significance. Overnight, Citibank wiped out its economist section as superfluous. Meantime, inside the Fed, the ancient Federal Reserve Board–MIT–Penn model of Modigliani, Ando, et al. kept being tweaked at the Bank of Italy and at home. For it, M did matter as for almost everyone. But never did M alone matter systemically, as post-1950 Friedman monetarism professed.

It was the 1970s’ supply shocks (OPEC oil, worldwide crop failures, . . .) that worsened forecasts and generated stagflation incurable by either fiscal or central bank policies. That’s what undermined Camelot cockiness—not better monetarism that gave better policy forecasts. No Tom Kuhn case study here at all.

Barnett: Let’s get back to your own post-1936 macro hits and misses, beliefs, and evolutions.

Samuelson: As in some other answers to this interview’s questions, after a struggle with myself and with my 1932–36 macro education, I opportunistically began to use The General Theory’s main paradigms: the fact that millions of people without jobs envied those like themselves who
had jobs, while those in jobs felt sorry for those without them, while all the time being fearful of losing the job they did have. These I took to be established facts and to serve as effective evidence that prices were not being unsticky, in the way that an auction market needs them to be, if full employment clearing were to be assured. Pragmatically and opportunistically, I accepted this as tolerable “micro foundations” for the new 1936 paradigm.

A later writer, such as Leijonhufvud, I knew to have it wrong, when he later argued the merits of Keynes’s subtle intuitions and downplayed the various (identical!) mathematical versions of The General Theory. The so-called 1937 Hicks or later Hicks–Hansen IS–LM diagram will do as an example for the debate. Hansen never pretended that it was something original. Actually, one could more legitimately call it the Harrod–Keynes system. In any case, it was isomorphic with an early Reddaway set of equations and similar sets independently exposited by Meade and by Lange. Early on, as a second-year Harvard graduate student, I had translated Keynes’s own words into the system that Leijonhufvud chose to belittle as unrepresentative of Keynes’s central message.

Just as Darwinism is not a religion in the sense that Marxism usually is, my Keynesianism has always been an evolving development, away from the Neanderthal Model T Keynesianism of liquidity traps and inadequate inclusion of stocks of wealth and stocks of invested goods, and, as needed, included independent variables in the mathematical functions determinative of equilibria and their trends.

By 1939, Tobin’s Harvard Honors thesis had properly added Wealth to the Consumption Function. Modigliani’s brilliant 1944 piece improved on 1936 Keynes. Increasingly, we American Keynesians in the Hansen School—Tobin, Metzler, Samuelson, Modigliani, Solow, . . . ,—became impatient with the foot-dragging English—such as Kahn and Robinson—whose lack of wisdoms became manifest in the 1959 Radcliffe Committee Report. The 1931 Kahn that I admired was not the later Kahn, who would assert that the MV = PQ definition contained bogus variables. Indeed, had Friedman explicitly played up, instead of playing down, the key fact that a rash Reagan fiscal deficit could raise V systematically by its inducing higher interest rates, Friedman’s would have been less of an eccentric macro model.

I would guess that most MIT Ph.D.’s since 1980 might deem themselves not to be “Keynesians.” But they, and modern economists everywhere, do use models like those of Samuelson, Modigliani, Solow, and Tobin. Professor Martin Feldstein, my Harvard neighbor, complained at the 350th Anniversary of Harvard that Keynesians had tried to poison his sophomore mind against saving. Tobin and I on the same panel took this amiss, since both of us since 1955 had been favoring a “neoclassical synthesis,” in which full employment with an austere fiscal budget would add to capital formation in preparation for a coming demographic turnaround. I find in Feldstein’s macro columns much the same paradigms that my kind of Keynesians use today.

On the other hand, within any “school,” schisms do tend to arise. Tobins and Modiglianis never approved of Robert Eisner or Sidney Weintraub as “neo-Keynesians,” who denied that lowering of real interest rates might augment capital formation at the expense of current consumption. Nor do I regard as optimal Lerner’s Functional Finance that would sanction any sized fiscal deficit so long as it did not generate inflation.

FIGURE 3. Left to right in back: James Tobin and Franco Modiqliani. Left to right in front: Milton Friedman and Paul A. Samuelson. All four are Nobel Laureates in Economics.

In 1990, I thought it unlikely ever again to encounter in the real world liquidity traps, or that Paradox of Thrift, which so realistically did apply in the Great Depression and which also did help shape our pay-as-you-go nonactuarial funding of our New Deal social security system. In economics what goes around may well come around. During the past 13 years, Japan has tasted a liquidity trap. When 2003 U.S. Fed rates are down to 1%, that’s a lot closer to 0% than it is to a more “normal” real interest rate of 4% or 5%. Both in micro- and macroeconomics, master economists know they must face up to nonstationary time series and the difficulties these confront us with.

If time permits, I’ll discuss later my qualified view about “rational expectations” and about “the New Classicism of Say’s law” and neutrality of money in effectuating systemic real-variable changes.

Barnett: What is your take on Friedman’s controversial view that his 1950 monetarism was an outgrowth of a forgotten subtle “oral tradition” at Chicago?

Samuelson: Briefly, I was there, knew all the players well, and kept class notes. And beyond Fisher–Marshall MV = PQ, there was little else in Cook County macro.

A related and somewhat contradictory allegation by David Laidler proclaimed that Ralph Hawtrey—through Harvard channels of Allyn Young, Lauchlin Currie, and John H. Williams—had an important (long-neglected) influence on Chicago’s macro paradigms of that same 1930–36 period. Again, my informed view is in the negative. A majority of the Big Ten courses did cite Hawtrey, but in no depth.

Before comparing views with me on Friedman’s disputed topic (and after having done so), Don Patinkin denied that in his Chicago period of the 1940′s any trace of such a specified oral tradition could be found in his class notes (on Mints, Knight, Viner), or could be found in his distinct memory. My Chicago years predated Friedman’s autumn 1932 arrival and postdated his departure for Columbia and the government’s survey of incomes and expenditures. I took all the macroeconomic courses on offer by Chicago teachers: Mints, Simons, Director, and Douglas. Also in that period, I attended lectures and discussions on the Great Depression, involving Knight, Viner, Yntema, Mints, and Gideonse. Nothing beyond the sophisticated account by Dennis Robertson, in his famous Cambridge Handbook on Money, of the Fisher–Marshall–Pigou MV = PQ paradigm can be found in my class notes and memories.

More importantly, as a star upper-class undergraduate, I talked a lot with the hotshot graduate students—Stigler, Wallis, Bronfenbenner, Hart—and rubbed elbows with Friedman and Homer Jones. Since no whisper reached my ears, and no cogent publications have ever been cited, I believe that this nominated myth should not be elevated to the rank of plausible history of ideas. Taylor Ostrander, then unknown to me, did graduate work on the Midway in my time and has kept copious notes. I have asked him and Warren Samuels to comb this important database to confirm or deny these strong contentions of mine.

Having killed off one 1930s’ Chicago myth, I do need to report on another too-little-noticed genuine macro oral tradition from the mid-1930′s Chicago. It is not at all confirmatory of the Friedman hypothesis, and is indeed 180 degrees opposed to that in its eclectic doubts about simplistic monetarism. Nor can I cogently connect it with a Young–Hawtrey influence.

You did not have to be a wunderkind to notice in the early 1930′s that traditional orthodox notions about Say’s law and neutral money were sterile in casting light on contemporary U.S. and global slumps. Intelligently creative scholars such as Simons and Viner had by the mid-1930′s learned something from current economic history about inadequacies of the simple MV = PQ paradigm and its “M alone drives PQ” nonsequitur.

Keynes, of course, in shedding the skin of the author of the Treatise, accomplished a virtual revolution by his liquidity preference paradigm, which realistically recognized the systematic variabilities in V. Pigou, when recanting in 1950 from his earlier bitter 1936 review of The General Theory, in effect abandoned what was to become 1950-like monistic Friedmanisms.

Henry Simons, to his credit, already in my pre-1935 undergraduate days, sensed the “liquidity trap” phenomenon. I was impressed by his reasonable dictum: When open-market operations add to the money supply and at the same time subtract equivalently from outstanding quasi-zero-yielding Treasury bills that are strong money substitutes, little increase can be expected as far as spending and employment are concerned. Note that this was some years before the 1938 period, when Treasury bills came to have only a derisory yield (sometimes negative).

Experts, but too few policymakers, were impressed by some famous Viner and Hardy researches for the 1935 Chicago Federal Reserve Bank. These authors interpreted experience of borrowers who could not find lenders as a sign that during (what we subsequently came to call) “liquidity trap times” money is tight rather than loose: Safe Treasury bills are cheap as dirt just because effective tightness of credit chokes off business activity and thereby lowers the market-clearing short interest rate down toward the zero level. Hoarding of money, which entailed slowing down of depression V, is then not a psychological aberration; rather, it is a cool and sensible adjustment to a world where potential plenty is aborted by failures in both investment and consumer spending out of expectable incomes (multiplier and accelerator, rigidity of prices and wages, etc).

Go back now to read Friedman’s article for the 1950 International Encyclopedia of the Social Sciences, where as an extremist he plays down (outside of hyperinflation) the effects of i (the interest rate) and fiscal deficits on V, to confirm that this Simons–Viner–Hardy Chicago oral tradition is not at all the one he has for a long time claimed to be the early Chicago tradition. (In his defense, I ought to mention that Friedman had left Chicago for Columbia by the time of the Viner–Hardy publications.) The commendable 1932 Chicago proclamation in favor of expanded deficit fiscal spending was itself a recognition of the limited potency of ∂(PQ)/∂M. In terms of latter-day logic, a consistent Friedman groupie ought to have refused to sign that 1932 Chicago proclamation. Meantime, in London, Hayek’s 1931 Prices and Production had converted the usually sensible Lionel Robbins into the eccentric belief that anything that expanded MV or PQ would only make the Depression worse!

Pages: 1 2

Related posts: