Edmund S. Phelps


Becoming an economist: From early preparation to my new direction
For decades, my research was driven by outstanding problems in macroeconomics: mainly growth theory and employment theory. Then, around 1990, my research turned to the study of economic systems and my development as an economist took on added dimensions. This biography will start with my student period and go on to my new direction.1 The two periods seem to me to be linked and I will be drawing some of the connections. I will recount too the intervening period when macroeconomics was in revolution and I was one of the revolutionaries. So this biography gives my account of the battles that started in the 1960s between, on the one side, those economists who wanted macroeconomic models to have lifelike actors whose expectations and beliefs were causal forces and, on the other side, those who did not.

A taste of creativity in Hastings
I am not sure that growing up give full weight to the role of the early education of economists in determining their direction and ambition. When my Prize was announced last October the town in which I grew up, Hastings-on-Hudson, N.Y., proudly claimed six Laureates! Newspaper writers asked whether there was “something in the water.”2 Most of those interviewed thought it was something in the school system. It turned out that four of those Nobels had been won by scientists who were not sons of Hastings, merely commuters to universities and research hospitals in New York City. The other two Nobel winners were genuine products of Hastings and its public school – the economist Robert Merton (son of Robert Merton, the Columbia sociologist, who was a dear friend) and me. I entered the school in 1939 and graduated in 1951. The best part of the high school in Hastings must have been the Music Department. Its orchestra and concert band did well in county competitions and the dance band formed by its students was the best in the region. I played lead trumpet in all of them. They were a gateway to appreciating the best professional musicians of the time. Music was always a part of my life. (One Sunday, at little more than age 8, I had been listening to the New York Philharmonic broadcast when programming was interrupted with news of Japan’s attack on Pearl Harbor.) In the postwar years I marveled at the musical imagination in the big band hits and the outsized personality of the famous soloists. I wanted to play like Harry James. Later I discovered radio station WBAI and was awestruck by the innovativeness of the modern jazz giants Hawkins, Parker and Gillespie. At Carnegie Hall and the Lewisohn Stadium I listened to William Vacchiano’s trumpet seeming to set the tone for the New York Philharmonic.3 I grew up, then, with towers of creativity around me. So I tried to be creative too. A musicology course, I remember, required us to harmonize a C major scale. I don’t suppose I did that very well but that I was able to do something interesting with it surprised me and made me believe that most of us sometimes find inspiration that we could not have imagined beforehand. The high school was also unusual in giving a course in “creative writing.” The assignments were my first experience at inventing a fictional world – a kind of modeling.

I reached the conclusion, though, that whatever my creativity in music or writing might have proved to be, it was not promising enough to warrant attempting a career as a performing artist or professional writer. My hunch was that I would be able to exercise the same or greater creativity in some other field. So I looked forward to the explorations of other fields in college.

Humanities and philosophy at Amherst
I attended Amherst College from 1951 to 1955. The first two years were a revelation. There were innumerable exchanges with brilliant classmates, among them the playwright Ralph Allen, the classics scholar Robert Fagles, and the composer Michael Sahl. Above all, there were some first-year courses that left a lasting mark. Decades later I could feel the full imprint they left on my outlook – in political economy and in my own life.

In English 10, no doubt as eccentric a course as was ever taught, the question was, “What is Amherst?” referring to the town of Amherst, Massachusetts. The answer gradually dawned on us after numerous class hours of being baffled and feeling stupid. There is no unique and true reality, only diverse representations of the town. In the diaries of Emily Dickinson (who lived there) the town is one thing, in a map of Massachusetts it is another thing. (By coincidence, at the Nobel ceremony I heard in the Citation of Orhan Pamuk that in one of his books he asks, “What is Istanbul?” The city could be described in terms of its bricks and mortar or its business life; Pamuk argued that another view came closer to Istanbul’s distinctiveness.4)

After so many weeks of building to the point, none of us would be likely to forget it. Certainly I kept in mind that many among a diverse set of perspectives on some mechanism or system may capture some aspect of the truth that the others miss. A nation’s economy is more than its markets, tastes, technologies and property rights. Capitalism is a system for grinding out GDP, yet it is simultaneously something totally different.

The first-year humanities course was enthralling. We read the Greek epics and dramas of ancient times, Cellini’s Autobiography from the Renaissance, Chaucer and, if I remember well, Cervantes. Why was Cellini there? Why were any of them there? It has only been in recent years that I have become conscious of what the intent must have been. We were being instilled with the philosophy called vitalism. Cellini’s Autobiography was there because of its open celebration of his success – his ambitions and his “making it.” We were being shown the value of a life in which sights are set high and determinedly pursued. Yale’s celebrant of vitalism Harold Bloom was not yet around then but Columbia’s Jacques Barzun was sending that message in his books.

Two other courses, one a first-year introduction to philosophy and a second- year course in American philosophy made a deep impression on me. In the first course we read just three philosophers – but what an inspired choice. There was Plato’s Republic, where the “shadows on the cave wall” suggested the inevitable incompleteness of human understanding, and some of his Dialogues, including the one on social responsibility. Next came Hume insisting on the role of the “passions,” not mere reason, and the importance of “imagination” as against some historical determinism. Last was the brilliant Henri Bergson. His upholding a life of “becoming” over that of mere “being” and his championing of “free will” over “determinism” made him the outstanding 20th century interpreter of vitalism. (He won a Nobel Prize in 1925.) In the second course the star was William James. His embrace of new problems and perspectives and his energy in expounding his ideas conveyed a vital way of living as well as a philosophy of vitalism.

To be sure, the many courses I took as part of my “major” in economics were important too but in a different way. Two courses with James Nelson and three with Arnold Collery were crucial to my decision to pursue economics further in graduate school. Both also provided me with information of use in deciding where to do my graduate work.

In autumn 1954 of my senior year, Nelson and Collery prevailed on Paul Samuelson to pay a visit to Amherst to meet and possibly recruit to MIT the four outstanding majors in economics that year. I had enjoyed enormously the brilliant style and the frequent humor of his textbook Economics: An Introductory Analysis and had read some of his articles too. His lecture was the most forceful I had heard – on Austrian business cycle theory, by the way. Of course I saw my interview with him as a privilege. It was the beginning of a long-lasting friendship. Among the economists I knew I admired Paul the most – admired him for his amazing breadth, unsurpassed clarity and penetrating insight – and loved him for his scrupulousness, seriousness and fairness. (When I asked him once why he alone in the world always referred to the natural rate as the “Phelps-Friedman natural rate” he replied, “because that’s the way it was.”) So I nearly burst with pride when in a packed hall with some eight Nobel Prize winners Paul gave the keynote address to the Festschrift conference for me in 2001. Later, I had the great pleasure of seeing a paper of mine in tribute to him appear in the 2006 Festschrift volume in his honor. Yet, as fate would have it, much of my work – the work after what Paul once called my “blue period” – has at times moved me far from his kind of economic model, most clearly my recent line of research on capitalism.

In the end I decided on graduate school at Yale rather than MIT. It was all I hoped for and more. Yet I will never know what inspiration I might have drawn from Paul and from Robert Solow, who was then bursting on the scene.

Two curents in graduate school at Yale
My courses and interactions with Yale’s greats in the last half of the 1950s – William Fellner, Jacob Marschak, Tjalling Koopmans, Gerard Debreu, Robert Triffin, Henry Wallich, James Tobin and Thomas Schelling – led me to deep parts of existing economic theory and resolved whatever doubts I had over whether economics was for me. Part of the graduate school experience is one’s fellow students, of course, and I recall interactions with Robert Aliber, Paul MacAvoy, and, from other classes, Duncan Foley and Sidney Winter.

At Yale, it is fair to say, the current of Keynesian economics ran very strongly even in the 1950s, when Yale was full of intellectual diversity. What I learned from the leader of the American Keynesians James Tobin, who very generously gave me an individual reading course, and, in my last year, from the young Arthur Okun became an important part of my toolkit – something I was to build on and to modify. Tobin was stunningly intelligent. After one of the more impenetrable-seeming talks given in the Tuesday Cowles Foundation seminar there would be a long silence until Jim quietly asked a question that invariably showed the rest of us what the argument had been or ought to have been. He was so lucid that if there was a weakness in his argument you had a chance of spotting it. When he did not understand something he would modestly ask whether somebody could “make sense” of it for him. And he was self-depreciating to a fault – not an ideal role model for those of us who would be joining the theory wars. Yet I think those who admired Jim greatly, which I did, also could not help noticing an unreasoning conviction in some of Jim’s beliefs about the economy’s workings.

Another current of economic thought at Yale appears to have had an even more enduring influence on the character of my work. Yale, as I noted in the Banquet speech concluding the Nobel ceremonies, had as many Europeans among its economics professors as Americans (or so it seemed) and as many non-Keynesians and Keynesians. Fellner, Marschak and Wallich made Yale a hotbed of the ideas that germinated in Mitteleuropa in the interwar years. I saw Schelling with Carnap on his desk and young Alan Manne was quoting Russell. From Schelling I grasped why it may be crucial to “abandon symmetry” in modeling decisions in circumstances of imperfect information where it is necessary to make conjectures.5 (I suspect that my 1983 paper on the prospect for a successful result from the central bank’s announced intention to bring about disinflation owed something to Tom’s pioneering paper on “surprise attack.”6) From Fellner I caught his fascination with reshaping probability theory to deal with decision making under Knightian uncertainty.7 And it was Fellner who imported to Yale the Austro-Hungarian idea, familiar to Mises, Hayek, Lerner and others, that came to be called the “natural” unemployment rate.8 At a 2000 celebration of 50 years of economics at Yale I saw that few of those present knew of Fellner’s influence: I had to tell them about the view of the world he taught and his contributions to economics.

My 1995 autobiography credited both the Amherst and the Yale economists for the stimulation and preparation they provided me. Yet it slighted the influence of Amherst’s education in the humanities and philosophy and Yale’s unique education in the real character of an enterprising economy on the view I took in my study of capitalism over the past 15 years. I am trying to rectify that in this autobiography.

Efforts at Yale to broaden growth economics
My very first job after the Ph.D. was at RAND. It was understood that one might work on one’s own projects in the time left after the company projects and I took advantage of the remarkable personnel there to talk at one time or another with Richard Bellman, Richard Nelson, Albert Madansky, John McCall, Alain Enthoven, Bart McGuire, Harvey Wagner and others. After lunch I usually headed for the library where I would read in the areas of operations research and dynamic programming. The Naval Logistics Quarterly was a special favorite. My RAND project involved a two-dimensional dynamic programming problem, which I managed to solve. (Kenneth Arrow remembered it a couple of years later, but too late to include it in a volume he was editing.) I was to benefit from this background a half-dozen years later when I began modeling the hiring and wage setting decisions of a decentralized firm. Nevertheless, I soon reentered the academic job market and left RAND in summer 1960, after a little more than a year there.

My academic position was right back at Yale, where I remained, apart from leaves, to the end of academic year 1965–66. My closest pals on the faculty were probably Gustav Ranis, Bela Balassa, David Cass and Arthur Okun. I had a goodly number of outstanding graduate students, including Mordecai Kurz and Seong Yawng Park, who wrote dissertations with me, and in the classroom Guillermo Calvo, Susan Rose Ackerman and Uwe Rheinhart.

My appointment was in the Cowles Foundation, so I had reduced teaching – a course load of 2 and 1, the same as now. A word about my teaching might be in order. In the first few years I worked as hard as needed to attain what I thought was a level of acceptability. I was amazed to hear Paul Steiger, the managing editor of the Wall Street Journal for the past couple of decades, say that I was the best teacher he had in Yale economics and was the reason he changed his major to economics. I may have exceeded my optimum level of effort. I was also surprised when Uwe Reinhart, the Princeton public policy economist, told me he keeps his notes from a graduate course I taught at Yale. My teaching was to decline in the future until I finally created a course to which I became increasingly devoted over the 1990s: a course on the political economy issues that drove many developments in the twentieth century.

In my first three years or so at Cowles, including a year away at MIT in 1962–63, I worked in the classical mode. Robert Solow along with Trevor Swan and some others were producing a fast-growing literature on economic growth – mainly from a policy point of view. Several papers of mine fit perfectly into that literature, most notably, the one on the golden rule and the one on accumulation of capital with a known probability distribution of returns. Some of these made me known in the profession at an early age. (Many explored the golden rule, including Solow, and Samuelson expanded the risky capital model, which sparked a wave of papers on intertemporal portfolio choice.) I took considerable pleasure in crafting them too.

My first step outside this box was a paper on technology formation. The Solow model supposed that new technologies rained down on the world economy like manna from heaven. I set out to model technological progress as produced, thus endogenous: a production function linked technology increase in the world to the volume of research. The function I posited had Isaac Newton’s “standing on the shoulders of giants,” later made famous by Bob Merton, which other modelers had not included. This led to a golden rule of research alongside the golden rule of accumulation. More strikingly, the model had the lesson that the world’s population growth must be credited with making possible much of the technological progress in the past two centuries.

I became restless, though, with the classical character of this growth economics as it had developed at MIT and Yale, with its postulates of perfect information and complete knowledge. Each new technology was instantly and costlessly implemented by producers and consumers, as if there was no need for managers to cope with change. There were not even any decentralized savers or workers in the theory. It was almost like classical physics. I was also guilty. I liked to say, evoking the two laws of thermodynamics, that Bob Solow’s result – an increase in the saving rate would increase output per worker – was the First Law of Economic Dynamics; and my golden rule result – it would increase consumption per worker only up to a point and thereafter decrease consumption – was the Second Law. Robert Summers (father of Larry Summers) told a story in his class at Yale of a puzzled student who, after seeing the equations of an economic model, asks the professor, “sir, where are the people in the model?” When a journalist asked me minutes after the announcement of my Nobel what had been the thrust of my work, I remembered that story and answered that I had tried to “put people” into macroeconomics. That was a bit much, since Solow too and others including David Cass had sooner or later put in some people, some consumers at any rate. We were all aware we did not live in a centralized economy. However, the mainstream growth model continued to make the classical postulates of perfect information and complete knowledge.

My move away from the classical began with a couple of works in the mid-1960s. A paper Richard Nelson and I published in 1966 recognizes that in a market economy any new technology introduced by an entrepreneur would not be immediately adopted by all managers or consumers, contrary to the costless and instantaneous implementation supposed by the standard growth model. Of course, entrepreneurs who fear such costs and lags may be scared off from developing innovations that they would otherwise see as profitable. Nelson and I went on to argue that a liberal arts education serves to enhance the capabilities of managers to evaluate new methods, thus to speed up diffusion of new technologies and possibly to encourage development of new innovations. This paper seemed to have a delayed fuse timed to go off in the 1990s, when it was picked up in books and papers by Robert Barro and Xavier Sala-i-Martin, by Philippe Aghion and Peter Howitt, and by Amar Bhidé.

An early exploration in a less classical vein was my book Fiscal Neutrality toward Economic Growth. It questioned the presumption that competitive markets do well from the point of view of intertemporal allocation, owing to the difficulty households must have in estimating their true wealth net of fiscal burdens. A neutral fiscal policy would undo the disequilibrating effects of such misestimation: it would gear people’s expectations of over-life tax burdens to what will be required to pay for public programs; thus it would tend to tax-finance public outlays other than for projects that will charge user fees. Running a budgetary deficit is apt to cause overconsumption and under-supply of labor. The book was a first try at thinking about mis-expectations – about economic disequilibrium and its correction. Whatever its importance may be, for me it was significant as a breaking away from some of the implicit assumptions of competitive equilibrium theory.

There was a story about that work. Shortly before publication I gave a Cowles seminar on my new theme: a good fiscal policy would collect just enough tax revenue so as to cause households to feel as rich as they really were – that is the presumption, at any rate. Jim Tobin was visibly annoyed. He insisted that we cannot understand the effects of fiscal policy without allowing for the role of money in the economy. I probed to see the basis for that belief but did not find any basis, at least none I perceived. (Jan Tumlir, a young colleague at the time, said it was the most scintillating exchange he had heard at Yale.) That seemed odd to me and somewhat worrying. Such resistance was to break out in full force in the 1970s when several neo-Keynesians railed against the natural unemployment rate. Later I thought I detected a similar spirit in some proponents of New Classical economics. This may be inevitable. In any field some tenets are very stubbornly held.9

Later, in a paper with Robert Pollak written at Penn and published in 1968, the subject was what households are to expect about the consumption of future descendants – or, possibly, their future selves. The assumption that “they” will consume as “we” in the present would like them to is not at all general: they might have a selfish streak, just as may do, and thus save less than we would like. But how much saving would that be? Pollak and I were able to work out a sort of game-theory solution – a step back to intertemporal equilibrium, in which future people save as they are expected to do. But it is not a good competitive equilibrium, as there is (in a well-defined sense) under-saving by every generation.

Bringing expectations to employment models at Penn
By the middle of the 1960s, I began to be aware that neither I nor anyone else was addressing what I felt ever since college was the most important challenge in economics: to integrate microeconomics and macroeconomics. Finally I decided to try to do it myself! I had already concluded that textbook micro, which was based on the classical model of perfect competition, truly was irreconcilable with anything like the prevailing macro. I suspected that any macro that we might find recognizable would have to be based on some different kind of micro – a micro in which, say, individual firms with their employees or customers were in imperfect communication with other firms and, more generally, the rest of the economy. (Later I was to think also of widely separated islands of workers, each island reachable from the others only by a slow boat.) I immersed myself, with a few breaks to do other things, in a project to rewrite the Keynesian economics of employment determination from the first half of 1966, when I started work on the project at London School of Economics (LSE) and Cambridge during a sabbatical leave from Yale, through 1971. The conference I organized on the subject took place in Philadelphia in January 1969, financed out of a threatening surplus in my National Science Foundation (NSF) research account. The conference volume Microeconomic Foundations – later known as “the Phelps volume” – was published in 1970. A monograph of mine that was a sort of sequel, Inflation Policy and Unemployment Theory, was written largely over 1970 and appeared in 1972. (Most of my papers in this area were published in my collection Studies in Macroeconomic Theory, Vol. 1, 1979.)

For the great bulk of this period, specifically from 1966–67 to 1970–71, I was a professor at the University of Pennsylvania. The advantage of Penn relative to my preceding home and maybe to my future home as well, was that there were not any established macroeconomists on the faculty and the younger ones there were not focused on employment or inflation, so I was not challenged to defend my thinking against the views of colleagues. This was also the disadvantage, of course, since there were few around who might alert me to mistakes. Still there were plenty of first class minds. I had friends in Karl Shell, Robert Pollak, Edwin Burmeister and Edward Prescott (from oldest to youngest). The first three all wrote a paper with me – Karl one about the incidence of public debt and Edwin one about which state agency to assign inflation control. (The paper with Bob was discussed above.) Ed Prescott gave me expert advice on my 1972 paper on what I dubbed “statistical discrimination” – the idea, introduced in my 1972 book on inflation policy and unemployment theory, that a person may be judged by the labor market in part by the group or groups in which he or she is categorized. I also engaged Ed in a discussion of that book’s digression on “routine stabilization.” At Penn there was also a brilliant undergrad, Steven Salop, who worked for several months on my 1968 paper on money-wage dynamics.

Confusion abounded at that time on the subject of employment determination. Keynes had said that, in general, one can expect “involuntary unemployment,” by which he (though no one before or since) meant a sort of under-employment that results from overestimates of prevailing labor market conditions causing the population to hold back some labor; the limiting case was called “full” employment, it being supposed that there could not be anything that might be called “over-employment.” If we departed from Keynes by allowing the existence of over-employment, resulting from underestimates of labor market conditions causing the population to offer supra-full amounts of labor, then the economy could be seen as ambling between underemployment and over-employment. Would there be a tendency toward recovery toward the “full,” or equilibrium level? He didn’t say yes, though he didn’t say no. (Nor did he say whether this equilibrium level might itself tend to be too low.) Keynes himself was not sanguine about the central bank’s ability to narrow employment swings around the equilibrium level, which might itself be subject to swings and shifts, of course. But in a time when employment and the price level had sunk to very depressed levels Keynes viewed interest rate cuts by the central bank as a means to pull up employment – and the price level with it.

Unfortunately, A. W. Phillips came upon an empirical relationship between the measured unemployment rate and the inflation rate – a negatively sloped curve, which was dubbed the Phillips Curve. There soon developed a neo-Keynesian aggregative model based on the postulate of such a “longrun Phillips Curve.” This model implied that a step-increase in the stock of money or its velocity would increase output and jobs at first but, as the price level rose, business activity would be driven back to equilibrium. But this same model implied the feasibility of a permanent increase of employment through successive increases in the money supply to accommodate successive increases of the price level – a vision beyond what Keynes saw to be the promise of demand management. This seemingly rock-solid Phillips Curve put the American Keynesians into a sort of euphoria, as if they had discovered atomic power. They, unlike Keynes himself, had long harbored the faith that the central bank was able to dial the average level around which employment would fluctuate – a claim the Continental monetary theorists Mises and Hayek had denied, arguing that the inflation rate would spiral upward if the central bank set too low an unemployment rate and spiral downward if the bank set too high an unemployment rate. With the establishment of an empirical Phillips Curve that was breathtakingly stable, there seemed to be no reason for the central bank to hold back from supporting low unemployment.

In the view I took in papers written in 1966 and 1967 (published in 1967 and 1968, respectively), it is a mistake to posit any such long-run negatively sloped Phillips Curve – to posit a long-run relationship running from long-run inflation rate to long-run unemployment. So a monetary policy shift that decreases the target inflation rate – the sort of event studied in my 1967 paper – should not be supposed to aggravate the level to which the unemployment rate tends in the long run; symmetrically, an increased tolerance for inflation should not be supposed to usher in a long-run decrease of unemployment. Even if there is a short-run Phillips Curve, such a curve takes the expected rate of inflation as a given; this expectations-augmented Phillips Curve will lie higher the greater is the given expected inflation rate. I then argued that a central bank policy to establish and maintain (un)employment at a level that would always represent, say, over-employment, so that the inflation rate would always exceed the expected inflation rate, would sooner or later cause the expected inflation rate to be rising and thus cause the economy to suffer an ever-higher Phillips Curve. This state of affairs could not be sustained.

The 1968 message, then, was that the central bank could not permanently maintain over-employment. The 1967 message was that it could engineer reduced inflation expectations and thus reduce actual inflation without causing permanent underemployment – but not without causing a temporary bulge of unemployment.

These two basic conclusions have apparently survived the controversies that arose over this analysis. The corollary was the policy message of the 1967 paper: the central bank’s main responsibility must be seen as controlling inflation – by means of a money-supply or interest rate policy aimed at managing inflation expectations. Inflation will still be capable of ups and downs but it cannot go far if the expected inflation rate is under control.

It is sometimes asked whether there was a Eureka moment in all this. The closest I came to that was when I realized early one morning in September 1966 that in the theoretical setting I was focused on, it is wage expectations – what employers and employees expect wages at other firms to be over the near future – not price expectations – their expectations of what the price level is going to be – that are crucial for firms’ wage setting and employees’ quit rates. (I took the poetic license of closing down workers’ price expectations by taking labor supply as perfectly inelastic.) The difference in setting and in the consequent expectations mechanism lay behind the contrast between my results and those of Milton Friedman, also published in 1968. His analysis was in fact directed to the natural level of labor-force participation while mine was actually directed to unemployment. Yet there were strong parallels in our conclusions and it was inevitable that we were soon yoked together – except for that minority of journalists who referred always to “Friedman’s natural rate of unemployment”!

I was not the only economist beginning to model unemployment determination from a micro point of view. Work by several others began to appear in manuscript or preliminary form over 1968, of which a paper by Dale Mortensen was closest in broad outline to what I was doing. There was also the paper on customer markets that Sidney Winter and I were producing that year. Thinking that there is strength in numbers I decided to try to organize a conference on the subject at Penn in January 1969, drawing on some unused funds in my (NSF) grant. A paper by Armen Alchian offered an overview of informational imperfections in product and labor markets. Focusing on the labor market were papers by Mortensen, Charles Holt, my 1968 paper (modified but not for the better) and, with a quite different model, the paper by Robert Lucas and Leonard Rapping. The Phelps-Winter paper and some others focused on the product market. This almost instant convergence of so many scholars on so new a subject was exciting to see and such openness to new ideas made me admire the profession. It was also something special to discuss questions about future directions with a group so remarkably talented. The exchanges with Bob Lucas, who was already mapping out his own journey, were especially interesting. We both expressed regret at one time or another that we did not have more conversations.

The conference volume Microeconomic Foundations appearing in March 1970 sent shock waves through the economics profession – certainly the younger generation. The blurb on the back of the book’s jacket by Robert Hall clearly suggested it marked a paradigm shift for macroeconomics. Right up to recent times I have heard economists tell how excited they were to have the book in their hands and to start poring through it. This publication was the biggest “high” in my scientific experience. A huge advance had been made but the war had really just begun!

There was stiff opposition from a great many “Keynesians.” It was more than I would have thought likely since, in my mind, my work was not radically opposed to Keynes. It clearly recognized, with Keynes, that the central bank has the responsibility of managing “effective demand” – to keep it under suitable control – and that doing so would help to foster employment stability. If a change in “effective demand” spontaneously erupted, an ideal monetary policy would act to close the gap. My work eventually pointed out that if a supply shock occurred, such as an overseas shock to the world price of energy, an ideal monetary policy might make some response. (Such a shock might wreak wage deflation, absent any response.) Though not a wanton interventionist, I did not presume that the right policy was always laissez-faire – known now as “free market” economics. I understood that every real-life market economy was neither efficient nor just – not even close to it – and there were some limited measures that the government might usefully do to improve its performance.

Some neo-Keynesians claimed my work could not be right, arguing that it implied “acceleration” of prices and wages whenever the economy was away from the “natural rate” while the data showed that a mounting rate of disinflation was not characteristic of the 1930s – the years of the Great Depression out of which Keynes’s theory was born. It is true that my 1967 paper implied mounting acceleration of the price level (i.e., an increasing inflation rate) if the central bank were to hold employment steady at a level above the “equilibrium” level – which was a constant in that first model, not the path it became in the 1968 paper. The expected inflation rate was rising, pushing up the actual inflation rate accordingly, because the actual inflation rate was always greater than the expected rate, owing to the postulate of “adaptive” expectations. That postulate was particularly reasonable in the context of that paper: if the central bank were to decide to try to drive down the expected inflation rate, credibility issues would require the bank to generate inflation rates that were steadily below expectations and thus to drive employment to a depressed level in order to convince an initially skeptical public of its determination to wring inflation expectations out of the economy. It was realistic to suppose expectations would be adaptive in such circumstances: that they would gradually recede in response to their constant disappointment. But adaptive expectations would not appear to be realistic in all other circumstances.

In a time of economic depression and resulting deflation it would not be realistic for the public indefinitely to expect the rate of deflation to be greater than it was in the last observation – certainly not once it was crystal clear that the central bank was not trying to increase the rate of deflation and thus had no desire to hold employment below its natural level. A point may very well have come in the decade, possibly around 1934, where the price level and employment too had sunk so far that people began to expect “reflation” and recovery. Then the price level might level off, the expectations of inflation just counterbalancing the deflationary force of the deficiency in effective demand. Yet that would be a disequilibrium condition, not a “low-level equilibrium” and in no way a disconfirmation of the hypothesis of the existence of a natural rate. (I remarked on this in a review article, “On Okun’s Micro-Macro System,” Journal of Economic Literature, September 1981.) However, the criticism by neo-Keynesians paid no attention to my 1968 paper in which, starting from a depressed level of employment, there would be an equilibrium path leading toward the medium-term natural rate. Along this path, despite massive joblessness, the actual inflation rate would not be below the expected inflation rate (nor above either). Thus the accompanying price level and general level of money wages might be flat or even rising.10

There was criticism too from the emerging New Classical School. This was directed not against my thesis of a natural rate, which the New Classicals embraced, nor against my insertion into macro modeling of wage or price expectations, which the New Classicals also embraced, but against the treatment of expectations behavior (by me and others) as adaptive. The argument, as I heard it, was this: if, whenever employment and prices or wages are pulled up into the high-employment region (above the natural level), the economy’s participants were to go on under-forecasting the inflation rate (beyond the inflation surprise that might well be caused by the initial shock driving up employment), that regularity would mean either the public was being obtuse in not soon grasping the understanding contained in the analyst’s model or the public has an understanding not given to the analyst – the public was using a model that differs in this regard from the model the analyst is using. Since it is impossible to believe that participants would commit the same forecasting error over and over again, it must be that it is the analyst’s model that is “wrong:” it has failed to pick up something that the participants’ model takes into account. The implication was that I and the other microfoundationists were using ‘wrong’ models. This was impressive. I appreciated the brilliance and the seriousness behind the argument. To the best of my recollection I never derided it or sought to defeat it by rhetoric alone.

The argument made my head spin. I understood where they wanted to go: they saw the immense analytical convenience for an analyst of imputing to the actors in their decision making a use of the very same model the analyst is using. (I had considered the “perfect foresight” case in my Fiscal Neutrality as had Robert Hall and others in the 1960s.) But I kept asking myself whether it followed that my “wrong” models are illegitimate or inferior to some other. How could a real-life analyst’s model not be wrong? I recalled that every model of Amherst has its limitations and many have their merits. A model abstracts from details and sometimes from complexities that are present in reality. So it may be common and almost inevitable that there are some enduring misalignments between what the analyst’s model says the participants would do if they knew and subscribed to that model and what they actually do. In some respects, some participants may very well be more knowledgeable than the analyst or the planner, as Friedrich Hayek taught.

A propos the issue of inflation expectations in particular, participants may suspect that, although employment has moved up into what might be supposed to be the zone above the natural level, the natural level may have moved even higher, so that employment, while perhaps historically high, is not actually higher (and may be even lower) than the natural level currently is. Such a structural shift might have preceded the demand shift or have occurred jointly and thus at the same time. So, on observing a movement of employment to historically high levels, people might reasonably give some probability-type weight to the possibility that the underlying shock, whatever it is, has pulled up the natural level of employment even more than the actual. (Papers I wrote with Hian Teck Hoon and with Gylfi Zoega between 2000 and 2003 showed how an investment boom, which surely increases effective demand, shifts up the nonmonetary equilibrium path too.)

A further point: I did not remember postulating a Cagan-Nerlove adaptive expectations equation as a general rule. (Such an equation is simply not in the model presented in my 1968 paper, which is very largely about equilibrium paths, not disequilibrium paths.) My thesis could not be fairly summarized as the idea that employment is slow to equilibrate because expectations are adaptive – as if this were a piece of psychology I was importing into macroeconomics. My thesis was that participants, not observing the simultaneous wage and price decisions being taken elsewhere, have to form expectations; and, with participants not knowing completely, if at all, the structure of the economy, the expectations formed are a causal force that have to be reckoned with when we analyze movements in (un)employment and movements in inflation.

I dwell on these two bodies of criticism because they manifested a schism in economists’ views of the economy. The pioneering figures of the modernist view – Knight, Keynes, Hayek, Morgenstern, Zeuthen and Fellner – did not take the economy’s structure, let alone the timepath of that structure over the future, to be completely known. The neo-Keynesians and the New Classicals were both on the other side of this schism, preferring mechanical models of economies of known structure, as Axel Leijonhufvud and Roman Frydman have long pointed out. There was therefore a deep affinity between them – an affinity at the level of methodology. Since this is my biography I have to say that this put me in a strange situation. I was close to them, having warm personal relations with Jim Tobin and Bob Solow as well as with Bob Lucas and Tom Sargent – relations that have survived our differences. But I belonged to neither school.

As I began my gradual departure from Penn I thought back more than once to 1968, a year in which I somehow published three seminal papers – “Money-Wage Dynamics,” Phelps-Pollak and “Population Increase” – and 1967 and 1969 were good too. I wondered whether this could last. It did not. But there were two more periods of unusual creativity ahead: my structuralist modeling from about 1986 to 1992, which depended on what I learned in the 1970s, and my work on capitalism theory from about 1998 to 2006, which depended on what I learned earlier in the 1990s.

From monetary macro to economic justice at Columbia
I left my professorship in the Economics Department at Penn for one at Columbia, starting in summer 1971. Although New Haven and Philadelphia had not been very far away, it was exciting to be right in New York City, with Bernstein at the Philharmonic, Balanchine at the New York City Ballet, Levine at the Met, and plenty of Stoppard and Pinter on or off Broadway. The move brought a momentous change in my personal life. At Columbia in early 1972 I met Viviana Montdor, a strikingly attractive and very cultivated young woman from Buenos Aires. I found her beautiful, smart and very insightful. Her background, which was so different from mine, fascinated me. We were married in 1974. She has been hugely supportive of my career. Viviana brought two children, Monica and Eduardo, to our marriage. Now we have seven grandchildren. I feel extremely fortunate.

Let me say that at Columbia, where my appointment has been for 36 years now, I have been fortunate to have as colleagues Bob Mundell and, in the present decade, Joe Stiglitz. We have had great fun together. In both cases, though, our friendship is based, I believe, on our mutual respect for each other’s work.

The immediate problem to be tackled at Columbia was that, after the Interwar period when it was right up with Chicago and Harvard, it did not have the breadth of first-rate talents that its tradition and location suggested it would have. The relative newcomers among us – the chairman Kelvin Lancaster, Phillip Cagan, Ronald Findlay and myself – set about to strengthen the Economics Department at Columbia. Within three years we recruited Robert Mundell and Carlos Rodriguez to international monetary economics, Guillermo Calvo and John Taylor to macroeconomics, and Phoebus Dhrymes to econometrics. In addition, there were Jacob Mincer and the young James Heckman manning labor economics and William Vickrey covering public finance. (Yet the Wall Street Journal wrote of a “decline” of Columbia’s economics in the 1970s!) That this renaissance happened at all is more important than the fact that it faded. Some left, including Heckman in mid-decade, Taylor at the decade’s end and Calvo in 1986. In the present decade three newcomers have worked with me: Max Amarante, Amar Bhidé and Richard Robb (the latter two outside the Department). In my Columbia years I have also been fortunate to have several extremely talented graduate students. Among those a part of whose dissertation was related to my work have been Janusz Ordover, Roman Frydman, Juan Carlos di Tata, Luigi Bonatti, Hian Teck Hoon and Gylfi Zoega. And I have basked in the friendship of Graciana del Castillo, Luis Campos Cunha, Alfredo Navarete and Agustin Garcia Lopez ever since their classes with me. Later, New York brought us our close friend Pentti Kouri, who has been a supporter of my career and a help in my work. With my move to Columbia I wanted to take up a new research theme. Every time I moved – from Yale’s graduate school to RAND, from RAND to Yale’s Cowles Foundation and from Yale to Penn – I took a new lease on life. Or, maybe, every time I needed a new lease I moved. When I arrived at Columbia in September 1971, with my book Inflation Policy and Unemployment Theory in final draft, I was itching to leave monetary economics and get back to “real” economics. But what exactly? I had a stroke of luck there. During 1969–70, when a visiting fellow at CASBS, my office was next door to that of John Rawls. My memories of our many conversations were still reverberating and these provided the inspiration for a new start once settled in at Columbia.

The research I began first, though, was not in Rawls’s field at all. Somehow I had come to the idea that, in an economy with imperfect information, there would be social benefits from private altruism and personal morality. (I had read Ibsen’s powerful play “The Enemy of the People” and its parable stayed with me.) It adds to enforcement costs if people hide taxable income, drive through red lights and generally conceal harms they are doing to others. It discourages experiment and change if sellers can be expected to misrepresent unfamiliar products and mislead potential buyers. I went to Eleanor Sheldon, then still at the Russell Sage Foundation, to ask her support for a conference on the subject. She had to fight for it, she later told me, but got the money. The conference ran two half-days. Samuelson, Kenneth Arrow, William Vickrey, Peter Hammond, James Buchanan and I were among the authors. (My paper extended the Phelps-Pollak model to continuous time, which changed some results.) Amartya Sen, Thomas Nagel, Guido Calabresi and Karl Shell were among the commentators.

The conference volume, Altruism, Morality and Economic Theory became a cult hit. To this day the occasional sociology student comes from France to ask me about the conference’s origins and its reception. (They may have known that it was not widely read.) This conception of the market economy was anathema to Chicago at that time. So I knew I was sticking my head in the lion’s mouth when at the invitation of George Stigler I decided to present my draft introduction to the conference volume before Chicago’s Law and Economics seminar. Entering the hall and taking my place, I looked up at steeply raked, semicircular rows of professors peering down at me and seeming to have me surrounded. For what seemed like hours, Gary Becker, Robert Posner, Stigler and others pounded away at me. I held my own, flagging only at the end when I grew tired or fed up. “You were doing so well,” one professor complained to me in a thick Viennese accent, “so why did you make that concession?” In my letter to George with my expenses and thanks I said that “now I know how Dr. Freud felt facing Vienna’s College of Physicians.” (I was comparing the two lecture experiences, of course, not comparing my thoughts on altruism to Freud’s theory of the unconscious.)

I was soon thinking again about Rawls. In papers written over the 1970s I explored the tax mix for a Rawlsian optimum, the inflation tax, optimal Rawlsian public debt and private wealth, and Rawls’s concept of economic justice and consequent issues for welfare economics. (Most of this work and earlier work in the same or a related vein is contained in my collection Studies in Macroeconomic Theory, Vol. 2, 1980.) James Mirrlees’s seminal 1970 paper on optimal wage-income taxation paper provided a necessary tool and the inspiration for an analysis of another problem: the optimal shape of the (generally nonlinear) tax schedule for wage income, which I went on to study in more than one setting. In these settings, I took the Rawlsian “least advantaged” to be an earner whose productivity is bound to be negligible, whatever tax policy may be. Thus this earner’s reward from work is essentially the size of the wage subsidy at the low end of the labor force. So Rawlsian justice meant raising the maximum possible tax revenue. This, it turned out, required setting the marginal tax rate on the last dollar of wage income of the highest earner equal to zero! (Efraim Sadka found the same result in a utilitarian framework. Joseph Stiglitz then analyzed a more general model in which the marginal tax rate at the top is negative!!)

How did Rawls view this result? Did he think it was immaterial? Not at all. He insisted that what is just cannot be determined with finality until “the consequences” of what analysis has provisionally pointed to as just have been explored and determined. I was to have two other significant areas of interchange with Jack over the years. The most fundamental was on the very meaning of his “justice.” A long letter I wrote to him from Amsterdam in 1980 said that his work was being taken as support for a universal basic income – a “negative income tax”, or “demogrant” – while his book clearly states in its first pages that the concept of justice it is concerned with is not some general social justice but rather economic justice in the structure of rewards for active contribution to society’s production. The matter lay there for quite some time. Some movement occured in 1985 when, writing to me about my just published textbook, he said its exposition of his theory of justice was entirely accurate. But that was not a public statement. Finally, in his last publication, one in the form of an interview, he underlined for those who had missed it this feature of his concept of justice. (He said that the book did not apply to “beach bums.”)

I mentioned work on the tax mix. Janusz Ordover, a young economist from Warsaw asked me to advise him on the last chapter of his doctoral thesis. (I was grateful to Ron Findlay for having introduced us.) Together we worked out the answer to the question of the optimal linear tax structure – the configuration of flat tax rates on wage income and the income from wealth – under various possible constraints involving maintenance of a subsequent steady state. When both private saving and public saving are unconstrained the analysis yielded simpler results! In that case the optimum tax rate on income from wealth is indeed equal to zero, as Corlett and Hague would have guessed, though that would not generally hold if the government’s algebraic surplus were constrained to be zero. (The taxation of wealth has been a continual obsession, surfacing most recently in December 2005, always without noteworthy results.) Through the algebraic haze could be seen another result: there exists a level of public debt and associated tax rate to sustain it that maximizes tax revenue. Thus, our paper (American Economic Review, September 1975) could be said to have characterized what it is like to be at the top of the Laffer Curve. For Rawlsians like us, it was good for society to be at the top of that curve. (Certainly it was good for society’s least advantaged participants.) Supply-siders felt that would be bad for innovation. To this day the issues remain unresolved.

Over the decade I also found time for the issues of that time in macroeconomics. Taylor and I, later joined by Calvo, built a series of models of the New Keynesian type. These constituted a sort of answer to the New Classical models coming out of the Midwest. In my 1968 paper I had pointed out that, for a model that would fit quarterly or annual data at all well, it might be important to take into account that firms’ managers cannot afford to spend every waking moment monitoring the economy with an eye to revising their posted wage rates and prices. The Columbia models were a “marriage of convenience” of that important point to the New Classical idea of rational expectations. The purpose was only to show that recoveries would be gradual, or sticky, even if rational expectations prevailed. For me, the models did not signify that the rational expectations hypothesis was a good idea, only that it was an expedient in a first round of modeling.

In fact, new differences between my work and the literature using rational expectations cropped up in those years, notably in my 1977 article “Indexation Issues” and its appendix, coauthored with Guillermo Calvo, “Employment-Contingent Wage Contracts” (Journal of Monetary Economics, 1977). The starting point was the observation that in a capitalist economy, more generally, in any enterprising economy, an element of the so-called state of the world has to be the subjective and only indirectly measured “visions” of the entrepreneurs. In such an economy there are no “rational expectation” of what the entrepreneurs are thinking – there are only their “animal spirits,” as Keynes said, referring to Plato. It followed that the neoclassical theory of the wage contract would be simply inapplicable to such a setting: In that model, the employment contract would link the number of employees who have the misfortune to be called up for work on any particular day to the “state” of the economy; but for the entrepreneur the state is something that he or she feels but cannot communicate to the employees nor would he or she be believed, owing to the temptation to pretend the state requires more labor input than it would if the contract were followed. I was surprised that the sad state of contract theory did not provoke something of a scandal. But apparently no one really cared very much what its implications were and how well or badly they matched up to any rough impression of reality.

A conference that Roman Frydman and I organized in 1981 contained another paper of mine that was different from rational expectations. Rational expectations means “rational” (in a manner of speaking) relative to a model, namely the analyst’s. I pointed to the existence of a more general case – an economy in which there prevails a pluralism of models: some actors in the economy operate on the monetarists’ model, others on the Keynesian model, still others on the supply-side model, and so forth. Here, I suggested, it would be wrongheaded for any of these actors to pretend that all the other actors follow a model the same as theirs; and the same goes for each analyst. Presumably analysts will find no use in the concept of rational expectations equilibrium in this case and will instead try to model the economy – in a stylized way, of course – as it is: to take account of the prevailing pluralism of models. I proposed that analysts of every persuasion will recognize that the monetarists in their model take account of the presence of Keynesians (and other dissidents) in the economy, that the Keynesians in their model take account of the monetarist (and other dissidents) in their midst, and so forth. (This is probably the best interpretation of the term “model-theoretic expectations” that Frydman and I sometimes used.) In my paper for the conference, however, I took only a small step in that direction by studying the case in which, for the sake of simplicity, everyone has the same model of the fundamental workings of the economy but, crucially, they do not know it: everyone believes that everyone else is operating on a different model. The conference volume, Frydman and Phelps (eds), Individual Forecasting and Aggregate Outcomes, was published in 1983.

Writing political economy: from Buenos Aires to Stockholm
When the 1980s arrived there was still an item left on my agenda: a contract with W. W. Norton signed in 1971 to produce a different kind of introductory textbook. Why did I do it? It is not unusual – at least it was not when I was young – that a textbook is written by someone who sees himself as an important voice of his times or who hopes that the textbook will make him one – Samuelson, Fisher, Marshall, … As a pioneer in the 1960s of what I called “modern economics” I felt a need to introduce college students to what this revolution was and to some of the insights to which it led: Here are two passages early in Part 6, the first of the textbook’s two modern parts:11

In the modern style of market analysis, … information and knowledge hold the spotlight. In this modern view, messages are expensive to send, hence information is costly to find, and secrets are cheap to conceal, hence knowledge is not completely shared in most or all markets. The resulting focus of this modern theory is on the roles of imperfect information and incomplete knowledge in the performance of the market economy – on the difficulties the market players have, as consumer and as employer or employee, in acquiring information and knowledge, and the consequences of those difficulties for the functioning of markets …

With the classical magic not there – no costless ‘auctioneer’ …, no unseen ‘regulator’ – the transactors are driven to collect their own information and to make their own arrangements … Thus some aspects of primitive exchange … are brought back: communicating terms, perhaps bargaining, watching out for cheating and other ‘moral hazards,’ aiming for cheat-proof incentive-based agreements, and so forth.

The payoff from any new theory, of course, lies in how it expands our understanding … The modern theory of markets, largely developed in the past two decades, has illuminated a whole universe of observations: the prevalence of the long-lived ‘firm,’ the mark-up of price above unit cost – hence a pure profit called ‘good will,’ undue hazards in products and occupations, reputable firms and deceiving ones, ‘job rationing’ and unemployment (even in equilibrium), discriminatory stereotyping in hiring and lending, layoff and seniority rules, and much else – a world of phenomena not understandable by neoclassical theory. (pages 380–381; italics in the original)

The inner need to produce the textbook was one thing, getting down to it was another. (Evidently I was not one of those Schumpeterian entrepreneurs who “get it done.”) Yet pressures to get it done were mounting. I was grateful to Donald Lamm for bringing me to Norton in 1961, they had made me their advisor in economics and had given me a nice advance. The time had come to deliver. At the same time Viviana was making the case that I ought to get out more into the rest of the world. The solution was a series of visiting faculty appointments overseas in Latin America and Europe where I could work mostly on the textbook and learn more about overseas economies. That also served to break up the large writing project into separate periods in varied settings and to replace the earnings that might have been made in other ways. In these trips Viviana would join me with a lag, accompanied by Monica and our precious dog Shaggy.

The first trip was a longish one in the early months of 1980 to CEMA in Buenos Aires at the invitation of Carlos Rodriguez, who earlier had been at Columbia. I remember to this day working on the first page and a half for two weeks, if not longer, in January 1980. (Odder still, I took a break somewhere in the midst of this effort in order to write what would be my paper for the Frydman-Phelps conference of 1981 and volume in 1983. I may have needed that break to avoid going crazy over pages 1–2.) The rest of the book went faster. But never fast. The problem was that there were few sections, let alone chapters, where I could simply imitate other textbooks. That summer I worked a term in the University of Amsterdam at the invitation of Jan Kremer – signed by the Queen, he said. After a year, I had done only three chapters of a projected 26, though. I was seriously behind now, so I worked harder in mid-1981 at the Getulio Vargas Foundation in Rio, invited by Enrique Simonsen; then even harder in summer 1982 at the University of Mannheim, where we enjoyed being with our dear friends Marlies and Jürgen Schröder in the hamlet of Juhöhe in the Odenwald. Now working frantically hard, I visited the European University Institute in summer 1983 at the invitation of Marcello de Cecco, in 1984 the University of Munich, invited by Edwin von Böventer, and finally, late that summer, Stockholm’s Institute of International Economic Studies, invited by Assar Lindbeck. Viviana and I celebrated completion of the manuscript with a trip to Patagonia at the end of 1984. The book was published under the title Political Economy: An Introductory Text in March 1985.

The book can be said to have been a success d’estime. It had the warm admiration of Partha Dasgupta, who got it into the Economics Tripos Part I at Cambridge. Jim Tobin declared it the most original introductory textbook since Irving Fisher’s. It found adoptions in a few premier schools such as Erasmus in Rotterdam, the Stockholm School of Economics, and the Institut d’Etudes Politique de Paris (in a superb translation by the French economist Jacques Le Cacheux). But it was not a commercial success. It was called pretentious, mannered, too difficult and so rich it was hard to teach from. This was about as painful an experience as I had in my career. Over the summer I would lovingly read portions of it, perhaps as part of a mourning process.

During a conversation a year later with Amartya Sen it struck me that in that long hot summer of 1985 in Rome I was not at all sure that I would find a way back into research. So many years had gone by without much time to do research or even to read it. I did reenter, though. What drew me back was a mounting curiosity to understand the deep macroeconomic puzzle posed by the 1980s experience: Continental western Europe was in its second depression of the 20th century – worse than that in the 1930s – yet one accompanied by elevated real interest rates rather than the depressed rates of the 1930s. Mises and Hayek may have had the answer, but if they had a model they had taken it to the grave.

Structuralism in Paris and Florence
With the extraordinary events in the Continental economies from the mid-1970s to the early 1980s on my mind, I began to question whether any explanation was provided by models of the New Classical type and those of the New Keynesian type, with their stationary stochastic structure and accompanying rational expectations. Was the massive slump on the Continent one great big disturbance term, to be followed by one big recovery? If so, is that an “explanation” of the “fluctuation”? I soon wondered whether any of the three monetary theories of employment (the New Classical, the New Keynesian, and the old neo-Keynesian) were up to explaining any of the big macroeconomic developments I had witnessed in the postwar decades right up to the mid-1980s. The emerging literature on Real Business Cycle theory, sparked by Frank Ramsey and David Cass and developed most of all by Edward Prescott, appeared to be of no help either. Was the rise of the world real interest rate originating in the U.S. a contractionary force in Europe, as suggested by RBC theory?

A change of scene may have served once again as the catalyst for starting a fresh page. A Columbia sabbatical enabled me to visit the Banca d’Italia in Rome for several months the summer of 1985, the think tank OFCE in Paris over December of that year, and the University Institute of Europe (IUE) in Fiesole (Florence) from January to August 1986. I knew just one economist in Rome, Luigi Spaventa, who had arranged for my position at the Bank. (The corporatist theoretician Enzo Tarantelli, whom I was looking forward to talking with, had been assassinated a few months earlier.) For a few weeks I treaded water, producing a modification of Tobin’s dynamic aggregative model in preparation for his Festschrift in autumn 1986. Finally, as my time at the Bank was dwindling, I became intrigued by the question: why in the early 1980s had Italy’s real interest rate risen so high and did that help to explain Italy’s slump? The Italian economists I talked to all thought that the high rates were “made in Italy” and I argued that the rise was imported from America. But I floundered about for a model with which to show that this rise had ill-effects on output and employment. This led in turn to conversations with Jean-Paul Fitoussi, whom I had met in previous years at IUE. He was Director of Research at the OFCE on his way to becoming President. We resolved to make this the next subject of our research.

Jean-Paul and I began modeling mechanisms from high imported real rates to decreased employment in December 1985 at the OFCE and largely completed this work at IUE the next March. This was a heartwarming collaboration and our friendship has only grown closer over the years. We gave our results in a paper at the BPEA meeting at Brookings in April and in a book, The Slump in Europe, published in 1988. It was exciting to have such a new set of ideas on offer. The thesis that real rates and real exchange rates impact on employment in unorthodox ways became the signature theme of Fitoussi-Phelps. There were plaudits from Jeffrey Sachs and Kenneth Rogoff and a skeptical assessment by Julio Rotemberg and Michael Woodford. In a debate with Michael in Capri around 1990 I argued that their analysis went amiss in testing for real interest rate while omitting real exchange rates.

Then, in 1988, I became intensely curious to see whether I could derive the same results and maybe more results from non-monetary models possessing some of the same mechanisms embodied in the models used in Fitoussi-Phelps. (To my later embarrassment I did not recall when I embarked on building a series of non-monetary models that a non-monetary version of my 1968 model of wage dynamics had been published by Steve Salop in 1979 and a non-monetary general-equilibrium version of the 1970 Phelps-Winter model of a customer market had been published by me and Calvo in 1983!) From 1988 to 1992, drawing on earlier papers with money from the 1960s, I managed to build a non-monetary general-equilibrium version of the customer model; I persuaded Hian Teck Hoon to collaborate with me in building a non-monetary version of the turnover-training model, or employee model – both published in 1992; and I built two sector models drawn from The Slump in Europe. Taken together, these models constituted what I dubbed the structuralist theory of unemployment. The distinctive character of the structuralist theory is not exactly to have rebased macro on non-monetary mechanisms; after all, RBC theory had done that. It is to have endogenized the (path of) the natural unemployment rate, to have integrated into the story some micro elements of business activity – acquiring customers and employees as well as plant and equipment – not usually at the center of models, and to have shown how valuations of each of these business assets are a key driver of the unemployment rate. In the resulting system, employment (or its growth) in an open economy is driven by the shadow price placed on the business asset, private wealth (or the income therefrom), the exchange rate, domestic productivity and the overseas real interest rate.

Hian Teck Hoon played a key role in the theoretical work. The phenomenal clarity he was able to bring to the turnover training model in the open economy and closed-economy cases made the two most crucial chapters transparent. Later he was to join me in clarifying the effect of technical progress on employment, a subject that had not been well treated initially, and in studying the effect of taxation on employment.

The work’s theoretical implications differed radically from both the Keynesian view and the RBC view of several causal forces: Contrary to the Keynesian view, aggregate demand is gone and the structure of demands is key: consumer demand and investment demand operate quite differently; private wealth in a (open) economy is unhealthy in having adverse effects not only on participation rates but also on employee conduct and loyalty. Contrary to the RBC theory, an increase in the overseas real interest rate, abstracting from any other overseas forces, is harmful for investment and employment, not beneficial. So is a decrease in domestic productivity growth, overseas growth being unchanged. We were later to extend our distinctive analysis to the effects of overseas forces causing a real exchange rate appreciation. (Subsequent papers in this project include ‘Growth, wealth and the natural rate: Is Europe’s jobs crisis a growth crisis?,’ with Hian Teck; ‘The rise and downward trend of the natural rate,’ with Gylfi; ‘Natural rate theory and OECD unemployment,’ with Gylfi; and ‘Lessons in natural-rate dynamics.’)

Gylfi Zoega, embarked on his dissertation at Columbia, had the idea of devising and performing statistical tests of this structuralist theory, taking advantage of the distinctiveness of several of its implications. This was no longer a fashionable activity so it took some getting used to. Nevertheless I got involved and at the OFCE in Paris we worked flat out over June 1992 to prepare the grand statistical regression, using OECD data on 18 nations, for a report of the results – whichever way they went. I remember my growing anxiety as the end of the month drew near. On our last day (a Friday if I am not mistaken) Gylfi pushed the button. I doubt I will ever forget my elation with the results. A few years passed until we produced, using a richer specification, another round of results. (As I write, we are beginning to prepare what will be the third round of statistical results.) It was a fantastic feeling to be in possession of truths that in the whole world only Hian Teck, Gylfi and I had. That was a source of confidence when in ensuing years it was necessary to go up against the conventional wisdom from other quarters. And how lucky I was to be able to share the experience with my great friends Gylfi and Hian Teck.

The three models and the statistical study were assembled over 1992–93 into a book published in 1994 under the title Structural Slumps. This was my second rewrite of “macro” and, for me, in many ways more satisfying than the rewrite of the Keynesian model that I had done in the second half of the 1960s. But what about the outside world? How important were the theory’s benefits and to what extent did the profession adopt its perspective?

At the level of historical understanding, this theoretical development served to underpin hypotheses linking the 1980s slump to a worldwide rise of real interest rates, the sharp transition to more moderate productivity growth in the European economies, and the growth of the welfare state to huge proportions, especially on the European continent. At a more general level, this work pointed to the crucial role for employment determination played by the values (also known as shadow prices) that firms place on the various sorts of business assets with which they operate: the employee with the needed firm-specific preparation, the customer, and nonhuman tangibles such as industrial plant and office facilities. This feature of the theory suggested that the prices of shares traded on organized stock exchanges might be serviceable as observable proxies for the mostly unobservable asset values, which opened up new statistical tests of the theory. (See the 2000 paper by Fitoussi, Jestaz, Phelps and Zoega, ‘Roots of the recent recoveries: Labor reforms or private-sector forces?’)

The structuralist models of unemployment movements turned out to be useful in coming to understand the inflationless booms in the late 1990s. In their thinking about the long wave of business expansions in the late 19th century, the German School under Spiethof and Cassel suggested that prospects of new industries or new methods requiring further capital, and this interpretation can be translated into an unexpected jump in the values that firms, looking to the new opportunities, place on one or more business assets. (An op-ed in the Wall Street Journal in April 2000 provides an introduction to this analysis. A trilogy of op-eds on the two great investment boom of the 20th century, with corrections and updating, appeared as a journal article under the title, “The Boom and the Slump: A Causal Account of the 1990s/2000s and the 1920s/1930s.”) I would add that I doubt that I would have been led to view the shadow prices that managers or owners place on the various business assets as independent, thus causal, had I not become increasingly accustomed to thinking about capitalist systems, with their visionary entrepreneurs and speculative financiers – which takes me to the last chapter of my research or at any rate the last chapter of this biography. I felt good to have experienced what I saw to be an unusually fruitful period of discovery running from age 55 (in 1988) to about 65. I was also proud that I was able to produce my magnum opus at a relatively late age. (I was 60 when Structural Slumps came out; but perhaps “60 is the new 50.”) At a presentation of what was essentially the OFCE report to a NBER meeting in Cambridge, Mass., in the summer of 1993, my discussant Julio Rotemberg was then very positive: “What more can you ask?” he said. The initial reception for the 1994 book was also enthusiastic. William Nordhaus, speaking as co-author of the Samuelson & Nordhaus textbook, told me “we loved your book.” The blurb Pentti Kouri wrote for the book’s jacket praised its radical break and pointed to its new results. The notices in The Economist and the Wall Street Journal were rave reviews (as one reviewer termed it himself) or close to it.

Yet the reviewers in the academic journals were noncommittal. Their reservation, as I see it, was that the methodology was novel; whatever it was, it was not exactly in the rational expectations mode. In my work, every shock, at least every important one, is de novo – totally unanticipated and unprecedented; business life is “one thing after another.” (It was only for simplicity that markets are supposed to put the economy on a (new) equilibrium path after each shock.) So the story might be called one of punctuated equilibrium. (A rational expectations approach would describe a stochastic stationary state in which, with known probabilities, random forces would establish which of the known potential “regimes” would be the next regime and how long it would prevail until giving way to one of the other potential regimes.) I knew that my method lacked the beauty of stochastic rational expectations equilibrium. But I liked its pragmatic and concrete qualities. I was to find this especially natural as I worked my way into capitalism theory. Furthermore, the structuralist framework appears to me to have had influence among practitioners – in discussions of the effects of Chinese saving on American asset prices, the effects of a weak exchange rate on the profit share increase, the effects of swollen government indebtedness on employment, and more.

“Inclusion” – 1990 to 2000
In the 1990s I became more and more concerned with the problem of “inclusion,” or “economic inclusion,” as it came to be called, in the United States and also in continental Europe. I got my start with it in 1990 when I had the opportunity to present a paper “Economic Justice to the Working Poor through a Wage Subsidy” at the Jerome Levy Economics Institute up the Hudson River at Bard College. After discussing my concept of economic justice and Rawls’s advocacy of maximizing the lowest pay rate, the paper discussed a graduated subsidy to employers aimed at pulling up wage rates at the low end of the distribution. There was not much more in the paper than that – no unemployment, no urban tensions, no social externalities, and the example given was simply sketched, not “costed out.” In my example, $3 an hour labor would then earn $7 an hour.

Later, in 1993, I was invited to be a visiting fellow in 1993–94 at the Russell Sage Foundation in mid-Manhattan, just 25 blocks from where I lived. This was the perfect place in which to carry out this project. Sage provided its visiting fellows with the opportunity to interact with Robert Merton, one of the very greatest social scientists of the century and a Columbia colleague with whom I had not had enough contact. I would send him my first drafts, which would come back with a profusion of comments, grave and minor, all in red ink. I could not ask for a better editor. He saw things about me that I had not seen. When I gave one of the weekly seminars, Eric Wanner was away, so Bob presided. At the end, everyone filed out except Bob who sat there musing until he finally spoke to me. I miss him to this day.

The book that finally came out of this was Rewarding Work, written between July 1995 and March 1996 and published in January 1997. It went “against the stream,” in a phrase Myrdal used, in three ways. Conventional thought held that jobs were a nearly unalloyed burden that is compensated only by the wages they bring in. In contrast, I saw jobs, as Marshall and Myrdal did, as the main source of most of their personal and certainly their intellectual development – maybe not in Soviet communism, where the rampant employee absenteeism was revealing, but certainly in highly enterprising economies, which well-functioning capitalist economies are. (“But surely there is not much mental stimulation and intellectual development provided by the jobs filled by disadvantaged people, is there?” I answer, “true, but for most people participating in the world of work, even in the bottom rungs, is more interesting and satisfying than the alternatives available if they do not get out of the house.”) I also saw jobs as providing people with the opportunity to do work in which they can take some pride (Veblen’s instinct of workmanship) and as providing whatever self-esteem may come from being self-supporting (Emerson’s self-reliance).

Conventional wisdom also held that, whatever the wages and the nonpecuniary rewards of work, free markets created the right amount of jobs. In contrast, I argued that many or most workers at the low end of the labor market find their pay so meager and perhaps demeaning (and possibly the non-pecuniary rewards too) that, if their economic situation is not dire, they may find it emotionally difficult to stay in the job for long. Or they may become so demoralized or distracted as to become inadequate in their jobs, so they are not kept on very long. In addition, statutory minimum wage legislation may make them unaffordable to law-abiding employers. In my criticism of the free market in low-end labor I also appealed to some old-fashioned values: the presence of many people unable to support themselves at normal standard tends to erode the place our culture gives to self support, engendering instead a culture of dependency. The presence of so many not engaged in the community’s business – the business of America is business, of course – tends to erode the place accorded to contribution and to engender instead an underground economy based on drug trade and violent crime. What will be the preparation of children for a fulfilling life if they grow up under such pathological conditions? Finally I did not shrink from adding that the social pathologies stemming from pay rates too small to provide a living and underemployment are Exhibit A in the populist attack on free enterprise, which is necessary for economic dynamism and helpful for the level of employment, thus prosperity, too. The upshot is that the members of society would be willing to pay something to raise low-end employment and pay.

Conventional thought also held that the right way for society, acting through its government, to address the poverty of the working age poor would be to offer money in the form of a universal income guarantee. It was explained that only social benefits of such a no-strings, work-unrelated, lump-sum nature would have the desired property of being neutral as between working and not working, joining and not joining. In scientific meetings in Europe and elsewhere I tried to explain how profoundly misguided such a position is. Instead, low-wage employment subsidies, or subventions, by imparting a non-neutral bias to participation, quitting and hiring decisions, would serve to raise employment and pay rates.

Of course I could not fail to bring in, despite the unpopularity of the very word injustice in some quarters, the self-respect member of society feel from seeing some of their tax money go to boosting the reward that the less fortunate receive from their part in the production of the nation’s output. When I was done I phoned Rawls to tell him about the book. I had to admit straightaway that I emphasized the negative social externalities that derive from sub-standard pay and under-employment. I did not emphasize the Rawlsian injustice of leaving pay and employment at the low end so far below what they could be. “You can’t in today’s climate,” he responded. This was quite a relief. I had only to bear the objection from the right that the book had too much on “belonging” and the objection from the left that it had too much on crime and violence.

Rewarding Work and a shorter 1995 report in French have had some success in Europe. In the year of its publication it was quoted at some length by Clive Crook in The Economist and endorsed by Martin Wolf in the Financial Times. I had the honor to speak about the ideas in the French Senate. Both France and the Netherlands adopted an assortment of plans to boost employment and pay at the low end, one component of which was been subsidies to employers. Neither country, though, adopted such subsidies with anything like the scale and universality envisioned by my plan. Now, in France, there has been a remarkable revival of attention to “the value of work” and to “inclusion,” and some have credited my work for helping to bring these ideas to politicians and to the public.

In contrast, these ideas have not been implemented in the United States, leaving aside the narrowly confined and not very cost-effective program called EITC, which sends checks most exclusively to women with low earnings in the current year (without regard to wage rates). In a marvelous review appearing one Sunday in January 1998 in the Washington Post the columnist Matthew Miller rebutted the objection that my plan, at 100 billions dollars per year, would “cost too much”. The Congress, he said, will certainly step up its annual spending by that much or more in a few years. And the Congress did – notably with a lavish new program of free pills for the elderly. Miller’s subsequent book The Two Percent Solution made my plan the centerpiece of a bouquet of legislative programs he proposed to meet a range of social problems in the U.S.

Rewarding Work was also the prelude to quite a lot of research. Hian Teck Hoon and I worked like dogs to mathematicize in our turnover/training model the optimum graduation of the employment subsidy. In November 1998 a quartet of highly technical papers were heard at a Russell Sage conference on alternative methods of helping low-wage workers. Several of my closest colleagues in this area came to give papers, including James Heckman, Dale Mortensen, Christopher Pissarides, Dennis Snower and Michael Orszag, as well as Hian Teck and myself. For the conference volume, Designing Inclusion, I wrote an introduction that is regarded by some as my best statement on the subject.

“Dynamism” in London, Rome and back in New York
In the early 1990s, with the Soviet Union in an economic crisis, I was drawn into thinking about the kind of economic system that would best replace the communist system. Going with Kenneth Arrow and Fitoussi on the 1991 mission to Moscow for the nascent European Bank for Reconstruction and Development and writing a report started me thinking.12 A year as a consultant to the EBRD in 1992–93 acquainted me with most of the issues. Years later I was led to the more general question: In any country, what criteria must be used in determining the best sort of economic system for it? This was to be a year of stretching myself into a new area. So it was a godsend to have in London the friendship and support of Beatriz and Philippe Aghion, Francesca and Chris Pissarides and Judith and Dennis Snower at this time.

When starting out I had no idea whether I would be well-suited for this research. But as I got farther into it I began to believe that there were elements of my background that did suit me to it. I am thinking particularly of Bergson and James, also the classic figures in the humanities who had made a deep impression on me.

The economic system that Marx dubbed “capitalism” held a fascination for me. The little I knew about Soviet communism mostly came from eastern European émigrés: Jan Tumlir, Bela Balassa, Janusz Ordover and especially Roman Frydman (the last two former students of mine). At Amherst and Yale I had read chapters of Hayek but more his writings on business fluctuations than his work on capitalism theory: what differentiates it from market socialism and communism. With occasional consultations with Frydman, Andrzej Rapaczynski, Jan Winiecki and Leszek Balcerowicz, I studied the key sections of Hayek’s seminal essays from the mid-1930s to the mid-1940s.13 This led me back to passages by Knight and Keynes and forward to Michael Polyáni. They seemed so modern and yet out of this time. I had never met any of them. My friend Emile Despres told me about a lunch he had with Keynes during the war. I missed Polyáni’s 1962 Yale lectures on personal knowledge. There was plenty of time to meet Hayek but I never took the considerable initiative it would have required. (Some other regrets in this category are not having talked to Marlene Dietrich when the plane we shared was forced by bad weather to land in Pittsburg and not having gone even once to Café Taci on singing nights in order to talk to Franco Corelli. But small talk was never my forte. And maybe not theirs.)

In my reading two core properties of capitalism struck me. First, capitalism, in allowing individuals to become sole proprietors of companies or at any rate shareowners (provided the stock market has real influence over a company’s management), opened a society to future directions that are not coordinated by anyone or any institution. And since there is no coordination, there will generally be a heterogeneity of views about how the economy works – about what the future consequences would be of any of a wide range of present actions, especially investment decisions. For this reason, our models of capitalism, if they are to be faithful to this fundamental and distinctive characteristic that the system has, must recognize that the system generates disequilibrium. They must be models of disorder, though it is a disorder that may be preferred to any “order” that might be imposed. (It seems to me that Hayek, who would have accepted this point in the 1960s, overlooked it when in the 1930s he referred to the “spontaneous order” in capitalist economies.)

Second, capitalism, in its tolerance of a diversity of views, allows more participants to venture into novel directions than would otherwise be possible: to pursue an original career, to adopt an untried portfolio strategy, to start a company to develop and market a new product, to pioneer the use of a product or method. This property makes well-functioning capitalist systems a wellspring, if not the wellspring, of Hayekian innovation: not the routine innovation of Schumpeterian entrepreneurs, who merely undertake “obvious” applications of known scientific advances, but innovations stemming from the conceptual originality of people in their business life. The prospects for the Hayekian innovations are clouded by their considerable ambiguity: no one can know beforehand with what probability they will be commercially successful. In a capitalist economy, therefore, present participants – at any rate those among them who hit upon successful innovations – influence or even shape the future: future know-how and goods in existence.

For me, this was heady stuff. I did not lose sight of capitalism’s faults. I appreciated that many economists in the history of the subject have expressed dislike of its headless directions and its doldrums. There was also the problem of inclusion, though that was not peculiar to capitalist systems: it was evident in Europe’s corporatist economies too. Little by little, I was drawn away from macroeconomics, where my research had been stuck for nearly two decades, to the economics – or political economy – of alternative economic systems: capitalism, market socialism and corporatism.

In my year at the EBRD, my job was to edit the first Annual Economic Outlook. At that time I had no original line, no new idea, of my own to impart. We – in my circle – thought that the “market economy” meant capitalism, no doubt buttressed by regulations and cushioned by welfare entitlements. In the chapter I drafted to introduce the volume I sought to clarify the theory of capitalism existing in the literature and to highlight ideas of Aghion, Kornai, Frydman and Rapaczynski that had surfaced at the EBRD.14 Here is a passage of the chapter arguing the long-termism of classic capitalism: The entrenchment of a manager is the most serious difficulty …

The entrenched manager will be more willing to risk covert practices for his own benefit than would a manager that owners could easily dislodged … [I]t is best that the business strategy of an enterprise be decided by its owners: the chance of huge gains motivates them to think big, the fact that they are not wedded to the enterprise (they can invest elsewhere) motivates them to think fresh, and the opportunity as owners to derive immediate benefit from expectations of far-future payoffs motivates them to think far ahead. In contrast, entrenched managers favor strategies that ensure the firm’s survival so that they can hang on to what they have.15

When John Flemming, who was head of the economic section, sent the page proofs of the Outlook for approval, something unexpected happened. The Vice President demanded my introductory chapter be removed, saying that the EBRD must be neutral about the sort of market economy eastern Europe would best “transition” to. John was able to save the piece only by making it the Annex! The privatizations in Russia and the Czech Republic operated to entrench managers and employees in their positions. It appeared that many economists would accept “transition” to the market socialism found in eastern Europe and many more were quite content to see Russia and the others adopt the corporatist system of continental western Europe. Much of my research since then has aimed to show these judgments to be misguided. It was one of the thrills of my career to have taken up the study of capitalism at such an exciting time. But I hoped eventually to make some theoretical contribution to the understanding of capitalist systems. How is a capitalist economy to be described – to be modeled? What, that is, do such economies do and how do they do it?

My research project in this direction, still in its infancy but nevertheless in progress, dates from May 1977. Luigi Paganetto, whose Faculty of Economics and Law at University Tor Vergata in Rome I had been visiting since 1989, asked me whether I would be willing to be play the role of adviser in a research project at the Consiglio Nazionale delle Ricerche to study the weaknesses of the Italian economy as it faces the European Union and the global economy. I began work in May 1997 and presented at CNR my last semi-annual report in May 2000. A compilation of the six reports was published in 2002 under the title Enterprise and Inclusion in Italy. It did not give birth to a model. But it was a start. Not that I did it alone. I owed much of my knowledge of the Italian economy to Luigi, to Stefano Micossi and Luigi Spaventa. A group of young economists working in Rome was assembled, some friends for a decade or more. Giovanni Tria and Ernesto Felli, whom I had known since the mid-1980s, were mainstays. Luigi Bonatti, Francesco Nucci and Alberto Petrucci were also important contributors.

In this period I hit upon an idea I liked – so much that I reported it four times: Financial Times in 2000, Economic Policy in 2001, Enterprise and Inclusion in 2002 and Aghion et al. in 2003. It was not about the endogenous offer and selection of new ideas, which I later denoted by the term dynamism. It was about vibrancy. Spiethoff and Cassel, the leaders of the turn-of-the century German School, saw innovation as a frictionless and unfailing response to advances in science and navigation. Schumpeter modified that, arguing that there is always a scarcity of entrepreneurial types able and willing to dedicate themselves to the challenge of developing a new product or method. But that is not all, I suggested. Before going ahead an entrepreneur has to foresee an adequate presence of Nelson-Phelps managers and, similarly, Bhidé consumers with the education and venturesomeness required to weigh, judge and operate the new product or method. Entrepreneurs may have to get through a thicket of government licensing and regulatory requirements. A broad stock market with high share prices where venture capitalists can cash in their investments in fledgling firms may be important. It turns out that of the 12 largest economies in the OECD are ranked according to an index of these attributes circa 1990, the ranking predicts very well how those economies ranked with regard to their responsiveness to the internet revolution that broke out in 1995. (Aghion and Peter Howitt also built on Schumpeter, of course.)

But I hoped to move away from this narrow, nearly classical perspective. At the Festschrift conference for me in 2001, Samuelson and I chatted before he gave his keynote speech. “You know, Schumpeter was not an Austrian,” he said. “Yes,” I answered, “I know.” I knew that very well. In Schumpeter’s system, there is no “Knightian uncertainty” faced by the start-up entrepreneur launching an innovation, which was central to Hayek and also Knight (for whom just about everything was uncertain), no “personal knowledge” of the entrepreneur that a financier has to face, which was familiar to Hayek and Polyáni, and no “moral hazard” for the financier that the entrepreneur will find chances for self-dealing, which would have been understood by Mises.16 Most fundamentally, in Hayek’s system, the entrepreneurs are people driven to implement unknown ideas of unknown promise; persons may have new commercial ideas that spring from the private knowledge and insights they acquire in the course of their business or job. As I saw it, advances in capitalism theory must model the economy as basically evaluating, selecting, developing, and spreading new commercial ideas. Every distinct good or method is the expression of a new idea.

The concept of a research center for developing capitalism theory began to emerge out of numerous conversations that Roman Frydman and I had in the late 1990s. Over the years I had greatly enjoyed our discussions as well as our warm friendship and this was one of our best ideas. We received strong encouragement from William McDonough in New York and David Lord Howell in London. At Columbia then president George Rupp and the provost Jonathan Cole gave me the green light. It would be inter-departmental, even inter-school, and allow members outside Columbia. Thus was born the Center on Capitalism and Society.

The project would not have been possible without a remarkable confluence of talent at the right place and the right time. Bruce Greenwald, Glenn Hubbard, Richard Nelson and Andrzej Rapaczynski were insiders from the beginning alongside Roman and myself. But we would have been too small a group. Our great luck was that Joseph Stiglitz had just arrived at Columbia and agreed to join the Center, Amar Bhidé, a world authority on innovation, had just come to the Business School, and Jeffrey Sachs came to Columbia a year later. Then I persuaded Janusz Ordover of NYU and Robert Shiller of Yale to join the roster. This body has remained at 11 but the Center has grown. Jean-Paul Fitoussi has become a Foreign Member. Several more have become advisors or regular attendees: Pentti Kouri, Robert Mundell, Richard Robb and Jürgen Schröder. Philippe Aghion and David Jestaz have been near-regulars.

Our inaugural conference, held at Columbia in 2004, was supported entirely by Jeff Sachs’ Earth Institute, which we became a part of in 2002. When not a single foundation would provide support during the years of the depressed stock market, Jeff came through for us. The 2nd annual conference was held in Reykjavik in cooperation with Tryggvi Herbertsson and Gylfi Zoega of Iceland University. The 3rd annual conference was held in Venice in partnership with Hans-Werner Sinn of Munich’s CESifo.17 The Center’s electronic journal Capitalism and Society was developed by Amar and launched in summer 2006.

In my efforts in recent years to contribute to capitalism theory these scholars have been a huge intellectual stimulus and also a great forum in which to try out my own ideas. Amar and Richard have been invaluable.

I would like to point to three papers of mine that strive to deepen the understanding of capitalism’s selection for development and for launch of new commercial ideas. These papers are all about dynamism. I use the term to refer to the volume of innovation of which the economy is capable (under given market conditions) weighted by the desirability of their direction and reach – the gifts of financial markets in deciding the innovations to back and of the product markets in deciding the innovations to adopt.18

One of these papers sketches a model in which, every day, each person with a new idea for a startup firm gathers in some known place to make a presentation to every one of the potential financiers for such entrepreneurial projects. I visualized a line of entrepreneurs going around a circle of financiers. It was delightful when Richard Robb reported he had learned that such an institution exists in the United States. Richard verified that the entrepreneurs do in fact circle around the financiers. My interests, however, were economic. What characterizes the set of projects that “make the cut” – that receive capital from some financier? Does the “optimism” of entrepreneurs raise market value of startups or just reduce the external finance required? Of financiers? What interested me was that by being precise about the interaction between the two sides of the market one could begin to tease out various implications. Modeling pays. (This paper is to appear in a conference volume of the Kauffman Foundation and the Max Planck Institute.)

A paper I prepared with Gylfi Zoega for the April 2004 issue of the CESifo periodical Forum weighed evidence on the importance of certain economic institutions for the degree of dynamism in a country’s economy. We looked first at the policy parameters of neoclassical theory. As in our 1998 paper, we found that inter-country differences in the tax rate on labor had little power to explain the differences in various economic indicators (productivity, unemployment rate, etc.). Inter-country differences in the replacement ratio provided by unemployment benefits were also statistically insignificant. In fact, the explanatory power of a great many policy variables – government purchase of goods and services, public capital stock, corporate profits tax rate and so forth – has been slight, even if sometimes statistically significant.

If our conception of the advanced economies is one centered not around consumption and leisure but instead around the attractions and rewards of business life – problem-solving, the discovery and development of talents, and the achievements that may result – then it is not surprising that these policy parameters, though important in the neoclassical perspective generally adopted by supply-side analysts, do not make much of a dent on unemployment and participation – as long as they stay in the historical range at any rate. It becomes hard to see why the neoclassical preoccupations with work-leisure substitution should be center-stage. Reducing the calibrations of the welfare state or cutting government purchases or adding to capital stocks will not make jobs far more engaging and rewarding, hence make participation in the labor force far more attractive and unemployment far smaller. Only modest results can be reasonably hoped for.

What institutions appear to matter for inter-country differences in performance? Presumably there are economic institutions the presence and high development of which affect performance indicators through their impact on dynamism. It is reasonable to hypothesize that organized stock exchanges, company law, suitable bankruptcy provisions, corporate governance spurring corporate performance, and schools preparing the population for business life all foster dynamism … and that economic institutions whose presence and force obstruct or impede dynamism – corporatist institutions that invest company employees, labor unions, communities and other interest groups with veto power to block or limit entrepreneurial ventures and shifts in corporate operations may choke off valuable innovations – all dampen entrepreneurial spirits and thus decrease dynamism.

Fortunately, our investigations found statistical support for as many of those hypotheses as we could test. Yet, in view of other studies, it appears that results in this area are fragile. Something is missing, it seems.

The third paper is more radical. A part of an economy is its economic institutions. Another part consists of the elements of the country’s economic culture. Some cultural attributes in a country may have direct effects on performance on top of their indirect effects via the institutions they foster. Values and attitudes are analogous to institutions – some are impeding, others enabling. Clearly, any explanation of the poor performance on the Continent that omits that part of the system must be of unknown reliability.

Of course, people may at bottom all want the same things. Yet not all people may have the instinct to demand and seek the things that best serve their ultimate goals. In the University of Michigan “values surveys” there is evidence of systematic differences in workplace values between the Continent’s Big 3 (Germany, France and Italy) and the U.S. and Canada. Take the values that might impact on dynamism. Asked what they look for when seeking a job, relatively few persons in the Big 3 reported that they want jobs offering opportunities for achievement; relatively few that they want chances for initiative in the job; and few that they look for interesting work (!). Relatively few are keen on taking responsibility, or exercising freedom in decision making, and relatively few are happy about taking orders. I like to joke that a country in which no one wants either to give or receive orders will find it difficult to start a firm capable of doing the novel and creative work required in developing an innovation.

The paper goes on to show that inter-country differences in the endowment of these values are very commonly statistically significant in explaining differences in standard performance indicators – more significant than many of the institutional variables usually accepted as quite important for performance.

Now I realize that this perspective, which looks very wide, is not deep enough. Where do differences in the set of economic institutions – differences in the economic system – come from? The same question may be asked about differences in workplace values. I have a suspicion about where they come from. In earlier work I organized my thinking around some of the intellectual currents of reaction on the Continent to the Enlightenment and to capitalism in the 19th century: the solidarism of protecting the “social partners,” the consensualism of blocking business initiatives lacking consent of the “stakeholders,” the anti-commercialism that revived in the 19th century, the conformism that militates against sticking out from the group, and the scientism that encourages looking to the state for social advance. I call this ideology corporatist though we could just call it Continental. Now I am inclined to hypothesize that the rise of the corporatist ideology allowed the erection of institutions harmful to dynamism: if some of the ill-effects were foreseen, their cost was not seen as high enough to deter building those institutions. I hypothesize further that the corporatist ideology was the root cause of the decline of economic values helpful to dynamism: Continentals, finding over the course of the 20th century that there was little call and little room for the exercise of freedom and responsibility in the workplace, learned not to prize those values and thus to stop looking for the satisfaction of those values in the workplace.

I would just add that in June 2006 I had the great pleasure of speaking at the Institut d’Etudes Politiques in Paris on how capitalism might be morally justified. Let me refer readers to the last section of my Prize Lecture where I discuss my ideas on that interesting and, I think, important subject.

In closing this section I will just recount how receptivity to the subject of capitalism has evolved in France. In 2000 a lecture of mine at Sciences Po was announced on some merits of capitalism. The attendance at the lecture was 5 persons, counting the moderator, Jean-Paul, and my wife Viviana. In 2005 I spoke at the Sorbonne on much the same topic to an audience of nearly 500. (The same number or more appeared at the June 2006 event.) I expect to be staying with this subject for quite awhile.

An immensely gratifying period
The 2000s have brought wondrous times. Long ago, in the early 1980s, when I was not yet 50, I was elected to the National Academy of Sciences (USA), to the American Academy of Arts and Sciences, and named to the relatively ancient McVickar chair at Columbia. That in the space of some two years. Then nothing of the sort happened again for nearly two decades. That was to change.

In January 2000 I was named a Distinguished Fellow of the American Economic Association. My fellow Fellow was Jack Hirschleifer, whose work I admired. This was a most relaxed event. The Chicago labor economist Sherwin Rosen read with glee the citation where it said “then the government tested the [natural rate] idea.”

In October 5–6, 2001, the Festschrift conference in my honor was held up at Columbia. It was an unforgettable event. As many said, it seemed that all the important macroeconomists in the world had come – nearly 140 in all – and this is in spite of the shocking and tragic events of 9/11 just weeks before. Roman Frydman was masterful as the main organizer while Philippe Aghion, Joe Stiglitz and Mike Woodford contributed in a range of important ways. Tobin, Lucas, Heckman and Solow each agreed to be a commentator at one of the four sections and Samuelson gave the keynote. Klein, Mirrlees and Mundell also came. (Sadly, Tobin as well as Merton and Rawls could not attend for reasons of health.) Many moments at this event took my breath away, almost literally. One of these came following my speech at the end of the first day. As I walked back to my seat to rejoin Viviana there was a deafening drone of applause that seemed endless. I felt a bond between me and each one of them, which was private but powerful.

The award of a Nobel Memorial Prize is a huge honor. To be the sole winner, as I was in 2006, is the greatest honor. My citation was, for me, a validation of the first half of my life’s work. What makes the Nobel so exceptional? For one thing, the Nobel award is so extraordinarily public – and on a global scale. With the press conferences and media coverage, a great many people come to know your name and think of you for a range of invitations. Thus the award may be life-changing, a watershed event. Nobel Week, with its ceremonies and festivities, represents a kind of elevation, which signals greater respect and correspondingly higher expectations. I have felt pressed to ramp up my performance in an effort to meet the elevated expectations.

It is a privilege to have all this stimulus and challenge. Toward the end of Nobel Week, at a gathering in the Nobel Foundation, Professor Gunnar Öquist, permanent secretary of the Royal Swedish Academy of Sciences and a man of dignity and seriousness, came over to say goodbye for the last time. As we shook hands he said: “Use it well.” I am trying to do that.

Edmund S. Phelps is McVickar Professor of Economics, Columbia University, and Director, Center on Capitalism and Society, Earth Institute, Columbia University.
1. Three earlier memoirs are “A life in economics,” in Arnold Heertje, ed., The Makers of Modern Economics. Aldershot: Elgar, 1995, “The origins and further development of the natural rate of unemployment,” in Rod Cross (ed), The Natural Rate of Unemployment: Reflections on 25 Years of the Hypothesis, Cambridge: Cambridge University Press, 1995, and “Recollections of my past research in economics,” Acceptance Speech, Beijing Technology and Business University, June 2005. These and other data, including my CV, are at www.columbia.edu/~esp2
2. New York Daily News, December 8, 2006.
3. A recent essay on the phases of the Boston Symphony remarked that Charles Munch in choosing Roger Voisin as principal trumpet “sealed the fate” of the orchestra.
4. That does not mean that economic descriptions are unimportant. Istanbul’s geography and access to water may be important to the city, even to its cultural character. Also, there may be a set of cities having a similar culture yet every one is distinctive in the character of its economy and the resulting rewards from work.
5. T. C. Schelling, “For the Abandonment of Symmetry in the Theory of Cooperative Games,” RAND , Santa Monica, 1958. Oddly, the bibliography of Schelling’s RAND papers that I found on the internet portrays the paper as showing that symmetry does not follow from “rational expectations” while the lesson I drew was that so-called rational expectations cannot be successfully defended by the argument that (in the model under study at any rate) it would be deducible if symmetry were present and were common knowledge.
6. T. C. Schelling, “The Reciprocal Fear of Surprise Attack,” RAND, Santa Monica, 1958.
7 W . J. Fellner, “Distortion of Subjective Probabilities as a Reaction to Uncertainty,” Quarterly Journal of Economics, 75, November 1961, 670-89.
8. W. J. Fellner, “Demand Inflation, Cost Inflation, and Collective Bargaining,” in P.D.Bradley (ed), The Public Stake in Union Power. Charlottesville: University of Virginia Press, 1959.
9. We cannot test and confirm everything in our model; there have to be some untested premises in any test of the other hypotheses. Hence, when a model fails it is indeterminate whether the problem lies in the hypotheses being tested or in the accompanying premises on which the model is grounded. So we are reluctant to scrap our untested premises. (The point is made in my Ph.D. thesis “A Test for the Presence of Cost-Push Inflation,” Yale Economic Essays, 1961. It is a theme in James Heckman and Richard Robb, “Alternative Methods for Solving the Problem of Selection Bias in Evaluating the Impact of Treatments on Outcomes,” H. Wainer (ed), Drawing Inferences from Self-Selected Samples, 1986. That paper’s reception by John Tukey is told with dry humor by Heckman in his Prize Lecture.) Thus, rational choice, in the sense of Samuelson’s revealed preference axioms or, perhaps less aptly, Savage’s axioms, is a premise that most of us economists are loath to give up, even though we recognize that there may be contexts in which we would do well to depart from it. But, we do not treat it as an absolute, never to be deviated from.
10. Another objection of some radical Keynesians was that the proposed microfoundations apparently assumed that supra-natural unemployment and sub-natural unemployment too would not occur but for the deficient response of the general price and money-wage level to contractionary or expansionary shifts in effective demand; and in the belief of these Keynesians it does no good for the price level to fall following a contraction of demand (and no good for it to rise following an expansion of demand). They explained that when the price level fell, the real burden of indebtedness at businesses would be increased, adding to insolvencies and failures, and thus weakening the amount of output demanded, particularly output for investment purposes. To a close approximation, they said, effective demand was perfectly inelastic with respect to the price level. In their view, this made the natural rate idea beside the point. But there was never any econometric evidence adduced in favor of that extreme position. Furthermore, effective demand could not remain deficient for long without the central bank’s taking the falling inflation rate as a signal to cut expected short-term real interest rates and thus to move by policy means the economy back toward the natural rate of unemployment. (Exchanges in 2006 with N. Gregory Mankiw and with Michael Woodford led to the result that, if the central bank uses a Taylor rule but misgauges the natural real rate of interest at too high a level, the inflation rate tends to sink to the level that is below the central bank’s target by just enough to drive the target real interest rate below the natural real rate by just enough to bring the unemployment rate back to the natural unemployment rate.)
11. Edmund S. Phelps, Political Economy: An Introductory Text. New York: Norton, 1985.
12. Kenneth J. Arrow and Edmund S. Phelps, “Proposed Reforms of the Economic System of Information and Decision in the USSR: Commentary and Advice,” Rivista di Politica Economica, v.81, November 1991.
13. Hayek, Collectivist Economic Planning. London: Routledge, 1935, The Road to Serfdom. London: Routledge, 1944 and Individualism and Economic Order. Chicago: University of Chicago Press, 1948.
14. Philippe Aghion argued the necessity of competition among firms; Janos Kornai the necessity for the state to break the habit of bailing out failing firms; Frydman and Rapaczynski the necessity of designing “voucher privatization” to create enterprise owners who would take control from powerful managers and employees.
15. ‘Arguments for private ownership,’ Annual Economic Outlook, EBRD Economic Review, September 1993. Quotation, p. 120.
16. William Baumol tells of Mises saying, “Vat is dee Austrian School bat me und Hayek?”
17. The website of the CCS is at www.earth.columbia.edu/ccs
18. Because a model capturing capitalism must be something different from the Spiethoff- Schumpeter model – something with some capabilities added (and some subtracted) – any such model requires added concepts, thus added terms.

From Les Prix Nobel. The Nobel Prizes 2006, Editor Karl Grandin, [Nobel Foundation], Stockholm, 2007

This autobiography/biography was written at the time of the award and later published in the book series Les Prix Nobel/ Nobel Lectures/The Nobel Prizes. The information is sometimes updated with an addendum submitted by the Laureate.

Copyright © The Nobel Foundation 2006

To cite this section
MLA style: Edmund S. Phelps – Biographical. NobelPrize.org. Nobel Prize Outreach AB 2023. Sun. 24 Sep 2023. <https://www.nobelprize.org/prizes/economic-sciences/2006/phelps/biographical/>

Back to top Back To Top Takes users back to the top of the page

Coming up

This year’s Nobel Prize announcements will take place 2–9 October. All announcements will be streamed live here on nobelprize.org.

See the full schedule
Announcement dates

Explore prizes and laureates

Look for popular awards and laureates in different fields, and discover the history of the Nobel Prize.