Before continuing, let me take a moment to remind EP readers of why this serial assumed the form it has – a form that may at times have seemed wayward.
My aim all along has been to explain what MIT-trained Ben Bernanke meant when he said in 2002, “You were right, Milton, we did it [caused the Great Depression], and, thanks to you, we won’t do it again.”
I aim, too, to show how it was that the MIT theorist who inherited Paul Samuelson’s office (and the professorial chair named for him) and a banking historian from Yale who, for a dozen years, spent a day a week in the risk markets at AIG-FP, collaborated to produce a uniquely satisfying explanation of what happened in the 2008 financial crisis – and the beginnings of an explanation of what happened as a result.
Submerging their differences, working in loose harness with a dozen other researchers, Bengt Holmstrom and Gary Gorton made a major contribution to twenty-first-century economics.
Without regular signposts, like this, much less a table of contents, not to mention the presence of various abrupt shortenings arising from the press of time, it has been easy to feel lost. Readers are doing as much for me as I am for them. This is a work in progress.
The punchline was conveyed at the beginning: the appearance since the crisis of a series of recent papers on the the role of debt in the economy, the development of opacity in banking, panics as information events. Their import is hard to understand without some context.
To that end we’ve gone back over a lot of ancient history – the long-forgotten rivalry of Adam Smith and Sir James Steuart; the eclipse or emptying-out of monetary theory and growth theory occasioned by the discovery of price theory; the development of the institutions of central banking institutions and related doctrine by Practicals, starting in the eighteenth century; the confusing catharsis of the Great Depression; the early stages of the “neoclassical synthesis” that emerged after World War II in the hands of Paul Samuelson, John Hicks, Franco Modigliani, and James Tobin; and the gradual reawakening of interest in monetary theory that began with Milton Friedman in the late 1940s.
When he moved to Harvard from Stanford in the ’60s, Kenneth Arrow brought with him awareness of various exciting developments that had occurred in the ’40s and ’50s at the Cowles Commission and RAND Corp. – especially the creation of an idealized blueprint of a competitive economy against which the details of the world itself could be compared. Almost immediately, Arrow himself furnished a demonstration of the power of the new investigative tool. His 1964 paper on the economics of health economic produced the related concepts of moral hazard and adverse selection.
Microeconomists had exciting new topic on which to work. The actors who populated their economic models began to behave strategically – playing for their own account, taking flamboyant risks or playing it altogether safe, competing with more cunning than before.
But that wasn’t all. The tools and formal methods that Arrow and others of his persuasion espoused would transform macroeconomics in the ’70, too. Herewith the conclusion of last week’s episode. I’ll make it quick.
When Arrow observed that Paul Samuelson hadn’t addressed one of the “major scandals” of price theory, the relation between micro and macro, he had something very specific in mind: the Phillips Curve, the trade-off between inflation and employment that New Zealand economist A.W. Phillips had identified a couple of years before.
The higher the inflation the lower the unemployment and vice versa, or so it seemed in the scatter-plot that accompanied Phillips’ article. Samuelson and his MIT colleague Robert Solow had conjectured in 1960 that the relationship might be exploitable as a policy instrument, a way of permanently lowering the unemployment rate by accepting a little more inflation.
Samuelson was dining on franks and beans with the president-elect on Cape Cod in those days, helping Kennedy choose his cabinet. Newspapers were full of the “New Economics.” The Council of Economic Advisers was contemplating “incomes policy,” meaning the broad governmental supervision of wage bargaining and corporate pricing decisions necessary to keep inflation in check. Two years later this would lead to a memorable confrontation with US Steel chief executive Roger Blough.
It all seemed to work for a time, even after President Kennedy was assassinated in 1963. When Walter Heller, his Council of Economic Advisers chairman gave a series of lectures at Harvard in 1966, he retraced the victories that accompanied the administration’s “willingness to use, for the first time, the full range of modern economic tools,” – the fight against a tax increase in connection with the Berlin buildup of 1961, the business tax cuts of 1964, the five years since 1961 of uninterrupted rapid growth without inflation.
Friedman had battered away at the consensus with every means at his disposal during those Camelot years. Capitalism and Freedom, his gloss on a broad spectrum of economic policies, including food trucks and charter schools, appeared in 1962; A Monetary History of the United States, the book with Anna Schwartz, in 1963. He advised the Goldwater campaign in 1964 and begun writing his column in Newsweek, opposite Paul Samuelson and Henry Wallich, in 1966.
But Friedman’s most powerful argument, because it was directed to the profession itself, was contained in his presidential address to the American Economic Association in December 1967. He had been skeptical of the Phillips Curve from the beginning; in 1967 he gave reasons. The tactic of exploiting such a trade-off might work in the short run, but gradually expectations would adjust. “Employees will start to reckon on rising prices of the things they buy and to demand higher nominal wages for the future,” he stated. Friedman’s argument rested on the assumption, dating back to Adam Smith, that (as Walter Bagehot had put it at an interval halfway between the one and the other), “there was a Scotchman inside of every man,” eager to earn, cautious to spend. About the same time, Edmund Phelps, of Columbia University, made a more complicated argument to the same effect; Arrow’s criticism added to the doubts.
By the end of the ’60s, the newly-elected Nixon administration was embroiled in crisis. Inflation was increasing at what then seemed like unacceptable rates– a 6 percent annual rate in 1970. Growth stalled, but prices continued to rise; a new phenomenon quickly dubbed “stagflation.” The social spending of the Great Society and the Vietnam War were generally blamed. But the Federal Reserve Board, too, was under vigorous attack for accommodating the spending.
By the time Harry Johnson, of the University of Chicago, gave his Ely Lecture to the American Economic Association in 1970, “The Keynesian Revolution and the Monetary Counter-Revolution,” it seemed well-established that something was missing in the Keynesian program.
They didn’t know the half of it. Another revolution was just beginning in technical economics, this one under the banner of a series of assumptions about human nature called “rational expectations.” And of course in the’70s, the global economy itself was in for a rough ride.
Friedman’s presidential address might have been electrifying, but it didn’t constitute much more than an assertion that people are not so easy to fool. Yet in the late ’60s, macroeconomists in the best universities, central banks, consulting firms and think-tanks were thriving, deriving statistical underpinnings for the top-down models of consumption, investment, and liquidity preference of the Keynesian by linking them to optimizing behavior of households and firms To describe the decision- making process of the agents involved, economists had developed the concept of “adaptive expectations.” Consumers and the executives of firms would extrapolate future prices from past ones, a little like driving forward by looking in the rear-view mirror. But using control theory, which was the main predictive tool economists had at their disposal at the time, adaptive expectations were easy to model. They were also easy to fool.
It’s no easy matter to explain rational expectations, but it is easy enough to say where it came from. It came from Pittsburgh, home to the Carnegie Institute of Technology (now Carnegie Mellon University), an institution so pragmatic in its founding impulse that its first classroom building was erected on a long, gentle slope, so that it could be used as a gravity-powered manufacturing factory in case its higher aspirations didn’t work out. In the 1950s it served as a refuge to some of those forced to relocate when the Cowles Commission was driven out of Chicago, notably Herbert Simon and Franco Modigliani (who had himself been chased from the University of Illinois, during its Red Scare, in 1952). And in late ’60s, Carnegie Tech was home to three remarkable young economists: Robert Lucas, Thomas Sargent and Edward Prescott. All three subsequently would receive Nobel prizes, as would Simon and Modigliani, but it is Lucas whom we think of as having been the foremost macroeconomist of the last forty years.
I have been calling the Arrow Debreu model an “idealized blueprint” of a competitive economy in earlier installments of this serial. Elsewhere I described it as an “infinite dimensional spreadsheet.” But that conversion to a highly dynamic tool didn’t begin to happen until Arrow, Debreu and others coupled it with Arrow’s “contingent claims” formulation of futures markets as a means of insuring against the uncertainty of the world. For if you imagined the existence of complete markets for goods and services in every possible state of the world – #2 grain if it rains in August, #2 grain if it doesn’t – then uncertainty is greatly alleviated, at least as a matter of description.
It was this project on which Lucas embarked with his former student Prescott (who had begun teaching at the University of Pennsylvania) in the mid-’60s, with a view to making better models of investment decisions. Before long it was clear that the method would serve in the Phillips curve controversy as well. Lucas and Prescott threw themselves into the newly-developed mathematics necessary to make the method work. With dynamic programming (and other ways of describing processes that unfold over time), economists could describe agents operating in a thoroughly chancy (stochastic) world. The very meaning of equilibrium would change; instead of connoting a system at rest, economic equilibrium had become whatever happened in an interdependent system. But it could be persuasively done with adaptive expectations.
As it happened, Carnegie had a team working on just this problem It’s a great story – I wish I had time to tell it here – how a team working on factory scheduling produced both Herbert Simon’s concepts of “bounded rationality” and “satisficing” (meaning good enough), and John Muth’s concept of rational expectations. Nor do I have time to explain in detail how Muth set out his idea in opposition to the time-honored “cobweb” model of the corn-hog cycle. (Jürg Niehans does it with six equations in A History of Economic Theory: Classic Contributions 1720-1980.)
It’s not that all expectations are the same, or identical to the model, or perfect. Rather word gets around; the Scotchman avails himself of whatever serves as social media, and, in the end, knows pretty much what the model-maker knows. There is no systematic gap between the forecast of the model and the people. They have become much harder to fool. Lucas put it this way to economist Michael Parkin:
You would never discover the idea… by introspection. Rational expectations describes something that had to be true of the outcome of a much more complicated underlying process. But it doesn’t describe the actual thought process people use in trying to figure out the future. Our behavior is adaptive. We try some mode of behavior. If it’s successful, we do it again. If not, we try something else. Rational expectations describes the situation when you’ve got it right.
Lucas’s argument came in four stages. First there was the paper with Prescott, “Investment Under Uncertainty.” Then came another, with his friend Leonard Rapping, of the University of Massachusetts, that became an important part of the micro-foundations of macro literature. “Expectations and the Neutrality of Money” appeared in 1972; in 1995 it was the foundation of Lucas’s Nobel Prize. And in 1976 appeared perhaps the most influential of all Lucas’s papers, known ever after as “the Lucas critique” – the Borges-like proposition that econometric models must incorporate the effects of the very changes they seek to evaluate, Gradually the idea morphed into a “law” named for British central banker Charles Goodhart. “When a measure becomes a policy target, it ceases to be a good measure.”
The rational expectations steam became a river. Sargent and Neil Wallace propounded a doctrine of “policy ineffectiveness.” Robert Barro restated an important old idea in modern terms: that because people are forward looking, taxation and public borrowing must have roughly equivalent effects. Finally, in “Rules rather than Discretion: The Inconsistency of Optimal Plans,” Norwegian economist Finn Kydland (yet another Carnegie student) and Prescott put credibility at the center of monetary policy, describing rate setting as a game between central bankers and market participants rather than an exercise in optimal control.
Keynesian economics reeled under the assault. So did the old –fashioned Chicago School. Samuelson and Solow worked on other topics – Samuelson on finance, Solow on monopolistic competition. Friedman dueled with Robert Mundell over the desirability of floating exchange rates, as the old Bretton Woods system of fixed rates broke down. Mundell stalked out of Chicago, but not before training the next generation of international economists in what became known as the monetary approach to the balance of payments, Rudiger Dornbusch, Jacob Frenkel, Michael Mussa, and Assaf Razin. In 1974 Lucas returned to Chicago, where he had been Friedman’s student. Friedman retired to California two years later.
MIT-trained Stanley Fischer taught at Chicago for several years, then returned to MIT, bringing the new problems and new tools with him. Dornbusch joined him there the following year. The two wrote a text together, becoming the public face, with Robert Gordon of Northwestern University, of sadder-but wiser Keynesian macroeconomic management. At Columbia University, John Taylor sought to square monetary policy with Keynesian sticky prices and rational expectations.
In the terms formulated by Chicago’s Harry Johnson in his 1970 Ely Lecture, rational expectations had everything it needed to have to become a successful revolution: it attacked a central proposition of a well-established orthodoxy; it appeared new, yet it assimilated most of the existing structure of thought; it seemed complex and hard to understand but was actually fairly easy to use; it offered major opportunities to those seeking to get ahead; and, finally, it gave econometricians something to measure.
By 1978 victory seemed complete. Lucas and Prescott travelled to a research conference at Bald Peak, New Hampshire, and, with Samuelson himself in attendance, administered a famous slapdown to the Keynesian program. Lucas restated Friedman’s 1948 “A Monetary and Fiscal Framework for Economic Stability” in strictly modern terms for his part of the program.
Yet if there had been a revolution, it didn’t last very long. In 1976, Robert Hall, of Standard University, surveyed the two key points on which opinions differed – why were there business cycles? what, if anything, could the government do about them – and concluded that the Keynesians and the monetarists were in the process of re-emerging as the salt-water and fresh-water schools, differently labeled mainly to reflect their shared commitment to the new methods and tools.
It would have been equally true to say that the profession was gradually dividing its house into wings devoted to mostly to practical engineering on the one hand, and mostly to advancing the aims of science on the other.
The Practicals had a different view of all this, naturally. Take Paul Volcker, for example. William L. Silber’s book, Volcker: the Triumph of Persistence (Bloomsbury, 2012), paints a vivid portrait of the career of a central banker who, as an undergraduate at Princeton University in the early ’50s, learned skepticism at the feet of Oskar Morgenstern.
Volcker was an official at the Federal Reserve Bank of New York when he first encountered Milton Friedman, coming to the bank to explain how the Federal Reserve had caused the Great Depression. “Milton was convinced he had found the gospel truth. I was skeptical of anyone so confident – whether Chicago or Cambridge.”
As Deputy Assistant Secretary for Monetary Affairs, under Douglas Dillon, he was the Treasury Department’s chief liaison in its dealings with the Kennedy Council of Economic Advisers. Volcker recalled:
It all sounded too easy. Push this button twice and out pops full employment. Equations don’t work as well on people as they do on rockets. I remember sitting in a [graduate school] class at Harvard listening to [the fiscal policy expert] Arthur Smithies say, “A little inflation is good for the economy.” And all I can remember after that is a word flashing in my brain like a yellow caution sign. “Bullshit.”
Volcker had to deal with managing the dollar through the withdrawal from the gold standard and the end of the Bretton Woods Treaty in the early ’70s. Called back by President Jimmy Carter in 1979 to serve as chairman of the Federal Reserve Board, Volcker led an epic battle against inflationary expectations over the next three years. It was to his old Treasury Department friend Frank Morris, long-serving president of the Federal Reserve Bank of Boston, not Friedman, to whom he turned for advice on targeting monetary aggregates.
But by standing firm against political pressures from every direction, Volcker probably did more to sell non-economists on the power of the rational expectations view than anyone in the profession. Sargent has written,
It goes without saying that the credibility that is essential under the rational expectations theory cannot be manipulated by promises or government announcements… [only a] once-and-for-all, widely believed, uncontroversial and irreversible regime change… can cure inflation at little or no cost in terms of real output.
The cost was not trivial. But Volcker demonstrated that expectations matter more than anything else.
The late ’70s moved swiftly for Kenneth Arrow. He still spent summers in California, more time at the Institute for Mathematical Studies in the Social Sciences (reconfigured in the early 1990s as the Stanford Institute for Theoretical Economics), where a great collision over principal-agent theory was slowly taking place between general equilibrium theorists and game theorists.
Minnesota’s Hurwicz, too, presided over a California series of workshop in the summers, the Decentralization Conferences at the University of California at Berkeley. Gradually researchers at both ends of the San Francisco Bay came to see they were working on the same problem from different sides. None of this was obvious to the public. But the exhilaration among leading economists was very great.
There were disappointments: New research showed that hopes for creating a clearly determined general equilibrium theory seemed to be at a dead-end. Arrow’s last student at Harvard, John Geanakoplos, took a job at Yale and began mortgage trading. Arrow returned to Stanford in 1979 and began to collect his papers.
Arrow had been at Harvard for eleven years. You wouldn’t say that the major developments in both micro and macroeconomics of the previous twenty years had happened under his baton. You might agree that what he had been was economics’ new magnifying lens.