A striking array of first-magnitude stars turned out last week when the Federal Reserve Bank of Boston convened a conference on Cape Cod to mark the fiftieth anniversary of the publication of the Phillips curve. That possible link between the unemployment and inflation rates, first observed as “an empirical regularity” by A.W. Phillips, has been at the heart of Keynesian economics for half a century. Plausible? Reliable? Chimerical? That’s what they came to debate.
Robert Solow, who, with his Massachusetts Institute of Technology office-mate Paul Samuelson, catapulted the mild-mannered New Zealander’s regularity to a (perhaps) dependable trade-off in 1960, was there. So was John Taylor, author of the eponymous central bankers’ rule-of-thumb target that more or less has replaced the Phillips curve as a guide to monetary policy, and which seems to have delivered remarkable stability in the process.
So were Federal Reserve chairman Ben Bernanke, charged with maintaining that stability; and a chorus of other distinguished central bankers (and shadow central bankers) as well: Stanley Fischer, governor of the Bank of Israel; Donald Kohn, vice chair of the Fed’s Board of Governors; Jurgen Stark, executive committee member of the European Central Bank; Lars Svensson, deputy governor of the Bank of Sweden; and Alan Meltzer, of Carnegie Mellon University.
And so was Christopher Sims, of Princeton University, the persistent skeptic, who at every turn voiced doubt that what he called the central tenet of “Phillips curve thinking” – that tightness of some sort or other in the real economy is the crucial determinant of money and inflation – is a useful way to organize economists’ understanding of the process.
All the stars were there, that is, but two – the two premier intellectual historian-interpreters of the post-World-War-II stabilization effort.
Christina Romer, of the University of California at Berkeley, preeminent historian among those who identify themselves as “New Keynesian” economists, didn’t attend. Neither did Thomas Sargent, of New York University, a leader of the New Classical school. (Both had been invited; each had more pressing commitments elsewhere)
Besides, the collision between the two had occurred earlier, in 2002, at a similar meeting, at Jackson Hole, Wyoming, “Rethinking Stabilization Policy,” sponsored by the Federal Reserve Bank of Kansas City. The differences of opinion between Romer and Sargent about the control of inflation, and between the extensive factions that they represent, are among the most complicated and interesting in all the discipline. Don’t expect the argument to be resolved any time soon.
The persistence of the controversy may, however, shed some light on an issue that is at once more pressing and more tractable. The economics department of Harvard University frequently boasts that it is the best in all of economics. Why, then, did President Drew Faust last month unexpectedly override its offer to Romer, indisputably one of the most widely-respected women in the profession?
It makes sense to recall, briefly, the beginning of the story. For practical purposes, that would be at another Federal Reserve Bank of Boston conference, this one in October 1978.
The first great collision between the Keynesian establishment that had dominated economics in the 1950s and 1960s, and the upstarts New Classicals, who would dominate it for the 1980s and 1990s, took place at the Bald Peak Colony Club, a private golf resort on Lake Winnipesauke, N.H. Paul Samuelson himself attended.
Robert Lucas gave a defense of Milton Friedman’s 1948 paper, “A Monetary and Fiscal Framework for Economic Stability,” emphasizing the support for Friedman’s ideas provided by recent work on the role of expectations. Edward Prescott gave a version of his work with Fynn Kydland that eventually would become, “Time to Build and Aggregate Fluctuations” – a brash new theory of the business cycle in which real factors, mostly changing technologies, played the leading role, rather than monetary shocks. These ideas had been percolating in Pittsburgh, Rochester, Chicago and Minnesota for several years. Thomas Sargent didn’t come, perhaps because he was beavering away on another epochal project, “The Ends of Big Inflations.”
For the next few years, the new classicals were everywhere, taking over and refining positions formerly known as “monetarist,” advancing explanations for why government policies to manage the business cycle in various ways might be expected to fail. The framework was set for a long struggle between the generations. Much of the argumentation took high-tech forms (the term itself was new in those days). Time was required to master the new mathematical techniques and their applications. Graduate students gradually chose up sides.
By the mid 1980s, it was common to distinguish between “freshwater” and “saltwater” economists, their loyalties generally lining up with the locations of universities where they worked. At schools located around the Great Lakes and along the great rivers in the middle of the country, freshwater economists professed faith in markets and the ingenuity of the common man.
Saltwater economists, in universities hard by the ocean in Massachusetts and California, saw many sorts of market failures, “bad equilibria” of output or employment, which they reckoned they could improve through government policy. By 1991, this had become “New Keynesian” economics, summed up in a two-volume collection of important papers edited by N. Gregory Mankiw, of Harvard, and David Romer, of Berkeley, both recent graduates of the Massachusetts Institute of Technology. The first volume was devoted to issues arising from imperfect competition and sticky prices, the second to coordination failures and various real rigidities.
The introduction was prepared by Benjamin Friedman and Lawrence Summers, both of Harvard University. The New Classicals in the ’70s had promised Americans the moon, they wrote: rapid and costless disinflation, longer expansions, faster growth. But events of the ’80s, in order to be “more easily understandable,” had inspired a return to essentially Keynesian models of wage and price stickiness.
Playing no part in the movement was Christina Romer (no relation, either, to Stanford’s Paul Romer). She was immersed in the business cycle. A series of exemplary articles in top journals tumbled from her pen: “World War I and the Postwar Depression: A Reinterpretation Based on Alternative Estimates of GNP”; “The Prewar Business Cycle Reconsidered: New Estimates of Gross National Product, 1869-1908”; “The Great Crash and the Onset of the Great Depression”; “The Cyclical Behavior of Individual Production Series, 1889-1984.”
By the early years of the twenty-first century, the conversation had taken a different turn. What economists were calling “the Great Moderation” had become the standard puzzle. Inflation had been almost completely subdued, with relatively little increase of unemployment; there had been only two relatively mild and short recessions in twenty-five years. Had the New Classicals won the argument? Had the Keynesians, Old and New, lost? The matter came to something of a head in 2, when the Kansas City Fed organized its meeting on “Rethinking Stabilization Policy.”
In the scene-setting paper, Romer and her macroeconomist husband and co-author, summarized what had become known as “the Berkeley View” of what had happened since 1945. (J. Bradford Delong, another Berkeley professor, had been among the first to give a version of it, in 1997.)
The evolution of economic understanding had not been a linear progression from dark to light, they said, but rather a more interesting story, in which policy-makers possessed “a crude but fundamentally sensible model of how the economy worked in the 1950s, [which gave way] to more formal but faulty models in the 1960s and 1970s, and finally to a model that was both sensible and sophisticated in the 1980s and 1990s.” And how had they gone astray? By embracing a naive reading of the Phillips curve and thinking they could reduce unemployment to very low levels at little cost by increasing the inflation rate. Gradually, Milton Friedman and Edmund Phelps observed that there must be a “natural rate,” below which unemployment couldn’t go, without kindling inflation. But it took most of the decade of the ’70s before policy makers (namely Paul Volckler) regained their ’50’s-style nerve in order to run unemployment up to levels that would curb an inflation that increasingly threatened the social fabric.
Commenting on their paper, Thomas Sargent painted a somewhat different picture. For one thing, the Romers had left out precisely the considerations he thought most relevant, “some of the most important and useful ideas” developed by economists of recent years, including rational expectations, commitment and time-consistency problems, reputation as a substitute for commitment, coefficient drift, to name only a few. (He had outlined what he viewed as the significance of those ideas in his1999 book, The Conquest of American Inflation.) Ungallantly, he added that the absence of these New Classical conventions from the Romers’ account was underscored by their presence in another recent book, this one by Alan Blinder, Princeton professor, former Fed vice chairman, and Old Keynesian, whose account of The Art of Central Banking “draws the reader into considered arguments about almost all of these ideas and how they constrain or inform monetary policy decisions. The Romers’ approach was literary and narrative, and “not tight enough to be statistically verified.” Sargent wanted for formal models of the Fed’s learning dynamics in order to be clear. The alternative was to risk a triumphalist version of events that might simply have been good luck, and that would give no warning that relief might be only temporary.
Some sharp exchanges followed in the discussion period. Christina Romer took exception, for example, to the view that her narrative approach was “literary.” Instead it was “science in a different form.” Serious sources, read systematically, consistently confronted with hypotheses, was a legitimate technique with which to inquire into policy-makers’ beliefs. Aside from Fed forecasts, what other data were there?
The differences of opinion only broadened in the next few years. In an address to the Economic History Association last year, Romer discussed “The Causes and Consequences of a Mistaken Revolution” (meaning Keynesian doctrines of the ’60s) and the joys of having learned in recent years from those mistakes. “Though there is a camp that emphasizes the role of good luck, I feel the contribution of good policy cannot be overstated.”
Sargent, for his part, returned to the Phillips curve with a vengeance in January, in his presidential address to the American Economic Association. “Evolution and Intelligent Design” is an ambitious, telegraphic (despite its length), and brilliant brief for the new learning and expectations literature in macroeconomics. It will take some years to be widely understood. But the difference between the narrative and model-base approaches to stabilization policies could hardly be more clear.
It is against this background that attempted explanations of the recent news from Harvard should be understood. Nobody is yet talking with much candor about President Faust’s decision to overrule the economics department offer to Romer, which was especially surprising since Harvard has been under fire for years for the relative paucity of females among its senior faculty; and Romer is highly respected within the profession. (The offer to her husband, the “trailing spouse,” from Harvard’s Kennedy School of Government, sailed through). I relate this background in hopes that it will spur greater clarity.
Two overlapping possibilities emerge. The first is a story of pure ideas, of the merits of high theory vs. narrative methods in present-day economics. Three of the most influential figures in the New Classical movement are Harvard PhDs – Robert Barro, Sargent and Christopher Sims. Barro is a member of the Harvard department and widely believed to have privately opposed the offer in the Ad Hoc Committee hearing that preceded Faust’s decision. (Harvard is all but unique in its requirement that the university president, review every offer.) Meanwhile, the Harvard department earlier this spring declined to promote Aleh Tsyvinski, a Minnesota PhD and specialist in dynamic fiscal policy, to a tenured position. And while there is nothing unusual about Harvard failing to promote its junior faculty, the decision could have serves as a rallying point for a faction, inside and outside the university, convinced that the New Classical movement is moving down the most promising path.
Another, somewhat related possibility is more sociological: it posits something like an “immune response” on the part of Harvard against economists acquired from nearby MIT. Harvard famously was eclipsed by MIT after Paul Samuelson decamped in 1940 to what in those days was called “Tech,” at the other end of town, in an argument over the future direction of economics. Robert Solow followed him a decade later. Together they attracted a seemingly endless stream of top students and junior faculty over the next thirty years. Meanwhile the Harvard department declined, at least until it made a series of notable hires in the late 1960s, led by Kenneth Arrow, Dale Jorgenson, Martin Feldstein and the late Zvi Griliches.
Harvard finally succeeded in reversing the flow in 1983, when it hired back its own PhD, Lawrence Summers, from the MIT faculty. (Earlier he had been an undergraduate there.) Already clearly destiny’s child, Summers helped attract MIT PhDs N. Gregory Mankiw in 1985, Lawrence Katz in 1986, David Cutler and Andrei Shleifer in 1991. And the presence of that growing corps of youthful talent, coupled with various dissatisfactions at the other end of town, led to a reverse exodus of senior faculty from MIT to Harvard, all scholars already in or near the Nobel nomination league – Eric Maskin in 1985, Robert Merton (to the business school) in 1988, Martin Weitzman in 1989, Oliver Hart and Drew Fudenberg in 1993. (In 2000, Maskin moved to the Institute for Advanced Study in Princeton.) Had they come, Christina and David Romer would have buttressed the younger cohort (both earned their degrees from MIT in 1986). In this view, Drew Faust simply finished what Larry Summers as president began: participating aggressively through the ad hoc proceeding to root out too many friends of friends.
Either explanation, if it suffices, would clarify Faust’s decision to overrule the offer to Christina Romer, and show that it arose mainly from the divisions within Harvard’s economics department, a mounting disdain for it within the university and its failure to remain in sufficiently close touch with the world without; and very little to do with the Berkeley professor. Romer’s reputation is intact. The irony is that today’s New Classical models may resemble those more formal but faulty models whose guidance marred the ’60s, while her somewhat old-fashioned methods are full of ’50s-style common sense. Low inflation and long-run fiscal balance are pretty good policy, after all.
In other words, in declining to hire Romer, Faust may have been wrong, for the right reasons. Harvard’s “New Keynesians” may have changed their banner — they are more likely to identify as behavioral economists today, as their research agendas have shifted — but their personal loyalties seem stronger than ever.
There is a much larger issue here, of course: Christopher Sims zeroed in on it at the Boston Fed conference last week. It is that financial markets, especially government debt, must be integrated into our understanding of monetary and fiscal policy if the wide swings of the last twenty five years of government deficits and the value of the dollar are to be understood. Nasty surprises await nations that assume bondholders can’t learn to anticipate the future. But that is an issue for another day.
As Sims put it, “the long-run value of Phillips curve theories may lie in the new flames that are emerging from [their] dying embers.