By the late 1990s the American economy was doing something it was not supposed to do. Unemployment kept falling — past 6 percent, past 5.5, toward 4 — and inflation, which the textbooks said should have been accelerating, kept drifting quietly down. People needed a name for it, and the name they chose was the new economy.
The new economy was really a bundle of three claims, each an answer to the same question: why can the economy run this hot without overheating? The first was that productivity growth had accelerated — output per hour was rising faster than in a generation, so costs and prices stayed tame even as the economy boomed. The second was that the NAIRU had fallen — the lowest unemployment rate the economy could sustain without igniting inflation was lower than it used to be, so there was more room to run. The third was that globalization and competition were holding prices down — cheap imports, a strong dollar, and firms that simply could not make a price increase stick. In the outside debate the three usually traveled together. Often the same optimists believed all three at once; they added up to a single conviction that the old rules had loosened their grip.
Inside the Federal Open Market Committee you can watch the three claims get argued out, meeting by meeting, and — this is the part that matters — priced, or not priced, into the one thing the committee controls: the federal funds rate. This essay reads the transcripts from 1993 through 2001 and codes what each participant’s argument implied for rates. Two things jump out before any of the channels do.
The first is that most of what the committee argued about was not the new economy at all. Meeting after meeting, the live question was ordinary and cyclical: the Mexican peso crisis, the Asian financial crisis and the collapse of Long-Term Capital Management, the Y2K liquidity scare, whether demand was softening, whether the real funds rate was too high for the moment. When a member pushed for easier policy than the committee was about to deliver, the reason was usually one of these — not a structural story about a changed economy.
The second is that when a genuinely structural argument was made, it was usually — though not always — a dovish one, a reason for patience. And among those structural arguments, the three new-economy stories were nowhere near equal in weight.
That ordering — productivity, then the NAIRU, then globalization — holds up whichever way you measure it. Productivity was not just the most persuasive of the three; it was the one members reached for most often, by a wide margin.
So the new economy, at the FOMC, was less a single idea than three loosely related ones that even their believers rarely joined up. The thread that runs through all of them is a trap in the sign. A structural claim only matters for policy once you decide which way it points for rates — and the new economy pointed both ways at once. Nowhere is that clearer than with the argument the committee leaned on hardest.
Channel one · the workhorse
Productivity: one fact, two opposite signs
The first channel is disinflation. If productivity is rising quickly, unit labor costs stay contained even when wages rise, and inflation stays quiet. That gives the Fed room — room to wait, room not to hike, even room to ease. On this reading, faster productivity argues for a lower funds rate than you would otherwise run. This was Alan Greenspan’s instinct for most of the decade, and he put it about as plainly as a central banker can:
If it is true, then clearly we can reach price stability with real interest rates lower than where they are now.Alan Greenspan, December 1995
The second channel runs the other way. Faster productivity raises the return on capital and the economy’s underlying growth rate — and therefore its equilibrium real interest rate, the level economists now call r*. It also makes people richer and readier to spend, lifting demand. On this reading, faster productivity means the appropriate funds rate is higher, not lower. Jerry Jordan of the Cleveland Fed was the clearest early voice for it:
We would expect a favorable productivity surprise, other things the same, to translate into higher real interest rates.Jerry Jordan, March 1997
These are not contradictions to be resolved so much as two real forces of uncertain relative size. The same sentence — “productivity is why we needn’t worry about inflation” — is dovish through the price channel and hawkish through r*. Which one a careful person emphasized was the whole ballgame, and it turned on a judgment no data could settle in real time.
Read chronologically, the committee’s productivity talk has a clear shape. From 1993 to 1996 it is almost entirely the disinflation channel; the argument is whether productivity has really risen — the official numbers were soft and confusing — and, if so, whether it buys room on inflation. The r* argument is essentially absent. It arrives in 1997, first from Jordan and then from Al Broaddus of Richmond, who became its steadiest advocate. Through 1998–2000 both run at once, with the hawkish reading peaking in 2000 as the stock market and demand surged. Then in 2001 the recession arrives and productivity falls off the rate agenda almost entirely.
The two readings sorted the room into two camps — with a couple of people, tellingly, in both.
Greenspan changes his mind
The single most interesting trajectory belongs to the Chairman. For most of the decade Greenspan was the leading exponent of the dovish reading, and the boom’s great optimist besides, forever insisting the true productivity numbers were higher than the measured ones. Then, in late 1999, he flips. At the November meeting he makes the r* argument himself:
Accelerating productivity exerts upward pressure on real long-term rates.Alan Greenspan, November 1999
and in early 2000 he adds the wealth-effect version, productivity lifting expected earnings, asset prices, and demand. The same fact that had justified patience for years was now, in his telling, a reason the economy might be running too hot. It is the clearest instance of a wider pattern: the committee did not so much disagree as migrate, and the Chairman migrated with it.
The staff saw it the other way
A note on the staff. The people quoted next are not voting members. They are the Board’s own economists — the authors of the staff forecast (the “Greenbook”), Mike Prell and then David Stockton, and the senior policy adviser Donald Kohn. They build the forecast the Committee argues over, but they do not set rates. They are worth separating out because their reading of productivity was strikingly uniform — and the opposite of the Chairman’s.
Here is the inversion at the heart of the story. Greenspan, the boom’s greatest optimist, read productivity dovishly. The staff, who started as its greatest skeptics, read it hawkishly once they came to believe it: faster productivity raises r*, so rates must eventually rise. Donald Kohn held that line from 1997 on, and was blunt that the disinflation was no license to ease:
Board staff · non-votingEasing policy under those conditions as inflation fell would risk greater price pressures down the road.Donald Kohn, senior staff adviser, December 1998
Board staff · non-voting… we end up with a higher equilibrium interest rate approximately equal to the size of the productivity shock.Donald Kohn, February 2000
David Stockton, by then running the forecast, tied the two together directly — “we are in the midst of an acceleration in productivity and the equilibrium real funds rate has to go up along with that” (February 2000). So the real fault line in the room was not “did productivity rise?” By 2000 almost everyone agreed it had. The fault line was higher rates — the staff, joined by the r* voters (Broaddus, Jordan, and Meyer at his hawkish meetings) — against patience, Greenspan and the dovish majority.
None of this was obvious at the time, because for years the fight was simply whether the speed-up was real. The staff were the institutional doubters — Prell unimpressed by “some loose thinking behind some of the more optimistic stories” (1994), Stockton suspecting “a statistical artifact” (1995). The Greenbook even cut its assumed productivity trend in early 1997, just as the real acceleration was getting underway, before ratcheting it up year after year. The capitulation, when it came, was gracious about who had been right:
Board staff · non-votingJust as the Chairman predicted, the incoming data have prompted us to raise our sights on the pace of structural improvements in output per hour. We’re now putting it at 3.2 percent …Mike Prell, March 2000
There is a quiet drama in that sentence: the optimistic Chairman dragging a skeptical staff to his view of the data — and the staff, once converted, turning his optimism into a hawkish conclusion he himself resisted.
Will it last? A separate question
Underneath the rate debate sat a forecasting question that shaped everything: was the speed-up durable, or a passing blip? This is logically separate from the rate sign — you can think productivity is disinflationary and doubt it will last — and the transcripts treat it that way. Bob McTeer of Dallas and, on balance, Greenspan leaned toward a lasting change; Larry Meyer led the skeptics, warning that the friendly numbers were borrowed from temporary supply shocks that would reverse.
Channel two · the contested middle
The NAIRU: a real minority, a fractured one
If productivity was the argument the committee agreed on — even the skeptics conceded the facts by 2000 — the natural rate of unemployment was the argument it never settled. And here is the first surprise: the people who believed the NAIRU had fallen were not simply the productivity doves under another name. The NAIRU split the very group that agreed on productivity.
Across the period, the committee’s modal view was that the conventional natural rate roughly held. A committed “it has fallen” view was a real but distinctly minority position, and genuine agnosticism was nearly as common as either.
The second surprise is that the two ideas were held side by side rather than joined. One might expect members to argue that faster productivity is why the natural rate fell. A few did — Gramlich above all, with Meyer and Jordan — but only about thirty times in nine years. The far more common move was to keep them apart: productivity restrains inflation given the natural rate; the natural rate is a separate question about the labor market. The workhorse and the contested middle ran on separate tracks.
Nothing shows the separation better than Larry Meyer. Meyer was one of the committee’s most active productivity doves — he made the disinflation argument repeatedly. Yet on the natural rate he was the framework’s great defender, insisting it still sat around 5½ percent and refusing to read low unemployment as proof it had dropped:
I do not believe … that the decline in core inflation experienced over 1996 suggests that NAIRU is below the 5.4 percent average unemployment rate for the year. I suspect that NAIRU, at least for the near term, is near 5½ percent.Laurence Meyer, February 1997
Against him stood the believers. Yellen made the structural case early and in textbook terms — that changes in the labor market “would shift the natural rate of unemployment and result in a persistent, not just a very transitory, decline” (1996). Gramlich was the careful, durable adopter of the late-1990s wave. At the other end, Jerry Jordan doubted the concept itself at nearly every meeting, and Tom Hoenig was the lone voice putting the natural rate above the staff’s. Greenspan, characteristically, declined to plant a flag: he conceded his productivity hypothesis was “the statistical equivalent of a falling NAIRU” but found saying so “not very helpful,” and set the concept aside in favor of watching the economy directly.
There is one more inversion worth marking, because it runs opposite to the productivity story. On productivity, the staff were behind — the last in the building to believe. On the natural rate, several members were ahead of the staff. Yellen and Alan Blinder pressed the case for a lower NAIRU in the mid-1990s; Gramlich carried it later; and the Greenbook’s own estimated natural rate drifted down across the decade to meet them. Here the committee, not the staff, moved first.
And yet — this is why the NAIRU is the contested middle and not a second workhorse — the explicit argument rarely drove a decision. It was raised about a fifth as often as productivity and carried only around a tenth of the committee’s dovish cases. A genuine belief, held by a genuine minority, ahead of the staff, and still not the engine of policy.
Channel three · the tailwind
Globalization: cited constantly, trusted by almost no one
The third story was the one members mentioned most casually and believed least. Global competition, a strong dollar, cheap and falling import prices, firms that could not raise prices without losing customers — the Beige Book was full of it, and it surfaced in more meetings than the explicit NAIRU argument did. Greenspan gave it its sharpest formulation:
Pricing power is gone, and the reason it is gone … is that competitors have sufficiently high rates of return to enable them to undercut price increases and go for market share, should some firm try to raise its prices.Alan Greenspan, March 1997
But mention is not conviction. When members were pressed on whether this restraint would last, almost none said yes. Of the roughly one hundred times the channel was invoked as actively holding prices down, only about one in eight treated it as durable. The rest tied it to the strong dollar and the Asian crisis — good luck that would reverse — which is precisely why it almost never drove a dovish vote. You do not cut rates on a tailwind you expect to turn into a headwind.
And it turned when the luck did. As the dollar eased and Asian demand recovered in 1999, members who had leaned on cheap imports for comfort began to warn of the reverse — Mike Kelley noting that “world commodity prices could firm … the dollar’s foreign exchange value could begin to weaken … all of this is potentially inflationary” (June 1999). Only Greenspan held to a durable version, tying global competition to his lasting productivity story — “the technologically driven process that is breaking down barriers to cross-border trade … simulates the old gold standard forces.” For everyone else it was a real-time observation, not a new-economy conviction.
What actually happened
The rate the committee chose — and the lesson we kept
So how did nine years of argument cash out? After cutting in the 1998 crisis, the Fed reversed hard. In six increases from June 1999 to May 2000 it raised the funds rate by about 1¾ percentage points, to 6.5 percent — the highest in nearly a decade. It tightened, in other words, straight into the boom the optimists were celebrating.
Why? Because the reading that won was not the dovish new-economy one. It was the r* reading — the staff and the small band of r* voters, who said that a real productivity boom raises the neutral rate and that easing on productivity-disinflation would “risk greater price pressures down the road” — reinforced by the persistence skeptics, led by Meyer, who never trusted that the friendly supply shocks would last. The patience camp had the Chairman and the numbers; the tightening camp had the argument.
And events sided with the tightening camp. Even with 1¾ points of hikes, inflation was rising, not falling. Greenspan’s own most famous statement of the dovish thesis came in a May 1999 speech crediting the productivity pickup as “an essential factor behind the slowing in inflation.” One month later the Fed began raising rates — and that same month marked the start of a steady climb in core inflation, roughly half a percentage point over the following year and still rising when the 2001 recession finally intervened. Absent that recession, it would likely have kept going. The productivity boom turned out to be both real and, on balance, a reason to raise rates — not to hold them down.
Which brings us to the lesson people now draw from all this. There is a popular story — revived lately in the case for cutting rates ahead of an AI-driven productivity surge — that Greenspan spotted the boom early, understood it meant fast growth without inflation, and therefore held rates down. It is a seductive story. It is also, on the record, backwards. Greenspan did see the acceleration early — earlier than his own staff, which is to his lasting credit. But he did not sit still. A month after his clearest public statement of the thesis, the Fed embarked on the most sustained tightening of the expansion, and inflation rose anyway.
The economics is why the legend cannot be right. Over any horizon that matters, productivity does not set inflation — monetary policy does. What faster productivity does is raise the neutral real rate, which means a central bank has to hold nominal rates higher, not lower, to keep inflation in check. In the short run the sign is genuinely ambiguous: a timing race between supply catching up to demand and demand — equity prices, investment, spending — running ahead of supply. In the late 1990s there was no way to know in advance which would win, and the FOMC’s transcripts are, in the end, a nine-year argument about exactly that race. The people who worried demand would win — the r* camp and the skeptics of the new rules — turned out closer to right than the patience camp whose memory we have kept.
So do we remember the right lesson? The record suggests the myth has the moral upside down. The late-1990s boom is not evidence that a central bank can let a productivity surge run without tightening. It is evidence that a real productivity surge is itself a reason rates may have to rise — and that moving quickly enough is about the most a central bank can do to keep inflation from doing what it did the last time.
Sources and method: built from the public FOMC meeting transcripts, 1993–2001 (Miguel Acosta’s parsed corpus of the Fed’s own records). The unit is one policy participant at one meeting; each was read and coded under fixed rubrics calibrated with Jason Furman across three channels — the productivity–rate implication (549 speaker-meetings), the natural-rate view (217), and global-competition pricing pressure (156) — plus a decomposition of members’ dovish arguments (76). Board staff (Prell, Stockton, Kohn) are quoted as context, not coded, since the datasets cover voting participants only; their quotes were checked against the same transcripts. Counts on the dovish decomposition are approximate — the boundary of “dovish relative to the action taken” is a matter of judgment — but the mix is robust. The economics of the closing section draws on Jason Furman’s writing on productivity, r*, and the Greenspan legend. Essay written by Claude with extensive guidance from Jason Furman; inspired in part by Alexander Wolman’s Richmond Fed economic brief (2026).