Stay informed with free updates

Simply sign up to the US inflation myFT Digest — delivered directly to your inbox.

The writer is a professor at Harvard University and a former chair of the White House Council of Economic Advisers

Calls from President Donald Trump and others to cut interest rates increasingly rest on a seductive reading of Federal Reserve history. In the late 1990s, the story goes, former Fed chair Alan Greenspan recognised a productivity boom early, understood it would allow faster growth without inflation and therefore resisted raising rates. If the Fed could do that back then, the argument runs, it should be able to do the same now in the face of an anticipated AI-driven productivity surge. Kevin Warsh, Trump’s nominee to chair the Fed, has recently revived this narrative.

It is an appealing story. It is also largely wrong.

Greenspan did, in fact, grasp earlier than most that productivity growth had accelerated. In a May 1999 speech, for example, he reiterated his argument that “a pickup in the growth of labour productivity — beyond the effects of the business cycle — appears to have been an essential factor behind the slowing in inflation”.

But that is where the legend departs from the historical record. Just one month later, the Fed raised interest rates by 25 basis points. Moreover, the month of Greenspan’s speech marked the beginning of a steady rise in inflation. Core PCE inflation increased by roughly half a percentage point over the following year and continued to climb at a similar pace until the 2001 recession was well under way. Without that recession, inflation would probably have continued to rise well above the Fed’s comfort zone.

That the Fed tightened policy aggressively during the late 1990s productivity boom should not be surprising once the economics are stated correctly. Over the long run, productivity growth does not determine inflation. Productivity reflects the economy’s real productive capacity; inflation reflects monetary policy choices. But sustained faster productivity growth does raise the economy’s neutral real interest rate. To prevent inflation, central banks must therefore maintain higher nominal rates.

The mechanism is straightforward. Faster productivity growth allows households to save less because they anticipate higher future income, while prompting businesses to invest more because expected returns rise. Both of these boost demand and push up real interest rates.

In the short run, an unexpected acceleration in productivity can influence inflation, but the direction is ambiguous. Greenspan’s hypothesis was that higher productivity allowed nominal demand to grow faster without igniting inflation. Because wages adjust less frequently than prices, this initially showed up as slower price growth rather than faster wage growth. That dynamic may well have characterised the early years after productivity began accelerating in the mid-1990s.

But there is a competing short-run effect that runs in the opposite direction. Anticipation of a sustained productivity boom can itself be inflationary, by lifting equity prices and household spending and by spurring business investment.

At bottom, this is a timing race: does demand surge ahead of supply, or does supply expand fast enough to accommodate demand without inflation? In the late 1990s — just as today — there was no clear way to know in advance which would dominate.

This brings us to the present, and to a crucial difference from the late 1990s. Over the past year, productivity has grown at a 1.9 per cent pace — modestly better than the last business cycle, but nowhere near the 3.9 per cent seen in the year before Greenspan’s 1999 speech. What we do have today is extraordinarily high expectations of productivity growth, already reflected in equity prices and business investment, boosting demand well ahead of any realised supply gains.

Moreover, inflation is currently running close to 3 per cent and has been above target for more than five years. We can ill-afford the risk of another half, let alone full, percentage point increase in the inflation rate that we experienced at the end of the 1990s.

The Fed may have good reason to cut interest rates more this year. But if so, it is much more likely to be in reaction to bad news, such as a deteriorating labour market, rather than the good news that an AI productivity boom has arrived.

And if the optimistic scenario does materialise, with unemployment stabilising and productivity finally accelerating, the true lesson Warsh should take from the late 1990s is not that the US central bank should stand pat. It is that the Fed may once again need to raise rates, even in the face of a productivity boom — and hope that acting quickly is enough to prevent inflation from doing what it did last time.