Paul Krugman

Wonking Out: What Would a Hard Landing Look Like?

A food bank for unemployed steelworkers in Braddock, Pa., in 1983.
Credit...Bettmann via Getty Images

Opinion columnist

Programming note: This newsletter will be off for the next two weeks, returning Jan. 4.

On Wednesday the Federal Reserve announced that it plans to make monetary policy a bit less easy out of concerns over inflation. Some of the media coverage made it seem like a road-to-Damascus moment, a sudden repudiation of everything Jerome Powell and associates have been saying. But if you read the Fed’s statement carefully, not so much: The Fed is still of the view that a lot of recent inflation is tr-tr-transit- … OK, we can’t use the T-word anymore, so maybe say that it’s fugacious?

The actual announcement says, “Supply and demand imbalances related to the pandemic and the reopening of the economy have continued to contribute to elevated levels of inflation.” Sounds fugacious to me. And the Fed’s projections still include a combination of falling inflation with falling unemployment.

There are good reasons for that optimistic assessment. But the Fed has been wrong about inflation so far. What if it’s wrong again?

Let’s revisit some history. Although we talk a lot about inflation in the 1970s, the real pain came in the 1980s, as Paul Volcker’s Fed tried to bring inflation down. It succeeded, but at an enormous price. Here’s the Fed’s preferred measure of underlying “core” inflation versus the unemployment rate:

Image
Credit...FRED

Getting core inflation down to around 4 percent from around 10 percent involved a huge surge in unemployment that took a long time to reverse: Despite “morning in America,” the unemployment rate didn’t get back to its pre-Volcker level until toward the end of the Reagan administration.

It’s not totally clear whether this huge wave of unemployment was necessary. Influential research using state-level data argues that the main factor in the Volcker disinflation was a big change in public expectations that could conceivably have happened without such a severe recession. But for now the working hypothesis for most economists is still that it took a nasty, sustained slump to end the inflation of the 1970s.

Are we looking at something similar in our future?

Probably not, for several reasons.

For one thing, despite high headline numbers lately, underlying inflation by the end of 2022 isn’t likely to be anywhere near 1980 levels. Standard measures are currently unreliable because of pandemic weirdness — who knew used cars could loom so large in the statistics? But possibly more robust measures like the Atlanta Fed’s “sticky price” inflation or the Dallas Fed’s “trimmed mean” suggest that a Powell disinflation, if it has to happen, would start from 3 percent or 4 percent, not Volcker’s 10 percent. In fact, the starting point for such a squeeze would be roughly the end point of Volcker’s squeeze, which raises the question of why we should even bother.

Back to that in a minute.

Even if we assume that we will have to get inflation down to 2 percent from, say, 3 or 4 percent, that’s only a third or a quarter of the 1980s adjustment. And there are reasons to believe that the cost would be even smaller than that comparison implies.

A number of economists have suggested that the current inflation looks more like 1946-48 than like the 1970s. Comparing the two episodes is tricky, in part because we don’t have standard measures of core inflation going back that far, and overall inflation was unstable as families still spent a third of their income on food. But one rough-and-ready way to get something like core inflation for the 1940s is to look at service prices, which were much less volatile than goods prices. Here’s what the disinflation looked like:

Image
Credit...Bureau of Economic Analysis

It looks like a 5- or 6-point decline in underlying inflation, roughly comparable with what happened in the 1980s. But the cost of elevated unemployment was much lower. Here’s the rise in unemployment over time from its initial level in the recessions that ended both inflations:

Image
Credit...Bureau of Labor Statistics

Comparable peaks, but much shorter duration in the 1940s. If we count the “point-years” of excess unemployment — one year of one percentage point elevated unemployment counts as a point-year — the Volcker disinflation cost 14 such point-years, the ’40s disinflation only a third as much.

Why the difference? By 1980 America had experienced many years of high inflation, so it was hard to persuade people that the inflation era was over. That wasn’t true in the 1940s, when some prominent businessmen were still betting on the return of the Great Depression.

So where are we now? Elevated inflation has been going on for less than a year, which suggests that it should be relatively easy to get rid of.

That is, to the extent we want to. The Fed’s 2 percent inflation target was set a couple of decades ago, based to a large extent on economic analyses that have turned out to be wrong — particularly the belief that situations in which even zero interest rates weren’t low enough to yield full employment would be vanishingly rare. There’s actually a very good case that we’d be better off on average with 3 percent or more inflation than we would be returning to 2 percent.

So there may be a real dilemma ahead. Suppose that we find ourselves a year from now with 3-ish percent inflation, and it’s clear that the Fed would have to impose at least a mild recession — costing hundreds of thousands if not millions of jobs, at least for a while — to get it back to 2. Would that be a price worth paying?

Right now the Fed is betting that this dilemma won’t arise. Let’s hope it’s right.