the economy

Larry Summers Was Wrong About Inflation

Photo: Tom Williams/CQ-Roll Call, Inc via Getty Imag

There aren’t many Americans who’ve benefited from the past two years of inflation. But Larry Summers is probably one of them.

In February 2021, the celebrated economist warned that Joe Biden’s $1.9 trillion American Rescue Plan was excessively large and that a sustained period of high inflation would follow in its wake. Summers’s argument was dismissed by White House officials and progressive policy wonks alike. Then consumer prices shot up to four-decade highs. Summers’s (already vaunted) stature within the business press rose. Even longtime critics felt compelled to acknowledge his foresight. The Biden administration started seeking his counsel.

From his newly elevated platform in June 2022, Summers delivered another prophecy of woe. During remarks before the London School of Economics, Summers declared that “we need five years of unemployment above 5 percent to contain inflation — in other words, we need two years of 7.5 percent unemployment or five years of 6 percent unemployment or one year of 10 percent unemployment.” In the Harvard professor’s estimation, U.S. policymakers had no choice but to deliberately throw millions of Americans out of work or else accept a steadily deepening inflationary crisis.

It is now clear that Summers was wrong.

His call for austerity was premised on the notion that only a sharp increase in unemployment could prevent a ruinous wage-price spiral. In reality, both wage and price growth have been slowing for months, even as unemployment has remained near historic lows. Summers’s failure to anticipate this outcome should lead us to reconsider just how prescient his analysis of the post-COVID economy ever was.

The core consumer price index (CPI) — a measure of inflation that strips out food and energy prices, which are unusually volatile — peaked in September 2022 at a 6.6 percent annualized rate. According to government data released Tuesday, that figure fell to 5.33 percent in May.

And that number actually overstates the amount of inflation in the contemporary economy. Shelter costs are the biggest service-sector component in the government’s inflation index, comprising about a third of the CPI. And its official estimate for the price of housing in the U.S. today is outdated. The reason is simple: Since tenants ordinarily sign 12-month leases on their housing units, the amount that many individuals are currently paying for shelter actually reflects price dynamics from many months ago, rather than current prices. And the housing market has cooled off considerably this year. Thus, as economist Adam Ozimek notes, if you replace the CPI’s estimate of housing costs with the Bureau of Labor Statistics’ estimate of real-time rent growth, you’d see that core inflation has fallen “substantially” since last fall.

When one looks at overall inflation (i.e., including food and energy), the annual rate of price growth fell to 4 percent last month, its lowest level since March 2021. Critically, the price that Summers was most concerned about — that of labor — has been growing at a decelerating pace for months. The Atlanta Fed maintains a tracker of median wage growth, which shows the three-month moving average of that figure at any given point in time. That average peaked 6.7 percent in June 2022 and now sits at 6 percent.

Last month’s jobs report, meanwhile, showed average hourly earnings rising at their slowest pace in nearly two years.

This deceleration in both price and wage growth occurred even as America’s unemployment rate remained essentially unchanged, hovering near historic lows.

Of course, inflation remains well above the Fed’s 2 percent annual target. But given that core inflation has fallen by more than a percentage point since September, in the absence of any increase in the unemployment rate, it is no longer credible to claim that nothing less than five years of 6 percent unemployment could “contain inflation.”

Summers’s analysis looks worse when one interrogates its underlying premises.

As the economist explained in an interview with Slate, his belief in the necessity of mass job losses to quell inflation was based on the notion that the “natural unemployment rate” in the contemporary economy is 5 percent. To understand the significance of this claim, we need to unpack the meaning of the “natural unemployment rate.” Orthodox economists believe that there is a tight relationship between unemployment and inflation. This is because the most costly input for the production of virtually every good or service is labor. And the unemployment rate largely determines labor’s cost: If unemployment is low, and jobless workers are scarce, then employers will need to offer generous wages in order to recruit and retain talent. If unemployment is high, by contrast, then firms can often find workers willing to accept low wages.

From this insight, economists extrapolate that within any economy there is always a certain level of unemployment that perfectly corresponds with the desired rate of inflation: An unemployment rate above this “natural” level will lead to excessively low inflation, while an employment rate below it will not merely lead to excessively high inflation but to a rapidly accelerating rate of inflation. Which is to say: If unemployment falls below the natural rate, firms will raise wages in order to attract workers, the rising cost of labor will then force firms to raise prices, rising prices will lead workers to demand higher wages, low unemployment will force firms to meet those demands, and the rising cost of labor will force them to raise prices in a vicious wage-price spiral.

For decades following the inflation crisis of the late 1970s, economists considered avoiding such a spiral the government’s chief macroeconomic obligation. Thus, when unemployment threatened to fall below the hypothetical natural rate, the Federal Reserve routinely raised interest rates in hopes of increasing joblessness, even if there was not yet any sign of excessive inflation.

In June 2022, Summers believed that the natural unemployment rate was 5 percent, roughly 1.5 percentage points higher than its actual level at that time. And since the Fed did not wish to merely avert accelerating inflation but actually lower it by multiple percentage points, Summers reasoned that it would need to raise unemployment well above the natural rate for a sustained period of time.

One year later, it is apparent that Summers’s view was fanciful. Allowing unemployment to remain near 3.5 percent has not only failed to trigger a vicious wage-price spiral but has actually proved consistent with substantially reducing both wage and price growth.

There are a few potential explanations for Summers’s mistake. One is that the reasoning behind his estimate of the natural unemployment rate was flawed. In 2019, the U.S. managed to enjoy a 3.6 percent unemployment rate without any acceleration of inflation. Summers assumed that this meant that 3.6 percent was the natural rate before the pandemic; had joblessness fallen any lower than that in 2019, inflation would have accelerated. He then posited that the true measure of labor market tightness is the ratio of job openings to unemployed workers. And in 2022, there were far more job openings than there had been in 2019, likely as a result of pandemic-induced “matching difficulties” between firms and workers. Thus, getting the jobless-workers–to–vacancies ratio back down to its “natural” 2019 level would require raising unemployment well past the pre-pandemic rate.

But there are at least two problems with this reasoning. One is that there is no basis for believing that 3.6 percent was the lowest unemployment rate that the pre-COVID economy could have sustained without inflation. Throughout the 2010s, economists routinely announced that we had reached the natural unemployment rate, only to see joblessness fall further without setting off inflation.

The second problem is that, historically, job vacancies have not actually correlated with rising wages or prices. So the assumption that the increase in the ratio of job openings to unemployed workers since the pandemic is hugely significant, indicating that the economy can no longer support unemployment below 5 percent owing to inefficiencies in “matching,” is dubious.

But it’s also possible that the entire idea of a “natural rate of unemployment” is misguided. That concept emerged out of the inflation crisis of the 1970s. And its model of the economy — in which low unemployment inevitably begets wage increases, which inevitably beget price increases, which inevitably beget further wage increases — might be an artifact of a very different political economy. Fifty years ago, more than a quarter of U.S. private sector workers were unionized. And many labored under collective-bargaining agreements that included cost-of-living adjustments that automatically increased their wages in response to rising prices. Today, only 6 percent of U.S. workers are unionized, and cost-of-living adjustments are far less prevalent in union contracts.

Under these conditions, workers may struggle to secure wage increases in response to inflation, even when unemployment is low. And indeed, between 2020 and 2022, the vast majority of U.S. workers’ wages failed to keep pace with inflation. Rather than high wages pushing prices upward, wages have been chasing, but never fully catching up with, price increases.

In an analysis released last month, economists at the San Francisco Fed found that wage growth has contributed only minimally to post-COVID inflation in the service sector, noting that businesses can absorb modest wage increases through lower profits, automation, or the discovery of new efficiencies.

Larry Summers was right to anticipate impending inflation in February 2021. But from the beginning, his analysis was predicated on the idea that excessive stimulus would lead to unsustainably low unemployment and thus wage-driven inflation. There has never much reason to believe that the labor market was the primary driver of post-COVID price growth. And at this point, it’s abundantly clear that, in 2023 America, a tight labor market will not inevitably trigger a wage-price spiral. We do not need to put millions of people out of work in order to contain inflation. Larry Summers was wrong to say otherwise.

Larry Summers Was Wrong About Inflation