No economist can drive a news cycle quite like Larry Summers. Ever since he correctly predicted the coming of Biden-era inflation during the heady, early days of 2021, the former treasury secretary has been treated as a seer by much of the financial and political world, much to the chagrin of online progressives who’ve long chafed at his moderate politics and largely believe he’s been given too much credit for what they see as a lucky guess. (My personal take on that matter: It’s complicated).
Over the past year, he’s used his media bullhorn to criticize the Federal Reserve for reacting too slowly in the face of rising prices, while warning that returning inflation to normal might require a recession and rising unemployment.
Late last month, Summers got more specific about just how much pain he thought it would take. “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,” Summers said during a speech at the London School of Economics. He added that the numbers were “remarkably discouraging” compared with the predictions of Federal Reserve leaders, who have suggested they can bring down inflation without pushing unemployment much above 4 percent.
Controversy ensued. Summers’ critics accused him of advocating a “draconian” plan to quell inflation that would require throwing more than 10 million people out of work. “It’s reckless to manufacture a recession that would devastate our most vulnerable,” Rep. Alexandria Ocasio-Cortez tweeted in response.
But where was Summers getting his numbers, exactly? And was he really saying we should aim to kill 10 million jobs, or was he just making a forecast about what he thought was likely to happen? It was a little hard to tell, because the original Bloomberg article that reported his comments didn’t explain how he arrived at his estimate. However, the London School of Economics finally posted a video of his speech last week, and I had a chance to chat with Summers over the phone about his “back of the envelope calculation,” as he describes it.
“A calculation like this is highly uncertain and every business cycle is different. So I certainly didn’t mean to suggest that there was any iron law of this being necessary,” he told me (“this” meaning a gigantic increase in unemployment.) “But I thought a calculation like this was probably a better approach for thinking about it than the Fed’s approach.”
In other words, Summers says he’s just trying to provide a reality check on Jerome Powell’s optimism. Whether this exercise has any major real world implications is a little complicated—but before we get to that, let’s talk about the math behind the uproar.
Summers isn’t using a fully fleshed out macroeconomic model, like a central bank would deploy. Instead, he’s doing some basic arithmetic using a few key variables. He argues that the simplicity is a virtue in this case, given how poor a job more complex and opaque models have done forecasting inflation over the past year or so. But to understand it, it still helps to be familiar with a few key technical-ish concepts.
First, economists will generally tell you that when the unemployment rate falls, the inflation rate rises, because workers have the power to bargain for higher wages, which then get passed on to consumers. When the unemployment rate rises, inflation should drop, since workers have less leverage in a weak labor market.
Most traditional economists also believe that the labor market and wider economy have a sort of Goldilocks point. This is theoretical concept is known as the NAIRU—or the nonaccelerating inflation rate of unemployment, if you don’t mind a mouthful. If the jobless rate falls below NAIRU, the labor market gets too hot and inflation accelerates. If it rises above NAIRU, the labor market cools down and inflation decelerates. If it settles at NAIRU, the labor market is just right and inflation stabilizes. The tricky thing about NAIRU is that it changes over time and there is no way to observe it in the moment; economists just kind of have to guess a number.
Second, there’s a long-standing concept called the “sacrifice ratio.” That’s how much extra unemployment is needed to reduce inflation by 1 percentage point.
You might notice that both of these concepts—NAIRU and the sacrifice ratio—involve the labor market. That’s not an accident; for most of the past four decades, economists have basically thought about inflation as an outgrowth of the unemployment rate. Pretty much everybody agrees that the inflation issues during Biden’s term have been a bit more complicated than that, as they’ve been fueled in part by massive, stimulus-fueled consumer spending that collided with crippling supply chain problems in key industries like automobiles. But Summers and others argue that, at this point, prices are also rising thanks to unsustainably low unemployment and fast wage increases. If the Fed wants to quell inflation, in his view, it will likely mean fewer people working.
Or, to put it another way: He thinks the problem is that the jobless rate is way below NAIRU. To fix things, the Fed will need to raise unemployment above that mark.
Now here’s how all this comes together (sorry for the blizzard of numbers):
• In his calculation, Summers assumes NAIRU is 5 percent—in his view, bringing down inflation will require unemployment to rise higher than that, at minimum. Currently, the unemployment rate is 3.6.
• He assumes the sacrifice ratio is 2, which is in line with a lot of historic estimates.
• Finally, he assumes the Fed will have to lower inflation by 2.5 percentage points in order to reach its official target of 2 percent annual price growth. That might sound a little low, given that the main Consumer Price Index is currently up 8.6 percent for the year, but setting aside volatile food and energy prices that have been thrown out of whack temporarily by the war in Ukraine, some measures show underlying inflation rising at closer to a 4 percent annual pace.
Multiplying 2.5 by 2 gives you 5, which is how many extra points of unemployment above NAIRU Summers believe we need to beat inflation. You can spread that over however many years you want, which is how Summers gets his prediction.
This is not a very long bit of arithmetic. But you can have a very, very long argument about basically every single piece of it. Some would argue that Summers is overestimating how much inflation is being driven by low unemployment, as opposed to supply issues or Ukraine. Some might think his estimate of NAIRU is too high, since wage growth already seems to be moderating while unemployment is nowhere near 5 percent.
Jerome Powell and other Fed officials believe that they may be able to cool off wage growth without raising actual unemployment if they can simply lower the very large number of job openings (Summers has argued that, usually, job vacancies only fall when unemployment rises). Johns Hopkins economist Laurence Ball, who did pioneering research on sacrifice ratios, told me he thought Summers was using the concept in a somewhat outdated way given how the economy has changed over time. “It’s not appropriate to take that off the shelf from 1990,” he told me.
What’s worrisome about Summers’ calculations, though, is that while they aren’t self-evidently correct, they aren’t self-evidently crazy either. He certainly could have stacked the deck by picking higher numbers (Jason Furman, the Harvard economist and former Obama adviser, has said the current sacrifice ratio is probably closer to 6 right now). Instead, he’s using some orthodox ideas to tell an at least somewhat plausible story about what the Fed might have to do if it really wants to push inflation all the way down to 2 percent. (In which case, maybe it’s a good idea to let inflation hover a little above 2 percent going forward? That’s just me spitballing).
Powell, for his part, has said that he does not want to “induce a recession.” But the Fed has already begun to hike interest rates aggressively, and he has said he will do more or less whatever it takes to bring inflation back down to normal. He has also implied that he wouldn’t go out of his way to avoid a recession if that’s what was required. Notably, when AOC asked the Fed chair at a recent hearing what he thought of Summers’ forecast, he didn’t exactly dismiss it out of hand.
“I understand how that number can be arrived at or derived, but I think there is so much uncertainty, and in particular, the answer is going to depend to a significant extent on what happens on the supply side,” he said. “If we get these supply side problems worked out, which I think is certainly going to happen in time, then we wouldn’t see anything like that.”
In other words, Powell is hoping things work out better than Summers thinks they will, but he won’t exactly guarantee it.
At the same time, Summers told me that he wouldn’t do much differently from Powell right now when it comes to interest rate policy. “I think they are broadly in the right kind of place,” he said. “I agree with Powell’s major tilt towards inflation reduction and I think it likely to be difficult in terms of the side effects, but I am certainly not advocating increases in unemployment as a strategy.” To some extent it feels like Summers is fudging here—there isn’t a huge difference between explicitly pressing for high unemployment, and pressing for a strategy that, in order to work, you expect will likely require a period of high unemployment.
But there also doesn’t seem to be much daylight between Powell and Summers regarding what the Fed should do at this point. Summers just thinks the outcome is going to be much more unpleasant for the rest of us.
