- Forecasting recessions is notoriously difficult. But we can look to historical data, as well as the behavior of the Fed, to make our best predictions about U.S. and global markets.
- Economist Peter Berezin believes we’re probably in for a downturn in 2024, but it may arrive when we least expect it.
- What about the rest of 2023? The labor market and the rate of deflation can give us important clues about the health of the economy.
Economists usually find it difficult to forecast recessions. Case in point: Last year, Peter Berezin concurred with the prevailing view that a recession was imminent in 2023. Of course, it hasn’t happened.
Recently, he argued that the next recession will probably come when we least expect it. And yet Peter’s assessment of the U.S. economy right now is upbeat, at least relative to many analysts’ consensus.
“Economists talk about long and variable lags, but sometimes investors forget just how long those lags can be,” says Peter, who serves as Director of Research and Chief Global Strategist for BCA Research. “The Fed only started raising rates last year. They’ve not even finished raising rates.”
“As soon as everybody says a recession is likely, it’s a pretty good bet that it probably isn’t,” says my colleague (and the host of this episode), Alan Dunne.
“That’s kind of the irony,” Peter replies. “Because so many people have been expecting a recession, long-term bond yields are not terribly high. They’re high by the standards of where they were a few years ago. But relative to where they were a decade or two ago, they are not especially elevated.”
When it comes to dynamic global markets in a rapidly changing world, expectations and reality are rarely in sync. On a recent Global Macro episode of Top Traders Unplugged, Alan and Peter discuss the state of the U.S. economy, their predictions for the near future and the best frameworks for predicting deflation.
Monetary policy in the age of the inverted curve
We generally don’t expect higher rates to have an immediate impact on the economy — and so far, they haven’t. The average mortgage rate is still lower than it was in 2019.
Why? Homeowners refinanced their mortgages during the pandemic. Most mortgages are 30-year fixed-rate. Only about 5% to 10% of homes are sold every year. Because of the prevalence of fixed-rate mortgages, household debt (which is dominated by mortgage debt) was low going into this Fed tightening cycle. Now, we have to wait for the impact of higher rates to make its way through the economy.
Meanwhile, the yield curve is incredibly inverted, and those longer-term bond yields matter the most for the economy.
“Ironically, you almost need more confidence that the economy will avert a recession to lift those yields to a level that actually could be restrictive for the economy,” Peter argues.
Alan wonders if that’s the reason why we’ve had those yields. “They haven’t really gone up that much, particularly given we’ve started quantitative tightening as well,” he notes. “And yields overseas — maybe not in Japan, but certainly in Europe — have gone up a decent amount. … Is there something different that’s keeping yields in check at the moment? Is it that kind of pervasive expectation of recession?”
Peter thinks that’s the main factor: “An inverted yield curve isn’t necessarily a recession indicator,” he explains. “All it really says is that the Fed is going to cut rates. But usually, the market thinks the Fed will cut rates because the economy is going to weaken by enough to require rate cuts, and that normally only happens during recessions.”
However, he contends that monetary policy is somewhat restrictive now. Job openings are declining, which tells us labor demand is declining as well.
“But there’s still a long way to go [to] get to a point where recession is likely,” Peter says.
The ‘immaculate disinflation’
Right now in the U.S., there are about 1.7 job openings for every unemployed worker — but the ideal number should be (naturally) closer to one job opening per unemployed worker.
“Until you get that number down, most people who lose their jobs are just going to walk across the street and find new work,” Peter says. “If unemployment doesn’t rise, you can’t have a recession. … A recession will occur, but it’s just going to happen later than most people are anticipating.”
What about the other big driver of markets: inflation?
“I suppose people have been looking for the immaculate disinflation,” Alan jokes.
Some economists believe that without a significant rise in unemployment, inflation won’t go down by itself. But Peter’s view is that inflation will still decrease even with a strong labor market. He has written quite a bit about the so-called “Phillips Curve” — the theory that the relationship between inflation and unemployment is nonlinear; that “it’s sort of kinked when you get down to very low levels of unemployment,” he explains.
“What that means in practice is that once the economy gets overheated, the aggregate supply curve gets really, really steep, which means that initially, as demand comes down … sliding down that steep portion of the supply curve, you get a lot of disinflation. But initially, you don’t get much in the way of higher unemployment. That’s what we’ve seen over the last year or so, this ‘benign disinflation.’ I suspect that can continue.”
Gimme shelter (numbers)
The forward-looking indicators for inflation right now are “fairly tame,” Peter says. He explains that the Fed divides inflation numbers into three categories: core goods, shelter and services excluding shelter.
The prices of core goods aren’t rising, as evidenced by the New York Fed’s supply chain pressure index, which has plummeted from a record high to a record low. “So no issues there,” he says.
Shelter inflation is dropping, too. Peter points out that the best leading indicator for shelter inflation is rent prices. Data from Zillow, CoreLogic, Apartment List show that rents aren’t rising.
On the services side of things, wage growth — the key driver of services inflation — is coming down as well.
“Average hourly earnings on a year-over-year basis have decelerated from around 7% to about 5%,” Peter notes. “If you look at leisure and hospitality, the deceleration in wage growth has been extremely significant.”
That’s a sign services inflation will fall as well. And because shelter inflation is falling and goods inflation is very low, overall inflation “probably still has further to decline,” he adds.
Kinky curves ahead
Alan circles back to Peter’s mention of the “kinked Phillips curve,” noting that the Phillips curve itself is a much-debated topic.
“If you were to go back a few years, there was this view of a very flat Phillips curve,” says Alan. “That was part of the Fed narrative around running the economy hot because [they] could do that.”
A nonlinear or kinked Phillips curve at very low levels of unemployment tends to coincide with inflation picking up quickly. But these are theoretical concepts, and economists’ models differ wildly. Alan asks Peter whether he has confidence in the Phillips curve framework — enough to base his investment strategies on it.
Peter says it’s a matter of intuition… and data.
“On the intuitive side, it’s very easy to understand why the Phillips curve is nonlinear,” he explains. “When unemployment is really high and starts to fall, there are still going to be a lot of workers eager to find jobs, so firms don’t have to raise wages. But when you get down to super low levels of unemployment — like 3% or 4% — at that point, you don’t have a surplus of workers.”
To compete in this kind of environment, companies must poach talent from other firms and offer higher wages. That’s why the idea that unemployment can fall without rising inflation doesn’t make sense.
When the Phillips curve relationship between unemployment and inflation ends,
inflation needs to start to rise in order to ration demand,” Peter adds. “You can’t just expand demand indefinitely. At some point, you run into these supply-side constraints.”
We can look to historical data for evidence of Peter’s arguments. In the first half of the 1960s, inflation was “remarkably stable” at around 2% or 2.5%, even though unemployment fell throughout that time — and the Phillips curve “looked to be very, very flat,” he explains. “But then 1966 comes around. We’ve got the Great Society programs … the Vietnam War starting to heat up; massive labor shortages. Boom! Within the span of six months, inflation doubles.”
We saw this phenomenon again during the COVID-19 pandemic.
“At one point in the first quarter of last year, the so-called ‘jobs-workers gap,’ this difference between labor demand and labor supply, got up to four percentage points. It was the highest on record. Labor demand exceeded supply by 4%. Not surprisingly, we got inflation.”
That makes Alan think about a big theme at Jackson Hole last year: that in the near future, aggregate supply may be more constrained because of deglobalization and/or lower labor force participation rates.
“All of that seems to have gone away,” Alan says. “We’ve had lower inflation. Was that wrong, or do you think those structural impediments are still there looking forward?”
Peter suggests we should consider supply/demand and labor participation separately.
“The pandemic-related supply constraints have largely gone away,” he says. “If you look at labor force participation, it’s still a bit depressed for older workers. But for prime-age workers — 25 to 54-year-olds — it’s actually higher now than it was in 2019. So that problem has largely disappeared. As I mentioned, the issues around the global supply chain have been largely resolved, although there are still component shortages in some areas. But that problem is quickly receding.”
Peter thinks the problem is structural. U.S. productivity growth has been “very weak since about 2005,” he notes. “In fact, [it’s] close to where it was in the 1970s.”
He argues that we’ll continue to face “fairly mediocre prospects for aggregate supply growth” unless productivity growth accelerates in a significant, structural way.
“Maybe AI is the ticket that boosts productivity growth. I certainly hope that that’s the case,” Peter adds. “The work that I’ve done suggests that it’s very likely to be extremely productivity-enhancing, but that could be a story more for the end of this decade [or] the 2030s, than the next, say, 18 months.”
Will the AI revolution save us? I agree with Peter: That’s a story for another time.
This is based on an episode of Top Traders Unplugged, a bi-weekly podcast with the most interesting and experienced investors, economists, traders and thought leaders in the world. Sign up to our Newsletter or Subscribe on your preferred podcast platform so that you don’t miss out on future episodes.