When it comes to forecasting a recession, economists today have a wealth of tools and data. Even so, it’s still more of an art than a science.
Most economists predict a downturn in the US this year, precipitated by the Federal Reserve’s barrage of interest-rate hikes aimed at combating inflation. Among those surveyed by Bloomberg, the consensus is that the effects of tighter credit on corporate investment and hiring, as well as consumer spending, won’t translate into a contraction in gross domestic product until the second quarter. Yet many anticipate the damage—at least from a jobs perspective—will be slight compared with earlier episodes.
Many countries define a recession as two consecutive quarters of negative growth for GDP, but the US delegates this assessment to a group of elite academics who meet in secret and typically take about a year to make a call. The verdict comes almost always well after Wall Street has widely recognized a recession.
The trick is catching one before it happens.
Economists, who have had some noteworthy misses on inflation and GDP in recent years, are the first to admit that forecasting the timing of a downturn is practically impossible. That’s despite a wealth of government and private data, including a slew of new high-frequency measures ushered in by the pandemic. The exercise is more akin to putting together the pieces of a puzzle, with each economic indicator filling in part of an image.
Right now, the puzzle pieces aren’t fitting together very well. The manufacturing sector is arguably already in a recession, and the housing market has slumped, yet factory and construction employment remain elevated. Fourth-quarter GDP numbers due on Thursday, Jan. 26, may only complicate the picture further, showing that consumers—the main engine of the economy—remained largely resilient but companies cut back.
“The signals are mixed in a way that we haven’t seen before,” says Claudia Sahm, an economist and the founder of Sahm Consulting in Arlington, Virginia. “People say, ‘Historically when this happens, that happens, and then we go into a recession.’ That’s a good starting place, but that shouldn’t be the end place for the analysis.” The former Fed economist came up with her own real-time recession test. Called the Sahm Rule, it holds that when the three-month moving average of the unemployment rate rises by 0.5 percentage point or more relative to the low in the previous 12 months, a downturn has begun. (The current reading doesn’t indicate a recession.)
Even though layoffs are making headlines, the overall strength of the US labor market continues to confound economists. Yet elsewhere in the economy, cracks are starting to show. Retail sales in December fell the most in a year, and several gauges of manufacturing activity show it contracting in the final quarter of 2022.
“The data is already turning negative broadly and for a number of months,” says James Knightley, chief international economist at ING. “It’s suggesting things are only going to get worse. That’s why I’m very fearful about a recession.”
The previous two recessions were the product of black swan events: the subprime mortgage crisis and a pandemic. But the coming downturn may be one of the most anticipated in years. “This is the most traditional of traditional recessions in a couple of decades—almost textbook,” Knightley says. “When the Federal Reserve has been hiking interest rates in the most aggressive cycle for 40-plus years, it does weigh on economic activity.”
Here’s a look at some of the data economists are tracking to figure out if we’re headed for a downturn. They range from the traditional—such as statistics on the labor market and retail sales—to the more offbeat—like demand for plastic surgery and fudged company earnings.
One tried-and-true recession indicator is an inverted yield curve. It occurs when yields on shorter-term US Treasuries rise above those on longer-term ones, which is the opposite of the norm. Inversions have preceded every recession since the 1970s, though there’s been at least one false positive, too.
The curve inverted a few times last year. And following the release of the latest jobs report on Jan. 6, which showed slower-than-expected wage growth, the gap between yields on 3-month bills and those on 10-year notes widened by a full percentage point for the first time in decades.
Economists are poring over labor market data, looking for signs of weakness. They’re not immediately apparent: Unemployment is at a half-century low, job gains each month are moderating but continue to beat expectations, and vacancies are still elevated.
Drill deeper and things look less rosy: December marked the fifth straight monthly decline in temporary help, a segment of the labor market that’s first fired in bad times. Also, hours worked have declined since March, another sign that demand for labor is easing (though it’s in line with the pre-pandemic level).
“We were in an exceptionally strong labor market, possibly the strongest that we’ll ever see in our lifetimes,” says Guy Berger, principal economist at LinkedIn Corp. “We’ve come back to what is more a conventional labor market, what we had in 2018-2019.”
Wall Street banks and tech giants including Amazon.com, Google and Microsoft have all announced big layoffs, but the accumulated job losses are minuscule when measured against the 153.7 million employed workers in the US labor market.
Goldman Sachs Group Inc. economists analyzed state-mandated layoff notices from a handful of the most populous states and found that while announced job cuts have mounted in recent months, the rate is still below the average in 2017-19. Moreover, the uptick is “unlikely to translate to an unusually large increase in net unemployment,” they wrote in a Jan. 13 report.
The pandemic added momentum to a trend in macroeconomic forecasting in which information from government releases is supplemented with smaller but more high-frequency data from private sources. Following the initial wave of lockdowns, analysts parsed numbers on restaurant reservations from OpenTable, movie attendance from Comscore Inc. and mobility data from Google to gauge how quickly Americans’ lives were returning to normal. Now they’re looking for evidence that households are dialing back on discretionary spending—but so far these indexes are not sounding any alarms.
Regional Fed banks, Wall Street firms and other institutions have invested a great deal of time and effort in developing their own GDP forecasting models. The Conference Board’s widely watched Leading Economic Index is signaling a recession is approaching. Bloomberg Economics’s model puts the chance of a recession in 2023 at 100%, though US economist Anna Wong says it’s likely closer to 80% given that consumers are doing comparatively well, despite the inflation shock. “We have a manufacturing recession, a housing recession, a tech recession. … Things are starting to add up,” she says.
The Philadelphia Fed’s state coincident index signals a recession is likely when at least 26 states show negative readings across a range of economic indicators. It’s been accurate for four of six major recessions. The index signaled a recession in October, but the data was later revised in the opposite direction.
“No forecast is perfect,” says Kevin Kliesen, an economist at the St. Louis Fed, who uses the data. “We try to look up a lot of data, and sometimes it’s tough. If the economy does fall into a recession, this is sure to be the most well-anticipated recession.”
US retail sales can be an early signal of waning consumer demand that can translate into an economic slump. The value of all goods purchased fell in December by the most in a year, with declines across many categories.
Stimulus checks, enhanced unemployment benefits and other relief measures deployed by federal and state governments to soften the hit from the pandemic initially boosted household savings. They’ve now fallen back to levels seen in 2005, which is being interpreted as a sign that Americans are dipping into rainy-day money to make up for the loss of purchasing power resulting from inflation.
They’re also buying more on credit. Total consumer debt—which includes student and auto loans, as well as credit card balances—rose the most since 2008 in the third quarter, New York Fed data show. Still, loan delinquencies remain below 2019 levels.
Nearly all the US chief executive officers polled by the Conference Board anticipate a recession this year. We should get a preview of how executives are preparing for that possibility as more companies report fourth-quarter earnings.
Among those that have done so already, Bank of America Corp. CEO Brian Moynihan expects a mild recession and Goldman’s David Solomon said there’s a reasonable chance of one this year. JPMorgan Chase & Co., the biggest bank in the US, is setting aside more than $1 billion in preparation for potential losses.
Investors are especially eager to hear what retailers and consumer-product companies such as PepsiCo, Target and Walmart have to say. Most begin reporting fourth-quarter results in February.
Those who make their living tracking the ups and downs of the economy often will own up to having their own alternative indicators.
Former Fed Chair Alan Greenspan used to keep tabs on men’s underwear sales. People buy this basic item year-round but put off new purchases in a severe downturn. Sales growth for menswear overall slowed in 2022 and is expected to remain constrained over the near term, according to Euromonitor. Beer and lipstick also have been cited as predictors of past recessions.
Professors from Indiana University and the University of Missouri recently published a paper in which they argue recessions are more likely when there’s a higher incidence of companies manipulating financial statements. They use the M-score, which famously caught accounting issues at Enron before the company’s downfall, to show that when companies misreport in public filings it likely leads to an industrywide misreading of the economic environment and overinvestment as competitor companies try to keep up with the fraudulent firm.
“The amount of misinformation in the economy has real implications,” says Messod Beneish, one of the authors. For 2023, their model predicts no recession, but it does show a slowdown in activity.
Michael Skordeles, senior US macro strategist at Truist Advisory Services in Atlanta, is keeping an eye on spending on elective cosmetic procedures such as rhinoplasty and face-lifts. The industry underwent a post-pandemic boom starting in 2021 as consumers had more disposable income, wanted a change after so many Zoom calls and could recover while working from home. In the past month, though, the appointment data Skordeles gathers from local doctors show that scheduling a procedure is getting easier, a sign that consumers are cutting back spending on nonessential items.
“Everyone is looking at the same government data,” Skordeles says. “So we need to start looking under other rocks for answers.”Read more: Stagflation Is the Dreaded Word No One Dared Speak at Davos
©2023 Bloomberg L.P.