Question 1

Why is it hard to predict a recession?

Accepted Answer

Because the data that would tell you we are already in one arrives late, gets revised, and sits next to data that says the opposite. GDP comes out six weeks after the quarter ends, and the print you get is revised at least twice. The monthly jobs number is surveyed in the week containing the 12th and reported three weeks later, and the prior two months usually get marked up or down by tens of thousands. The leading indicators that ostensibly run ahead of the cycle, yield curve, permits, PMIs, consumer confidence, do sometimes call the turn. They also false-positive regularly. By the time the signals converge clearly enough to stake a career on, the thing is typically already happening. The honest answer is that nobody predicts recessions well in real time, including the institutions whose job it is to predict them.

Question 2

What are leading economic indicators?

Accepted Answer

Data series that tend to move before the broader economy turns. The Conference Board publishes a Leading Economic Index that bundles ten of them, including the yield curve, weekly jobless claims, new orders for capital goods, building permits, consumer confidence, and the stock market. The logic is that some decisions happen earlier in the chain. Builders stop applying for permits before they stop pouring concrete, and they stop pouring concrete before the construction crews get laid off. Purchasing managers cut orders before factories cut headcount. Consumer sentiment dips before the paycheck actually gets smaller. Any one of these can give a false signal. The whole basket usually turns together a few quarters before the economy does, but the lead time varies from three months to a year, and you only know the exact number after the fact.

Question 3

What is the NBER recession dating committee?

Accepted Answer

The Business Cycle Dating Committee at the National Bureau of Economic Research. Eight economists, sitting at a handful of elite universities, who meet irregularly and ratify recession start and end dates by consensus. They do not use the popular two-quarters-of-negative-GDP shorthand. Their working definition is a significant decline in economic activity, spread across the economy, lasting more than a few months, and they look at payrolls, real personal income, industrial production, and retail sales as the core indicators. They publish a short memo with a date and a brief justification. They do not forecast. They do not issue warnings. They ratify history. The committee is the reason you can look up a recession's official start date in a textbook, and the reason nobody tells you the recession has started until well after you have started feeling it.

Question 4

Why do recessions only get declared after they start?

Accepted Answer

Because the committee is deliberately late. Declaring a recession that turns out to be a false alarm is worse for the institution than declaring one six months after the fact. The data the committee uses gets revised routinely, sometimes substantially. A soft patch can look like a recession in real time and get revised away on the next release. So the committee waits. It waits until the signal is unambiguous across every series it watches and until the revisions have mostly settled. In practice that wait is typically six to eighteen months from the peak. The 2008 recession started in December 2007 by the committee's eventual dating; the committee did not announce that start date until December 2008. The 2020 recession was two months long and the committee announced it had started in June 2020, by which point it was already over. This is the feature, not the bug. The date needs to be right more than it needs to be early.

Question 5

What is a soft landing?

Accepted Answer

The outcome where the Fed hikes rates enough to cool inflation without tipping the economy into recession. Growth slows, the labor market loosens a little, inflation drifts back toward 2%, and unemployment rises a tick without spiraling. Historically it is rare. The Fed has raised rates meaningfully eleven times in the postwar era, and the number of times that cycle ended cleanly in the soft-landing column is small enough that economists argue about whether any of them fully count. The usual mechanism for how a soft landing turns hard is the same one that makes recessions hard to call in the first place. The data looks fine until it does not, the lag is long, and by the time the slowdown is visible in the numbers the credit cycle has already turned. A soft landing is what every Fed chair wants in the press conference and what the data only occasionally lets them have.

Indicator	Type	What it tells you
Yield curve (2s10s)	Leading	Inverts months before a recession starts. The bond market voting that the Fed will have to cut into weakness.
Building permits	Leading	Construction slows before broader layoffs, because builders stop pouring foundations the moment financing gets expensive.
Jobless claims (weekly)	Leading	An early bellwether. Pink slips show up in the unemployment filings long before they show up in the monthly jobs report.
ISM Manufacturing PMI	Leading	Survey of purchasing managers. A reading below 50 says factory orders are shrinking right now, not next quarter.
Consumer confidence	Leading	Households cut spending on the stuff they can delay before they lose jobs. The vibes lead the paycheck.
Unemployment rate (U3)	Lagging	Rises after the recession has already begun. Firms exhaust every other lever before they start cutting headcount.
Core CPI	Lagging	Prices keep rising after demand has softened. Wages, rents, and service prices move slowly in both directions.
GDP revisions	Lagging	The official numbers get clarified a year later. Two rounds of BEA revisions can flip a quarter from growth to contraction.

Why recessions are hard to call.

6 recessions since 1980. Every one of them was officially dated in hindsight.

Some data turns early. Most data turns late. The recession lives in the gap.

Recessions, answered.

Why recessions are hard to call.

6 recessions since 1980. Every one of them was officially dated in hindsight.

Some data turns early. Most data turns late. The recession lives in the gap.

Recessions, answered.

Get the briefing in your inbox.