Every few years the same argument breaks out. Growth stalls, the headlines warn of a downturn, and someone points to two straight quarters of shrinking output and declares the country is already in a slump. Someone else insists it is not. Both sound equally certain. So who decides if the US is in a recession, and on what evidence?
In the United States, no government agency holds that power. A committee of eight economists at the National Bureau of Economic Research, a private nonprofit, makes the call. It looks for a significant decline in economic activity that is spread across the economy and lasts more than a few months — and it decides only well after the fact.
That gap between the popular test and the official one is where most of the confusion lives. Two negative quarters is a serviceable rule of thumb. It is not the definition, and in 2022 the difference turned into a national shouting match.
Who actually makes the call
The Business Cycle Dating Committee has existed in its current form since 1978. Its members are academic macroeconomists, appointed by the president of the National Bureau of Economic Research, and the group works quietly — no press conferences, no scheduled votes, just an announcement once it has reached a conclusion. Valerie Ramey has chaired it since 2024, succeeding Robert Hall, who led it for decades.
The NBER itself is not part of the government. It is a research organisation founded in 1913, best known for the working papers economists circulate before formal publication. Its authority over recession dating is convention rather than law: the government’s own statisticians, the financial press and the textbooks all settled on its dates as the official record.
No statute grants it the job. Everyone simply defers to it.
Why a committee of people rather than a formula? Because a recession is a judgement about the shape of the whole economy, and no single equation captures that reliably. A rule mechanical enough to run on its own would either fire on every wobble or miss slumps that arrived through an unusual door. The committee weighs a messy picture, which no formula can do.
Why two negative quarters isn’t the definition
The two-quarters rule has a real origin. In 1974 the economist Julius Shiskin, then head of the government’s statistics bureau, offered it in a newspaper article as a quick, rough signal of a downturn. It was meant as shorthand for the public, not as a technical standard, and it stuck precisely because it is easy to check against a single published number.
The trouble is that gross domestic product is a single, volatile, heavily revised number. It can fall for reasons that have little to do with a broad slump — a swing in inventories, a quirk in the trade balance — while jobs and household incomes keep rising. A downturn that leaves employment growing is not what the word recession is meant to describe.
There is a further wrinkle. GDP has a twin, gross domestic income, which measures the same economy from the earnings side and should in theory match it. The two often diverge, sometimes sharply, and the committee looks at both. When output says one thing and income says another, the tidy two-quarter story falls apart on its own terms.
This is exactly what happened in 2022. Output shrank in the first quarter and again in the second, meeting the popular test. Yet employers added millions of jobs and consumer spending held up. The White House argued the economy was not in recession, opponents accused it of moving the goalposts, and — as NPR reported — the two-quarter rule had never been the real one.
The BEA, the agency that actually produces the GDP figures, says as much in its own glossary of terms: identifying a recession with two negative quarters does not always hold, because the NBER weighs monthly data, especially employment, alongside output.
What counts as a recession
The committee’s definition rests on three tests, often summarised as the three D’s. A downturn has to satisfy each to some degree, though an extreme reading on one can offset a softer reading on another.
- Depth — the decline has to be significant, not a rounding error in a noisy series.
- Diffusion — it has to be spread across the economy, not confined to a single industry or region.
- Duration — it has to last more than a few months, which rules out a brief shock that reverses almost at once.
Those trade-offs are not hand-waving; they decide real cases. The recession that began in February 2020 lasted barely two months, far short of the usual duration, yet the committee dated it anyway because the collapse was so deep and so broad that depth and diffusion overwhelmed the short clock. The rules bend, but only in the direction the evidence points.
The six numbers the committee watches
Rather than lean on GDP alone, the committee tracks a spread of monthly indicators that together describe how much the economy is producing, earning and spending. No single one is decisive; they are read as a group, and the “real” figures among them are adjusted for inflation using the same price indexes that decide how inflation is measured.
| Indicator | What it captures |
|---|---|
| Nonfarm payroll employment | Jobs added or lost across most of the economy |
| Real personal income less transfers | What households earn, stripping out benefit payments |
| Real personal consumption expenditure | What households actually spend |
| Industrial production | Output of the nation’s factories, mines and utilities |
| Manufacturing and trade sales | Goods moving through wholesale and retail, adjusted for prices |
| Household employment survey | Employment measured from households rather than from employers |
Employment usually carries the most weight, because a genuine, broad slump almost always shows up as job losses. Income and spending measure the same pressure from the household side. Industrial production and trade sales catch the goods economy, which can turn before services do. When these move together and downward, the case for a recession becomes hard to argue with.
When the indicators disagree with one another, the committee simply waits.
Why the call comes months late
The lag is the feature people find hardest to accept. The committee has announced turning points anywhere from four months to nearly two years after they actually happened.
The February 2020 peak was called in June 2020, unusually fast; the trough of the 1991 recession was not confirmed until late 1992, some twenty-one months later. The 2008 downturn was declared that December but dated to a start of December 2007 — a full year after it had already begun.
There are two reasons for the delay. The first is the data itself. The monthly figures the committee relies on are estimates when first published and are revised for months afterward, sometimes substantially. Calling a recession on numbers that later get rewritten would produce false alarms and embarrassing reversals, and the committee would plainly rather be late and right than early and wrong.
The second reason is definitional. A recession is a turning point, and you cannot identify a peak with confidence until you have seen enough of what came after it to know activity genuinely declined rather than merely paused. Recognising the top of a hill requires walking some way down the far side of it first.
So the official recession is always, by design, a verdict on the past.
Does the recession label actually change anything?
Legally, almost nothing hangs on the announcement itself. No benefit switches on, no tax changes, no automatic spending kicks in the moment a recession is declared. Federal programmes that expand in downturns, such as extended unemployment benefits, are triggered by their own statutory measures — usually a state’s unemployment rate crossing a set threshold — not by anything the committee says.
Severe recessions also test the financial system directly. The wave of bank failures in 2008 and 2009 is the reason deposit protection, and the question of how FDIC deposit insurance works, stops being fine print and becomes something ordinary savers actually think about.
What the label carries is weight of a different kind. It shapes how voters judge an administration, how firms plan their hiring, and how confident households feel about spending. A downturn that is never formally named still hurts; a recession that is officially dated becomes a fixed historical fact that campaigns and textbooks argue over for decades.
This is also why the timing can feel political even when the process is not. A ruling that lands during an election, or years after the pain has passed, satisfies almost nobody — but that is the price of a method built for accuracy rather than for speed.
How to read the economy while the committee waits
If the official answer always arrives late, the practical move is to watch the same indicators the committee does, rather than the GDP headline that set off the last false alarm. Payrolls, real income and consumer spending, all released monthly, give a fuller and timelier picture than a single quarterly output number ever will.
It also helps to remember what the dating exercise is and is not for. It settles the historical record. It does not tell you, in the moment, whether your own job or your business is about to feel a slowdown — for that, the raw monthly data is far more use than the label.
In practice, that means watching whether payroll growth stalls at the same time as real spending turns down. That pairing, rather than a single quarterly GDP print, is closest to what the committee itself treats as an early warning.
The same logic runs through your own finances. Knowing that recessions are only confirmed in hindsight is an argument for preparing before one is announced rather than after, which is really a question of how big an emergency fund should be and how much of it sits somewhere you can reach without penalty.
By the time the committee finally rules, the hard part is usually over.
