What is a recession? The definition and indicators that matter

A recession is a broad, significant decline in economic activity that lasts more than a few months. That is the working definition used by the National Bureau of Economic Research's Business Cycle Dating Committee, the group whose dates are widely used for US recessions.

You may have heard a simpler definition: two consecutive quarters of falling real gross domestic product. It is a useful warning sign, but it is not the official test used by the NBER. The committee studies production, employment, income, and spending, then dates the peak when an expansion ended and the trough when the contraction ended.

That distinction matters. A single weak report can move markets and dominate headlines without proving that the whole economy is in recession.

Why two negative GDP quarters are not enough

Real GDP measures inflation-adjusted economic output. The Bureau of Economic Analysis publishes GDP estimates every quarter, with an advance estimate followed by later revisions. Two straight declines plainly deserve attention. They show that total output shrank across a six-month stretch, at least according to the data available at that time.

The shortcut has three problems.

First, GDP is one measure of a large economy. The NBER also considers real gross domestic income, or GDI, which measures the income generated by production. GDP and GDI should describe the same activity in theory, but measurement differences can make them disagree for a while.

Second, small quarterly declines may be too shallow or narrow to qualify as a recession. In the other direction, a sudden and severe collapse can qualify even if it does not last two full quarters. The February-to-April 2020 contraction is the obvious example.

Third, quarterly GDP cannot identify a turning point by month. The NBER publishes monthly peak and trough dates, so it needs monthly data as well.

The committee's own recession definition and methodology focuses on three ideas: depth, diffusion, and duration. How large was the decline? How widely did it spread? How long did it last? There is no fixed formula that turns those answers into an automatic call.

The recession indicators the NBER watches

The NBER says it generally consults six monthly series when dating business cycles:

  • Real personal income excluding government transfers.
  • Nonfarm payroll employment.
  • Real personal consumption expenditures.
  • Wholesale and retail sales adjusted for price changes.
  • Employment measured by the household survey.
  • Industrial production.

The list is not a mechanical scorecard. The committee does not assign a permanent weight to each series, and it may consider other evidence. Still, these measures offer a sensible way to check whether weakness is spreading beyond one corner of the economy.

Jobs

Payroll employment estimates how many jobs employers added or lost. The household survey is the source for the unemployment rate and provides a second view of employment. Both appear in the Bureau of Labor Statistics' monthly Employment Situation release.

One soft payroll report does not settle the recession question. Revisions matter, as do the unemployment rate, average weekly hours, temporary employment, and the share of industries adding jobs. Several months of broad deterioration carry more information than one surprise.

Our guide to reading a weak jobs report explains why bonds, stocks, gold, and Bitcoin can react differently to the same labor data.

Household income and spending

Real income strips out inflation, while the NBER's preferred income measure also removes government transfers. The goal is to see what households are earning from regular economic activity.

Real consumer spending shows whether households are buying more goods and services after accounting for price changes. Income can weaken before spending does if people use savings or credit to maintain purchases. That gap cannot last forever, which is why the two series make more sense together.

Inflation can muddy the view. A store may report higher dollar sales even when it moved fewer goods. The Inflation Calculator shows the same problem at a household scale: nominal dollars and buying power are not interchangeable.

Production and sales

Industrial production covers manufacturing, mining, and utilities. It does not capture the entire service-heavy US economy, but it can expose weakness in factories and other rate-sensitive sectors.

Inflation-adjusted wholesale and retail sales provide another check on demand. Falling sales in one industry may reflect a specific product cycle. Declines across many industries, paired with weaker production and employment, are harder to dismiss.

Leading signals are warnings, not recession declarations

Some popular recession indicators try to warn before the broad decline is obvious. They should not be confused with the evidence used to date a recession after the economy turns.

An inverted Treasury yield curve occurs when some short-term yields rise above longer-term yields. It has preceded several US recessions, but the lead time varies and false signals are possible. Lending standards, building permits, new unemployment claims, consumer expectations, and new manufacturing orders can also weaken ahead of the broader economy.

The Sahm Rule looks for a sharp rise in the three-month average unemployment rate relative to its low over the previous 12 months. It was designed to recognize a recession quickly after labor conditions deteriorate. It is not the NBER's definition, and a threshold crossing does not force the committee to announce a recession.

No indicator gets to skip context. A yield curve can invert while hiring remains firm. GDP can contract while payrolls and real income rise. The useful question is whether several independent measures are beginning to tell the same story.

Recession versus depression

A depression is an unusually deep and prolonged economic contraction. Unlike recession dating, the United States has no standing committee with a numerical depression threshold. The term is usually reserved for extreme episodes such as the Great Depression, not an ordinary business-cycle downturn.

That makes "recession versus depression" a difference of severity and duration, not two adjacent rungs on an official scale. A recession does not automatically become a depression after a set number of months or a particular GDP decline.

Why recession calls arrive late

Economic data take time to collect and revise. The NBER also waits for enough evidence to distinguish a real turning point from ordinary noise. By the time it announces a peak, the recession may have been underway for months. By the time it dates a trough, the recovery may have started.

This delay is deliberate. The committee is maintaining a historical chronology, not issuing a real-time trading signal. Its dates can help researchers compare cycles consistently, but they are not designed to tell investors what to buy or households when to change a financial plan.

A practical way to read recession headlines

Start with the source and period. Is the claim based on one monthly report, two quarterly GDP estimates, or a pattern across several series? Check whether the number is adjusted for inflation and whether prior months or quarters were revised.

Then separate three different claims:

1. Growth is slowing. 2. One part of the economy is contracting. 3. A broad recession has begun.

Those statements are not synonyms. Slower growth can still be growth, and a housing or manufacturing slump can occur while services and employment expand.

The cleanest read comes from watching output, jobs, real income, real spending, production, and sales together. If only one line is falling, the story is probably narrower than the headline. If most of them turn down and stay down, recession risk is harder to wave away.

Browse the Economy section for more explainers on inflation, Federal Reserve decisions, jobs, and housing.

Educational only. This article explains public economic data and recession dating. It is not personalized financial, tax, or investment advice.