What Is a Recession? The Two-Quarter Rule and NBER Definition

In everyday language, a recession means that economic activity has fallen across a meaningful part of the economy. For the United States, the National Bureau of Economic Research (NBER) describes a recession as a significant decline in activity that is spread across the economy and lasts more than a few months. Its Business Cycle Dating Committee considers the overall pattern rather than treating one data release as a definitive test. The committee maintains a U.S. chronology; it is part of a private nonprofit research organization, not a government agency. {source:nberBusinessCycleDatingFaq}

A recession is the contraction phase between a business-cycle peak and trough. The peak marks a high point in economic activity before a broad decline; the trough marks a low point before activity begins to rise again. This phase description is different from saying that the economy is simply “bad,” that growth is slow, or that people face financial hardship. A country can have weak growth without a broad contraction, and a recession can affect industries and households unevenly.

The word is also not a measure of how severe every person’s experience feels. A recession can involve falling production, employment, income, sales, or other indicators, but no single household’s circumstances establish the state of the whole economy. Nor does the NBER maintain a separate formal category in its chronology for a “depression”; that word is commonly used for a particularly severe period of economic weakness. {source:nberBusinessCycleDatingFaq}

The label therefore depends on both a concept and a measurement process. Before comparing headlines, ask which country, institution, series, and time period the claim refers to. In the United States, NBER’s retrospective dates serve as the widely used business-cycle chronology. In other countries, statistical agencies or analysts may communicate downturns differently.

The familiar two-quarter rule says that an economy is in recession after real GDP falls for two consecutive quarters. The rule is easy to summarize using quarterly output data, which helps explain why it appears in news coverage. The Bureau of Economic Analysis (BEA) says that this often-cited identification is not an official U.S. designation. The Office for National Statistics (ONS) in the United Kingdom acknowledges the two-quarter convention when communicating a “technical” recession, while also explaining its limitations. These uses do not make the rule a universal definition. {source:beaRecessionGlossary} {source:onsEconomicCycleCommunication2022}

The shorthand focuses on the direction of quarterly real GDP, not on how large the decline is or how widely it is shared. Two small negative readings may tell a different story from a steep, broad decline, yet the rule counts the number of quarters rather than the broader pattern. It also depends on GDP estimates that can be revised. A result close to zero may change sign when source data are updated, so the initial sequence of reported quarters is not always the final historical account.

The rule is not a necessary condition for the NBER’s U.S. chronology either. NBER notes that some U.S. recessions did not include two consecutive quarters of falling real GDP. A significant downturn can be evident across monthly activity measures even when the quarterly GDP pattern does not match the shorthand. Conversely, two consecutive GDP declines alone do not automatically settle whether a broad U.S. recession occurred. NBER also considers the size and spread of the decline and gives weight to other indicators. {source:nberBusinessCycleDatingFaq}

Use “technical recession” with care. It usually signals that a speaker is applying the two-quarter convention in a particular country or statistical context. State that convention and jurisdiction rather than implying that every economy or recession-dating body applies it in the same way. The phrase describes a reporting rule; it does not, by itself, summarize the full business-cycle evidence.

The NBER’s traditional definition centers on three dimensions: depth, diffusion, and duration. Depth asks how large the decline in activity is. Diffusion asks whether weakness is spread across industries and parts of the economy rather than concentrated in one sector. Duration asks how long the contraction lasts. The committee considers each dimension to some degree, but it does not publish a fixed numerical formula that mechanically combines them. In an unusually severe and widespread downturn, strength in those dimensions may matter even if the decline is brief. {source:nberBusinessCycleDatingFaq}

This framework is designed to identify a broad turning point, not to reward a specific GDP threshold. A small decline in two consecutive quarters may not represent a significant economy-wide contraction. A sharp fall that affects many activities can be informative even before several quarters of data accumulate. Those examples clarify the logic of the criteria; they do not establish a real-time rule for declaring a recession.

The committee also examines a range of indicators, including monthly measures of real personal income less transfers, payroll employment, real personal consumption, manufacturing and trade sales adjusted for price changes, household-survey employment, and industrial production. It does not follow a fixed weighting formula for these series. Instead, it evaluates whether the evidence describes a meaningful decline in aggregate activity and where the turning points occurred. A single industry’s contraction or one weak report is not enough to establish that the entire economy is in recession. {source:nberBusinessCycleDatingFaq}

This U.S. chronology is retrospective. NBER waits for enough evidence to judge a peak or trough and for standard data revisions to arrive; it does not promise an immediate call after a particular release. The method is useful for preserving a consistent historical timeline, but it should not be mistaken for a real-time alarm or a forecast.

NBER dates U.S. business-cycle turning points by month. A peak is the month when economic activity reaches a high point before a significant decline. A trough is the month when activity reaches a low point and begins a sustained rise. Under NBER’s convention, the first month of a recession is the month after the peak, and the last month is the trough month. The following expansion begins after the trough. {source:nberBusinessCycleDatingFaq} {source:nberUsBusinessCycleDates}

These dates describe the chronology of a phase; they do not identify the precise day or moment when the economy crossed into recession. For example, if a committee later identifies a monthly peak, that date marks the turning point in the retrospective record. It does not mean the committee knew in real time that the peak had occurred or that every measure began falling on that exact date.

The committee announces its judgments only after examining the available information. There is no fixed announcement delay. By waiting for additional data and revisions, it can make a more reliable historical determination, but the date may be announced well after the peak or trough month itself. The chronology is therefore best used to describe past cycles, not to provide an immediate status signal. {source:nberBusinessCycleDatingFaq}

A monthly chronology and quarterly GDP estimates can also point to different turning-point quarters. GDP is one important measure, but it is quarterly, while several indicators used to locate monthly peaks and troughs are reported more frequently. The month and quarter of a turning point often align, but their relationship is not a rule that readers should infer from a single GDP release. {source:nberBusinessCycleDatingFaq} {source:nberUsBusinessCycleDates}

GDP measures the value of final goods and services produced over a period. Real GDP adjusts for price changes so output volumes can be compared across periods. Because it is a broad production measure, it is central to recession analysis. But GDP is quarterly, estimated from many source data, and subject to revisions; it also does not capture every dimension of economic well-being or every monthly movement in activity.

NBER considers quarterly real GDP alongside real gross domestic income (GDI). GDP measures production from the expenditure side, while GDI measures incomes earned and costs incurred in producing that output. In national accounting the two are conceptually equal, but in practice they differ because they rely on largely independent source data. NBER gives real GDP and real GDI equal weight when assessing quarterly production, rather than assuming one early estimate settles the question. {source:nberBusinessCycleDatingFaq}

Monthly evidence helps fill the timing gap. The committee’s indicators include income, payroll jobs, consumption, sales, and industrial production. These series can move differently because they measure different parts of the economy, are released on different schedules, and may be revised. Employment may continue to change after output has reached a trough, while an industry-specific decline may not reflect the broader economy. The goal is to assess a coherent pattern, not to require every measure to turn at once. {source:nberBusinessCycleDatingFaq}

That is why “GDP fell” and “the NBER dated a recession” are related statements, not interchangeable ones. GDP is an important measure and may show contraction, but NBER’s U.S. chronology asks whether the wider evidence indicates a significant, broad decline and identifies its monthly turning points. The agency glossary likewise notes that the two-quarter rule is not the official U.S. designation and that the NBER considers multiple indicators. {source:beaRecessionGlossary}

Economic data are estimates built from source information that arrives over time. BEA’s advance quarterly GDP estimate is released using some incomplete or provisional source data. Second and third estimates follow as more detailed and comprehensive information becomes available. Later annual or comprehensive updates can incorporate further data or changes in methods and definitions. This process does not mean every early number changes substantially, but it does mean the estimate and its vintage should be identified. {source:beaGdpReleaseAdditionalInfo}

Revisions matter especially near turning points, when reported growth may be close to zero. A small change in an estimate can change whether a quarter is recorded as slightly positive or slightly negative. Revisions can also alter the measured size of a contraction or the historical timing implied by early data. The two-quarter shorthand is particularly sensitive to this sign change because its simple test counts consecutive declines.

NBER’s dating process is designed to use more information than the earliest GDP releases. Its FAQ says the committee waits for sufficient evidence and standard revisions, with no fixed number of months before an announcement. That delay is a feature of retrospective chronology: it reduces dependence on a single preliminary reading. It also means that the committee’s eventual business-cycle date should not be treated as a live indicator available at the moment the downturn begins. {source:nberBusinessCycleDatingFaq}

When summarizing a figure, specify the source, release vintage, reference period, and whether it is real or nominal GDP. If you compare two estimates, make sure they use the same country, frequency, seasonal adjustment, and measure. A headline about negative GDP growth can describe a release accurately while still being incomplete as a claim about the broader business cycle.

A recession describes the direction of change between a peak and a trough, not whether economic activity has recovered to its previous highest level. Once activity reaches a trough and begins to rise, the next phase is an expansion under the NBER chronology. The level of output can remain below its prior peak during part or all of that early expansion. {source:nberBusinessCycleDatingFaq}

Consider a simple hypothetical path. An index of output rises to 100, falls to 94, then begins increasing to 95 and 97. The decline from 100 to 94 is the contraction; the increase from 94 onward is an expansion, even though the index has not yet reached 100. These numbers are only an illustration, not historical or current data. A recovery in the rate of change and a recovery in the level are separate milestones.

This distinction prevents two common misreadings. First, continued weakness after a trough does not necessarily mean that the contraction is still underway if activity is broadly rising. Second, the start of an expansion does not mean households or firms have returned to their former income, employment, or production levels. Different indicators can regain prior peaks at different times, and some may not follow the same path.

The same level-versus-change distinction appears in other macroeconomic measures. An output gap compares actual output with an estimate of sustainable potential output; it does not date recession peaks and troughs or replace the NBER chronology. For that separate measure, see What Is the Output Gap? Potential GDP Explained.

Before accepting a recession claim, identify the jurisdiction and the institution using the term. Ask whether it is a two-quarter technical convention, a statistical agency’s communication, an NBER-dated U.S. contraction, or another local business-cycle measure. Countries differ in their statistical systems, typical growth patterns, and institutions for dating turning points. The ONS discusses the two-quarter convention in the UK context while also emphasizing its limitations and the value of broader indicators. Do not assume that this reporting convention is a single international definition. {source:onsEconomicCycleCommunication2022}

Then check what the claim measures. Is it real GDP or a different activity measure? Are the data quarterly or monthly, preliminary or revised? Does “recession” refer to a decline in output, a broad contraction across activity, or a technical label used for communication? Stating those details makes two claims easier to compare without pretending they answer exactly the same question.

Finally, distinguish a definition from an indicator or an outlook. The Sahm Rule is a labor-market indicator developed to flag a particular pattern; it is not the NBER’s definition of recession. See What Is the Sahm Rule? Formula, Trigger, and Limits. A soft landing describes an outcome in which inflation pressure eases without a severe downturn; it is not another way to date a recession. For that distinction, read What Is a Soft Landing? Meaning, Indicators, and Recession Risk.