What is a Recession? NBER Rules, Technical Recessions, and Modern Debates

What is a recession? The textbook shortcut, two consecutive quarters of falling real GDP, is the answer most people learn first. It is also wrong, or at least incomplete. The United States has had recessions with two negative quarters and two quarters without a recession. The official definition, used by the body that actually dates US recessions, is broader, judgment-based, and built around a committee of economists who weigh several indicators rather than apply a single rule.

This gap between the popular definition and the official one matters more than it sounds. It shaped public debate in the United States in 2022, when GDP fell for two quarters, but the National Bureau of Economic Research declined to call a recession. It explains why the unemployment rate, not GDP, often arrives first in real-time recession discussions. And it determines when fiscal stimulus, monetary easing, and automatic stabilizers are politically justified. Understanding the difference between a technical recession and an official one is the foundation for almost everything else written about predicting downturns.

Two Conflicting Definitions of Recession

The popular definition is simple: two consecutive quarters of negative real GDP growth equals a recession. This rule of thumb is sometimes attributed to economist Julius Shiskin, who in a 1974 New York Times article proposed several quantitative tests for identifying recessions, including the two-quarter rule. It stuck because it is easy to apply, easy to communicate, and works most of the time.

The official definition, used in the United States, comes from the NBER’s Business Cycle Dating Committee. The committee defines a recession as “a significant decline in economic activity that is spread across the economy and lasts more than a few months.” Three words in that sentence do most of the work: significant (depth), spread (diffusion), and more than a few months (duration). The committee weighs several monthly indicators and makes a judgment. There is no formula.

The two definitions usually agree. In every US recession since the Second World War except one, the two-quarter rule and the NBER call have eventually converged. The exception is 2001, when the NBER declared a recession that lasted from March to November, but real GDP fell in only one quarter, not two consecutively. The reverse case, two negative quarters without an NBER recession, is rarer but happened in 2022.

Why Two Definitions Exist

GDP is a powerful but imperfect measure. It is released quarterly with a lag, revised repeatedly for years afterward, and captures only part of what economists mean by “economic activity.” Employment, industrial production, consumer spending, and personal income all move with the business cycle and are released monthly, which is much closer to real time. A definition built only on GDP would miss the labor market entirely. A definition built on a committee weighing several monthly indicators picks up downturns that the GDP data alone may obscure.

The trade-off is between simplicity and accuracy. The two-quarter rule is mechanical, transparent, and replicable. The NBER approach is judgment-based, slower, and contested at the margins, but it has historically matched what economists, businesses, and households recognize as recessions.

How the NBER Actually Decides

The NBER’s Business Cycle Dating Committee is a group of eight academic economists, currently chaired by Robert Hall of Stanford. They do not meet on a calendar. They convene when the data warrant it, deliberate privately, and release a single dated statement when they reach consensus. The committee dates the peak (the start of the recession, when the economy stops expanding) and, later, the trough (the end, when the recession bottoms out). The expansion begins again from the trough.

The committee’s reading of “significant, broad-based, and durable” rests on five monthly indicators. None is decisive on its own. All are watched together.

Table 1. The Five Monthly Indicators the NBER Business Cycle Dating Committee Weighs. Source: NBER Business Cycle Dating Procedure.

Indicator	What It Measures	Source Agency
Real personal income less transfers	Household earned income, adjusted for inflation and stripped of government transfers	Bureau of Economic Analysis
Nonfarm payroll employment	Jobs across the economy excluding farms	Bureau of Labor Statistics
Real personal consumption expenditures	Household spending on goods and services, adjusted for inflation	Bureau of Economic Analysis
Real wholesale and retail sales	Volume of goods sold through wholesale and retail channels	Census Bureau
Industrial production	Output of manufacturing, mining, and utilities	Federal Reserve

Real GDP and gross domestic product are also examined, but quarterly data sit awkwardly with the monthly framework. The committee treats GDP and GDI as supporting evidence rather than the primary trigger. Employment, in particular, carries heavy weight. Robert Hall has noted that the committee gives “particular weight to two measures that have been shown to be highly correlated with the business cycle: nonfarm payroll employment and a comprehensive measure of monthly real personal income less transfers.”

Three D’s: Depth, Diffusion, Duration

Economists summarize the NBER’s criteria as the three D’s. Depth is the size of the decline. A small dip in one indicator is not enough; the fall must be substantial. Diffusion is the breadth across sectors. A recession in one industry is not a recession in the economy. Duration is how long the decline lasts. Brief disruptions, such as a hurricane that knocks out one month of data, do not count.

The three D’s are not independent. A very deep decline can qualify even if it is short, as in the two-month US recession of 2020 caused by the pandemic. A modest decline can qualify if it is wide and lasts long enough, as in the 2001 recession. The committee uses judgment to balance them.

The 2022 Case That Forced the Debate Open

In the first two quarters of 2022, US real GDP fell by 1.6 percent and 0.6 percent at annualized rates. By the two-quarter rule, that was a recession. Headlines and political commentary said so. The NBER did not. Employment kept rising at over 400,000 jobs per month. Real personal income held up once stimulus-related distortions were stripped out. Industrial production grew. Consumer spending stayed positive. None of the committee’s five monthly indicators signaled a downturn.

The GDP decline turned out to be heavily driven by inventory swings and trade flows, components that move sharply quarter to quarter and tell little about underlying activity. Subsequent GDP revisions weakened the case further; one of the two declines was eventually revised closer to zero. The episode demonstrated, in real time, why a single-variable rule based on a noisy and revised data series can mislead. It also exposed how politicized recession terminology has become.

International Definitions of Recessions

The US is unusual in having an independent academic committee date its business cycles. Most other countries rely either on the technical two-quarter rule or on their national statistical office’s judgment. The Euro Area Business Cycle Dating Committee, hosted by the Centre for Economic Policy Research, plays a role similar to the NBER’s for the eurozone and uses a comparable multi-indicator approach. The United Kingdom typically uses the two-quarter rule, with the Office for National Statistics providing the GDP figures. Japan’s Cabinet Office uses a coincident indicator index. Australia famously avoided a “technical recession” for nearly thirty years until 2020, partly because its definition was tied to the two-quarter rule and its quarterly growth rarely turned negative twice in a row.

These institutional differences mean cross-country comparisons of recession frequency need care. A country that uses stricter rules will appear to have fewer recessions, even if its underlying business cycle volatility is similar.

Modern Debates on Defining Recessions

The traditional definitions are under pressure from three directions.

The Sahm Rule and Real-Time Detection

Economist Claudia Sahm proposed a simple, real-time recession indicator in 2019: a recession has likely started when the three-month moving average of the unemployment rate rises by 0.5 percentage points or more above its low in the previous twelve months. The Sahm Rule has correctly identified every US recession since 1970 and does so within months of the peak, far ahead of an official NBER call. It is now widely tracked, including by the Federal Reserve Bank of St. Louis, which publishes a real-time version.

The Sahm Rule is not a replacement for the NBER definition. It is a fast-moving warning indicator that uses one variable, the unemployment rate, to flag what is most likely a recession in progress. The 2024 episode, in which the Sahm Rule was triggered but the labor market then stabilized, showed that even well-designed indicators can produce false alarms when the labor market is recovering from unusual shocks. Sahm herself has cautioned against treating the rule as mechanical.

Growth Recessions and Slowdowns

Some downturns are not deep enough to register as recessions but are deep enough to feel like one. The concept of a “growth recession,” introduced by Solomon Fabricant and popularized by the Conference Board, refers to a period when growth slows sharply but remains positive, often producing rising unemployment. Growth recessions do not appear in official chronologies but show up clearly in labor market data. The mid-1990s in the US and several episodes in the eurozone fit this pattern.

K-Shaped Downturns and Distributional Realities

Recessions used to affect most of the population in broadly similar ways. Recent downturns, especially the 2020 pandemic recession, have been highly uneven. Some sectors collapsed while others boomed; some income groups lost jobs while others gained from rising asset prices. This pattern, sometimes called a K-shaped recovery, raises a question the traditional definitions do not address: how do we describe an economy that is in recession for half the population and expanding for the other half?

The aggregate indicators used by the NBER and the two-quarter rule both treat the economy as a single object. When the distributional pattern is sharply uneven, the aggregate may understate how many households are experiencing a downturn. This is a measurement debate, not a definitional one, but it is increasingly central to how recessions are discussed.

Characteristics of Recessions in Data

Recessions share recognizable patterns even when they differ in cause. Real GDP falls, peak to trough, by an average of around 2 to 3 percent in the post-war US record, with deep recessions such as 1973 to 1975, 1981 to 1982, and 2007 to 2009 falling further. Unemployment rises, typically by 2 to 5 percentage points. Industrial production drops. Consumer spending decelerates, with discretionary categories falling first. Business investment, especially in structures and equipment, falls sharply. Inflation typically slows, though the 1970s episodes paired recession with high inflation, producing stagflation.

The duration is usually short. Post-war US recessions have averaged about ten to eleven months from peak to trough. The longest in the official chronology, the Great Recession of December 2007 to June 2009, lasted eighteen months. The shortest, the pandemic recession of February to April 2020, lasted only two. Expansions, by contrast, have averaged about five years and have grown longer over time. The 2009 to 2020 expansion was the longest on record at 128 months.

Table 2. Post-War US Recessions: Duration and Peak-to-Trough Real GDP Decline. Source: NBER Business Cycle Dating Committee and Bureau of Economic Analysis.

Peak	Trough	Duration (months)	Real GDP decline (%)
November 1948	October 1949	11	-1.7
July 1953	May 1954	10	-2.5
August 1957	April 1958	8	-3.0
April 1960	February 1961	10	-1.6
December 1969	November 1970	11	-0.6
November 1973	March 1975	16	-3.2
January 1980	July 1980	6	-2.2
July 1981	November 1982	16	-2.7
July 1990	March 1991	8	-1.4
March 2001	November 2001	8	-0.4
December 2007	June 2009	18	-4.0
February 2020	April 2020	2	-9.1
Post-war average		10.3	-2.7

Okun’s Approximation

Change in unemployment is approximately proportional to the gap between actual GDP growth and trend growth, with the Okun coefficient β typically estimated near 0.4 to 0.5 for the US.

Okun’s Law captures the rough quantitative regularity that links GDP to unemployment during cycles. It is an empirical relationship, not a structural one, and it shifts over time, but it explains why employment and GDP move together closely enough that the two-quarter rule and the NBER definition usually agree.

Policy Responses to Recessions

Once a recession is recognized, policy responses follow predictable patterns. Central banks cut interest rates to stimulate spending and investment, often aggressively. Governments deploy countercyclical fiscal measures, including expanded unemployment benefits, direct transfers, and increased public spending. Automatic stabilizers, such as falling tax receipts and rising welfare payments, kick in without legislation.

The lag between the start of a recession and policy response is one of the recurring problems in macroeconomic management. By the time the NBER declares a recession, it may have been underway for six to twelve months. Central banks try to act on real-time indicators rather than wait for the official call, but they too are working with imperfect data and uncertain signals. This is part of why monetary policy is sometimes said to operate with “long and variable lags,” a phrase Milton Friedman used to describe how policy changes today affect the economy only with delay.

The Importance of the Definition

Recession terminology shapes public expectations, business decisions, election outcomes, and policy debates. A recession declared too late may delay the response. A recession declared too eagerly may produce unnecessary panic. The discipline of weighing several indicators, examining depth and diffusion, and waiting until the evidence is clear is what gives the NBER definition its authority, even at the cost of slowness.

The two-quarter rule will continue to dominate news headlines because it is easy to explain in one sentence. The NBER framework will continue to be used by serious analysts because it captures more of what recessions actually are. Understanding both and the gap between them, is the foundation for following any current debate about whether an economy is, or is about to be, in recession.

Explains

Conclusion

What is a recession has two answers, and the gap between them is the point. The two-quarter rule is a useful shortcut that works most of the time and fails in the cases that matter most. The NBER’s depth-diffusion-duration framework is slower and judgment-based but captures what economists, businesses, and households actually mean when they describe a downturn. The 2022 US episode showed why a single-variable rule based on noisy and revised data can mislead, and why a multi-indicator approach built around employment, income, production, and spending tends to hold up better.

Recession terminology is not a technicality. It shapes when central banks cut rates, when fiscal stimulus is politically defensible, when automatic stabilizers do their work, and how households and firms adjust their plans. Real-time indicators such as the Sahm Rule have made early detection sharper, and distributional questions raised by K-shaped downturns have widened the conversation beyond aggregates. The traditional definitions remain the foundation, but the modern debate around them is where the most useful analysis now happens.

Frequently Asked Questions

Thanks for reading! If you found this helpful, share it with friends and spread the knowledge. Happy learning with MASEconomics