What Is a Recession? The Official Definition and Live 2026 Indicator Dashboard

A recession is a significant decline in economic activity that spreads across the economy and lasts more than a few months. The commonly repeated two consecutive quarters of negative GDP growth rule is a rough proxy, not the official US definition. The National Bureau of Economic Research (NBER) officially dates recessions using a broader set of indicators including GDP, employment, personal income, industrial production, and wholesale-retail sales. The key word is significant - the decline must be meaningful, not just a brief dip.

The NBER's Business Cycle Dating Committee officially declares US recessions. The Committee is a group of eight academic economists that typically announces the start and end of recessions 6-18 months after the turning point, once sufficient data has been revised and confirmed. The committee is based in Cambridge, Massachusetts and has dated business cycles since 1978. Their authority is widely accepted by government, media, and financial institutions.

As of 2 October 2026, NBER has not declared a recession. The last NBER-dated recession was the two-month COVID recession of February-April 2020. The trigger-based real-time indicators remain healthy: the Sahm rule reading of -0.07 sits far below the 0.50 trigger, the yield curve is positive at +0.41 (10Y-2Y) after a prolonged inversion, and initial claims remain historically low, with the 4-week average at 200k; no publisher sets a recessionary level for that series, and the eight onsets since 1967 ran from 210k to 438k. ISM manufacturing PMI eased to 54.5 in September but held a ninth straight month above 50. The August jobs report was firmer than the soft summer: nonfarm payrolls rebounded 162,000, well above the roughly 53,000 expected, while the unemployment rate held at 4.1%. The clear exception is sentiment. Conference Board consumer confidence fell 6.7 points to 81.9 in September, with the forward-looking Expectations Index down to 63.6, well under the 80 line the Conference Board associates with recession ahead. The Federal Reserve raised its target range to 3.75-4.00% on 16 September, its first hike since 2023, citing inflation at 3.4% rather than any recession risk. The New York Fed yield-curve model puts the 12-month recession probability near 14%, which is close to the average reading across that model's own published history rather than an unusually low one.

Recessions typically result from one or more of four mechanisms: demand shocks (sharp declines in consumer or business spending), supply shocks (disruptions like the 1973 oil crisis or COVID lockdowns), financial crises (the 2008 subprime collapse), and policy errors (excessive monetary tightening, as in the 1981 Volcker recession). In practice, most recessions involve more than one cause reinforcing each other. The 2008 recession, for example, combined a financial crisis with a demand shock from collapsing housing wealth.

According to NBER data, the average post-WWII US recession has lasted 10.3 months peak to trough. Pre-WWII recessions averaged 18 months or longer due to the absence of automatic stabilisers and deposit insurance. The shortest NBER-dated recession was the COVID recession of February-April 2020, lasting just 2 months. The longest post-WWII recession was the 2007-2009 Great Recession at 18 months. The 2001 recession lasted 8 months.

There is no universally agreed numerical definition, but economists typically describe a depression as a recession that is both significantly deeper and longer. Common thresholds used include more than 10% real GDP contraction or more than three years' duration. The 1929-1933 Great Depression saw US GDP fall approximately 30% and unemployment reach 25%, the only US event clearly meeting all criteria. The 2008 Great Recession, while severe, contracted GDP by 4.3% over 18 months, well short of depression scale.

The research-backed priorities are: first, build a 3-6 month emergency fund somewhere liquid and interest-bearing, which in practice means a high-yield savings account rather than a checking account; second, pay down high-interest credit card and personal-loan debt, where the Federal Reserve's G.19 release put the average rate on card accounts assessed interest at 22.15% in May 2026; third, diversify income sources where possible; fourth, keep long-term investment contributions running. On that last point the record is reasonably clear: total return was positive over the 12 months after 11 of the 13 NBER troughs since 1945, and positive over 3 years after all 13, so pausing contributions to time a re-entry has usually cost money.

The historical record argues against it, with one honest caveat. Measured from each of the 13 NBER recession troughs since 1945, S&P 500 total return was positive over the following 12 months after 11 of them, averaging +20.3% (median +19.5%). The two exceptions were October 1945 (-7.1%) and November 2001 (-18.2%). Over longer windows the record is cleaner: all 13 troughs were followed by a positive cumulative 3-year return, and all 12 with a full five years behind them were positive over 5 years. Market timing attempts generally underperform dollar-cost averaging because markets usually fall in advance of the recession call and recover before it ends. Investors within 5 years of retirement may still want to rebalance toward bonds or cash to manage sequence-of-returns risk.