The business cycle: expansions, peaks, recessions and troughs

The business cycle is the economy's habit of alternating between periods of growth and periods of decline. Output, jobs, incomes and profits rise together for several years, reach a high point, fall for a few months or quarters, bottom out, and start rising again. Economists call the high point a peak, the low point a trough, the rising stretch an expansion and the falling stretch a contraction or recession.

The word "cycle" is a little misleading. Nothing about the pattern is regular. Since 1945 US expansions have lasted anywhere from 12 months to 128 months, and recessions from 2 to 18. What repeats is the shape, not the timing. This page explains the phases, how long they have actually lasted, what tends to end an expansion, and how the labor market tracks the whole thing. How stock prices relate to the cycle is covered in the section hub, How the economy affects stocks: growth, inflation, jobs and rates, and the sector-by-sector pattern in How sectors behave across the business cycle: cyclicals vs. defensives.

The four phases

The figure shows the textbook picture: output swings around a rising long-run trend. Real economies are lumpier, but the vocabulary is the same.

• Expansion. Output, employment and income grow. Early in an expansion, growth is often fastest, because idle workers and factories can be put back to work cheaply. Later it slows toward the economy's long-run trend.
• Peak. The month in which activity reaches its high before turning down. A peak is identified only in hindsight; at the time it looks like ordinary slow growth.
• Contraction. Activity falls across many industries at once: production, sales, hiring and income all decline. A contraction that is deep, broad and lasts more than a few months is a recession.
• Trough. The low month, after which activity starts rising again. The trough marks the end of the recession, not the end of the damage: output and employment can take years to get back to their previous highs.

Notice that the phases are defined by the direction of activity, not its level. The first month of an expansion can still have high unemployment and depressed output. That is why "the recession is over" often sounds wrong to people living through it.

Who decides where the phases begin and end

In the United States there is no government agency that declares recessions. The dates come from the Business Cycle Dating Committee of the National Bureau of Economic Research (NBER), a private, non-profit research organization. The committee looks at a range of monthly measures of real activity, especially real personal income excluding transfers and nonfarm payroll employment, and picks the month of each peak and trough. It works retrospectively and usually announces a turning point many months after it happens. The method, the indicators and why the "two negative quarters" rule of thumb is not the definition are covered in What is a recession? How the NBER dates US recessions.

How long expansions and recessions last

Using the NBER's monthly peak and trough dates, here are the US cycles since World War II.

The first row's expansion is measured from the October 1945 trough. Three things stand out.

Expansions got longer. The four expansions before 1960 averaged under four years. Three of the last four ran for more than seven: 92 months in 1982–90, 120 months in 1991–2001 and 128 months in 2009–20, the longest on record. Economists debate why; more stable inflation, a larger service sector and better inventory management are common explanations, though none is settled.

Recessions are usually short, but not always mild. The NBER puts the post-war average contraction at 10.3 months. The 2020 recession lasted only two months, yet it produced the largest jump in unemployment in the data.

There is no clock. A long expansion is not "due" to end, and a short one is not safe. The 1980 recovery lasted just 12 months before the 1981–82 recession began.

As of the NBER's published table, the most recent turning point is the April 2020 trough. No peak has been announced since, which would make the current expansion 78 months old through October 2026, if one is not dated later.

What ends an expansion

Expansions rarely die of old age. They usually end because something pushes spending down across the economy at once. The post-war record shows a handful of recurring triggers, often in combination.

• Tight monetary policy. When inflation runs high, the Federal Reserve raises interest rates to cool demand, and sometimes it cools it into a recession. The effective federal funds rate averaged 19.1% in June 1981, one month before the 1981–82 recession began. See The Federal Reserve: structure, mandate and policy tools and How Fed decisions move markets: expectations, surprises and yields.
• Energy shocks. A sharp rise in oil prices acts like a tax on households and businesses. The 1973 Arab oil embargo roughly quadrupled crude prices, according to the Federal Reserve's history of the period, and the US entered a recession in November 1973. See Oil prices and the economy: WTI, Brent, gasoline and inflation and Stagflation: high inflation, weak growth and the 1970s lesson.
• Financial crises. When banks and credit markets seize up, lending stops and spending follows. The 2007–09 recession is the clearest modern example; see The 2008 financial crisis: housing, Lehman, TARP and the 57% fall.
• Asset busts. When a boom in one kind of investment collapses, the spending it supported disappears. The 2001 recession followed the end of the technology investment boom (The dot-com bubble: the 1995–2002 boom and the 78% Nasdaq fall).
• External shocks. The 2020 recession was caused by the COVID-19 pandemic and the shutdowns that followed it (The COVID crash of 2020: the fastest bear market and recovery).

Two internal dynamics amplify whichever shock arrives. Inventories: when sales slow, companies cut orders to work down unsold stock, so production falls faster than demand. Credit: lenders loosen standards in good times and tighten them in bad, so borrowing rises and falls with the cycle. Both mean small shocks can grow into broad downturns.

The cycle in the labor market

Unemployment is the clearest way to see the cycle in a single line. The chart shows the same shape repeating: the rate drifts down through each expansion, then jumps in a recession.

The rise is large every time. Measuring from the lowest rate in the year before each business-cycle peak to the highest rate within two and a half years after it, the smallest increase since 1948 was 2.3 percentage points (4.8% to 7.1% around the 1960–61 recession). Typical increases were much larger:

The last column shows the other regularity: unemployment usually peaks after the recession ends. It peaked four months after the June 2009 trough, 15 months after the March 1991 trough and 19 months after the November 2001 trough. Companies wait to be sure a recovery is real before they hire. That makes unemployment a lagging indicator, and it is why the The Sahm rule: a recession signal from the unemployment rate, which flags a recession when unemployment starts rising quickly from its low, focuses on the early part of the increase rather than the peak.

Payroll employment tells the same story in head counts. Nonfarm payrolls peaked at 138.4 million in December 2007, the month the recession began, and fell to 129.7 million in February 2010, eight months after the recession ended. In 2020 they fell from 152.3 million in February to 130.4 million in April, a loss of 21.9 million jobs in two months. See the The jobs report: how to read nonfarm payrolls and the two surveys and The unemployment rate: U-3, U-6 and labor force participation pages for how these numbers are collected.

How to read where we are in the cycle

No single number tells you the phase, but a few groups of data, read together, give a reasonable picture.

• Leading data turn before the economy: weekly Jobless claims: reading the weekly unemployment insurance report, building permits in the Housing data: starts, permits and new home sales explained, new orders in the PMI explained: how purchasing managers’ indexes work, the slope of the yield curve and stock prices themselves.
• Coincident data move with it: payrolls, real income, industrial production and real sales.
• Lagging data confirm it afterwards: the unemployment rate, inflation, and the duration of unemployment.

The Leading, lagging and coincident indicators: what turns first page explains the three groups and how reliable each has been.

Pitfalls and limitations

• Phases are only visible in hindsight. A peak looks like ordinary slow growth at the time. The NBER took 12 months to announce the December 2007 peak.
• "Late cycle" is not a forecast. The 2009–20 expansion was called late-cycle for years before it ended, and it ended because of a pandemic, not overheating.
• National averages hide local cycles. Manufacturing, housing, energy-producing states and individual industries can be in recession while the whole economy is not.
• The textbook wave is smooth; data are not. Monthly figures are noisy and revised (see Seasonal adjustment and data revisions: why the numbers change), so a single weak month rarely marks a turn.
• Markets run on their own clock. Stocks often turn months before the economy does, so knowing the phase of the cycle is not the same as knowing what prices will do.

Business cycle terms compared

How Finabase does it

Finabase does not track or chart the business cycle in the app: there are no GDP, employment or recession-dating pages. The unemployment chart on this page is our own, drawn from public BLS data with the NBER's published dates.

The app does show how markets are pricing the cycle day to day. The sector page lists the 11 S&P 500 sector indexes with returns from one day to one year, so you can see whether cyclical or defensive sectors are leading. The market overview shows the indexes, futures, breadth and the day's movers together, the Treasury table shows the yield curve as a table of yields, and the economy topic in Discover collects news about the latest releases.

Frequently asked questions

What are the four phases of the business cycle?

Expansion, peak, contraction and trough. Activity grows during an expansion, reaches its high at the peak, falls during the contraction (a recession if it is significant) and bottoms at the trough before the next expansion begins.

How long does a business cycle last?

There is no fixed length. Since 1945, US expansions have averaged about 64 months and recessions about 10, according to the NBER, but individual expansions have ranged from 12 to 128 months and recessions from 2 to 18.

What phase of the business cycle are we in?

As of the NBER's latest published dates, the US has been in an expansion since the April 2020 trough; no later peak has been announced. Because the NBER dates turning points only in hindsight, a recession can be under way for months before it is declared.

Does unemployment rise before or after a recession starts?

It usually starts rising around the start of a recession and keeps rising after it ends. In 2007–09 the rate was 5.0% in the December 2007 peak month and peaked at 10.0% in October 2009, four months after the trough.

Summary

The business cycle is the recurring, irregular swing of the economy between expansions and contractions, with peaks and troughs dated in the US by the NBER. Post-war expansions have averaged a little over five years and recessions under a year, but the range is wide and there is no clock: expansions end when a shock, such as tight policy, an oil spike, a financial crisis, a bust or a pandemic, pushes spending down across the economy at once, and inventories and credit amplify it. Unemployment is the clearest mirror of the cycle: it has risen at least 2.3 points in every post-war recession and usually peaks after the recovery has begun.