Business Cycle

The business cycle is the recurring movement of overall economic activity between periods of growth and periods of contraction. In the United States the reference dates are maintained by the Business Cycle Dating Committee of the National Bureau of Economic Research, which describes the object it maintains plainly: "The chronology identifies the months of peaks and troughs of economic activity. Expansions are the periods between a trough and a peak; recessions are the periods between a peak and a trough." NBER adds that "Expansion is the normal state of the economy; most recessions are brief."

Who does the dating, which indicators the committee weighs, how long the announcement takes, and why the two-quarter rule of thumb is not the test are all covered on the page for recession. This page is about the chronology itself: how it is shaped, what its own numbers show, and the questions it cannot answer.

The first thing to correct is the number of phases. Popular diagrams show four, running expansion, peak, contraction, trough, and that four-box picture belongs to the market cycle, which describes asset prices. The official business-cycle chronology has two phases and only two. A peak and a trough are not phases at all; each is a single month marking the boundary where one phase ends and the other begins. NBER's own data page states the boundary rule in one line: recessions "start in the month after a peak in the business cycle, and end in the month of the trough." So a recession does not begin in the peak month, and the peak month is the last month of the expansion that preceded it.

The second is that the chronology comes in two versions that do not always line up. The committee makes a separate determination of the calendar quarter of a turning point, using quarterly measures including the expenditure-side and income-side estimates of real gross domestic product, which are not available monthly. Usually the quarter contains the month. Sometimes it does not: NBER says "generally, the peak or trough quarter contains the peak or trough month, but there are exceptions, with 2019Q4 the most recent," and in its table the February 2020 peak is paired with 2019Q4 and flagged in red, under the note "Red indicates that the turning point quarter does not include the turning point month." Anyone comparing a quarterly series with a monthly one across that turning point is comparing two different dates for the same event.

The third is the most practical, and it is a property of the document rather than of the economy. The chronology is a finished list of past turning points. Read today, its most recent peak is February 2020, its most recent trough is April 2020, and the table itself carries the line "Business cycle data last updated: 03/14/2023." A household cannot look up its own present in it. The chronology answers "when did the last one turn?" with authority and answers "has this one turned?" not at all.

The chronology is a table, and its summary rows are averages of its own entries, so the arithmetic can be checked by hand. Taking the cycles from 1945 to 2020, NBER reports an average contraction of 10.3 months, an average expansion of 64.2 months, and an average full cycle measured trough to trough of 74.5 months. Those three figures are consistent by construction, since 10.3 + 64.2 = 74.5. The share of an average postwar cycle spent in recession is therefore 10.3 divided by 74.5, or about 14 percent.

Run the same calculation across the whole table, from 1854 to 2020, and the answer is very different. Contractions averaged 17.0 months and expansions 41.4 months, for a trough-to-trough cycle of 58.4 months, so 17.0 divided by 58.4 is about 29 percent. On NBER's own averages, a smaller share of the postwar period has been spent in recession than of the record as a whole. All of these are historical figures from the chronology as last updated in March 2023, not projections.

Averages also hide the range, which is wide enough to make any single average a poor planning input. Among the postwar cycles in the same table, contractions run from 2 months, for the February 2020 peak to the April 2020 trough, to 18 months, for the December 2007 peak to the June 2009 trough. Expansions run from 12 months, following the July 1980 trough, to 128 months, from the June 2009 trough to the February 2020 peak. Two of NBER's own entries differ by more than a factor of ten.

What the chronology is good for

• It is a single, dated, stable reference for when U.S. economic activity turned, which makes historical comparisons across decades possible.
• Turning points are located to a specific month, not just a quarter, because the committee works primarily from monthly measures.
• Its published summary rows let anyone compute how the balance between expansion and recession has changed over time, rather than taking a commentator's word for it.

What it cannot do

• It says nothing about the present, by design. Its most recent dated turning point is April 2020.
• It is a chronology of economic activity, not of asset prices, and prices can turn at different times than output and employment do.
• It gives one national answer per turning point, so a region, an industry or a household can be having a different experience than the label describes.
• Its averages summarize a very wide range, which is why "the average expansion lasts about five years" is a description of the past rather than a schedule for the future.