What a Bull Market Cycle Tends to Look Like

A bull market is usually defined as a rise of 20% or more from a market low, and no authority rules on it. Measured on S&P 500 closing prices from 1957 to October 2026, ten completed bull markets lasted from about 21 to 148 months, with a median near 55 months and a median gain near 108%. Valuations rose in every one of them, and the end of each was clear only after the next large decline had begun.

Key takeaways

The 20% rise from a low is a market convention, not an official definition; Reuters describes it as the most commonly accepted one and says even that is open to interpretation.
The ten completed S&P 500 bull markets since 1957 lasted 21.4 to 147.6 months (median about 55) and gained 48.0% to 582.1% in price (median about 108%); two of them make up 43% of all the months.
Every completed bull contained a pullback from a high set after the low, from −9.6% to −19.9% (median −14.2%), so a bull is not a straight line.
The Shiller CAPE valuation rose between the start and the end of all ten completed bulls (median 14.1 to 22.9), but it ended anywhere from 9.7 to 43.2, so no single level marked a top.
Six of the eleven bulls began during an NBER recession, with the market low 1 to 6 months before the NBER trough; a bull's peak, by contrast, was confirmed only by a 20% closing decline, a median 7.8 months later.

What counts as a bull market

A bull market is a sustained rise in prices. Reuters reported in 2023 that there is no set definition of a bull or bear market, or any regulatory body that declares one. It called a 20% rise off a low the most commonly accepted definition, though even that is open to interpretation.

Every figure here is the article's own calculation from daily S&P 500 closing prices in Yahoo Finance's ^GSPC history, from March 4, 1957 to the close of October 14, 2026. A bear market is a fall of 20% or more from a record close (counting from March 1957), which happened 11 times in that span. A bull market runs from the low close of one bear market to the last record close before the next bear market begins. Every bull below rose far more than 20%; the smallest gain was 48.0%.

Yahoo is a commercial, secondary source. The St. Louis Fed's FRED SP500 series holds only about ten years of daily history, and S&P Dow Jones Indices restricts its reproduction. Over the 2,512 days both share, closes differed by at most 0.12%, and both give the same gain for the two latest bulls. Earlier bulls rest on Yahoo alone. Prices exclude dividends; months are days divided by 30.44.

Ten completed bulls and one running

The first bull starts at the October 1957 low because the index began in March 1957, so the rise before that low is not measured. The last row is still running: as of the October 14, 2026 close the S&P 500 stood 1.3% below its latest record close, set on October 6, 2026, and no fall since the October 2022 low has reached 20% on a closing basis.

Start (low close)	End (peak close)	Months	Price gain	Largest pullback inside
Oct 22, 1957	Dec 12, 1961	49.7	+86.4%	−14.0%
Jun 26, 1962	Feb 9, 1966	43.5	+79.8%	−10.5%
Oct 7, 1966	Nov 29, 1968	25.8	+48.0%	−10.1%
May 26, 1970	Jan 11, 1973	31.6	+73.5%	−13.9%
Oct 3, 1974	Nov 28, 1980	73.9	+125.6%	−19.4%
Aug 12, 1982	Aug 25, 1987	60.4	+228.8%	−14.4%
Dec 4, 1987	Mar 24, 2000	147.6	+582.1%	−19.9%
Oct 9, 2002	Oct 9, 2007	60.0	+101.5%	−14.7%
Mar 9, 2009	Feb 19, 2020	131.4	+400.5%	−19.8%
Mar 23, 2020	Jan 3, 2022	21.4	+114.4%	−9.6%
Oct 12, 2022	Running (record close Oct 6, 2026)	47.8	+118.6%	−18.9%

The last column is the deepest closing fall from any high set after the low, before the bull ended. It cannot reach 20% by construction, since a 20% closing fall would end the bull. For 1987–2000 it was the July to October 1990 slide (368.95 to 295.46, −19.9%).

Measure (10 completed bulls)	Lowest	Median	Mean	Highest
Length, months	21.4	about 55	64.5	147.6
Price gain	+48.0%	about +108%	+184.1%	+582.1%
Gain in first 12 months after the low	+21.4%	+35.9%	+43.5%	+74.8%
Largest pullback inside	−9.6%	−14.2%	−14.6%	−19.9%

Averages hide how uneven the sample is. The two longest bulls (1987–2000 at 147.6 months and 2009–2020 at 131.4) add up to 279 of the 645 months, or 43%, and they produced the two largest gains. Six of the ten at least doubled the index; the other four gained between 48.0% and 86.4%.

Early phases were usually brisk. The first year after the low delivered +21.4% to +74.8%, the three largest in 1982–83 (+58.3%), 2009–10 (+68.6%) and 2020–21 (+74.8%). A strong first year did not set the total: the 1987 low produced only +21.4% in its first year and still ran 147.6 months. Nine of the ten also had a closing pullback of 10% or more; the exception, 2020–22, peaked after 21.4 months.

Valuations across each bull

The cyclically adjusted price-earnings ratio (CAPE) divides an index's price by the average of its last ten years of inflation-adjusted earnings. Robert Shiller's published data set carries monthly values back to 1871. The table reads the CAPE in the month of each bull's low and peak; Shiller's prices are monthly averages of daily closes, so these approximate the exact dates. The last row uses the file's October 2026 value, based on the October 2 close and an estimated consumer price index.

Bull market	CAPE at start	CAPE at end	CAPE multiple	Avg. real earnings multiple	Real price multiple
1957–1961	14.1	22.0	1.56	1.05	1.64
1962–1966	16.8	23.7	1.41	1.12	1.57
1966–1968	18.8	22.2	1.18	1.08	1.27
1970–1973	14.0	18.7	1.34	1.05	1.41
1974–1980	8.7	9.7	1.10	1.06	1.17
1982–1987	6.6	18.3	2.76	0.93	2.56
1987–2000	13.4	43.2	3.23	1.25	4.03
2002–2007	22.0	27.3	1.24	1.26	1.56
2009–2020	13.3	30.7	2.31	1.54	3.56
2020–2022	24.8	36.9	1.49	1.06	1.58
2022–running	27.1	40.7 (Oct 2026)	1.50	1.22	1.84

CAPE rose in all ten completed bulls, from a median of 14.1 to a median of 22.9, a median multiple of 1.45 (range 1.10 to 3.23). The level at the end, though, ran from 9.7 (1980) to 43.2 (March 2000), so the same reading served as an end in one bull and a start in another: about 22 was the end of the 1957–1961 bull and the start of the 2002–2007 one. The October 2026 reading of 40.7 is above the end value of every completed bull except 2000. The table records levels; it does not say when a run stops.

The columns also separate two sources of price gain, because a price is earnings times the multiple paid for them:

real price multiple = CAPE multiple x average real earnings multiple

2009-2020: 3.56 = 2.31 x 1.54

real price multiple   = real index at the peak month / real index at the low month
CAPE multiple         = CAPE at the peak month / CAPE at the low month
avg. real earnings    = the ten-year average of real earnings in the CAPE denominator

Across the ten completed bulls, the change in CAPE accounted for a median of about 80% of the logarithmic real price gain, ranging from 49% (2002–2007) to 108% (1982–87, when the ten-year average of real earnings actually fell 7%). The ten-year average moves slowly and lags the earnings companies report in a given year, so the split shows valuation against smoothed earnings, not current profits.

Campbell and Shiller's NBER working paper on valuation ratios, which used annual US data from 1871 to 2000, found that price-earnings and dividend-price ratios did poorly at forecasting later dividend, earnings or productivity growth and were mainly informative about future price changes. That is a long-horizon statistical finding; it does not say when the current bull ends.

Bull markets and the NBER calendar

The National Bureau of Economic Research (NBER) dates US business-cycle peaks and troughs, and it describes a recession as a significant decline in economic activity spread across the economy and lasting more than a few months. Its table of expansions and contractions lists 12 recessions from 1948 to 2020. The table sets each bull's low against those dates and counts months from each bull's peak to the next NBER peak.

Bull market	State at the market low	Months from stock peak to next NBER peak
1957–1961	In recession; NBER trough Apr 1958 (6 months later)	96
1962–1966	Expansion	46
1966–1968	Expansion	13
1970–1973	In recession; NBER trough Nov 1970 (6 months later)	10
1974–1980	In recession; NBER trough Mar 1975 (5 months later)	8
1982–1987	In recession; NBER trough Nov 1982 (3 months later)	35
1987–2000	Expansion	12
2002–2007	Expansion (NBER trough Nov 2001, 11 months earlier)	2
2009–2020	In recession; NBER trough Jun 2009 (3 months later)	0 (same month as the Feb 2020 NBER peak)
2020–2022	In recession; NBER trough Apr 2020 (1 month later)	None; no NBER peak listed after Feb 2020
2022–running	Expansion	Running

Six of the eleven bulls (1957, 1970, 1974, 1982, 2009 and 2020) began while an NBER recession was under way, and each low came 1 to 6 months before the NBER trough (median 4 months). The 2002 low, by contrast, came 11 months after the NBER trough, and four bulls began well inside an expansion.

At the other end, six of the ten completed bulls peaked 0 to 13 months before an NBER recession began (median 9 months): 1968, 1973, 1980, 2000, 2007 and 2020. After the other four (1961, 1966, 1987 and January 2022), no recession began for at least 35 months, or, after 2022, at all in the NBER table. A bull peak often preceded a recession, but did not need one. Dates here are calendar months, so a lead of 2 months can mean 1 to 3 months of actual distance.

Why a bull's length stays unknown

The 20% test works backward. A peak becomes a peak only after the index has fallen far enough to count as a bear market. For the ten completed bulls, the first close 20% below the peak came a median 7.8 months after it, ranging from 0.7 to 14.8 months. During that interval the bull is still running by the convention.

Completed bulls give a wide range, not a typical length: 21.4 to 147.6 months, a sevenfold spread across only ten observations. One loose pattern appears: a lower CAPE at the start tended to go with a longer bull (a rank correlation of about −0.7), and the two longest began near 13. It is not clean. The 1982 bull began at the lowest CAPE, 6.6, yet ran 60.4 months, far shorter than the 1987 and 2009 bulls that began near 13, and ten runs are too few to rely on. Recession state and first-year gain show no clear link: bulls that began in a recession ran a median 55.1 months, those that began in an expansion 51.8. None of these figures forecasts how long the current bull will last or how it will end.

Frequently asked questions

Is the current bull market the longest on record?

Not in this data. From the October 12, 2022 low to the October 14, 2026 close is 48.1 months, longer than four of the ten completed bulls and shorter than the other six, the longest being 147.6. It is a length so far, not an estimate of what remains.

How much of a bull's gain is already gone when a 20% rise confirms it?

Very little, in hindsight. For the ten completed bulls, the first close 20% above the low came a median 2.3 months after it (range 0.5 to 9.1 months). Measured from that 20% line, a median of about 81% of the eventual gain in index points still lay ahead (range 58% to 97%). Nobody at the 20% mark could know that, since the run could also have faded.

Do the gains in the table include dividends and inflation?

No. They are S&P 500 price changes, before dividends, taxes, fund fees and inflation. For 2009–2020 the index rose 400.5% (a multiple of 5.00), while the Shiller data's inflation-adjusted price, using monthly averages, rose by a multiple of 3.56, about +256%. The site's investment growth and purchasing power calculators can show the same effect on a personal amount.

Why do published bull-market dates and counts differ?

Because the rule is a convention, as Reuters noted, and choices change the answer. The closing-basis 1990 decline was 19.9% and does not split the 1987–2000 bull. In 2025 the closing-basis fall from February 19 to April 8 was 18.9%, but Yahoo's daily high and low fields give −21.3% from the February 19 intraday high (6,147.43) to the April 7 intraday low (4,835.04), which an intraday rule would count as a bear market.

Why can't NBER recession dates be used to spot a bull's end as it happens?

NBER's committee works retrospectively. Its dating page says it tends to wait to identify a peak until a number of months after it has actually occurred and likewise after a trough, with no fixed number of months. The date is a record of the past, not a signal.

A checklist for reading bull claims

Check the definition: which index, closing or intraday prices, and whether the start is the low or the date the 20% rise was confirmed.
Check what the gain includes: price only, or dividends reinvested; nominal or after inflation.
Ask where the start valuation sits: a CAPE of 6.6 and of 27.1 were both starting points in the table.
Treat the length of the current run as a count so far, and compare it with the 21 to 148 month range rather than with an average.
Pullbacks of about 10% to 20% were normal inside past bulls (median largest −14.2%); the 20% ceiling comes from the definition, since a deeper closing fall ends the bull.