Seasonality Patterns in US Markets
What Seasonality Means for Markets
Seasonality refers to recurring patterns in market returns that correspond to specific calendar periods. Unlike fundamental factors like earnings or economic growth, seasonal patterns suggest that time of year, day of week, or proximity to holidays influences stock returns independently of other factors.
Researchers have documented numerous seasonal patterns in US markets spanning decades of data. Some patterns appear statistically significant in historical analysis, meaning they occur more frequently than random chance would predict. However, statistical significance does not guarantee future persistence, and many calendar effects have weakened or disappeared once widely publicized.
Understanding seasonality helps investors evaluate claims about timing strategies while maintaining realistic expectations about their practical value.
"Sell in May and Go Away": Historical Evidence
The best-known seasonal pattern suggests investors should avoid stocks from May through October, when returns have historically lagged the November-April period. The saying "Sell in May and go away" captures this pattern, which has roots stretching back to British markets in the 19th century.
What the Data Shows
Analysis of S&P 500 returns from 1950 through 2023 reveals a measurable difference between seasonal periods:
- November through April: Average monthly return of approximately 1.2%
- May through October: Average monthly return of approximately 0.4%
Compounded over many years, this difference becomes substantial. An investor holding stocks only during November-April would have significantly outperformed a May-October strategy over most long historical periods.
The pattern appears across multiple developed markets, not just the US. UK, European, and Japanese markets have shown similar seasonal weakness during summer months. This international consistency suggests the pattern is not merely a US-specific anomaly.
Why the Pattern Might Exist
Several explanations have been proposed:
Vacation effect: Institutional trading activity declines during summer months as traders and fund managers take vacations. Lower trading volume may reduce buying pressure and price momentum.
Earnings seasonality: Corporate earnings announcements cluster in January-February (fourth quarter results) and July-August (second quarter results). The strongest positive surprise potential may occur around the January reporting period.
Institutional fund flows: Many institutional investment decisions occur in the fourth quarter as portfolios are repositioned for year-end reporting. This creates buying pressure in November and December.
Important Caveats
Despite historical evidence, the Sell in May pattern has significant limitations:
The pattern is inconsistent year-to-year. Some years see strong summer gains while winter months disappoint. The 2020 pattern inverted, with May-October delivering strong returns during the post-COVID recovery.
Transaction costs and taxes erode potential gains from seasonal switching. Selling in May triggers capital gains taxes and transaction costs that reduce net returns. The pattern's historical advantage may not survive these frictions for taxable investors.
The pattern has weakened in recent decades. Since 2000, the difference between seasonal periods has been smaller than in earlier decades. Widespread awareness of the pattern may have led investors to act earlier, reducing the anomaly.
January Effect and Tax-Loss Harvesting Rebound
The January effect describes the tendency for stocks, particularly small-cap stocks, to outperform during the first weeks of January. This pattern has been among the most studied calendar effects in finance.
Historical Observations
Research documents small-cap stocks outperforming large caps by approximately 3-5% on average during January, concentrated in the first two weeks of the month. The effect was most pronounced from 1925 through the 1980s and appeared across international markets.
Small-cap value stocks and recent losers showed the strongest January outperformance, suggesting a connection to year-end tax-loss harvesting.
Tax-Loss Harvesting Explanation
The dominant explanation links January outperformance to year-end tax-loss selling:
-
December selling: Investors sell losing positions in December to realize capital losses that offset gains elsewhere in their portfolios, reducing tax liability.
-
Price pressure: Concentrated selling of losing stocks depresses their prices below fundamental value during December.
-
January buying: When selling pressure ends in January, prices recover. Investors may also reinvest in similar positions after the 30-day wash-sale period expires.
This explanation predicts that recent losers should outperform most strongly in January, which matches the historical pattern.
Diminished Effect in Recent Years
The January effect has weakened substantially since the 1990s. Several factors may explain this:
- Awareness effect: Once the pattern became widely known, investors began buying December losers earlier, reducing the January opportunity.
- ETFs and institutional trading: Index funds and institutional investors that do not engage in tax-loss harvesting now dominate trading, reducing the proportion of tax-motivated selling.
- Year-round tax management: Sophisticated investors now harvest losses throughout the year rather than waiting until December.
Recent decades show January returns that are statistically indistinguishable from other months for small-cap stocks. The effect may persist in less-liquid stocks and certain international markets, but should not be relied upon as a trading strategy.
Pre-Holiday Effect
Stock markets have historically delivered above-average returns on trading days immediately before market holidays. This pre-holiday effect appears in US data spanning nearly a century.
What the Data Shows
Research by Lakonishok and Smidt found that returns on the trading day before holidays (such as Thanksgiving, Christmas, New Year's, Independence Day, and Labor Day) averaged approximately 0.5% compared to 0.05% for ordinary trading days. This represents a ten-fold difference in daily returns.
The effect was strongest before major holidays like Thanksgiving and Christmas, with more modest patterns before other holidays. Pre-holiday weeks also showed above-average returns, though less dramatically than the day immediately preceding the market closure.
Possible Explanations
Short covering before extended weekends may contribute. Traders reluctant to hold short positions over multi-day market closures may buy to cover, pushing prices higher.
Reduced trading volume as institutional traders take early departures may allow positive sentiment to dominate. With fewer sellers active, even modest buying pressure moves prices higher.
Psychological effects may also play a role. Holiday anticipation improves mood, and research has linked positive mood to increased risk-taking and buying activity.
Practical Limitations
The pre-holiday effect, while statistically measurable, generates only a few trading opportunities per year. Building a strategy around 6-8 pre-holiday days annually is impractical. Transaction costs and bid-ask spreads would likely consume any excess returns.
Additionally, the effect may have diminished with electronic trading and 24-hour global markets that reduce the significance of US holiday closures.
End-of-Month and Turn-of-Month Effects
Markets have shown above-average returns during the last trading days of each month and the first trading days of the following month. This turn-of-month pattern appears in both US and international data.
Historical Evidence
Research documents that returns during the four-day window surrounding month-end (last day of the month plus first three days of the new month) account for a disproportionate share of total monthly returns. Some studies found that this short window captured nearly all of the market's positive return, with the remaining days averaging near-zero returns.
Contributing Factors
Institutional fund flows create buying pressure around month-end. Pension funds, 401(k) plans, and other institutional investors often invest new contributions at the beginning of each month. Payroll timing concentrates these flows around the turn of the month.
Portfolio rebalancing also clusters around month-end as institutional investors adjust positions to maintain target allocations.
Window dressing by fund managers, who buy winning stocks before month-end reporting, may contribute as well.
Durability Concerns
Like other calendar effects, the turn-of-month pattern has been extensively documented and publicized. Evidence suggests the effect has weakened in recent years as markets incorporated this information. Current data shows less dramatic differences between turn-of-month and mid-month periods.
Earnings Season Patterns
While not strictly a calendar effect, earnings announcement seasons follow predictable quarterly patterns that influence market behavior.
Timing and Concentration
US companies report quarterly earnings in concentrated periods following quarter-end:
- January-February: Fourth quarter results (prior year)
- April-May: First quarter results
- July-August: Second quarter results
- October-November: Third quarter results
The first two weeks of each reporting season see the highest concentration of announcements, with major companies often reporting within a narrow window.
Volatility Patterns
Implied volatility typically rises before major earnings announcements and falls afterward as uncertainty resolves. This pattern creates predictable option pricing dynamics, with premium expansion before earnings and contraction after.
Index-level volatility also tends to decline during earnings season as company-specific news reduces correlation among stocks. Individual stock volatility rises, but diversified portfolio volatility may fall.
Drift Patterns
Research documents post-earnings announcement drift, where stocks that beat expectations continue outperforming for weeks after the announcement, and stocks that miss continue underperforming. This drift suggests markets underreact to earnings information initially.
However, acting on earnings drift requires rapid execution and incurs transaction costs that may eliminate profits for retail investors.
Statistical Significance and Practical Limitations
Before acting on any seasonal pattern, investors should understand important limitations.
Data Mining Concerns
Researchers examining enough calendar periods will inevitably find some patterns that appear significant by chance. With 12 months, 52 weeks, and approximately 250 trading days per year, the opportunities for spurious correlations are substantial. Many published patterns fail to replicate out-of-sample.
Transaction Costs and Taxes
Seasonal trading strategies require frequent transactions that generate costs and tax liabilities. A pattern delivering 1% monthly excess returns before costs may deliver negative returns after accounting for trading frictions and short-term capital gains taxes.
Weakening Over Time
Calendar effects tend to weaken once published and widely known. Arbitrageurs and systematic traders exploit documented patterns, pushing prices to reflect the information and reducing future opportunities. Many patterns that worked in the 1970s and 1980s have disappeared in recent decades.
Sample Size Limitations
Even with decades of data, the number of independent observations for annual patterns remains small. January has occurred only 74 times since 1950. This limited sample makes it difficult to distinguish genuine patterns from random variation.
Key Takeaways for Investors
Seasonal patterns in US markets have historical statistical support but significant practical limitations. The Sell in May pattern, January effect, pre-holiday returns, and turn-of-month effects all appear in historical data but have weakened in recent years.
Transaction costs, taxes, and the tendency for patterns to diminish once known reduce their practical value. Investors should not build portfolios primarily around calendar effects.
Understanding seasonality provides context for market commentary and helps evaluate timing claims. When someone suggests selling in May or buying in January, you can assess their claims against the actual historical evidence and its limitations.
For most investors, maintaining a consistent long-term allocation strategy remains more reliable than attempting to exploit seasonal patterns that may or may not persist.