Seasonality and Cycle Studies

Seasonality—the tendency for markets to produce different returns at predictable calendar intervals—shows up in portfolios as buying into "favorable" months on autopilot, selling every May because a 200-year-old British saying told you to, and ignoring the current macro environment because "the pattern says so." The historical data is real: the S&P 500 returned roughly +7% from November through April versus +2% from May through October since 1950. But here's the practitioner truth that the almanac crowd glosses over: an investor who stayed fully invested since 1975 turned $1,000 into over $340,000, while the disciplined "Sell in May" follower ended up with roughly $64,000. The practical lesson isn't that seasonality is useless—it's that calendar patterns are context tools, not trading systems.

Why Calendar Patterns Exist (And Why They Erode)

Before you memorize a seasonal calendar, you need to understand why these patterns form in the first place. They aren't magic—they're behavioral footprints.

Institutional cash flow cycles: Retirement contributions, year-end bonuses, and corporate buyback blackout periods create predictable waves of buying and selling. January sees fresh 401(k) allocations. November sees window-dressing purchases as fund managers scramble to show winners in annual reports.

Tax-motivated trading: December selling (to harvest losses) creates January buying opportunities. This was the primary engine behind the January Effect for decades.

Earnings calendar clustering: Q4 and Q1 earnings seasons tend to carry stronger revision momentum, which concentrates positive surprises in the November-April window.

Attention and vacation cycles: Trading desks thin out in summer. Lower liquidity means less institutional price support—and more vulnerability to shocks (think August 2015's flash crash or August 2024's yen carry trade unwind).

The pattern that holds: every seasonal pattern traces back to human behavior or institutional plumbing. When the underlying behavior changes (tax law reforms, algorithmic trading adoption, globalization of markets), the pattern shifts. The Halloween Indicator worked beautifully from 1950-2000. It's been far less reliable in the algorithmic era—because the machines front-run the humans who created the pattern.

Calendar pattern → Behavioral cause → Publication → Arbitrage → Decay

That causal chain is why serious practitioners treat seasonality as a probability tilt, not a trade signal.

The Halloween Indicator: What the Data Actually Shows

The most famous seasonal pattern divides the year into two halves: November through April (the "favorable" window) and May through October (the "unfavorable" window). Here's the long-term scorecard:

The point is: the favorable period isn't just slightly better—it captured the vast majority of compounded returns over 74 years. A dollar invested only during November-April since 1950 grew to roughly $1.2 million. The same dollar invested only during May-October grew to about $25,000.

But before you rush to implement this, consider what happened recently. May through October 2020 returned +18.6% as post-COVID stimulus flooded markets. May through October 2024 produced solid gains as AI momentum carried tech stocks higher. You'd have missed both rallies sitting in cash.

The test: if your seasonal strategy forces you to miss a +18% rally every five to seven years, does the pattern's average edge survive those misses? For most investors, the answer is no—the opportunity cost of being wrong in any single year overwhelms the marginal edge from being right on average.

When the Pattern Works Best

The Halloween Indicator has its highest hit rate when combined with other signals:

• Weak May-October period + bear market technicals = high-confidence seasonal weakness (think 2001, 2002, 2008)
• Strong November-April period + bullish breadth = high-confidence seasonal strength
• Strong macro tailwinds overriding seasonality = ignore the calendar entirely (2020, 2024)

Why this matters: seasonality is a tiebreaker, not a primary signal. When your technical and fundamental picture is ambiguous, seasonal tendencies can nudge your positioning. When the macro picture is screaming a clear direction, the calendar becomes noise.

The January Effect: Fading But Not Dead

The January Effect—small-cap stocks outperforming large-caps in January—was one of the most robust anomalies in academic finance. The keyword there is "was."

The classic data (1926–2000):

• Small-cap January average: +5.4%
• Large-cap January average: +1.3%
• The spread concentrated in the first five trading days

The post-publication data (2001–2025): The effect has weakened substantially. Tax-loss harvesting now starts earlier (November instead of December), algorithms front-run the January bounce, and ETF flows have smoothed the mechanics that created the pattern. Some years the effect still appears—beaten-down small-caps and tax-loss candidates do tend to bounce in early January—but the edge is inconsistent and smaller than transaction costs for most investors.

The 2025-2026 window illustrated this perfectly. The Santa Claus rally (last five trading days of December plus first two of January) failed for the third consecutive year. The Nasdaq dropped roughly 1% and the S&P 500 fell about 0.6% during the traditional rally window. Small-caps, weighed down by tariff uncertainty and AI-driven market concentration, didn't deliver the expected January bounce.

The practical point: the January Effect is a case study in anomaly decay. Once an edge gets published in academic journals, popularized in financial media, and coded into algorithms, it gets arbitraged away. This pattern repeats across virtually every calendar anomaly—the more famous the pattern, the less reliable it becomes.

The Presidential Cycle: Real Pattern, Small Sample

Stock returns have historically varied by presidential term year. The data (going back to 1896) shows a consistent hierarchy:

The pre-election year (Year 3) stands out dramatically—only 3 negative years since 1950. The logic is intuitive: administrations front-load painful policy changes in years one and two (when the next election is distant), then shift to stimulus and market-friendly policy in year three to build momentum heading into the election.

The 2024 election year confirmed the pattern beautifully—the S&P 500 returned +23.3%, well above the historical election-year average of +8.2%. That was the third-best election year start since 1926.

But here's the practitioner caveat (and it's a big one): you have fewer than 20 complete presidential cycles in the modern data set. That's a sample size that would get laughed out of any statistics classroom. A single outlier year—2008's financial crisis landing in year 4, for instance—can warp the averages significantly.

The takeaway: the presidential cycle is interesting background context, but never size a position based on which year of the term you're in. The pattern is too noisy and the sample too small to carry real predictive weight on its own.

Sector Seasonality: Where Calendar Patterns Have Real Teeth

While broad market seasonality has eroded, sector-level seasonality remains more durable because it's driven by actual business cycles rather than pure investor behavior.

Energy: Seasonal strength runs roughly late January through early May, aligning with the transition from winter heating demand to summer driving season. Energy stocks have historically gained 8-12% during peak demand windows. The pattern weakens when oil supply dynamics (OPEC decisions, shale production) dominate.

Retail and Consumer Discretionary: Strongest from August through December, driven by back-to-school spending followed by holiday shopping. Retail stocks have historically surged 15-20% during this window in strong consumer years. Q4 earnings for retailers often set the tone for the entire following year.

Technology: Tends to show strength from October through January, coinciding with product launch cycles (new iPhones, holiday electronics sales) and enterprise budget refreshes at fiscal year-end.

Financials: Often strongest in Q1 when loan origination picks up and trading desks capitalize on new-year positioning.

The point is: sector seasonality works better than broad market seasonality because it reflects real economic activity, not just investor habits. When heating oil demand rises in winter, energy companies actually earn more money—that's fundamental, not behavioral.

A Practical Sector-Seasonal Framework

You don't trade sector seasonality in isolation (that would repeat the same mistake as blind "Sell in May" followers). Instead, use it as a rotation tilt:

1. Identify the current seasonal sweet spot for each sector
2. Cross-reference with fundamental momentum (is the sector actually seeing improving earnings?)
3. Check relative strength (is money flowing in or out?)
4. Only tilt toward sectors where all three align

This approach turns seasonal data from a blunt instrument into a precision filter.

September: The Market's Worst Month (And What To Do About It)

September deserves special attention because it's the only month with a negative average return for the S&P 500 since 1950:

• Average September return: -0.5%
• Positive Septembers: only 45% of the time
• Some of the worst market crashes have clustered in September-October (1929, 1987, 2001, 2008)

Why September is weak remains debated (mutual fund fiscal year-end selling, post-vacation institutional repositioning, Q3 earnings anxiety), but the pattern is remarkably persistent.

The real play isn't selling everything on August 31st. It's tightening your risk management heading into September:

• Trim your weakest positions before September (the ones you've been meaning to sell anyway)
• Widen your expected volatility range—don't panic at a 3-4% drawdown in September, because that's historically normal
• Treat September weakness as a potential buying opportunity for the strong November-April window ahead

Why this matters: September's historical weakness actually creates one of the best seasonal entries of the year. If you can stay patient through September-October weakness, you position yourself to ride the historically strong November-December-January cluster.

The Anomaly Decay Problem (Why Published Patterns Fade)

Academic research on calendar effects reveals a consistent meta-pattern: anomalies weaken after publication. A 2001 study by Sullivan, Timmermann, and White argued that many calendar effects are the product of data mining—test enough calendar slices and you'll find "significant" patterns by chance alone.

More recent research (2024) shows mixed results:

• Calendar effects persist in emerging markets (where fewer algorithmic traders exploit them)
• Effects have largely disappeared in large-cap U.S. equities (the most scrutinized market on earth)
• Small-cap indices still show residual seasonal patterns (less arbitrage capital targets these stocks)

The causal chain: Academic discovery → Media popularization → Algorithmic exploitation → Pattern erosion → Occasional resurgence when everyone stops watching

This cycle typically runs 10-15 years from publication to significant decay. The January Effect was documented in the early 1980s and was largely dead by the late 1990s. The Halloween Indicator was formally published in a 2002 paper and has been notably less consistent since roughly 2015.

The lesson worth internalizing: treat every seasonal pattern as having a half-life. The older and more famous the pattern, the less you should rely on it. Fresh, less-publicized patterns (sector-level seasonality, options expiration effects) tend to be more durable because fewer traders crowd into them.

Seasonality Application Checklist (Tiered)

Essential (prevents the biggest mistakes)

These four items stop you from blindly trading the calendar:

• Never use seasonality as a standalone signal—always confirm with at least one fundamental or technical factor
• Know the base rate before acting—a "favorable" period that works 65% of the time still fails 35% of the time
• Calculate your opportunity cost—what do you give up sitting in cash during "unfavorable" months? (historically about +1.8% average, but occasionally +18%)
• Check the macro override—is there a dominant theme (pandemic recovery, AI boom, financial crisis) that dwarfs seasonal tendencies?

High-Impact (systematic integration)

For investors who want to incorporate seasonality into a broader process:

• Use September weakness as a buying window—build watchlists in August, execute in late September or October
• Tilt sector exposure seasonally—overweight energy in Q1, retail in Q4, but only when fundamentals confirm
• Adjust position sizing by season—run slightly larger positions during November-April, slightly smaller during May-October (a 5-10% tilt, not an all-or-nothing switch)
• Track presidential cycle context—note which term year you're in as background, not as a trade trigger

Optional (for active traders and research nerds)

If you want to go deeper into seasonal analysis:

• Monitor options expiration week effects (third Friday of each month)—quad-witching months (March, June, September, December) see 30-50% volume spikes and elevated volatility
• Track first-five-days-of-January as a barometer for the full year (historically, as January's first five days go, so goes the year about 70% of the time)
• Compare current-year seasonal performance against the 10-year average to identify when patterns are holding versus breaking down

Next Step (Put This Into Practice)

Pull up a chart of the S&P 500 and overlay the current calendar position against the seasonal pattern.

How to do it:

1. Identify today's date on the seasonal calendar—are you in the favorable window (November-April) or the unfavorable window (May-October)?
2. Check how the current year's performance compares to the seasonal average for this point in the cycle
3. Note the presidential cycle year (2026 is Year 2—midterm year, historically the weakest)
4. Look at September through October as your next potential seasonal buying opportunity

Interpretation:

• Current performance tracking above seasonal average: macro factors are overriding seasonality—don't fight the trend
• Current performance tracking below seasonal average: seasonal headwinds may be in play—tighten stops and reduce marginal positions
• Current performance wildly divergent from seasonal norms: something bigger is driving the market—ignore the calendar entirely

Action: If you're approaching the May-October window, don't sell everything. Instead, review your weakest three to five positions and ask whether you'd buy them today at current prices. Use seasonal awareness as a forcing function for portfolio hygiene—not as an excuse to go to cash.