What it means
The rule is the trading version of an old saying: sell in May and go away. Followers invest for the six months starting around Halloween, then move to cash or bonds for the summer half-year.
The claim is not folklore alone; it is one of the most studied patterns in market history, known formally as the Halloween indicator. The academic foundation is a 2002 paper by Sven Bouman and Ben Jacobsen in the American Economic Review, titled The Halloween Indicator, Sell in May and Go Away: Another Puzzle.
Examining stock market data across dozens of countries, much of it spanning a century or more, they found the winter-half outperformance was widespread, economically large, and unexplained by standard risk models. In many markets the November-to-April period delivered most or all of the long-run return, with May-to-October contributing little.
That evidence makes the Halloween strategy the most respectable of the calendar anomalies, yet respectable is not the same as tradeable. The pattern is an average across many decades and many countries; individual years violate it regularly, and some of the strongest summer rallies have arrived exactly when the rule said to be out.
Averages describe history, not next year. The strategy also pays real costs that the raw comparison ignores, because selling every April and buying every November realises trading spreads and, in taxable accounts, capital gains taxes year after year.
Long stretches out of equities mean missed dividends and missed compounding if the summer happens to rise. Once these frictions are priced in, the edge shrinks materially, and no consensus explanation has emerged for why the pattern should persist once widely known.
Explanations offered over the years include summer liquidity thinning as traders and bankers holiday, bonus-season investment flows early in the year, and seasonal shifts in risk appetite. None has been proven.
A pattern without a mechanism can vanish, the fate of several other calendar effects once publication made them famous. For managers, the strategy's real lesson is about evidence discipline.
A pattern can be statistically strong across a century of data and still be a poor basis for next year's decision, because averages hide variance and because publication can kill the anomaly itself. Seasonality belongs in the notebook of things to notice, not in the rulebook of things to trade.
In practice
Real-world examples.
Example
In the Bouman and Jacobsen study, a majority of the countries examined showed winter-half returns dominating summer-half returns, in some cases by margins of several percentage points per half-year.
Example
An investor following the rule in a year when markets rally 15 percent from May to October sits in cash and watches the strategy's long-run average fail in real time.
Example
A taxable investor who sells every April pays capital gains tax annually; over a decade the tax drag erases much of the seasonal edge the backtest promised.
Formula
Calculation
No single formula applies; the rule of thumb is to hold equities November through April and hold cash or bonds May through October. An illustrative friction check shows why the edge shrinks, using hypothetical figures. On a $100,000 portfolio, an average seasonal edge of 3% a year is worth $100,000 x 3% = $3,000 before costs. If trading spreads cost 0.25% on each of two trades a year, that is 2 x $250 = $500, and tax of 20% on a $6,000 realised gain is $1,200, so the net edge is $3,000 - $500 - $1,200 = $1,300, before counting dividends missed over the summer.Case study
Seen in the real world.
Fictional example: Tamsin Vell, the fictional finance head of a family investment company in Wellington, ran the Halloween rule against her own balanced portfolio using twenty-five years of returns. The backtest showed a genuine edge: the November-to-April window had contributed nearly all of the portfolio's equity gains. Before acting, she priced the frictions: two round trips a year at current spreads, tax on realised gains, and the dividends forgone each summer. The edge shrank by more than half, and in five of the twenty-five years the rule would have missed rallies worth more than the entire remaining margin. The committee recorded the finding, kept the portfolio invested year-round, and adopted a weaker version of the insight instead: they stopped making major new commitments during thin August liquidity, a process discipline rather than a seasonal trade.
Watch out
Common mistakes.
- Treating the historical average as a forecast. The Halloween pattern describes many decades on average; any single year can, and often does, break it.
- Ignoring taxes and trading costs. Twice-yearly portfolio turnover creates drag that backtests routinely omit, and it can consume most of the measured edge.
- Assuming a published anomaly persists. Once a pattern is famous, money trading it can arbitrage it away, which is the fate of several other calendar effects.
Questions
People also ask.
Is the Halloween effect real?
The historical pattern is well documented: Bouman and Jacobsen's 2002 American Economic Review paper found winter-half outperformance across most countries studied. Whether it will continue, and whether it survives costs, are separate questions with no settled answer.
Why would stocks do better in winter?
Proposed explanations include summer liquidity droughts, calendar-year investment flows, and seasonal risk appetite, but none has been proven. The lack of an agreed mechanism is itself a reason for caution.
Should I actually trade it?
For most investors, no. The edge is an average, the frictions are certain, and the cost of being out during a strong summer is high. If anything, use the pattern as a reminder that liquidity and behaviour have seasons, not as a timing rule.
From the founder's library
Take it further with the book.
Build your financial confidence beyond this definition. Shihan's full-length guide, Accounting Fundamentals, takes the same plain-English approach and turns it into a complete, practical playbook for non-finance managers, business owners and students - with chapter-end quiz answers and presentation slides included.
25% off with code MMHQ25, applied at checkout. Priced in USD - checkout may show the equivalent in your local currency.
View the book and save 25%