The January effect on IWM: small caps trailed an ordinary hold in 15 of 25 Januaries

The January effect, tested on IWM

The January effect is one of the oldest calendar claims in finance. Rozeff and Kinney (1976) found that an equal-weighted index of New York Stock Exchange shares earned far more in January than in other months, over 1904–1974. Keim (1983) traced much of the extra return of small companies to January, and to its first few trading days in particular. Since then the claim has usually been repeated in its small-cap form: small stocks do best at the turn of the year.

This page tests one fixed, tradeable version of it on IWM, the ETF that tracks the Russell 2000 small-cap index, and compares each year with holding IWM itself.

IWM began trading in May 2000, so its first January is 2001. Each trade is compared with an ordinary IWM hold of exactly the same number of sessions, averaged over the history before the research cutoff.

How this differs from the original studies: they measured indexes of individual stocks, often close to close, and often small stocks against large ones; the strongest days were the first few of January. This page measures one ETF, open to open, over the whole month, against its own ordinary exposure. A negative result here does not refute every version of the claim.

The average January lost money; an ordinary month made it

IWM rose in 10 of 25 Januaries. The average January returned −0.18%; an ordinary IWM hold of the same length averaged +0.81%. So the rule trailed ordinary exposure, and because its own average was below zero, it also trailed simply staying out of the market for the month.

Every January that beat holding also made money, and every one that made money beat holding: the 10 good years and the profitable years are the same years.

10 Januaries above holding, 15 below

The years January beat holding were 2001, 2004, 2006, 2007, 2012, 2013, 2018, 2019, 2021 and 2023. The best, 2019, returned +13.00%, +12.16 pp above holding. The worst, 2009, returned −12.24%, −13.04 pp against holding.

The median January trailed holding by 1.43 percentage points, more than the average did. A few large good years, 2006, 2019 and 2023, hold the average up; the typical year was weaker than it.

Larger wins, but more losing years

The negative average is not a story of large losses. IWM's 10 rising Januaries gained more on average, +5.73%, than its 15 falling ones lost, −4.12%. What put the average below zero is that losing Januaries came more often. Against holding, the years above and below were the same size on average, +4.92 pp and −4.92 pp, and there were more of the second.

Payoff = mean positive ÷ absolute mean negative. Sum ratio = sum of positives ÷ absolute sum of negatives; for gross returns that is a profit factor only under an equal starting amount in every trade, not a cash account. For excess it is a descriptive ratio against an average benchmark, not the result of a hedged trade.

Below holding in both halves, and without its worst year

A claim that once held and then faded would show a strong early period. This one does not, within IWM's history: the first 12 Januaries (2001–2012) averaged −1.41 pp and the last 13 (2013–2025) −0.59 pp. The four to 2025 averaged −1.90 pp, and 10 of the 22 overlapping four-year windows averaged above zero.

The check most generous to the rule is to remove its worst year. 2009, in the middle of the financial crisis, trailed holding by 13.04 percentage points. Without it the other 24 Januaries still averaged −0.48 pp.

The original findings come from decades before IWM existed, so whether January once worked for small stocks is a question this ETF's history cannot answer.

SPY and QQQ trailed their own holds in January too, and IWM did not beat SPY. See the three on the same years.

Large caps trailed holding in January too

On the 25 Januaries all three ETFs share, 2001–2025, each against its own matched hold:

All three averages are below holding. SPY's January returned −0.002% on average, close to zero, and trailed its hold by 0.75 percentage points. QQQ came nearest: −0.25 pp on average, with a median year just above holding, +0.06 pp, so its typical January was roughly an ordinary month and a few weak ones set the average.

The small-cap version of the claim compares small stocks with large ones directly. Set IWM's January return against SPY's in the same year and IWM was ahead in 11 of 25 years, and behind by 0.18 pp on average. That is a plain difference of two gross returns, not a hedged position and not a risk-adjusted comparison: the two ETFs have different volatility, so a small-cap premium cannot be read from it either way.

After the research cutoff, and what comes next

The history above stops at 31 December 2025. Trades completed since then are kept apart from it and listed here as they close, with the rule's next dates. They are not a forecast made in advance: the cutoff marks where the research data ends, not a date when this rule was chosen and recorded.

Every IWM January, open to inspection

All 25 trades, including the 10 that beat holding. The holding column changes with the window's length.

What this result cannot tell you

25 annual trades is a short record. The excess is measured against an average of historical holds, so it is not a risk-adjusted return. No significance test is supplied with this evidence. The test differs from the original studies in instrument, window and price convention, so it shows how one tradeable version of the claim fared, not whether every version is wrong.

A below-holding average does not make the opposite a tested strategy: going short IWM in January, or staying out of the market, would need its own evaluation. All figures are gross of costs. Path risk is not measured: open-to-open returns over a month do not show how far a January fell along the way. The cause of any calendar pattern is not identified here; tax-year selling and year-end portfolio changes are explanations offered for the original effect, and nothing here tests them.

A sharper question

The original findings put the effect in the first few trading days of January, and in small stocks against large ones. A fixed window of a few sessions from the year's first open, on IWM, would test that part directly. It is a different rule, and its result would need to be read on its own evidence, not borrowed from this page.