A Declared-Dead Effect That Refuses to Die
The turn-of-the-month effect — the tendency of equity markets to post above-average returns around the month boundary — is widely considered "arbitraged away." But a new 2026 study and a look at our own data paint a different picture: the effect is alive — it has merely shifted its time window.
Here we summarize the current research and test it against 20 years of S&P 500 data.
What Is the Turn-of-the-Month Effect?
Since the late 1980s, finance research has documented that a large share of monthly equity returns is concentrated in a few days around the month-end — the last trading day of a month and the first days of the next. A widely cited study on this is by John McConnell and Wei Xu (SSRN). If you want to dig into the basics, our article The Turn-of-the-Month Effect Explained covers why exactly these days stand out statistically.
The common explanation: at month boundaries, salaries, savings plans and institutional flows enter the market, and large funds rebalance their portfolios.
The New Research: The Effect Persists
In a 2026 study published in the Journal of International Financial Markets, Institutions and Money, Nuri Volkan Kayaçetin examines the turn-of-the-month effect across roughly 30 countries over 1994–2023 (DOI). His finding: the effect persists in nearly all markets studied — with Japan as the exception. According to the study, the average return at the turn of the month is around ten basis points, versus essentially zero on ordinary days. As a mechanism, Kayaçetin points to infrequent rebalancing and a deferred risk premium ("risk deferral").
A practitioner analysis by the pseudonymous quant author QuantSeeker (February 2025) adds the missing piece (source): the classic, narrow window (last trading day plus the first three days) is no longer statistically significant for U.S. equities — likely arbitraged away. The broader window, by contrast (three days before to three days after the month boundary), still shows a significant premium of roughly 5 to 12 basis points.
The Test With Our Data
This exact distinction can be reproduced on SeasonAlpha. The chart below shows the average cumulative return curve of the S&P 500 around the month boundary (t0 = last trading day), over the past 20 years.
And the result confirms the research strikingly well:
| Window | Avg Return | Hit Rate | t-statistic | Significant? |
|---|---|---|---|---|
| Classic (t0 to t+3) | +0.14% | 58% | 1.18 | no |
| Broad (t−3 to t+3) | +0.50% | 62% | 3.15 | yes (p < 0.01) |
Across 245 month boundaries, the narrow window no longer delivers a statistically reliable edge (t = 1.18). The broad window, with a t-statistic of 3.15, is clearly significant. The month-boundary effect hasn't disappeared — it simply starts earlier, a few days before month-end.
Why Does the Narrow Window Vanish?
This is economically logical: the better known and tighter a pattern, the sooner traders exploit it until the excess return is gone. The broader window is harder to trade "cleanly" — and that is exactly where the effect survives. Kayaçetin's mechanism fits: if part of the premium is paid as a deferred risk premium, it cannot be fully arbitraged away.
What Does This Mean for Investors?
Three sober conclusions:
- An effect that "vanishes" is often just mismeasured — the window decides.
- Even the broad window averages half a percentage point — interesting as context, but no free lunch. Transaction costs and dispersion eat much of it.
- Seasonal patterns belong in a toolbox, not at its head.
On seasonalpha.ai you can check this yourself: open Turn-of-Month, pick a ticker and the window — the significance gauge (t-value, p-value, win rate) instantly shows whether the effect holds up.
Conclusion
The turn-of-the-month effect is a case study: current research (Kayaçetin 2026) and our own S&P 500 data agree that it persists — but in the broader time window, not the classic narrow one. Anyone taking calendar effects seriously must look closely at how they are measured. Try it yourself on seasonalpha.ai.
Sources
- Kayaçetin, N. V. (2026): Infrequent rebalancing, risk deferral, and equity returns at the turn of the month. Journal of International Financial Markets, Institutions and Money. DOI
- QuantSeeker (2025): Turn-of-the-Month Strategies: Do They Still Work? quantseeker.com
- McConnell, J. J. & Xu, W.: Equity Returns at the Turn of the Month. SSRN
- Own calculation: SeasonAlpha, S&P 500 (^GSPC), 2006–2026, 245 month boundaries.
Frequently Asked Questions
Is the turn-of-the-month effect still real?
Yes — both a 2026 study (Kayaçetin) and our own S&P 500 analysis find a statistically significant effect, but in the broader window (three days before to three days after the month boundary), not the classic narrow one.
Why doesn't the classic narrow window work anymore?
It was likely arbitraged away: the better known and tighter a pattern, the faster the excess return disappears. In our data the narrow window, with a t-statistic of 1.18, is no longer significant.
Can I build a strategy on this?
Be careful. The average premium is small, and transaction costs and dispersion erode it quickly. It is more sensible to treat the effect as one of several building blocks, not as a standalone signal.
Does this hold outside the U.S. too?
According to Kayaçetin's study, the effect persists across roughly 30 countries — with Japan as a notable exception. On seasonalpha.ai you can check the DAX, the Dow and many other indices yourself.