There is a statistic that circulates every few months on trading forums: nearly all of the stock market’s long-run gain has come while the market was closed. Buy at the close, sell at the open, and you supposedly captured the whole move while day traders churned for nothing.
The statistic is real. The lesson most people take from it, that you should simply hold every night and collect, does not survive trading costs. The more useful lesson is subtler, and it favors traders who hold overnight selectively rather than mechanically.
This post covers what the research on overnight vs intraday returns actually found, how it compares with the evidence on day trading, and how to use the overnight effect without fooling yourself.
Overnight vs Intraday Returns: What Four Studies Found
The overnight-return question has been studied through the SPY ETF, S&P 500 futures by the hour, and individual stocks sorted into factor portfolios. Each found a real overnight effect, and each found a catch.
| Study | Sample | Headline finding | The catch |
|---|---|---|---|
| Bartolini, State Street Global Advisors (2020) | SPY, Jan 1993 – Jan 2020 | Close-to-open cumulative return of 717% vs. 12% open-to-close; buy-and-hold returned 627% | After bid-ask costs, the every-night strategy fell to −32% |
| CXO Advisory | SPY, Jan 1993 – Aug 2017 | Overnight beat intraday on every weekday, on average | Differences were small and did not hold in the most recent subperiod |
| Boyarchenko, Larsen & Whelan, NY Fed (2020–2023) | S&P 500 futures, 1998 – 2020 | The 2–3 a.m. ET hour earned more than 60% of a 5.9% annualized return | That hour has averaged close to zero since 2021 |
| Lou, Polk & Skouras, Journal of Financial Economics (2019) | U.S. stocks, 1993 – 2013 | Momentum profits were earned almost entirely overnight | Value, profitability and investment premiums came mostly intraday |
Figures as reported by each source; sample windows differ and are not directly comparable.
What the Research Rules Out: Holding Every Single Night
The State Street study is the most quoted, and it contains its own correction. On paper, owning SPY only from close to open grew 717% over 27 years. But that required a round trip every session, roughly 6,800 of them, and each one paid the bid-ask spread. Once costs were included, the result swung to a 32% loss.
The edge was also thin day to day. The overnight strategy outperformed on only 53% of days, and a paired test of the two return streams produced a p-value of 0.06, just short of the usual threshold for statistical significance.
One fair caveat: much of that sample predates 2001, when U.S. stocks still traded in fractions and spreads were far wider. Today SPY usually trades with a one-cent spread, so the −32% overstates what a modern trader would pay. The principle still holds. A strategy that trades every night needs an edge larger than its cost, every night.
The sharpest version of the edge has faded
The New York Fed research found that from 1998 to 2020, the single hour from 2 to 3 a.m. Eastern, when European markets open, produced more than 60% of S&P 500 futures’ annualized return. The cause was dealer inventory. When the U.S. session closed with heavy selling, dealers absorbed the imbalance and were paid to carry it as prices drifted back up.
In July 2026, the same authors reported on the NY Fed’s Liberty Street Economics blog that this hour has averaged close to zero since 2021. Volatility and overnight volume barely changed. What changed was the closing imbalance: the dispersion of end-of-day signed volume shrank from 6.5% to 2.9% as algorithms spread order flow more evenly through the day. Less leftover inventory meant less to be paid for holding.
“Overnight returns” are several forces, not one
Lou, Polk and Skouras found that momentum profits accrued overnight while value and profitability premiums accrued intraday. An index-level overnight average is the sum of different investors pushing in different directions, and any one of those forces can weaken on its own.
Overnight Swing Trading vs Day Trading
If trading costs sink an every-night strategy, they sink day trading faster. A day trader making five round trips a day pays the spread five times as often, and fishes in the session that historically paid the least.
| Selective overnight swing trading | Active day trading | |
|---|---|---|
| Round trips | A few per week | Several per day |
| Cost drag | Low | High, and it compounds with every trade |
| Historical session return (SPY, 1993–2020, before costs) | 717% close-to-open | 12% open-to-close |
| Edge needed per trade | Moderate, over fewer trades | Must clear costs on every trade |
| Main risk | Gaps you cannot exit during | Overtrading and slippage |
| Time commitment | Planning around the close | Watching screens all session |
The research on individual day traders is sobering. In a study of Taiwanese day traders, Barber, Lee, Liu and Odean found that fewer than 1% could predictably profit after fees. In Brazil, Chague, De-Losso and Giovannetti found that about 97% of people who persisted in day trading futures for more than 300 days lost money.
That does not make overnight trading easy. The trade-off is real: the overnight trader gives up the ability to react during the move. A stop order does not protect you when the market opens well past it. You trade less and pay less, but you accept gap risk you cannot manage in the moment.
How to Put the Overnight Effect to Work
The research does not say to hold every night. It says the overnight session carries real return, that the return is conditional, and that costs and gap risk decide whether you keep any of it. That points to a clear approach.
- Be selective, not constant. The NY Fed found overnight reversals were strong after selloffs into the close and weak after rallies. Hold when the setup gives you a reason, and stay flat when it does not.
- Know what moves the open. Futures trade through the night, and overnight catalysts such as earnings and 8:30 a.m. economic data can reset the open. Our walkthrough of ES futures fair value and the SPY open explains why a futures print and a gap are not the same number.
- Size for the gap, not the stop. Decide how large an adverse gap you can absorb, then size the position so that gap is survivable. Our study of gap-fill rates by gap size shows how much outcomes change when you sort by conditions instead of averages.
- Plan before the close and pre-commit. Write down the thesis, the size and the exit before you hold. SpyTomorrow’s pressure readings are built for this step: one input for planning a trade in advance, not a signal to follow blindly.
- Keep costs low. Use liquid instruments and limit orders. With options, the bid-ask spread can quietly consume a thin edge.
- Audit your edge on recent data. A published pattern can fade, as the 2–3 a.m. drift did. Our look at how much a 16-for-20 scorecard really tells you explains why short records deserve skepticism, including ours.
The Honest Summary
Historically, a large share of U.S. equity gains accrued outside regular hours, and that finding is well documented. What the research does not support is holding every night as a strategy. The gross figure disappears after costs, the sharpest hourly version has faded since 2021, and the average hides several competing forces.
What the research does support is being selective. Compared with day trading, a few well-planned overnight holds a week carry far lower costs and draw on the session where returns have historically concentrated. The price is gap risk, which you manage through sizing and planning rather than stops. The average overnight return was never the trade. The conditions around it are.
This article is for educational and informational purposes only. It does not constitute financial advice, investment recommendations, or a solicitation to buy or sell any securities. Past performance and historical data do not guarantee future results. All trading involves risk, including the potential loss of principal. Always do your own research and consult with a qualified financial advisor before making trading decisions.
