NSPY - ETF for the "Night Effect," because stocks perform better when degens don't poke them much

An ETF to capture the “Night Effect” :elf:

NightShares, a new entrant to the ETF space led by a highly experienced team that has built and grown multiple successful ETF brands, is today launching its first two ETFs: the NightShares 500 ETF (NSPY) and the NightShares 2000 ETF (NIWM).

These ETFs seek to capture value from the “Night Effect,” a persistent phenomenon whereby overnight markets have historically outperformed the daytime trading session on a risk-adjusted basis.

The strategy is based on academic research dating back to 2008 and these ETFs are designed for investors to efficiently capture this potentially better investor experience than a simple buy and hold approach that has heretofore dominated the industry. The NightShares 500 ETF seeks to return the night performance of a portfolio of 500 large cap U.S. companies. The NightShares 2000 ETF seeks to return the night performance of a portfolio of 2000 small cap U.S. companies.

(Source: Press Release)

Some additional information from their three-pager:

The Rationale Behind the ETF

It’s not 100% clear why this “night effect” happens, though there are some good hypotheses. Like most interesting things in life, it’s likely a combination of a bunch of things. The FT did a piece on it six months ago:

“Quant” funds that use algorithmic or systemic strategies take advantage of the bigger impact that trades can have when markets are closed and liquidity is thinner. They aggressively buy shares they already own, driving their price higher.

Then, as markets open and trading conditions improve, they can gradually ditch the purchases without undoing all of their earlier impact. By the end of the day, Knuteson says they should be left with a slightly higher-valued portfolio. Systematically doing this, day-in-day-out, would produce the pattern of overnight gains and gentle intraday declines, he argues.

Given his DE Shaw background, Knuteson’s theory is certainly more intriguing than the usual conspiracy theories that clog up the internet. DE Shaw declined to comment. But how plausible is it? Not very.

First, the US stock market is only officially open for a fraction of the hours in a day. But you can in practice trade equities around the clock. George Pearkes at Bespoke Investment Group has calculated that if one adjusts for the different lengths of trading windows, then average intraday and overnight returns are not that dissimilar.

Second, a New York Federal Reserve study of S&P 500 futures patterns showed that returns actually spike most notably between 2am and 3am in New York. Tellingly, this is roughly when European traders get to work, and not what Knuteson’s theory would suggest.

Third, a quarter of US corporate earnings releases are published right after the market closes, and another 60 per cent before trading starts in the morning. Most companies tend to beat estimates and therefore enjoy subsequent price spikes in the overnight trading session, helping explain the phenomenon.

Nightshares itself shares the following papers:

Here’s another piece that was a good read: The Magic of Overnight Stock Market Returns (I) - systematic individual investor

All are very interesting reads.

The Play

The obvious play is to go long this ETF. After all, two of the core reasons for this effect are: a) the Euros are responsible, and b) Option flows are responsible, which, when combined have a certain subjective appeal to it.

Nevertheless, since we try to go beyond just looks here, it’d be helpful to corroborate some of these findings. Or stress test it somehow. One of the reasons for sharing this is to ask the community to do just that.

To start off with, one of the reasons I am not sure we will continue to see this effect is: all the papers seem to be from the decade long bull market. We turned a corner six months ago and have gone bearish. So need to see if overnight gains still occur or it’s more overnight losses now.

Any other hypotheses to test?

nspy spy night-effect


Have downloaded SPY data for the last five years and put this spreadsheet together:

When I look at all five years, the shape of the returns resembles what they have, though the return profile is much less. They report > 300% returns, while I get ~ 60%. This is at least in some part because dividend data is missing. But it is also possible I’m making a calculation error. Shared the spreadsheet above so you can check and see what that might be.


Overall, the thesis seems to hold that day-trading, i.e. buying at open and selling at close, and then compounding that effect day over day, results in pretty flat returns. (Blue line). On the other hand, overnight trading, i.e. buying at close and selling at next day’s open, also compounded day over day, has consistent returns.

That COVID-period move is interesting though - look at how strongly the overnight market reacted to the bad news.

Market regimes change though. I tried to capture this in a rather crude way by resetting the index to 100 at the beginning of every year. While market regimes don’t change on the calendar year, they do allow us to separate time slices somewhat. If someone has better markers, I’m happy to try to build those in.


Here, something very interesting emerges

  • 2018: The Night Effect holds - Overnight holding (green) clearly outperforms Day holds (blue) by miles. And it’s not so much that Days gain less, but more so that Days actually lose money while overnight makes it. We do not see this again for other normal years.
  • 2019: Both Days and Overnights have positive returns, but for about 2/3 of the year, day actually outperforms overnight, which is the opposite of the Night Effect
  • 2020: Look at that COVID plunge! Seems to suggest Night Effect works in both ways then - the market flushes overnight with just as much zeal when things are bad.
  • 2021: The Night Effect holds. Unlike 2018, both Day and Overnight go up.
  • 2022: Both are falling, and its not clear which is falling more than the other.

Here’s a zoom-in for 2022 that shows Day and Overnight moving in total lockstep:


(Legend missing because :GoogleSheets: …)

If there is an error in the calculations, all this is moot and we’ll have to rethink things. But for now it looks like:

  • The Night Effect is not as ubiquitous as it was made out to be.
  • It works great when the times are good
  • Overnight holds bleeds out worse than Day holds when the times are bad

If we wanted to play this, and we likely do because we think we can time the market ( :sunglasses:), it would make sense to go long NSPY when we are in a bull market again, and might actually want to short this in the current bear market.

Assuming this has better risk-adjusted returns compared to the leveraged SPY ETFs, VIX etc. Which is a whole other story.


Doesn’t look like market really cares about this one. Minimal trade volume so far.



Trade ideas like this that are so broad and simple rarely outperform buy and hold in the long run.

Like there was another one, “sell in may and go away” that would’ve costed you a fortune if you did that the last decade. Even this year, one of the shittiest markets in decades, SPX was barely up in may.