Abstract Stocks

Like Night and Day

Phil Mackintosh
Phil Mackintosh Nasdaq Chief Economist

Some in the industry are talking about making stocks trade 24 hours a day, like stock futures and currencies. But for now, stock markets are typically only open for a fraction of the day. That’s good for traders, as it lets them focus on prices and liquidity for a small part of the day and lets them sleep soundly on completed trades or hedged positions.

However, just because the market is closed doesn’t mean valuations don’t change.

Today, we look at returns from holding stocks intraday compared to overnight. The difference is (quite literally) like night and day!

Most trading happens during the day

When we look at trading activity in the market, we see that almost all trading, over 95%, happens during the market's official “open” hours.

Although after-hours trading is allowed, there is very little trading of stocks overnight. And the majority of what does trade after hours occurs in the hour before and after the market is officially open, which is also when some U.S. economic data and most earnings are released.

In fact, a lot of the trading that happens after 8 p.m. when the SIP stops reporting is reported to the SIP at 4 a.m. However, in total, that adds to less than 0.2% of the typical ADV.

Chart 1: Most trading happens intraday


Chart 1

Given this – and the economic theory suggests that prices move because of supply and demand – you might expect prices to move much more in the day than at night.

However, that’s not true.

Day volumes dominate, but night returns are higher!

Interestingly, if we model the performance of a day-hold and a night-hold strategy for the S&P 500, we see that overnight returns have been significantly higher than day returns. That’s despite most of the trading happening during the day.

In fact, over the more than 30 years in this chart:

  • Intraday returns are barely positive (+12%).
  • The buy-and-hold strategy has increased almost 20-fold.
  • Overnight returns, therefore, have also increased close to 20-fold.

To some, this difference in returns may seem unexpected, given that trading happens during the day. However, there are some very obvious explanations as to why this happens.

Chart 2: Historic returns of day and night hold strategies, with and without trading fees


Chart 2

In the chart, we compare:

  1. Buy-and-hold strategy (green line), which represents the total return of holding SPY, the oldest ETF in the U.S.
  2. Day strategy (orange line), which represents buying the market open and then selling the market close, holding uninvested cash overnight.
  3. Night strategy (blue line), which represents buying the market close and then selling the market open, holding uninvested cash during the day.

Note that the sum of the day and night strategies equals the buy-and-hold return.

Don’t try this at home!

Although this overnight-only strategy might look attractive, we would remind investors that trading is often harder than it looks. For a start, a strategy as active as this, where you buy and sell your whole portfolio each day, would likely have high annual trading costs.

In the dashed line above, we add trading costs of just 1 cent per trade (Chart 2). Remembering that SPY is currently almost $500, that adds to less than ½ basis point (or 0.004%) per day (both trades).

The dashed line shows that even these small trading costs almost totally eliminate the historic gains of the overnight strategy. If a lot of people did this trade, costs could easily be higher – especially after paying commissions and SEC fees.

Since 2001, this hasn’t worked consistently

It’s also important to understand the power of compounding.

Starting this strategy back in 1992, when SPY was first launched, seems to bias the results. During the formation of the tech bubble, the overnight strategy consistently outperformed, adding to what may have turned out to be an unassailable lead.

If we instead look at rolling 12-month returns, we see that investors joining this strategy after 2001 would have far less predictable results. In fact, on this basis, the day strategy leads 30% of the time (orange bars in chart 3).

Interestingly, that’s consistent with two ETFs that were assigned to this theme. Both were closed recently after failing to gain assets, but they also significantly underperformed in the year before they closed.

Chart 3: Rolling 12-month difference in returns shows outperformance is not consistent


Chart 3

It’s the economy, stupid?

The data in Chart 3 also shows that many of the longer periods of “day” outperformance seem to align with some important macro events (shaded areas).

That makes some sense. Bad macro news from foreign markets would make all stocks fall overnight. Often, U.S. rescue packages are announced during the U.S. day, boosting that day's returns.

Covid is an interesting case study. A period of mostly bad news for stocks, with some significant updates and shocks coming out of Europe and Asia – during their trading hours – that made U.S. stocks fall after hours. Still, the largest stimulus and rescue plans came from in the U.S. and were mostly announced during U.S. day.

In short, news affects stock valuations – and in an efficient market, it should be priced into stock returns when it happens, not hours later.

There is a lot of stock news after hours too

It’s not just macro news that matters. As we recently showed, company earnings and other material disclosures are also, overwhelmingly, released after hours. That data also suggested that company earnings are, on balance, positive for stocks, too, which would contribute more to positive returns after hours.

In fact, a lot of the trading we see after-hours occurs in earnings stocks on the day they report.

Chart 4: Earnings are all outside of market hours


Chart 4

Sometimes, there is a logical explanation for something unexpected

With so much price movement in the quiet of the night, it might be easy to think that something spurious is going on.

However, with an understanding of how stock valuations work and when important news happens, we see that sometimes prices can move without significant supply and demand and trading. It reminds us of another study we did where we found that some ETFs can tick actively all day, adjusting their prices to new news without ever trading (not even to arbitrage a mispricing).

Although the proportion of stock returns coming after hours is disproportional to the level of trading, this shows it can all be explained by something as fundamental as news leading to updates in stock valuations.

Looks like the market is efficient after all!

Latest articles

Info icon

This data feed is not available at this time.