What Overnight Gaps Tell us about the Stock Market
I’ve had overnight gaps on my mind.
As the graph below shows, volatility in the overnight market (like the overall stock market) is at historical highs. What happens after the market closes has never been so important to investors.
This is all made extra important by the fact that what’s happening in the stock market overnight (close to open) vs in the daytime (open to close) matters. In recent history, bulls have played at night and bears in the day. And I’m convinced that by better understanding how the daytime and overnight markets interact, we can better predict the markets.
Now, a lot of folks (including myself) have been guilty of saying that changes in the daytime and overnight markets are not correlated, because on the surface, they aren’t. Correlation between the overnight and following daytime is a tiny -7.2% and between the daytime and following overnight an even tinier -5.7%.
But that low correlation hides the fact that what happens overnight does influence what happens in the daytime when the overnight market is very volatile (i.e. opening up or down with big gaps) like it is at this very moment.
The table above shows the correlation between overnight and subsequent daytime changes on the SPY (S&P 500 ETF) following “small” and “large” gaps up/down in up/down trending markets from early-1993 to present.
Geek Notes: (1) I’ve defined a large gap as greater than the trailing 1-year standard deviation of daily changes on the S&P 500, which in today’s market equals a gap of about +/- 0.9%. (2) I’ve used my zigzag method to define up and down trending periods. The zigzag method is forward-looking, but remember, the point is to understand what’s happening in up/downtrends, not to predict up/downtrends.
KEY OBSERVATIONS
- The idea that the overnight market (close to open) doesn’t influence the daytime market (open to close) is usually correct, but NOT when the overnight market is moving violently like it has recently.
- Large gaps down exhibit a relatively strong negative correlation to subsequent daytime changes indicating a tendency to reverse some of the overnight gap in the daytime. I haven’t shown it in these statistics, but the larger the gap down, the higher the average return and the more negative the correlation (but also the higher the volatility) of the daytime reversal.
- Large gaps up do not exhibit a consistent influence in all market conditions. Also not reflected in these statistics is that as the size of the gap up increases, the correlation to the daytime market becomes more and more asymmetrical based on the broader trend. In up trending markets, these very large gaps up exhibit strong follow-through (positive correlation), but in down trending markets, very strong reversal (negative correlation).
In a nutshell, how the market performs in the overnight and daytime markets is generally unrelated, but that isn’t the case when the overnight market gaps significantly up or down. Depending on the direction of the gap and the broader trend, the market can have a strong tendency to reverse or follow-through on the overnight move.
Happy Trading,
ms
P.S. I’ve intentionally ignored a very important factor: volume. Food for a future post.
P.S.S. For narrower, more focused analysis of specific gaps as they happen, tune in to Quantifiable Edges. QE does what I think is the best job on the Net at this. You’ll notice that QE’s conclusions and my own differ in many instances. It’s important to note that QE is using a long-term moving average to identify the broader trend (which is how a trader would approach the problem if trading in real-time) while I’m using my forward-looking zigzag method. Remember, I’m not creating a trading strategy; I’m trying to understand the mechanics of the market when it’s actually in an up or downtrend regardless of whether the trader has recognized it yet.
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Follow-Through, Stock Market Mechanics | 4 Comments
I’m not exactly sure what you are saying.
If the overnight market is down 200 Dow points, and the next day’s DOW is down 190 points, Is that a high correlation because down night then down day, or is there little correlation because down 200 at night vs UP 10 from the overnight?
Thanks
RE to Mark: The analysis is looking at overnight and daytime changes separately. So in your example, if the market gaps down 200 points and then recovers 10 points (to close 190 points down from the previous close) that would indicate some degree of negative correlation.
Hope that helps.
michael