Another three-day weekend is sneaking up on us. The result for those of us that have an unhealthy obsession with VIX is to talk about how a three-day weekend provides a headwind for VIX performance. It has been awhile since I looked at the impact of holiday weekends so I downloaded data and got to work.
The reason we believe in the three-day weekend effect stems from how VIX is calculated. VIX is a calendar day measure, so when we are approaching a day where the markets will be closed the result can be a bit of extra downside pressure for VIX. Once the markets get back to trading VIX gets a little tailwind.
First, I took all trading days for VIX from January 2, 1990 through February 10, 2017 and pulled out all days where there was no interference of a holiday. For instance, only looking at Monday performance when the previous trading day is a Friday and the following day is a Tuesday. Using these days, I tallied up the number of days VIX rose and how often it was down on the day. The results appear on the table below.
The first column is a sum of all non-holiday impacted trading days and the total number of days VIX was higher and lower. VIX is slightly more likely to drop in value on any one of these days than to rise. Anyone that follows VIX is probably not terribly surprised by this. VIX tends to spike quickly to the upside and grind lower. This is a nice quantification of something we all ‘know’ with respect to the behavior of VIX and implied volatility.
Note on non-holiday Mondays VIX is up 63% of the time and lower 36%. For those doing the quick math, there are a few unchanged days in there so the total is less than 100%. More importantly VIX rises on Monday after the impact of two non-trading days benefitting from that tailwind I mentioned above.
Finally, note that on non-holiday Fridays VIX is down 62% of these observations. That’s the result of that headwind I noted with respect to the market being closed for a couple of days.
My next test involved looking at long weekends. We can have three day weekends that run from Thursday to Monday or from Friday to Tuesday. I segmented both types of long weekends out for the data below. On the left side are data for the pre-holiday Thursday and Friday on the right side we see how VIX performs the first trading day after a long weekend.
Pre-holiday Thursdays see VIX rise 37% of trading days and 40% of pre-holiday Fridays experience higher VIX. These numbers were honestly surprising to me. Since they are basically in line with non-holiday Fridays. It may just be that there is no true three-day weekend effect, we are just seeing the normal end of week impact, but paying more attention to it because of the pending holiday.
Finally, Monday and Tuesday after long weekends tell a different story. 85% of post three-day weekend Mondays and 74% of Tuesdays have seen a higher VIX. This is higher than the 63% number associated with non-holiday Mondays. My theory on this is if there is an extra headwind going into long weekends it may kick in two days before the long weekend. For now, that’s a theory, but I’ll probably follow up with a little more digging into the data soon and definitely report back in this space.