Actually I don’t dislike Mondays more than any other day, but
that is a title that I can get a funny picture to go with.
In the last few blog posts I wrote about how equity
options don’t fully account for the weekend and that there is edge in selling
on Friday. In this post I’m going to briefly look at how we can exploit the
same effect in the VIX.
It is well known that the VIX tends to be up on Mondays.
This effect has been consistent, and highly statistically significant, since
1990. The average return by day of the week is shown in Figure One.
Figure One: The average (log) VIX returns by weekday.
There is a structural reason for this. The VIX is based on
calendar time. It uses actual days to expiration to calculate the variance swap
it is based on. So if an option is priced at 5 on Friday afternoon, and opens
at 5 on Monday, the VIX calculation thinks that implied volatility has to have
increased because no time decay has occurred even though time has passed.
But the options’ market realizes that weekends do actually
exist so they tend to anticipate lower volatility on Fridays. The VIX drops on
Friday due to this. If traders perfectly accounted for the amount of volatility
on weekends there would be no trading opportunities, but the last two blog posts suggested this isn’t true. So let’s look at the daily returns of VXX.
Figure Two: The average (log) VXX returns by weekday.
The fact that all days have negative returns shouldn’t be a
surprise. Much has been written about the decay of VXX (and some of it is even
sensible!). But what might be more surprising is that Mondays have the worst
returns of all. The day the VIX is up the most corresponds to the day when VXX
is down the most.
There are two reasons for this. The reason VXX “decays” is
that the VIX futures curve is generally in contango, so keeping a 30 day
duration generally involves buying higher priced futures and selling lower
priced futures. Doing this results in a daily loss. On Monday VXX essentially
has three days of losses.
But the other reason is also important. The option market
makers don’t perfectly account for the weekend. If they did, there would be no
edge in selling options over the weekend. They actually attribute more value to
the volatility of Saturday and Sunday than they should. This is probably
nothing more than risk aversion. There is always the chance of a major event
happening on Saturday or Sunday and market makers don’t like the chance of this
happening when even the overnight markets are closed.
Direct trading takeaways from this:
- · If VIX is an input to any trading decisions you make, account for the Monday effect.
- · If you can handle the perceived risk, selling index options or VXX over the weekend is a good trade.
Very interesting read. Concept seems so logical that you would expect everyone to take advantage of this and the market to sort of correct itself because of this.
ReplyDeleteGreat post Euan,
ReplyDeleteI took your suggestion for a spin and back-tested this strategy on XIV and VXX. My first hypothesis was that if I look at Friday close to Monday open trades I would get better results. This came out to be absolutely wrong. On both products close-to-close produced better PnL (I am not sure why, maybe market makers correct their bid-ask throughout Monday?).
My second hypothesis was that the strategy should do OK across the board. But the equity curve seems to have really picked up during the last year and a half. During 2011-2013 we are pretty much flat. I am still puzzled by this.
Thanks,