On the Rarity of a 20% One Day Drop in the VIX
Back in February and March, the wild swings in the VIX inspired me to write extensively about my research into VIX spikes and topics related to volatility extremes. Most of these are spikes higher, as fear is not something that is usually extinguished during the course of a single trading session, just as one slain dragon is not likely to put a village back at ease. After all, what is that the chance that there was only one dragon out there?
There was an instance, however, in which the VIX nearly dropped 20% in one day in March, prompting a retrospective look, at the three previous instances of single day 20% drops in the VIX that I titled “Elast-o-VIX.”
Now that I know enough blog-related HTML to be dangerous, I have updated that earlier graph and added second graph that depicts yesterday’s 23% drop in the VIX in the context of the past week. If the past is prologue, then does it matter if the current VIX drop was not preceded by a VIX runup of similar proportions? That remains to be seen, of course. Past instances notwithstanding, the VWSI currently stands at +4, suggesting that the VIX is more likely to move up from here rather than down in the next week or two.
[With only three previous data points to draw upon, I feel compelled to note that we are a long distance from anything resembling statistical significance, but sometimes historical voyeurism is more interesting than statistical significance, particularly when we are talking about once or twice a decade events.]
9 comments:
Another difference is that the other 20% drops happened in relatively low VIX environments, whereas this one is relatively high, being above its long run average.
Not sure whether that affects the analysis or not.
Bill,
Notwithstanding the excellent work here, your charts generally get chopped off by the sidebar on a 1024 pixel screen.
Regards,
CB
Good point, David, and one I had intended to mention, but forgot about. (Sometimes trading gets in the way of blogging...)
The absolute numbers are important here, I believe that a sharp upward and/or downward VIX move in a low VIX context is more likely to be received as a random stray event (bullet?) in an environment that is generally positive for equities. On the other hand, a sharp VIX move in a high VIX cycle is more likely to be just one of several moves in a larger volatility cluster -- and have a more ominous feel to it as well.
All this is just my opinion, but I think your point is an excellent one.
Cheers,
-Bill
Hi Charles,
Thanks for the input. I recently had this to say about the issue you raise:
I've gone back and forth on linking vs. embedding and for now I think I'm going to go with embedding, unless there is a groundswell of opinion against it.
There are a couple of workarounds I can offer you, though. If you use Firefox or Flock, just right click on the image and try "View Image." Also, "Copy Image Location" or "Properties / Location" will allow you to paste the original image into a new window. In IE it is a little more unwieldy, as I believe you have to use the "Properties" command to copy and paste the address into a new window.
Thanks for the comment; I hope this helps.
Cheers,
-Bill
VIXMaster Bill, to put even more context to past occurrences of a large one-day drop in the VIX, could you tell us what the VWSI was in the past cases? I suspect that, given the mean-reversion thesis of the method of calculation, it would have been slightly positive or neutral in the past cases, but your confirmation would be welcome and appreciated.
Also, yesterday's case may be somewhat anomalous, given settlement of the SEP VIX futures and options on the morning following the FOMC announcement.
Today's SPY/SPX and VIX candlesticks, assuming they hold up through the close, suggest to me that we could have short-term reversals on them this week. An inability of the bulls to sustain yesterday's rally today, as well as lower volume, could lead to some profit taking before the end of the week.
Of course, my analysis may be colored by my position.... and yes, I should have my head examined! :)
Felix,
Good observations all around. FWIW, we are of similar minds about the FOMC announcement anomaly, as well as the fairly high likelihood of seeing a short-term reversal, though the bulls appear to have some kick left as I type this.
Regarding previous VWSI levels at the end of the day following the 20% drop, your hypothesis is right on. The 2006 VWSI was an even 0, the 1994 level was -2 and the 1993 level was -1, nothing that suggested any large move was likely to follow.
Now go get that head examined before the markets decide you are right...
-Bill
Master Bill,
Thank you for your supportive comments, and especially for the additional VWSI data.
As we agree on the mean-reversion nature of the VIX, I suppose it is natural that we both feel a snapback to higher levels is likely. As I am currently long VIX vegas, I'd also like to see higher implied volatility levels in the VIX, though I'll admit it's unlikely to go much over 100% soon unless the market gets fearful and bids up VIX premiums.... I'm probably being too greedy here. :)
And yes, I'll go get my head examined whether the market agrees with me or not! :D
Best regadrds,
Felix
Hmmmm... close was a little stronger than I expected, and volume was kinda bullish. Still, the VIX and all the major indexes appear to be at or beyond the two sigma bounds of their Bollinger Bands, so I'm still expecting some snapback towards the means tomorrow (or at least by the end of the week).
Best regards,
Felix
It´s no huge problem. It gets solved by reducing the screen to 80% in Firefox or Opera. I just thought you might like to know, seeing as 1024 screens have not yet fallen off the face of the earth.
Thanks,
CB
Post a Comment