Someone has to be wondering about this, even if I haven’t been asked yet.
Here is the reasonably concise answer:
In December 2006, I began wondering when we were going to have a correction from the July rally (I’m still wondering) and set about to look at a bunch of sentiment and other contrary indicators that might support my case. One that particularly caught my attention was the VIX, not just because of the obvious negative correlation to the SPX (see graph below), but also because options are actively traded in the VIX and could provide a more leveraged play than say the QID in a nasty downturn. I did some Googling and concluded that if there are any VIX ‘experts’ out there, they are certainly not a publicity-seeking bunch.
So I download the VIX price history data to Excel and started playing around with it. I posted some observations on an InvestorVillage forum and decided that a blog would be a much more convenient way to archive my thoughts.
That’s about it.
I find it ironic that I now have a VIX blog and am now thinking regularly about implied volatility issues, as I am not an “options guy” per se, having traded almost exclusively equities for the past 20 years, with the exception of an occasional outright purchase of a call or a put, and a rare covered call play. Only in the past month have I made my first foray into (bear call) spreads.
So that is why I call this a “learning laboratory of sorts.”
Initially, I thought I might tackle a wide variety of sentiment-related issues, but since Zen’s Market Insights, HeadlineCharts, and others do such a good job of covering the waterfront there, is not a high priority for me at this stage.
I will, however, experiment with the “and More” portion of this blog soon enough. Hell, espn has never figured out what to do with the “e” at the beginning of their name, so I am not worried about how long it takes me to sort it out.
Thanks to all who have contributed here in one way or another.
As always, comments and suggestions are welcome.
Did you see Justin Lahart's column in the Wall Street Journal today? (2/13) There was a little paragraph in there that mentioned "buying and selling volatility" -- maybe you could clarify: "The VIX and other measures of implied volatility are low, in part, because investors are selling put and call options - "selling volatility" in Wall Street parlance. That helps to drive option prices - and implied volatility even lower..." As I understand it, buying and selling volatility refers to specific option trades that do not depend on directional movement in the underlying stock position, only in a change in volatility. Justin, though, seems to define it as selling puts and calls.
ReplyDeleteHi Dominic,
ReplyDeleteA quick answer is that you are mostly correct in that being short volatility (e.g., selling straddles and strangles) means -- for the most part -- that the price of the underlying is not important. Sellers are betting on the underlying moving less than the implied volatility priced into the options. Range-bound markets would be an ideal scenario in which to sell volatility.
Lahart is also correct, at least to an extent, in that aggressive selling on a large scale could force options prices down and with it the implied volatility component of prices, particularly if there were expectations of continued aggressive options selling.
What I don't know is how much selling it would take to substantially move the market prices and implied volatility. This would be a good question for Adam of the Daily Options Report.
Bill, about the blog, I think you have created a niche and you're good in what you are doing so heck, who is worried :) It is insightful and i guess the by-product "fame" is just a click away as there aren't many out there talking about Volatility...
ReplyDeleteCheers and profitable trading via VIX :)
OptionPundit
Thanks pundit!
ReplyDeleteBill Luby! How the heck are ya? Welcome to the blogosphere!
ReplyDeleteBR
Thanks for the welcome, Barry. I must say that perhaps more than any other blog, The Big Picture has helped get me up to speed in this strange virtualverse.
ReplyDeleteHow much of the VIX is the product of recent actual price volatility and how much of the VIX is the product of the emotional content of options traders? Of the portion that is emotional content, how much of that emotion is the result of recent price changes in the index itself, i.e., falling price instills fear? (ANOVA)
ReplyDeleteIf the VIX is a product of price movements (up/down + volatility) in the index, then would it be simpler to use price movement to predict corrections?
If the VIX is a product of price movements (up/down + volatility) in the index, then could those movements be modeled to predict trade-able changes in the VIX?
Bill,
ReplyDeleteI'm not sure how I missed your comments in real-time, but your questions were excellent ones then and will remain so for a long time going forward.
I will attempt to answer them in this blog at some point going forward, but here are my quick thoughts.
I buy your thesis that the VIX may be largely the aggregation of:
1) pure rate of price change in indices (or individual components)
2) emotional reaction strictly to the speed and magnitude of the price change (lagging indicator)
3) emotions unrelated to recent price changes (leading indicator)
I think the problem is that these are very different animals rolled up into one aggregate beast: #1 is relatively constant; #2 is rather predictable, but has some multiplier/accelerator components that kick in from time to time (e.g., around 2/27, the rate of change in the VIX vs. the rate of change in the SPX accelerated dramatically); #3 vascillates wildly and is sometimes also affected by #2
Theoretically, #1 can be modeled fairly easily, #2 can be modeled with a little more effort, and #3 is the elusive sentiment/fear/greed component -- very hard to model and predict, but worth the effort, given the potential for unlocking even a small piece of the kingdom.
Some great questions. Sorry the reply -- if not the answer -- took so long.
So hypothetically, if 1) is a function of the ATR or STDEV of the daily index movements, and 2) is a function of the ROC for some period of days prior to the VIX reading, then a model of 1) and 2) could be a predictor of VIX, and the "error term" could be used (probably with more accuracy than the VIX itself) as a "leading indicator."
ReplyDeleteWorth a try?
I think you and I are on a similar wavelength here.
ReplyDeleteI suspect, however, that #2 has extremely fat tails (kurtosis) and that in attempting to model this accurately, the fat tails dynamic and the #3 'error term' (I love it that we can refer to sentiment as an 'error term') distinction will probably get blurred.
I think modeling the error term component will be the heavy lifting. Of course it can just be calculated looking back at historical data, but going forward, it sure would be nice to be able to distinguish between #2 and #3.
Please tip off VIX and More if the NO DooDahs research team has any conclusions they are interested in sharing.
Cheers,
-Bill
What if the "price rate of change" and "reaction to price rate of change" components were rolled up into one factor called "dumb money"? Clearly this factor would be reactionary and momentum-driven which happen to be two common characteristics of speculators, aka lagging indicator. What if there already exited an indicator that could sum up this sentiment? *cough...total put:call ratio*
ReplyDeleteMy theory is that options market makers (some more than others) are among the most astute traders in the financial world because of the inherent risks in their product. They keep their ears very close to the Street and eyes on the order flow. They are often among the first to notice if paper is in the market and adjust options prices accordingly based on their interpretation of future volatilty. If you can tease out this data real-time as the volatility indexes get out of sync with price action, you have quite an edge...
This method has worked *extremeley* well for me.