More Record Lows in the ISEE
I haven't bothered posting about the ISEE lately not because nothing of interest is happening with that index, but because the story hasn't changed. In fact, the ISEE has set new record lows in the 20 day simple moving average for the past two days and the 50 day SMA for the past three days.
The VIX may look placid, but be careful not to conclude too much from that one data point. Options buyers may not be panic-buying puts and driving up volatility readings, but their relative interest in buying calls vs. puts is much lower than it ever has been during the six years for which the ISE has published ISEE sentiment data.
As I mentioned a month ago, history suggests that the markets usually struggle after we have a lackluster VIX combined with investors who are reluctant to buy calls.
7 comments:
I like the ISEE methodology much better than the traditional put/call ratio methodology, but I feel it still suffers from similar defects... without knowing the context of the opening long purchase (of the call or put), it doesn't indicate whether true sentiment of the customer.
For example, suppose I owned a million shares of some equity (I wish!) and bought 10,000 puts as a hedge... the ISEE would go down, indicating bearish sentiment, but my net position is synthetically 10,000 long calls, suggesting that I am bullish..
Or suppose I sold a 90/100 call spread, 5000 times (bought 5000 calls at a strike price of 100 and simultaneously sold 5000 calls at a strike price of 90). That is a bearish position (short call vertical), but if I understand the ISEE methodology, only opening long call and opening long put transactions are included in the calculation (i.e. the 5000 calls at 100), which would make the ISEE slightly positive.
Of course, most retail traders don't play these (synthetic) games, and at extreme values, the ISEE (and even the tradtional put/call ratio) is probably accurate to a reasonable degree.
What I find interesting is the relationship between the cash VIX and the near-month VIX futures; today, they closed at 25.02 and 26.05, respectively. Earlier in the day, the FEB VIX futures were as much as 1.75 higher than the cash VIX!
Now bear in mind that the FEB VIX futures settlement price is based on the opening transaction prices this coming Tuesday morning. It suggests to me that futures traders are betting that VIX values will spike up by a little over a point after the three-day holiday... maybe forecasting a gap down in the SPX? Either that, or they forgot that the settlement date was pushed back by a day (due to the Good Friday holiday in March), or they're nuts/dumb.
I think I should discount the last two possibilities, since dumb or forgetful traders don't last in the futures market for very long.
Depending on what happens, I will either be quite pleased or rather unhappy on Tuesday...
Felix,
some statistics concerning the behavior of the VIX before compared to after regular and long weekends (including a holiday).
Going all the way back to January 2, 1992, there were 842 weekends. On 678 (= 80.52%) of them, the VIX opened higher the first day after the weekend compared to its closing value the day before the weekend. On 164 (=19.48%) weekends, the VIX opened lower.
The average gain (the delta between the opening quotation the day after the weekend and the closing value on the last trading day before the weekend) was 0.77 points, the maximum was 11.36 points (!).
I already posted these statistics on Adam's Daily Options Report.
I will add an additional statistic (but it takes some time) if there is any correlation between how the VIX opened after the weekend (higher or lower) and the behavior of the S&P 500 itself (that means did the S&P 500 really opened lower than it closed the day before the weekend 80.52% of the time in correlation with a VIX which opened higher 80.52% of the time ? I really doubt that, but "believing" and "guessing" is nothing which I would rely on trading in the markets)
So a premium of ($)1.03 concerning the VIX Feb.'08 future versus the cash VIX before a long weekend is nothing to write home about (taking into account that the average "gain" is 0.77 on regular weekends with very little long weekends included), and selling such premium (shorting VIX Feb.'08 future) seems -a least from a historical perspective- a bet with only average odds and risk/reward (at best).
And especially with respect to the front month in VIX futures which almost always have the highest volume and open interest I would take special care that I have an extraordinary edge before I enter any short position going into a (long) weekend before expiration. VIX front month futures regularly expire on Wednesdays, so there is almost always only a short window between Tuesday's closing and Wednesday's opening where something surprisingly can happen (if one is short the expiring VIX front month future), but this time you have 3 additional days in between which justifies for some additional risk premium in VIX Feb.'08 futures as well I think.
Frank
Yikes, these guys know their stuff. Time for me to brush up.
Was anyone a little stunned by the fact that the indicative VIX has been wrong for the last month? That might explain the large spread between the Feb futures closing on Friday and the cash VIX.
Not sure if the link will work but see the info circular on the CFE site.
http://cfe.cboe.com/framed/PDFframed.aspx?content=/publish/CFEinfocirc/CFEIC08-002.pdf§ionName=SEC_ABOUT_CFE&title=CBOE%20-%20CFEIC08-002%20Revised%20Closing%20Indicative%20Values%20for%20VIX,%20RVX,%20VXN%20and%20VXD
BTW does anyone know where I can find the Feb settlement value? It doesn't appear to be on the CBOE site yet.
Great comments, all. Let's see if I can help here.
Felix, the ISEE calculation will not incorporate any spread trades or orther simultaneous transactions in their model: "Unlike traditional put/call ratios, ISEE is based on opening long customer transactions only."
Frank, thanks for the data and opinions.
Anon, the VIX SOQ has its own special ticker: VRO. I Yahoo Finance page probably won't excite you, but you can get info on VRO at Yahoo Finance: VRO -- which shows a settlement of 25.51.
I hope this helps.
-Bill
Thanks.. It looks like the "real" closing cash VIX on Friday was 25.54, vs the widely quoted close of 25.02. So that was only about a .5 point premium of the Feb VIX contract close (26.05). I had noticed the huge premium earlier in the day but was too chicken to put a huge position on since I wasn't sure why it was so large.. apparently some sheneganians were going on..
Frank, thank you for compiling and sharing your data. It's very interesting, and there seems to be something there (that I don't quite understand yet).
That said, it seems that the FEB settlement for the VIX was one of the <20% cases, in that the 'revised' close of the cash VIX on Friday (per the CFE information circular) was 25.54, and the open today was 25.39, just a little lower.
However, the settlement at 25.51 was higher than the open (despite the SPX gapping up on the open and the S&P e-mini futures up over 10 pts. at 9:30 am EST), very close to the revised closing price. I'd be tempted to suspect that the VIX futures traders knew about or anticipated that the CFE would revise the VIX upward, but if that was the case, they got the direction right but the magnitude wrong.
I agree with Anonymous' comment about being stunned that the cash $VIX has been wrong since Jan 14th 2008 -- it's difficult for me to believe that the CBOE/CFE only discovered the error on Friday. And they posted CBOE Information Circular IC08-026 (also CFEIC08-002) after trading hours on Friday.... perhaps it doesn't matter though, since one trades the VIX futures directly, not the cash $VIX.
And I really shouldn't complain... got lucky and was pleased with the settlement price and VIX action today. :)
Regarding the ISEE calculation, I guess I interpreted the quote a little differently, Bill. I was thinking that opening a spread position would be considered as opening one or more long positions (which would be counted by the ISE methodology) and one or more short positions (which would not be counted). But your interpretation would of course make for a more valid indicator. I guess it could depend on how the transaction data is reported to the ISE.
Felix aka thenakedtrader
Post a Comment