I consider the VIX:VXV ratio to be an indicator in the making. Since the VXV is only six months old it is still too early to give it the robust indicator seal of approval, but that doesn’t mean you shouldn’t pay attention to it.
Looking at the chart below, the only previous time the VIX:VXV ratio gave a bearish signal comparable to the current one was just as the markets were about to move down in dramatic fashion from late December to late January.
At the very least, the bulls should consider some downside protection in the current market environment. I suspect the bears are preparing to pounce very soon…
noticed that in 12/07, the mkt rallied for a few days b4 the rollover. be interesting to see what happens in 3-5 days
ReplyDeleteI thought that was particularly interesting -- and one of the reasons why I have been excited about this ratio.
ReplyDeleteOf course there is nothing like options expiration week to muck up a good indicator.
FWIW, I didn't see much in the way of weakness in today's rally.
Cheers,
-Bill
Bill,
ReplyDeletebut there is a significant difference between now and the end of December 2007: At the end of December 2007 (Dec. 21, 2007) front month futures concerning the S&P 500, Dow Jones Ind., Nasdaq 100 and Russel 200 volatility indexes (VIX, VXD, VXN and RVX) were trading close to (VIX) or at (VXD, VXN and RVX) their highest premium ever over their respective volatility indexes :
December 21st:
Future vs Vola. Index: Premium
VX Jan.'08 vs. VIX: +4.63
DV Jan.'08 vs. VXD: +4.70 (all time high)
VXN Jan.'08 vs. VXN: +5,25 (all time high)
RVX Jan.'08 vs. RVX: +4.94 (all time high)
Due to the fact that front month futures always have the highest volume (regularly several thousands contracts a day, a little bit over 3,000 today), highest margin requirements and highest volatility (risk) one can assume that these futures are mainly traded by the so called "smart money" (commercial and large traders), either for hedging or speculation purposes.
This time any kind of premium in the respective front month futures is either non-existent or very small (VIX: +1.15; VXN: +0.20), which would at most -if any- indicate a short term and/or temporary reversion to the mean (10 day moving average).
My personal opinion: If in doubt I'd always prefer to look at the behavior of VIX front month futures instead of any kind of VXV/VIX ratio (especially with respect to the missing amount of sample data). And this time VIX front month futures doesn't indicate that the smart money expects any kind of a larger meltdown in the market (which is of course no guarantee that it won't happen).
Best regards
Frank
(with my apologies for not being a native speaker)
Always good to hear from you, Frank.
ReplyDeleteFirst things first: your English is significantly better than that of most native speakers, but if you insist on the disclaimer, I will be polite and ignore it :)
Regarding the futures spreads, you make some excellent points about watching the front month, none of which I am going to try to dispute.
It is interesting to note the divergence between the VIX-VXV and the VIX futures premium at the moment.
At this point I am content to watch and see whether the VIX:VXV ratio or the front month VIX futures premium turns out to be the better predictor in this cycle. Of course one example won't settle the issue, but it will make for some more lively discussions and analysis going forward.
Cheers,
-Bill
since the topic is now somewhat different. how about any divergence at the other extreme? when the vix/vxv ratio peaked in mid january and mid march, were there any significant difference between the front month vix contract and the vix/vxv ratio.? maybe at this extreme, it may or maynot shed some light about this issue? I'm not an options trader (stocks only) so any option pros can look into this?
ReplyDeleteJimmy
bill,
ReplyDeleteWhats the "tell" in the vix:vxv ratio ? I use the vix alot for short term trading and I would like to get your insights here. Thanx
dave
Hi Dave,
ReplyDeleteRegarding tells on the VIX:VXV, my primary interpretive reads at this stage (I consider all of this to be a working hypothesis) on the ratio are:
1. Distance between ratio and 1.00
2. Distance between ratio and 10d SMA (or other MAs)
3. I use a couple of momentum indicators (i.e., Wms %R)here too and they are also showing an "oversold" condition
4. TBD
In terms of additional confirmation, I'd like to see the ratio begin to reverse itself and head up before I start to scale from a smaller to a larger short position.
There is also the issue, as Frank pointed out above, of whether the ratio is consistent with the VIX futures spreads.
Cheers,
-Bill
Bill,
ReplyDeletein order to verify if and to what extent the VIX front month futures could be a (or may be "the") telling indicator for the further direction of the cash VIX itself, I checked the behavior of the cash VIX over the next couple of days whenever the respective VIX front month future was trading at least 2 standard deviations above or below its all time average premium over the cash VIX:
The respective values are:
VIX front month future vs. cash VIX: +0.35 (all time average premium since introduction of VIX futures)
1 SD: 1.04
Premium range therefore: < +0.35-2*1.04 ; > 0,35 + 2*1.04 (< -1.74 ; > +2.44)
Date: Premium VIX front month future vs. cash VIX ; VIX the next day (or days)
04/17/2008: +2.53 ; VIX almost unchanged over the next couple of days
03/19/2008: -2.51 ; VIX -3.30 the next day
03/14/2008: -2.13 ; VIX +1.08 the next day, but -5.37 the following day
01/25/2008: -2.33 ; VIX -1.67 two days later
01/23/2008: -2.62 ; VIX -1.34 the next day
01/22/2008: -3.28 ; VIX -3.23 two days later
12/19 - 27/2007: between +2.74 and +4.63 ; VIX +4.70 on 01/02/2008 (vs. 12/21/2007)
11/12/2007: -3.45 ; VIX -6.99 one day later
10/18/2007: +2.59 ; VIX +4.46 one day later
10/17/2007: +2.53 ; VIX +4.42 two days later
08/15/2007: -1.86 ; VIX -0.68 two days later, VIX -4.34 three days later
(*) 08/09/2007: -3.22 ; VIX +2.82 one day later , VIX unchanged two days later
08/03/2007: -2.55 ; VIX -2.22 two days later
08/01/2007: -4.87 ; VIX -5.45 one day later
(*) 07/31/2007: -3.00 ; VIX +3.15 one day later, but VIX -2.30 two days later
07/27/2007: -4.10 ; VIX -3.30 one day later
06/26/2007: -2.30 ; VIX -3.36 one day later
03/13/2007: -2.79 ; VIX -0.86 one day, -1.70 two days later
03/05/2007: -3.48 ; VIX -3.67 one day later
03/02/2007: -3.42 ; VIX +1.02 one day, VIX -2.65 two days later
(*) 02/28/2007: -1.92 ; VIX +0.40 one day, VIX +3.19 two days later
02/27/2007: -3.50 ; VIX -2.89 one day later
06/13/2006: -3.03 ; VIX -2.35 one day later
(*) 06/12/2006: -2.75 ; VIX +2.85 one day, VIX only +0.50 two days later
... (to be continued)
(*) dates when the VIX went at least temporarily the opposite direction of the VIX front month future's premium
And on the few occurrences when the VIX went at least the next day in the opposite direction of the VIX front month future's premium, the cash VIX was almost always at unchanged levels or went the "correct" direction (the direction of the premium) a few days later.
So at least with respect to historic data it seems that there is a very high probability that VIX front month futures are an excellent telling indication for the future direction of the cash VIX itself.
Best regards
Frank
(and thanks for the flowers concerning my language skills)
Very nice work, Frank.
ReplyDeleteThis will obviously be an area of ongoing research interest for me.
Cheers,
-Bill
Yes, agree with you Bill - Frank's research is certainly interesting. My problem is that I don't have access to data on the front month contract going backwards to test - because it should be relatively easy to test given that data.
ReplyDeleteHi Damian,
ReplyDeleteYou can get the historical VIX futures data directly from the CBOE: CBOE VX futures - historical data
It seems as if you might get to this before I do. If you do, I'd be interested in hearing what conclusions you come up with.
Cheers,
-Bill
Funny - I just found that link myself - sadly, on the page they have a chart showing the VXV back to 2002 (probably simulated) but I wish I had that much data. I'll get on with testing a simple system based on it and give a result. But it's pretty easy to see that a system buying when the ratio is 1.08 or high and selling when it hits 0.82, or on a stop-loss will probably do ok.
ReplyDeleteSorry - just realized you were pointing me to a different site - I was looking at VXV data. This data will be useful as well.
ReplyDeleteSince I'd always be comparing front month data to the VIX, I believe I only need the last month of the contract's existence - so I might spend some time constructing a composite symbol that only takes data from the date of the last expiration to the end of the contract life.
Doing a bit of work on this - Fred - what do you mean by "all time average premium"?
ReplyDeleteDo you mean Average(VIX Future - VIX cash) across all quotes?
damian,
ReplyDeletecorrect, I took the arithmetic mean of (respective VIX front month future - cash VIX) since introduction of VIX futures.
Best regards
Frank
Interesting stuff - still getting my data together but in my initial explorations in Excel there's some interesting stuff. Before I setup a system, it looks to me like (not a shocker based on Frank's notes) you'd get 1-2 day advantage - which is certainly useful. I'll have to see if it will have a longer-term usefulness.
ReplyDeleteThanks for the update, Damian.
ReplyDeleteI'm adding this research project to my R&D to do list...
-Bill
Bill, I saw your post of this article on Alpha and was intrigued enough to follow it to your blog. If you view your vix:vxv ratio with bollinger bands, you will notice that on ALL 3 prior occasions that the ratio moved by .02 points outside the band, the market moved hard in the other direction.
ReplyDeleteMonday of course was the fourth time this ratio moved by .02 outside the band so you're right, this is a real test and it will be interesting.
I'm also intrigued by Frank's front month futures signal but while it did predict the December crash, it apparently did not predict the earlier crashes in July and November. It does appear to be a great short-term signal however so I've bookmarked that futures page and will monitor it closely.
Thanks to both of you guys for the excellent work. I've been aware of the VIX but I've never used it as a market signal so I'll be following your site with interest.
Jeff
I admire your work and check your site frequently. Hindsight clearly shows your "strong bear signal" of May 12 was not exactly accurate. But do you have a reliable indicator for us this coming week? Do you feel the bullish trend since March has ended?
ReplyDeleteHugh
Hi 2win,
ReplyDeleteI'm glad to see someone reads the archives!
Regarding the VIX:VXV ratio, this ratio has not been accurate since it started generating sell signals in July.
In my own investing and in my newsletter, I combine a number of signals including some related to volatility (like VIX:VXV), breadth, volume, market sentiment (put to call ratios, etc.) and some more traditional charting techniques.
For a number of reasons I am bullish on equities right now, but would flip to bearish should the SPX close below the 1015 level.
I hope this helps.
-Bill