How to calculate the effects of backwardation

Knowing the 1st and 2nd vix futures values and the days remaining to expiration of the 1st future you can calculate a weighted average sum of those two futures to obtain the 30 day future for any given day. I gathered the data necessary to do this here.

If backwardation and contango did not exist the VXX price would be exclusively determined by that 30 day future (since both the 1st and 2nd month would be the same).

Therefore to compare the VXX with the 30 days weighted average future will tell you how much backwardation (or contango) has influenced.

Here below are the values since the 29 July, just the day before backwardation started:

Date VXX 30-day weighted future
07/29/2011 — 23.41 — 21.10
08/01/2011 — 22.41 — 20.66
08/02/2011 — 23.97 — 22.17
08/03/2011 — 24.08 — 21.93
08/04/2011 — 28.89 — 26.61
08/05/2011 — 30.31 — 27.39
08/08/2011 — 34.78 — 32.41
08/09/2011 — 31.26 — 26.91
08/10/2011 — 35.17 — 30.51
08/11/2011 — 33.78 — 29.09
08/12/2011 — 34.13 — 28.87
08/15/2011 — 32.18 — 27.21
08/16/2011 — 32.87 — 27.21
08/17/2011 — 33.53 — 27.95
08/18/2011 — 40.47 — 33.21
08/19/2011 — 42.55 — 34.65
08/22/2011 — 43.86 — 35.80
08/23/2011 — 41.77 — 33.72
08/24/2011 — 40.90 — 32.96
08/25/2011 — 41.73 — 33.74
08/26/2011 — 41.07 — 32.83
08/29/2011 — 38.50 — 30.79
08/30/2011 — 38.90 — 31.53
08/31/2011 — 38.96 — 30.86
09/01/2011 — 39.53 — 31.28
09/02/2011 — 41.90 — 32.91
09/06/2011 — 42.72 — 33.72
09/07/2011 — 41.19 — 31.93
09/08/2011 — 41.86 — 32.64
09/09/2011 — 45.83 — 35.14
09/12/2011 — 46.25 — 35.95
09/13/2011 — 45.85 — 35.79
09/14/2011 — 44.60 — 34.42
09/15/2011 — 42.38 — 32.65
09/16/2011 — 41.55 — 31.99
09/19/2011 — 42.94 — 33.15
09/20/2011 — 42.88 — 32.97
09/21/2011 — 45.21 — 33.75
09/22/2011 — 49.84 — 38.40
09/23/2011 — 50.34 — 38.76
09/26/2011 — 48.60 — 37.05
09/27/2011 — 47.64 — 36.23
09/28/2011 — 50.73 — 38.55
09/29/2011 — 49.86 — 37.55
09/30/2011 — 53.37 — 40.52
10/03/2011 — 56.84 — 43.13

As you can se VXX increased 142.8% (from 23.41 to 56.84) while the 30-day future increased 104.4%. The 38.4% increase difference (142.8-104.4) is exclusively due to backwardation.

This does not imply that the VXX should not be shorted, just that you better know about this tailwind to avoid being caught off-guard. Here I wrote about under what conditions I thought that the VXX is a safer short candidate.

See here below a graphical version of the above data:

As you can see VXX goes higher due to backwardation much more than what is explained by the futures increase.

This was calculated using un updated version of the data found here.

Cheers!
jrv

About jrv

I was born in Spain and lived in Belgium, Chile, France, USA, Argentina among other places. Currently I am trying to settle down in a wild place. I am "retired", even though now I dedicate more hours "working" for my investments than I ever worked in the real labor market where I used to work in IT and Banking. I am a family man, I have a lovely wife, several sons and one step daughter. I have humble tastes, I like to stay home and read about companies and investments. I started investing at 25 before the internet bubble exploded. I did not know much about investing and liked technical analysis so my results were pretty bad. Fortunately I did not have much to lose. Some years later in 2006 bored of doing only real state investments and with quite a lot of money saved I opened an account in a cheap and excellent online broker and started again. This time I did not want to commit the same mistake, so I decided to follow a model. I heard that Warren Buffett was the best at making money via stocks so I started by reading a lot about him, all of his shareholders letters and several of the books that he recommended. I learned a lot, started applying his investing principles and reading a lot of 10K's. Digested news from lots of different sources. Basically I started buying very good and cheap companies and holding them for ever if possible and if nothing changed fundamentally. When the housing crisis started I was more than 75% cash. At that time I identified good companies at incredibly cheap prices so I invested most of my savings in stocks. In less than I year I doubled. By the second semester of 2009 I turned my software company into an investment vehicle and dedicated myself full time to it. My wife and I decided to change our lifestyle and moved from Belgium to the beach in a wild country. The goal was to keep fixed costs low in order to be able to live with a minimum 6-8% yearly return but specially to move away from the inhuman life of civilization and to have finally some peace and sunny weather to concentrate better on investing. Now I can think and study about companies 60 hours a week and I am doing great. I can finally do what I want full time and can proudly say that I have never been so happy, specially also with my just born 4th son, my other great kids and my sweet wife who supports me fully while I study most of the day and patiently wait for the opportunity to make a swing ! You can learn a bit more about my portfolio by viewing it at www.kuchita.com/view/sumo.php or you may learn more about me and my family by following the link "Author's site" from the menu above.
This entry was posted in General and tagged , , . Bookmark the permalink.

The Intelligent Investor Blog Posts Feeds

Receive posts by email?

15 Responses to How to calculate the effects of backwardation

  1. Fabian Escalante says:

    Argh!!!! I’ve just closed a losing position after reading your article. Do you know of anything like VXX but without this strong backwardation/contango effect?

  2. jrv says:

    Not really you can directly buy / sell options or futures on the vix on the Chicago market. But they do not behave like stock options/futures because the VIX cannot be bought so they are settled in cash at expiry unless you can trade them before expiry.

    But you need to understand their dynamics quite well to get into that.

  3. david says:

    jrv,

    What is your prediction how long will it take for vxx to go back to 2x. I also shorted vxx and hedged with Dec call. It seems vxx can run like this beyond that.

  4. jrv says:

    If the volatility goes to 25 by mid December the vxx could reach 30 by then. The trick is to unload the hedges when they are not needed anymore like that you can ride your short the way down. So even if it goes up it is not a problem if you are 100% hedged, you do have to make sure you unload the hedges when the probability of vxx going higher is low and that could very well happen before December.

  5. david says:

    Thanks, jrv. Do you think XIV is a good vehicle to play with down trend of VXX? From your simulated long term chart, it seems contango did not eat a lot from it.

  6. kainvest says:

    Contango is XIV’s friend. It adds returns to XIV as XIV is the opposite instrument of VXX. So arguably buying XIV is kind of equivalent of shorting VIX.

  7. jrv says:

    depends if you’re hedged or not with vxx, the amount you plan to put in relation with how much cash you have as collateral, if you’re ok with the underperformance of xiv, and if you’re ok investing in an etf where you cannot price its net asset value…

    see my comment here:

    http://investing.kuchita.com/2011/09/10/when-should-the-vxx-be-shorted/#comment-2247

  8. Jake says:

    jrv,

    How much of an effect does the daily rolling on front and back month vix futures have compared to if the contracts were only rolled at the start and end of the month? When I try and calculate the once a month rolling, I find the results(profits on XIV, loses on VXX) to be much worse than the constant rolling. I am not sure if I am doing these calculations correctly though. I have mainly been a value investor who now has to deal with some math due to this 2x and 3x leveraged etfs and rolling futures etfs. The reason I ask about this, is I am interested in trying to trade the VIX futures similarly to the how the XIV trades them, but I am wondering if I could get away with rolling them once a month instead of daily.

    Thanks,

    Your blog is great,

    Jake.

  9. jrv says:

    it should depend on when you are doing it, if by then there is contango or not, and if you are buying the 1st month and selling the second (short volatility) or if you are buying the second and selling the first (long volatility), another problem is that you can lose if the bid/ask spreads are too big.

    i think rolling every day has a different performance than once a month (which is more of an approximation to the trend) it could be that because of doing it once a month you lose or gain depending on what happens in between.

    the question can only be answered by looking at past data and backward testing it, but that would give no assurance as if it would replicate in the future

    i think that overtime it should be quite similar, one or the other, but in the near term it could have big differences because by doing it monthly you’re taking a snapshot. Its similar to buying a stock once a month over time it would be basically the same but not in the short term.

    for example if you buy a stock one a month you could lose if you missed buying in the minimum or you could gain if you did not buy at the top, but overtime it should be the same since it would average out

  10. Kainvest says:

    On 9/22, both VIX and VXX jumped 10.x%. Is there that much backwardation?
    Can your calculation verify jump in VXX?
    Can VXX deviate from its intrinsic value (NAV)?

    thanks,
    Kai

  11. jrv says:

    The calculation corresponds almost exactly to the market price, it does fluctuate on any given day (same days a little bigger, others smaller) but hardly ever much.

    You can download the data and compare the calculated and market vxx price, they are basically the same except a small variable percentage that depends on daily offer/demand dynamics

  12. tangledweb says:

    What I find really weird here is backwardation starting at the end of June, a month before the market started to fall. That seems atypical in my experience.

  13. tangledweb says:

    Ah, misread. The world is so much clearer now. I could have sworn it said 29 June before.

  14. Pingback: VXX data since VIX futures available: March 2004 until the 4th October 2011 | The Intelligent Investor Blog

Comments are closed.