As previously posted I have now gathered all the futures data necessary to calculate the VXX (and of course used it to calculate it) since futures data are available (26th March 2004).
Here you can download the data and use it to make your own graphs like for example a contango / backwardation graph or to calculate the devastating effects of backwardation in periods of high/increasing volatility..
Note that the data is calculated until the 11th May 2012, if you want to calculate new VXX values you will have to implement the formula as explained here. It is good that you implement how to calculate the data in order to have a better understanding of the VXX dynamics. The futures data is available from the CBOE site here and you can use it to keep on updating the futures data since I may not update it regularly.
If you do not have time to implement the pricing formulas, or want future forecasts, or a sell signal model, I sell for 35 US$ the same spreadsheet with a data model for historical and future forecasted data (based on public VIX futures data) including:
1) Pricing formulas.
2) Forecasting model for future values, including market randomness.
3) Latest data update, and future updates if you don’t manage to do them.
4) Contango and backwardation data.
5) Shorting model, parametrizable, which gives automatically sell signals.
6) Support on forecasting, selling signals, and VIX futures updates.
Besides the historical and modelling data it has the advantage that you can play with future values or understand exactly how backwardation and contango affect the VXX price. You can also use it to make forecasts of how the VXX will be affected depending on 1st month and 2nd month future prices. It includes VXX price forecasts based on VIX futures data with which you can play with to estimate what the VXX price will be depending on different scenarios. That’s actually why I did it because I could not find anyone who had done a VXX pricing and forecasting model based on VIX futures (like it should). It’s basically the tool I use to guide my trading. (anyone interested in sharing modelling or trading ideas feel free to contact).
The short model gives you selling signals. It is essential in order to support your trading system. You can very easily fine tune the parameters to obtain more or less selling signals. It’s parameters depend on momentum trends (between fast and slow moving averages) for the VIX futures and the contango and backwardation, combined with absolute values for both of them. Basically it gives sell signals if the backwardation is disappearing and losing momentum but the VIX futures (based on a contract weighted average of the 1st and 2nd vix futures) are increasing but are already very high. That simple mode gave selling signs on every year since 2008 until 2011. You would have not shorted at the peak but quite close and most importantly: after the VXX peak was reached! I talked just a bit about it in the following post: http://investing.kuchita.com/2012/04/19/vxx-shorting-model-sell-signals/, but the best is to use it to grasp its power.
You can make the payment via paypal to my email jvelasco@gmail.com and let me know and I’ll send you the excel spreadsheet with the latest data and the formulas, I’ll gladly give you e-mail support should you need it, alternatively if you are in Europe you can send a transfer for the equivalent amount to a European bank account (just ask if needed).
Here below I created two graphs using it and showing the VXX values for two different time frames:
From mid October 2005 until around end July 2011:
As you can see you can get severely damaged by shorting before a VIX period of high volatility. Due to the increase of the futures value amplified by the backwardation effect on the VXX. Or if you are long you could multiply your money by 3 to 5 times, relatively fast if you buy before a crisis as severe as the great recession.
But on the other hand history shows, in the complete VXX graph here below, that due to contango the VXX is in a downward trend so if you can wait long enough, and if margin calls do not ruin you, you could eventually make money shorting VXX at almost any time if you can wait long enough, years eventually are needed for the short to work out:
That is if history holds, we do not know if it will. So do not short the VXX and blame me if you are ruined even after holding 10 years your short 
.
The VXX started to be calculated from the moment when there were VIX futures available in 2004 and in that time the futures were in a persistent and long period of contango up to 2007, that basically eroded most of VXXs value. Also the VIX was at very low levels and even going lower. Therefore since its birth and the following years the VXX went abruptly down. It had some recoveries for the 2008/2009 crisis and in 2007 but it never recovered to the original levels, and I have my serious doubts it will ever because most of the time the VIX futures are in contango and that erodes the VXX value much more than what it regains in the rare but strong periods of backwardation. You can also see the same data by looking at the S&P VIX short-term futures index which is computed exactly like the VXX, therefore it is proportional to the VXX having just a different scale factor.
The long term historical trend of the VXX is down but it can have violent spikes that multiply its value so its not as easy to trade it as it initially would seem.
Hope this data is useful and if you find any interesting patterns by analyzing it please do not doubt to let me know !
Cheers!
Juan
Pingback: VIX vs VXX historical graph | The Intelligent Investor Blog
Pingback: VIX contango – backwardation historical graph | The Intelligent Investor Blog
Pingback: Contango/Backwardation VIX Graph | The Intelligent Investor Blog
Pingback: XIV graph and data since VIX futures available (2004) | The Intelligent Investor Blog
Pingback: How to calculate the VXX price and how does backwardation and contango influence it | The Intelligent Investor Blog
Pingback: When should the VXX be shorted ? | The Intelligent Investor Blog
Pingback: How to calculate the effects of backwardation | The Intelligent Investor Blog
It would be interesting to see an historical simulation of TVIX, the 2x levered ETN, as well. It’s decay, and pops, should be incredible.
Great blog! Very interesting stuff… I wish I would have read it before!
I am trying to understand by replicating your spreadsheet.
cheers
Louis
Hi! Loved your blog, and I thought I would take a crack at your spreadsheet. I wanted to try to do this from scratch by downloading the historical data and seeing if I could make a sheet of the first two months of data. I noticed some irregularities in the data, and noticed that your sheet may suffer from the same problem.
The CBOE historical data is very incomplete for many dates. There are places where the values you list in month 1 and month 2 for dates are actually not month 1 and 2, but rather some other spread of months, like month 1 and 3. For example, 8/16/2005 is a date where there is no M2 month data, so the 2nd month data in my sheet (and your sheet) is actually M3 data. Feel free to contact me if you have questions about this.
Hi Ragu, I am aware of that problem, I used the two front end months, if one of the first months was missing in the CBOE site, I took the next.
Cheers!
Juan
But when you used two front end months that were further away than normal due to a missing month you didn’t use the correct number of days. You used, say, 16 days when the data was actually for 16+30 days.
Thanks for pointing that out Nick, I updated it.
It affected 3 months on 2005 and 1 of 2004. For 2004-2005 the model remains basically the same except that the vxx decay is slightly slower.
The model was not affected for the other years.
Cheers!
Juan
This is what I wrote to one of my readers and pricing model buyers in reply to what I thought about the current volatility market and the historical high contango (I post t here because it might be helpful to others):
Hi Michael,
Here I send you the file, you find instructions on how to keep it updated on the 2nd spreadsheet tab, do not doubt to ask if you have doubts.
Contango has been high because the 2nd vix future is much higher than the 1st month vix future, mainly because the volatility fell very fast and the 1st future followed but not the second yet, so the market thinks in the long term that volatility could spike, but if it doesn’t spike fast the 2nd future could fall and the contango become smaller and at the same time causing the vxx to fall more.
So I think that the only way to make money is with the volatility to spike soon, otherwise if it stays contango will make the vxx fall a bit every day and a lot in the long run, and if the volatility falls more then the vxx will have a double fall due to contango and the futures falling.
So honestly I think the only way you can make money is if the volatility has a spike soon, by that I mean in at most 2 weeks, otherwise contango could make the vxx fall another 10% in that period.
If volatility traders do really believe that recovery is coming a long position has no chances and you should get out soon. If on the other hand there is a scare in the coming days the vxx could spike and you should take the chance to get out. If you are very confident that a big scare will come then of course you can make a fortune but for that you need a big market scare. If volatility stays as it is then you can be sure that vxx will relentlessly go down. So it all depends on what you think regarding how scared or confident the market is about the recovery.
If contango narrows because the 2nd future falls then the vxx will fall even more, if it narrows because the 1st future goes up then the vxx will go up. What matters is the average 30day vix price, that’s what dictates the vxx price mostly. So it all depends on what you think again, contango will go down probably yes but it depends how, if it’s the 1st future going up and approaching the 2nd one then the vxx will go up, in the other case it will go down.
I personally, as you maybe know, am short, since long ago, but I am about to sell because the risk is high that the vxx will go up. On the other hand I would not bet on that either, so at current levels I am basically neutral and I would stay out, or at most I would short but just a small position or I would be long only if and really only if I thought a scare would come soon. Otherwise Id stay out of it.
Hope I could be helpful !
Good luck!
Juan
Hi Juan, thanks (as ever) for sharing these comments. In the past months I have been shorting vxx with monthly limited risk positions. I was able to manage minor spikes of up to 25% by dynamically adjusting my positions. With VXX declining at roughly the same rate as AAPL was rising it’s been a very profitable trade. To manage major spikes I limit position size so there will be enough cash left to profit from a subsequent vol collapse. Based on analysis of your contango chart I think this trade will be profitable for as long as VIX contango persists.
It would be nice to hedge a major VIX spike, but contango makes a direct hedge too expensive. A more cost effective approach might be allocating 10% of reg T margin to (June) dollar index futures (UUP) with a stop loss. Appreciate any thoughts on hedging of VIX spikes.
Groeten!
Hi Max, congratulations! I have not investigated too much on hedges for a volatility spike, dollars or treasury bonds could do the job but if volatility drops they might drop too, probably less so it could be a nice hedge. I’m also inclined to think that a good strategy is being permanently short and adding even more short positions on spikes. The key is not to be short on a big amount that could do you harm. I am inclined to think that’s a good strategy because even though I acknowledge that the VXX could triplicate or more suddenly it would not matter too much if you are already making a lot of money on your shorts and that would leave you money to short more on a good time: when the vxx is high.
Anyways a very good market knowledge is also probably of more help than anything else, if you can really understand the market and foresee major problems you probably have an edge trading volatility but that is something quite difficult to do, I don’t think I can or am inclined to do it so I focus my investing efforts more on single companies than on the whole market. Maybe some people can understand the general market well enough to profit from it, for example I think George Soros was quite good at understanding it and profiting from it, in his top times.
I leave the VXX as a side trade because I think I understand better than the average trader how it operates technically and recognize that it passes most of its time in contango and I believe that will not change, there will be periods of bacwardation but they will most probably pass too and what’s more important: they give you the opportunity to short.
In conclusion I like the idea of being permanently short on a position small enough to avoid having pain if it goes deeply against you and that leaves you money to add on spikes, the higher the spike the better to add more. At these levels though I am not adding, I might even close my short positions one of these days, I’m not sure yet bout that, but what I’m sure is that I am not shorting, as of now, more than what I already have.
Cheers!
Juan
With VIX at 15, I couldn’t resist spending 1% of my portfolio on VIX June 30 calls, sufficient to hedge my monthly VXX shorts for April, May and June in case of a vol spike. Will let you know how my Q2 plays out. I follow your stock picks with great interest but I have not been able to find the time for the value investing approach myself. Just long AAPL call spreads because support over the past year has been incredible. I wonder what could bring it down?
Hi Max, indeed at this levels a market scare or pullback could make the vxx increase quite much, probably not a bad idea to be hedged if you do not spend too much on it because on the other hand contango relentlessly erodes its value …
Cheers!
Juan
Apple sales/margins on relatively new products have been absolutely amazing, congratulations on that !
Pingback: VXX shorting model (sell signals) | The Intelligent Investor Blog
Pingback: VXX represents a VIX future expiring in exactly one month | The Intelligent Investor Blog
Hi Juan,
Great site! Quick question … any idea what causes posted dollarweights at:
http://www.ipathetn.com/us/product/VXX/#/dollarweights
to differ from calculated values. I did a similar calculation as your spreadsheet, and conclude the May 12 future should have 69.7% dollar weight on 4/26/2012 instead of 62% as published. Do these values stray from each other often, and if so, do you know any reason why or if it is cause for concern?
Hi Cameron, thanks !
It could be because they report the weights at the begining of the day and I do it at the end of the day. So one day is shifted.
The calculations I made come from their prospectus, I start with 100% of first future month VIX contracts and sell a fraction of them every day until none are left by the end of the month, just as stated in their prospectus. It’s hard to make a mistake in such a simple calculation and predict the VXX right. If it was incorrect it would have an impact and the VXX calculated price would differ, so it probably is OK.
Cheers!
Juan
I was wondering if the discrepancy has anything to do with this statement in the prospectus:
http://www.ipathetn.com/static/pdf/vix-prospectus.pdf
In addition to the transactions described above [mechanics of rolling] the weight of each index component is also adjusted every day to ensure that the change in total dollar exposure for the index is only due to the price change of each contract and not due to using a different weight for a contract trading at a higher price.
Not sure what the implication is for the relative dollar weights on each contract. The current published dollar weights seem at odds with the calculation based on rolling only.
Cheers,
Cameron
I don’t understand that sentence. I just rolled. If it had an impact then the pricing of the model and the market should differ, since that does not happen I conclude it has no major importance.
Ok I understood I think: They say that you basically do not use a different weight for a contract trading at a higher price. Which is how it has to be. So you weight based on the number of contracts exclusively and not on the futures price. That’s correct. If you start with 100 contracts every day that passes you sell the same fraction of contracts and no more or less because the futures price goes up or down. So if one day the futures are high you will sell more dollars if another day they are low you will sell less, but the amount of contracts you sell is fixed.
The apparent discrepancy comes because you cannot compare the weights directly because they show DOLLAR weights on their site but I show number of contracts. The dollar weights do not change linearly because they depend on the futures price, whereas the contract weight change in a fixed amount every day.
If you want dollar weights to be able to compare then you need to multiply the number of contracts by the futures price on the spreadsheet, that will tell you how many dollars are in each type of contract and you’ll be able to compare apples with apples.
Cheers!
Juan
This post below explains the discrepancy between dollar weights and # of contract contract weights. You can go from one to the other easily anyways:
http://investing.kuchita.com/2012/04/27/vxx-represents-a-vix-future-expiring-in-exactly-one-month/
Hi Juan,
I am having some trouble locating historic data so was hoping you might know where I can find it.
First, in your spreadsheet I see historic data for VXX-IV. Is this calculated by you or do you know of a website where this info can be downloaded (into excel)?
I’m also looking for historic data for the following:
1. S&P500 VIX Short Term Futures Index (^VXX) daily closing prices. I believe this is the index VXX is most highly correlated to.
2. Daily historic weightings of ea futures contract for ^VXX
3. UVXY-IV historical data. (I’d like to do some reasearch on this vs VXX and ^VXX.
Any idea where I can download this info?
Thanks!
Hi PC,
I downloaded all the VIX futures data since March 2004, its available on the CBOE site here: http://cfe.cboe.com/Products/historicalVIX.aspx
Every link represents a VIX future expiring on a particular month. You can download all that as text and open into excel.
You cannot download daily historic weightings of futures contracts for the VXX, you can calculate their relative values, that’s what I did. You can get the dollar weightings on a day to day basis but not the historical data.
I also do not know where you can download 1) or 3) since I have not used it.
Cheers!
Juan