Profiting from Market Volatility with the “Anti-VIX” ETN ZIV

  • The “Anti-VIX” ETN ZIV is designed to increase in value when the volatility of the S&P 500 decreases, as measured by the prices of VIX futures contracts.
  • The model buys ZIV only during up-markets when the VIX > 17 and rising, otherwise during up-markets it buys either QLD or DDM, or IEF when upmarket conditions are absent. 
  • A backtest of the model from Jan-2011 to Jul-2017 produced a high 60% annualized return with a maximum drawdown of -16% with only 41realized trades.

The VelocityShares Daily Inverse VIX Medium-Term ETN (ZIV) is linked to the inverse (-1x) of the daily performance of the S&P 500 VIX Mid-Term Futures Index less fees and expenses. It has a medium-term time horizon of about 5-month.

The author of Backtesting VelocityShares’ ZIV inverse volatility ETN to 2004 was “surprised at how volatile, and how low this hypothetical ZIV went during the recent bear market—losing 80% of its value from 2007 to 2008.” He concludes that “ZIV and XIV appear to be bull market only instruments and not suitable for buy and hold.” 

Our own backtest to 2007 of ProShares Short VIX Short-Term Futures (SVXY) described in Profiting from Market Volatility with the “Anti-VIX” ETF SVXY indicated that a hypothetical SVXY would have lost about 90% of its value from 2007 to 2008, with subsequent further high drawdowns of 75% in 2011, 56% in 2015, and 68% in 2016. It follows that Anti-VIX instruments should only be held during up-stock market periods and are not useable as long-term investments.

The inception date of ZIV was 11/30/2010, from which date data is available at Portfolio 123. There is no extended data for ZIV on P123 prior to inception. For QLD and DDM the inception date was 6/19/2006. The start date of the model is 1/2/2011.

Performance of ZIV

Figure-1 shows the performance of ZIV from Jan-2011 to end of Jul-2017. Buy-and-hold ZIV produced an annualized return of 29% with a -41% maximum drawdown in Oct-2011, and further large drawdowns followed; 30% in 2015, and 35% in 2016. It is evident that ZIV should only be held during up-market periods to avoid these losses.

width="640"/(click to enlarge)

Performance of a ZIV–Cash Model

Figure-2 shows the performance of ZIV when timed with the iM-Composite Market Timer and the VIX. The simulated annualized return from Jan-2011 to end of Jul-2017 was 32% with a maximum drawdown of -22%. Although the return is similar to buy-and-hold, the drawdowns are much better for the timed model.

width="640"/(click to enlarge)

This model buys ZIV only when the Composite Market Timer Score is greater or equal to 50; i.e. when up-stock market conditions prevail.

It sells ZIV

  1. If ZIV has lost more than 5% from a recent high after it was bought,
  2. or if the VIX becomes less than 17 and the 3-day moving average of the VIX is below the 20-day moving average at the same time,
  3. or if the 3-day moving average of the VIX becomes less than 13.5.

The model is rebalanced weekly on the first trading day of the week using closing prices, and transaction costs were assumed as 0.1% of the trade amounts to account for commission and slippage.

There were 14 periods when the model was in ZIV, about 65% of the backtest period; otherwise it was in cash.

 

iM VIX-Timer with ZIV – 2011

Buy- and Sell-Rules for the iM VIX-Timer with ZIV

This model’s rules include stop-loss rules for all ETPs except for IEF.

The model buys ZIV only during up-markets when the VIX is higher than 17 and rising; otherwise during up-markets when the VIX is lower than 17 it buys either QLD or DDM, or IEF when upmarket conditions are absent.

It sells ZIV when

  1. ZIV has lost more than 5% from a recent high after it was bought,
  2. or when the 3-day moving average of the VIX becomes less than 13.5.

It sells DDM or QLD when

  1. DDM or QLD have lost more than 5% from a recent high after they were bought,
  2. or when down-stock market conditions exist, as indicated by Composite Market Timer Score being less than 50, and when up-market conditions do not exist based on analysts’ Current Year Consensus Estimated Earnings for the S&P500 stocks.

It sells IEF when the Composite Market Timer Score is greater or equal to 50; i.e. when up-stock market conditions prevail.

Performance of the iM VIX-Timer with ZIV

Performance of the ZIV–Cash model with a 32% annualized return and a reasonable maximum drawdown of -22% is very satisfactory. However, a much higher return with much lower drawdown can be achieved by switching between ZIV, and either QLD or DDM, or IEF.

The model uses a simple ranking system which assumes trading ETPs, rather than investing for longer periods. The one factor system is based on the price changes over a short period. The idea being that ETPs which have experienced a decline over a short period will bounce back, reverting and doing better than ETPs which have not declined in this way.

Simulated performance from Jan-2011 to end of Jul-2017 is shown in Figure-3 and Table-1. The annualized return is about 60% and the maximum drawdown is a reasonable -16%. The model is rebalanced weekly on the first trading day of the week using closing prices, and transaction costs were assumed as 0.1% of the trade amounts to account for commission and slippage.

width="640"/(click to enlarge)


width="640"/(click to enlarge)

The trading summary is in Table-2. There were only 41 realized trades. The biggest loss of any trade was 4.8%, 81% of all trades were winners and the average holding period was about 55 days. The longest holding period was 33 weeks and 4 trades had the shortest holding period of one week (not shown in the table).

width="640"/(click to enlarge)

The risk measurements are in Table-3.

width="640"/(click to enlarge)

Following the model

At iMarketSignals we will provide weekly signals of the iM-VIX Timer with ZIV (starting before end August 2017 and a Gold subscription is required).  Currently the model holds DDM, which it bought on 3/27/2017.

A similar model, the iM-VIX Timer with SVXY, can be followed by subscribing to it at Portfolio 123 when it will be opened for subscription at the beginning of Nov-2017.

Disclaimer

Note: All performance results are hypothetical and the result of backtesting over the period 2011 to 2017. No claim is made about future performance. Backtesting involves optimizing parameters by looking at past data. Even if parameter values may be optimal going forward, future returns may generally not be as high as past returns.

 

 

 

 

Posted in blogs, featured
53 comments on “Profiting from Market Volatility with the “Anti-VIX” ETN ZIV
  1. dscott says:

    Hi,I am curious to know what the returns/drawdowns look like if IEF was replaced with SDS.
    Thanks.

    • geovrba says:

      Replacing IEF with SDS produces much higher drawdown.

      Period 01/01/11 – 07/31/17

      Benchmark SPDR S&P 500 ETF Trust
      Universe All ETFs

      Quick Stats as of 7/31/2017
      Total Return 2,000.70%
      Benchmark Return 124.14%
      Active Return 1,876.57%
      Annualized Return 58.85%
      Annual Turnover 635.02%
      Max Drawdown -36.89%
      Benchmark Max Drawdown -18.61%
      Overall Winners (33/45) 73.33%

  2. Viking says:

    Hello Georg
    Quite an original combination and cautious risk management rules. Any idea when you intend to allow subscriptions?

  3. RV says:

    Hi Georg,

    Another fantastic system from iMarketSignals making the Gold subscription now worth even more. A big thanks. 60% CAGR with such a low max. drawdown of 16% is just fantastic.

    I am trying to understand the logic and mechanics and am able to broadly understand why IEF would be an escape in down market conditions and why XIV, QLD, and DDM would be used to get juicier returns during up markets. I am struggling with the selection of ZIV though, and hoping you could help throw some light on the logic for timing of when to enter ZIV.

    At the tail end of a bull market, would the VIX NOT start from low levels of under 12 and then cross above 17 first before rising to higher levels. Why would the system buy ZIV at VIX level of 17 and rising? The “and rising” part is what I am trying to understand.

    Thanks again for all you do.

    Regards,

    RV

    • geovrba says:

      Basically we want to buy ZIV during up-markets only and when the VIX is rising and already quite high (>17), because any pull-back of the market is likely to be only a correction and the VIX will decline again to lower than 17.

      We sell when the 3-day moving average of the VIX becomes less than 13.5.
      There is also a stop-loss provision in case we were wrong and the correction is really the start of a bear-market.

      You can read many comments on the sister model with SVXY which was published at Seeking Alpha.
      https://seekingalpha.com/article/4095356-profiting-market-volatility-anti-vix-etf-svxy

      • RV says:

        Hi Georg,

        Thanks for providing some color to this concept of buying ZIV at Vix>17 in an up market.

        Can you please tell us:

        1. How many times in the past 6 years ZIV got stopped out and the trade lost money and how many times ZIV made money?

        2. Only for the ZIV purchase, would a modified rule do better than the existing rule of “buy ZIV when Vix>17 and rising”. The modified rule could be “Buy ZIV in an up market, when Vix has risen above 17 and now crossed below 16.99”? All the other rules for DDM, IEF as well as sell rules for ZIV will remain the same. If you could test or think about it, comment on this “modified” rule and/or guide us, that would be great.

        Thanks again. Much appreciate your creativity and the great service, you folks at iMarketSignals provide.

        RV

        • geovrba says:

          Open ……………. Close ……………. Days ……… Pct … Note
          11/07/2016 …… 11/21/2016 …… 14 …… 3.00% …… sma(3,0,$VIX)<13.5
          07/06/2015 ...... 07/20/2015 ...... 14 ...... 8.00% ...... sma(3,0,$VIX)<13.5
          10/13/2014 ...... 11/17/2014 ...... 35 ...... 15.20% ...... sma(3,0,$VIX)<13.5
          02/03/2014 ...... 03/17/2014 ...... 42 ...... 6.60% ...... Stop-Loss
          06/17/2013 ...... 07/22/2013 ...... 35 ...... 8.50% ...... sma(3,0,$VIX)<13.5
          06/04/2012 ...... 01/22/2013 ...... 232 ...... 110.50% ...... sma(3,0,$VIX)<13.5
          04/16/2012 ...... 05/29/2012 ...... 43 ...... -4.80% ...... Stop-Loss
          03/14/2011 ...... 06/20/2011 ...... 98 ...... 8.80% ...... Stop-Loss
          01/03/2011 ...... 03/07/2011 ...... 63 ...... 10.20% ...... Stop-Loss

          • RV says:

            Amazing. Never would’ve guessed there was only 1 instance where the ZIV trade made a loss out of 9 trades. Even that 1 loss was very small of 4.8%.

            Thanks Georg.

          • RV says:

            Hi Georg,

            Would you be willing to post the remainder of the 32 trades with DDM, QLD, & IEF? On some strategies you’ve given us all the trades and it’s helpful to understand the dynamics.

            I noticed that there is no negative year for this strategy from 2011 to 2017, but the max. drawdown is 15.9%. Was wondering when the max. drawdown occurred and also how the other trades did.

            Thanks a bunch!

            RV

        • geovrba says:

          Can’t do No.2.

          • geovrba says:

            Max D/D of about -16% occurred in 2011 and 2012

            Symbol ……………. Open ……………. Close ……… Days ……… Pct
            IEF …… 03/06/2017 …… 03/27/2017 …… 21 …… 0.90%
            QLD …… 12/05/2016 …… 03/06/2017 …… 91 …… 25.50%
            IEF …… 11/21/2016 …… 12/05/2016 …… 14 …… -0.80%
            ZIV …… 11/07/2016 …… 11/21/2016 …… 14 …… 3.00%
            IEF …… 09/26/2016 …… 11/07/2016 …… 42 …… -1.90%
            QLD …… 09/12/2016 …… 09/26/2016 …… 14 …… 2.00%
            IEF …… 09/06/2016 …… 09/12/2016 …… 6 …… -1.00%
            DDM …… 08/01/2016 …… 09/06/2016 …… 36 …… 1.90%
            IEF …… 07/11/2016 …… 08/01/2016 …… 21 …… -0.80%
            QLD …… 06/27/2016 …… 07/11/2016 …… 14 …… 17.10%
            IEF …… 11/30/2015 …… 06/27/2016 …… 210 …… 6.20%
            QLD …… 11/16/2015 …… 11/30/2015 …… 14 …… 4.40%
            IEF …… 09/14/2015 …… 11/16/2015 …… 63 …… -0.90%
            QLD …… 08/24/2015 …… 09/14/2015 …… 21 …… 12.70%
            IEF …… 07/20/2015 …… 08/24/2015 …… 35 …… 2.60%
            ZIV …… 07/06/2015 …… 07/20/2015 …… 14 …… 8.00%
            IEF …… 06/08/2015 …… 07/06/2015 …… 28 …… 0.60%
            QLD …… 05/04/2015 …… 06/08/2015 …… 35 …… -2.30%
            IEF …… 12/22/2014 …… 05/04/2015 …… 133 …… 0.10%
            QLD …… 12/08/2014 …… 12/22/2014 …… 14 …… 0.20%
            IEF …… 11/17/2014 …… 12/08/2014 …… 21 …… 0.20%
            ZIV …… 10/13/2014 …… 11/17/2014 …… 35 …… 15.20%
            DDM …… 08/04/2014 …… 10/13/2014 …… 70 …… -2.80%
            QLD …… 04/07/2014 …… 08/04/2014 …… 119 …… 23.10%
            ZIV …… 02/03/2014 …… 04/07/2014 …… 63 …… 9.60%
            QLD …… 07/22/2013 …… 02/03/2014 …… 196 …… 26.60%
            ZIV …… 06/17/2013 …… 07/22/2013 …… 35 …… 8.50%
            QLD …… 04/08/2013 …… 06/17/2013 …… 70 …… 13.40%
            IEF …… 04/01/2013 …… 04/08/2013 …… 7 …… 0.40%
            QLD …… 01/22/2013 …… 04/01/2013 …… 69 …… 3.50%
            ZIV …… 06/04/2012 …… 01/22/2013 …… 232 …… 110.50%
            IEF …… 05/29/2012 …… 06/04/2012 …… 6 …… 1.30%
            ZIV …… 04/16/2012 …… 05/29/2012 …… 43 …… -4.80%
            DDM …… 01/30/2012 …… 04/16/2012 …… 77 …… 4.60%
            IEF …… 08/29/2011 …… 01/30/2012 …… 154 …… 3.10%
            QLD …… 08/08/2011 …… 08/29/2011 …… 21 …… 14.40%
            IEF …… 07/05/2011 …… 08/08/2011 …… 34 …… 5.50%
            QLD …… 06/20/2011 …… 07/05/2011 …… 15 …… 15.40%
            ZIV …… 03/14/2011 …… 06/20/2011 …… 98 …… 8.80%
            IEF …… 03/07/2011 …… 03/14/2011 …… 7 …… 1.00%
            ZIV …… 01/03/2011 …… 03/07/2011 …… 63 …… 10.20%

  4. numejak says:

    Hello Georg! With back-testing data only available from 2011, how do you gauge the resilience of this model to recessions, the likes of the last two in 2001 and 2008, and so-called Blackswan events? And second, It would seem to me that during a recession linked bear market that holding SH instead of IEF would be more profitable. Have you or can you analyze this scenario? Thank you for your wonderful work. Jake

    • geovrba says:

      That is why we also published a backtest with SVXY from 2007 onward. The SVXY model switched to IEF during the 2008/9 down-market. In IEF from7/2/2007 to 3/23/2009.

      Black Swan events cannot be captured by any model.

      Also SH did not improve performance over IEF.

  5. RV says:

    Georg,

    Thanks for providing all the 41 trades. Truly an amazing and creative use of your timing models (which you folks have developed with so much hard work over the years). This model is just fantastic. The Gold subscription monthly prices should be raised by the addition of this model :-)

    If the 16% Max. Drawdown occurred in 2011 then it probably occurred when the model was holding IEF, right? The amazing thing is that the max. drawdown probably occurred during a trade, because the losing trades have such small losses.

    Could you please tell us, what instrument was the model holding in 2011 0r 2012 i.e., QLD or IEF when the 16% max. drawdown occurred? How long did it take for the model to recover from this drawdown? It appears it recovered very quickly, in few weeks or maybe just a couple of months at the most.

    Thanks again for a superb addition to the portfolio of already great models like Combo5, iM-5 ETF Trader, etc. for Gold subscribers.

    RV

  6. RV says:

    Hi Georg,

    Can you please share what months and year the max. drawdown of approx. 16% occurred? What was the model holding at that time? QLD, DDM, ZIV, or IEF?

    Thanks

  7. RV says:

    Hi Georg,

    When will we (Gold subscribers) start getting signals for this model/strategy?

    RV

  8. 13qe says:

    What are the returns and drawdown if leveraged ETFs are not used when the model is out of ZIV and in stocks? Thanks!

    • geovrba says:

      Using the unleveraged ETFs DIA and QQQ instead of DDM and QLD:

      Period 01/02/11 – 08/21/17

      Benchmark SPDR S&P 500 ETF Trust

      Quick Stats as of 8/19/2017
      Total Return 664.22%
      Benchmark Return 120.45%
      Active Return 543.77%
      Annualized Return 35.91%
      Annual Turnover 494.07%
      Max Drawdown -15.20%
      Benchmark Max Drawdown -18.61%
      Overall Winners (25/35) 71.43%
      Sharpe Ratio 2.20

  9. RV says:

    Have performance updates for this strategy started? Where can I find them?

    RV

  10. randyfloyd says:

    Hi, i just would like to clarify please…

    the stop-losses are built into your system, I don’t need to set them myself in my own accounts. is that correct?

    thanks

  11. bud787 says:

    Best trading model so far…pls explain how the two largest “draw downs” of ZIV (-15.2% and -15.7%) are possible when the stop loss rules sell when price declines by 5%…thanks.

    • geovrba says:

      The model only trades on the first trading day of the week and not daily. Thus a price decline of 5% could occur early in the preceding week and the model would only react to it until a few days later when it trades. By that time the decline can be a lot more.

  12. numejak says:

    Where or when can we find the model’s out of sample performance YTD weekly updates?

    • RV says:

      Georg – I’ve also requested this info and no one seems to be responding. Can you folks please tell us what area of the website are you posting performance of the ZIV timing model? Thanks.

      • geovrba says:

        Performance data is now available as of 9/26/2017.

        iM-VIX Timer with ZIV:
        The model’s out of sample performance YTD is 42.4%, and since inception 42.4%. Over the same period the benchmark SPY performance was 12.9% and 12.9% respectively. Over the previous week the market value of the iM-VIX Timer with ZIV gained -0.40% at a time when SPY gained -0.32%. A starting capital of $100,000 at inception on 12/30/2016 would have grown to $142,412 which includes $293 cash and excludes $578 spent on fees and slippage.

  13. Goldbeard says:

    What would be an optimum combo of Anti Vix Ziv & 5ETF models? Could you provide a percentage split between the two models that provided the optimal risk-return profile? And then provide the performance numbers for that combo? Thank you.

    • geovrba says:

      50:50 combo of Anti Vix Ziv & 5ETF models.

      Period 12/29/06 – 09/18/17

      Benchmark SPDR S&P 500 ETF Trust

      Quick Stats as of 9/18/2017
      Total Return 7,278.34%
      Benchmark Return 120.22%
      Active Return 7,158.12%
      Annualized Return 49.36%
      Max Drawdown -14.17%
      Benchmark Max Drawdown -55.19%
      Sharpe Ratio 2.69
      Sortino Ratio 4.24
      Correlation with SPDR S&P 500 ETF Trust 0.32

  14. Goldbeard says:

    Very good, thank you!

  15. RV says:

    Hi Georg,

    It’s pretty clear when the model invests in ZIV or IEF.

    How does the model decide when to invest in DDM vs. QLD?

    For example, currently why is this model invested in DDM and NOT QLD?

    Thanks

    • geovrba says:

      The model uses a ranking system: 1-Factor Ranker.
      The ranking system’s approach assumes trading ETFs, rather than investing for longer periods.
      The one factor system is based on the price changes over a short period. The idea being that ETFs which have experienced a decline over a short period will bounce back, reverting and doing better than ETFs which have not declined in this way.

      • RV says:

        Georg –

        Thanks for explaining the 1-factor ranking for ETFs.

        I also have a question about the 5% pullback exit trigger rule.

        If the selection ETF (currently DDM) falls 5% from its peak within a week, then would we exit at the end of the week even if it recovers fully during the week? Or is the 5% based on closing the week down at least 5%?

        Thanks

  16. chrissmit says:

    Georg

    In you post of Aug 9th the following trades are listed:
    QLD …… 06/27/2016 …… 07/11/2016 …… 14 …… 17.10%
    QLD …… 08/24/2015 …… 09/14/2015 …… 21 …… 12.70%

    In both cases the VIX was well above 17 on the trades’ start dates. Should those trades not have been with ZIV instead of QLD?

    Chris

    • geovrba says:

      The model buys ZIV only during up-markets when the VIX is higher than 17 and rising.

      The model takes its data from Friday’s closing prices, 3 days before the buy dates. The Up-market conditions may not have been met on those dates. Also the ranking system comes into play to decide what to buy.

  17. TDCARLSON says:

    Georg,

    What would the results look like for CAGR and DD if you used the same rules, but used XIV in place of ZIV?

    Tom C

  18. cleanmaster94 says:

    Hi, you wrote this in the comment section on your seeking alpha article;

    “OK, so lets see how this model performs with RSP or SPY instead of SVXY, and the unleveraged ETFs DIA or QQQ instead of DDM and QLD. The annualized return would have been 24% with a max D/D of -12.5%. Total return was about 8x that of SPY, and Alpha (annualized) =21.68%. For the model, as presented, Alpha =65.43%. That is a lot more Alpha coming from the leverage and SVXY.

    Period 01/02/07 – 07/31/17
    Benchmark SPDR S&P 500 ETF Trust

    Quick Stats as of 7/31/2017
    Total Return 890.74%
    Benchmark Return 116.55%
    Active Return 774.19%
    Annualized Return 24.21%
    Annual Turnover 507.13%
    Max Drawdown -12.54%
    Benchmark Max Drawdown -55.19%
    Overall Winners (40/56) 71.43%
    Sharpe Ratio 1.99”

    Is there anyway you can backtest it further using SPY instead of ZIV/UVXY and QQQ/DIA instead of the leveraged ETFs? 24% return with 12 % drawdown is ntohing to sneeze at; would like to know how the model did during the dot com bear market as well. Thanks

    • geovrba says:

      iM VIX Timer with [SPY | (DIA|QQQ)] | IEF

      From Jan-2000:

      Benchmark S&P 500 (SPY)

      Quick Stats as of 10/29/2017

      Total Return 2,608.84%
      Benchmark Return 144.59%
      Active Return 2,464.25%
      Annualized Return 20.33%
      Annual Turnover 607.89%
      Max Drawdown -13.60%
      Benchmark Max Drawdown -55.19%
      Overall Winners (72/110) 65.45%
      Sharpe Ratio 1.64
      Correlation with S&P 500 (SPY) 0.22

  19. RV says:

    Georg –

    Thanks for explaining the 1-factor ranking for ETFs. This is a fab system. I need your help in understanding this system in further detail.

    In that context, I have a question about the 5% pullback exit trigger rule.

    If the selection ETF (currently DDM) falls 5% from its peak intra-week, then would the system exit at the end of the week, even if it recovers fully by the end of the week? Or is the 5% based on closing weekly prices for ETFs? In other words, DDM would have to close down at least 5%?

    Thanks for all you do for us, the subscribers.

    • geovrba says:

      The system checks the percentage difference from highest closing price on any day since the position was started to the closing price on the last day of a week (usually Friday). If the percentage is smaller than the one specified then the position gets sold on the first trading day after Friday.

  20. numejak says:

    Georg –

    I am looking around the web, can you tell me where I could find “analysts’ Current Year Consensus Estimated Earnings for the S&P500 stocks?” I am looking back at your past trades on this system trying to understand it all. Thanks!

  21. vman says:

    Hey Georg and Anton
    I read in several places that ZIV has the potential to become almost worthless during a BlackSwan type of event…..hmmm. Any thought on this? Do. you suggest always keeping a stop loss under ZIV?

    Thanks
    Jimmy V

    • geovrba says:

      From model description:

      This model’s rules include stop-loss rules for all ETPs except for IEF.

      It sells ZIV when
      ZIV has lost more than 5% from a recent high after it was bought,
      or when the 3-day moving average of the VIX becomes less than 13.5.

Leave a Reply

With reference to Section 202(a)(11)(D) of the Investment Advisers Act: We are Engineers and not Investment Advisers, read more ...
By the mere act of reading this page and navigating this site you acknowledge, agree to, and abide by the Terms of Use / Disclaimer
Share This