An Unusual 4 ETF Portfolio…That Seems to Work

  • SumoMe

I know I repeat it a lot but the purpose of this blog is not to offer recommendations but rather to share ideas.  So here is one that I am not quite sure about but am keeping an eye on.

The FourNonCorr Portfolio

Somewhere awhile back I started looking at trying to pair non correlated – or even inversely correlated – securities in a portfolio that had the potential to outperform the overall market. What follows is what I refer to as the FourNonCorr Portfolio.  For the record I do not trade this portfolio with real money.  I am still trying to figure out if there is something to it or not.  But given that it has outperformed the S&P 500 by a factor of 3-to-1 (granted, using hypothetical results) since December of 2007, I figure it might be worth monitoring for awhile.

The portfolio consists of four ETFs:

Ticker FXE – Guggenheim CurrencyShares Euro Trust

Ticker UUP – PowerShares DB US Dollar Index Bullish

Ticker TLT – iShares Barclays 20+ Yr Treas. Bond

Ticker XIV – VelocityShares Daily Inverse VIX ST ETN

The monthly charts for each appear in Figure 1.1Figure 1 –The Four ETFs in The Four NonCorr Portfolio (Courtesy AIQ TradingExpert)

As you can see there is a lot of “zigging” by one accompanied by “zagging” for another.  No surprise that when the Euro rises the dollar falls and vice versa. Also, TLT often seems to move opposite XIV. That is essentially the purpose of these pairings.

Figure 2 displays the correlations between the four ETFs in the portfolio (using AIQ TradingExpert Matchmaker function from 8/31/2012 through 8/31/2017 using weekly data).  A reading of 1000 indicates a perfect correlation, a reading of -1000 indicates a perfectly inverse correlation.

FXE UUP TLT XIV
FXE (913) 77 (13)
UUP (913) (117) 43
TLT 77 (117) (234)
XIV (13) 43 (234)

Figure 2 – Correlations for the FourNonCorr Portfolio ETFs (Source: AIQ TradingExpert)

Clearly there is a whole lot of “not correlating much” going on.

Results

For testing purposes I used monthly total return data for each ETF from the PEP Database from Callan Associates.  The one exception is ticker XIV which did not start actual trading until December 2010.  For January 2008 through November 2010 I used index data for the index that ticker XIV tracks inversely (S&P 500 VIX SHORT-TERM FUTURES INDEX). Actual XIV ETF data is used starting in December 2010.

As a benchmark, I also tracked the cumulative total return for ticker SPY (that tracks the S&P 500 Index).

Figure 3 displays the cumulative percent gain or loss for both the FourNonCorr Portfolio and ticker SPY.3Figure 3 – Cumulative % gain/loss for The FourNonCorr Portfolio (blue) versus SPY (red); 12/31/2007-9/30/2017

Year-by-year results appear in Figure 4

  4 NonCorr SPY Diff
2008 (6.0) (37.0) 31.0
2009 26.1 26.4 (0.3)
2010 45.2 14.9 30.3
2011 (1.3) 2.1 (3.4)
2012 34.3 15.8 18.5
2013 19.3 32.2 (12.9)
2014 5.3 13.5 (8.2)
2015 0.6 1.3 (0.8)
2016 21.0 11.8 9.2
2017* 24.4 14.1 10.2

Figure 4 – Year-by-Year Results

The results by the numbers appear in Figure 5.

4NonCorr SPY
Average 12mo % +/- 17.8 11.2
Median 12mo % +/- 14.9 15.0
Std. Deviation 17.1 16.8
Ave/Std. Dev. 1.04 0.67
Worst 12mo % (11.9) (43.2)
Max. Drawdown % (17.8) (48.4)

Figure 5 – By the numbers

All told The FourNonCorr Portfolio:

*Gained +334% versus +110% for SPY since 12/31/2007

*Experienced a maximum drawdown of -17.8% versus-48.4% for SPY

Thoughts

On paper, The FourNonCorr Portfolio looks pretty decent, particularly compared to the S&P 500 Index.  But you will recall that I stated earlier that I don’t actually trade this portfolio with real money.  Why not?  A few concerns:

*Interest rates tend to move in long-term waves up and down.  How beneficial will it be to have TLT in the portfolio if and when interest rates embark on a long-term wave up?

*I don’t entirely trust ticker XIV.  Because of the way it is built it seems to have the benefit of upward bias due to contango in the VIX futures market (the opposite of ticker VXX – please Google “VXX” and/or “contango” for an actual explanation) it also holds the potential to sell off in shocking fashion.  Using the index data as I did in order to replicate hypothetical performance from Jan 2008 through Nov 2010, XIV declined a stunning -72% between the end of May 2008 and the end of November 2008. It also experienced a -60% decline in 2015-2016. Need to give some thought to adding a security that is even capable of that to a permanent portfolio.

*On the flip side, XIV has been the driving force for gains in recent years and shows a cumulative gain of +416% since 12/31/2007.  If (and when?) we ever do see a bear market and/or a significant pickup in volatility will XIV have a large negative influence on performance?  That seems to be the $64,000 question.

Summary

As a thought experiment, The FourNonCorr Portfolio shows a pretty decent track record and seems to hold some interesting promise.  As a real money, real world experience – questions remain.

Stay tuned, tinker and experiment if you wish,and don’t be too quick to “dive in.”

Jay Kaeppel

Disclaimer:  The data presented herein were obtained from various third-party sources.  Whilne I believe the data to be reliable, no representation is made as to, and no responsibility, warranty or liability is accepted for the accuracy or completeness of such information.  The information, opinions and ideas expressed herein are for informational and educational purposes only and do not constitute and should not be construed as investment advice, an advertisement or offering of investment advisory services, or an offer to sell or a solicitation to buy any security.

 

3 thoughts on “An Unusual 4 ETF Portfolio…That Seems to Work

  1. Jay,
    Interesting idea, thank you for sharing. I ran the numbers going back to 2004 using simulated data for XIV based on VIX futures prior to 2011. I also used $USD, $XAU (StockCharts data) and TLT. Keeping equal weights and starting account of 1,000 the end result is 5,800 and max DD is -33% (based on monthly closing prices). Removing $USD and $XAU and using 50% of the account split evenly between XIV and TLT gives end result of 5,600 and max DD of -34%. So I think TLT and XIV combination is interesting but the $USD and $XAU cancel each other, hence may not add much value.

    Another option is to use TMF instead of TLT. Even though XIV is unleveraged but it acts approximately as 3x SPY. Therefore, XIV would be best used with TMF and perhaps we could use ZIV, its less volatile sister, with TLT. I used TMF data since the inception in 2009 and before that I used a quick and dirty simulation of triple TLT. The TMF and XIV combination final equity was 12,250 with max DD of -23% since 2004. Keep in mind this needs to be repeated with a more accurate analysis, every month I just took an average of % gains for the two symbols, which is not exactly the most accurate way to do it but it looks promising.

    1. Interesting ideas, I appreciate you sharing your thoughts. Like I said in the article, there seems to “something there”, but do not want to be too quick to “take the leap”. XIV – and also TMF as you mentioned – can add alot of potential to a properly constructed portfolio. They can also add a whole lot of volatility and risk. Need to get a good handle on the risk side before getting overly excited about the potential profit side. Jay

Comments are closed.