CHRISTMAS PROMOTION
LTOptions at a 33% discount during the Year End Holidays.
Tell me More

Monday, January 19, 2015

ETF Rotation Systems to beat the Market - SPY + IWM + EEM + EFA + TLT + TLH + DBC + GLD + ICF + RWX

This is the fifth and last chapter of the ETF Rotation Systems to beat the Market series. If you haven't done so, you may want to read the previous chapters:

- ETF Rotation Systems to beat the Market - American Equities
- ETF Rotation Systems to beat the Market - Global Equities
- ETF Rotation Systems to beat the Market - American Equities + TLT + GLD + IYR + EEM
- ETF Rotation Systems to beat the Market - SPY + EFA + IEF + GLD + ICF

Today we are going to take the concept of diversification a little farther. We will exploit the idea of the fourth chapter materialized in the SPY + EFA + IEF + GLD + ICF portfolio. That is, we will create a portfolio that uses the same asset classes: Equity (US), Equity (International), Bonds, Real Estate and Commodities but this time we will choose two symbols for each category instead of just one.

These are the ETFs of choice:
US Equity:
SPY - SPDR S&P 500 ETF Trust
IWM - iShares Russell 2000 Index (ETF)

International Equity:
EFA - iShares MSCI EAFE Index Fund
EEM - iShares MSCI Emerging Markets Index

Bonds:
TLT - iShares Barclays 20+ Yr Treasury Bond
TLH - iShares Lehman 10-20 Yr Treasury Bond

Commodities:
DBC - PowerShares DB Com Index Tracking Fund
GLD - SPDR Gold Trust

Real Estate:
ICF - iShares Cohen & Steers Realty Maj.
RWX- SPDR Dow Jones International Real Estate

Let's analyze the system when it invests in the two best performers each month. 50% of the capital goes to each one of the two best ETFs. The formula to evaluate the ETFs was discussed in the Introduction of the series. The system is always 100% invested. No cash position.

Here are the results:
(Click on image to enlarge)
CAGR: 18.8% vs 9.5% SPY
Volatility: 15.3% vs 19.5% SPY
Worst draw-down: -13.8% vs -55.2% SPY
Sharpe ratio: 1.06 vs 0.44 SPY

This is a great result. 18.8% Compound Annual Growth Rate for a 12 year back-test period is nothing to sniff at. Let alone with a maximum draw-down of only 13.8%.

Now, let's do another test but instead of investing in the 2 best ETFs each month, the system will invest in the best 3 ETFs each month. 33% of the capital will go to each one of the selected ETFs. Again the portfolio remains 100% invested at all times. No cash position. In a pool of 10 instruments to select from, choosing 3 shouldn't cause a huge loss of performance as the third best ETFs should still be a pretty good one. I would also expect less volatility, more stability in the equity curve over time. Let's see:

(Click on image to enlarge)
CAGR: 19.5% vs 9.5% SPY
Volatility: 13.5% vs 19.5% SPY
Worst draw-down: -16.5% vs -55.2% SPY
Sharpe ratio: 1.22 vs 0.44 SPY

Nice improvement in terms of Compound Annual Growth Rate. The volatility of the portfolio also decreased but the worst draw-down was a bit higher. Still, great ETF Rotation System with an outstanding Sharpe Ratio of 1.22. I feel pretty happy with this portfolio and would trade it with my own money (if only I had more).

Surprisingly adding a cash filter rule to this system doesn't provide improvements. The filter rule, as a reminder, is the rule by which, if one of the selected instruments to be invested in is trading below its 10 month moving average, that portion of the capital sits in cash or very short term bonds (SHY). No improvements were obtained. The addition of the filter rule always decreases the returns a little bit and the goal is to see a significant reduction in the worst draw-down and the volatility of the portfolio. While it did decrease the returns, the rule didn't provide a reduction in the volatility nor the worst draw-down numbers to justify its addition.


My final take on ETF Rotation Systems and some things to consider
ETF Rotation systems have been quiet a discovery for me. What really impresses me is the fact that your particular choice of ETFs is not as important for outperforming the markets as it is to diversify your selection using different asset classes. This says a lot about how powerful the strategy is. Positive long term results will not depend on your ability to magically pick the best instruments of the future, which is great. During my hundreds of tests, portfolios seemed to always easily beat the market as long as several asset classes were used. Any portfolio of ETFs that covered US Equity, International Equity, Bonds, Real Estate and Commodities outperformed the S&P500 regardless of the specific ETFs I chose. This reinforces my confidence in the strategy. The idea in principle is solid and the particular choice of ETFs is not as determinant for your long term success as long as you are diversified across multiple asset classes.

One thing to keep in mind when designing your portfolio is the management fees of your ETFs of choice. ETFs are usually cheaper than Mutual Funds by far, but still something to look at. Obviously the smaller the management fee the better. Also, look at possible dividends. If your ETFs of choice are dividend payers then that's better. By the way all the test results shown in the series include dividends.

If you have a small portfolio, make sure trading commissions costs are reasonably controlled. For example if your ETF Rotation system chooses the 2 best ETFs each month, assume that you will make 4 transactions every month. Two in order to sell your current ETFs plus two in order to buy the best two ETFs for the next cycle. That is a total of 48 transactions per year. You may end up trading much less than that. You may find your system riding an ETF for an entire year without having to sell it to purchase another one, but when designing your system you must take into account the worst case scenario and the maximum possible number of trades per year. In this case, 4 trades per month means 48 per year. A broker charging only $5 per transaction would cost you $240. If you have a small $10,000 portfolio, that trading cost represents a 2.40% drag in the performance, which is unacceptable. Either find a broker with cheap commissions for trading ETFs or wait until you have a larger amount of money in order to implement an ETF Rotation system where your worst trading costs per year will be less than 1% of the portfolio balance.

ETFs rotation systems have great potential for outperforming a simply "long the market strategy" in the long run and to do it with less volatility and smaller draw-downs. There is plenty of evidence to support that statement.

If you want to expand your knowledge on the area of Rotation Systems and consult deeper research encompassing multiple decades of market history (even more than a century), I suggest you read A Quantitative Approach to Tactical Asset Allocation. It's an eye opener.

This is the end of the ETF Rotation Systems to beat the Market series. I hope you had as much fun reading it as I had writing it and I hope it is useful to someone out there.

Disclaimer: Data and back-testing via ETFReplay.com

For all the details on how the ranking works, read: ETF Rotation - Free Ranking evaluation tool


Interested in this Series?
Here are all the chapters:
1. ETF Rotation Systems to beat the Market - American Equities
2. ETF Rotation Systems to beat the Market - Global Equities
3. ETF Rotation Systems to beat the Market - American Equities + TLT + GLD + IYR + EEM
4. ETF Rotation Systems to beat the Market - SPY + EFA + IEF + GLD + ICF
5. ETF Rotation Systems to beat the Market - SPY + IWM + EEM + EFA + TLT + TLH + DBC + GLD + ICF + RWX


Go to the bottom of this page in order to see the Legal Stuff

28 comments:

  1. This is really good stuff. I have never heard of this before until you introduced it in these series. I hope you consider demoing this strategy on your website as it meets the criteria for a lazy trader.

    ReplyDelete
    Replies
    1. I have certainly considered demoing it as I also think it meets the criteria of a lazy trader. It matches the mission and goals of this site. I don't have the capital to do it, I would like to apply at least $50,000. I have really thought about it but I feel like my personal time is already so tight. We'll see. I may stop tracking my dividend investments and track this instead. I can also track it with paper money just as an exercise for everybody. I'll think of something but it definitely deserves its space. No promises so far, but definitely thinking about doing something.

      Regards,
      LT

      Delete
    2. Thanks LT. As you know, I am primarily a credit spread trader but I need to diversify my trading strategies into something that is conservative, consistently profitable and low maintenance. I believe this ETF rotation system fits that criteria nicely. I will be doing my own research into this strategy this year. After my research and backtesting, I will use real money to test it out. Thanks again for letting us know about this very interesting strategy.

      Delete
    3. No problem Jonathan. I do think that when you have a significant amount of capital, whatever that means to you, you should diversify into other strategies and not be a one trick pony. Hence my approaches to passive investing & automated Forex trading. ETF Rotations in particular are a great vehicle because it is incredibly scalable and I wouldn't be afraid to put a good chunk of money there.
      Good luck on your research and let me know how it all goes.
      LT

      Delete
  2. I certainly hope you will continue to track the dividend investments here. Perhaps if time permits you can do a $50k paper trade version & the live dividend investments. Assuming of course that time permits. Thanks for all the great knowledge you are sharing and opinions.

    ReplyDelete
    Replies
    1. Thanks Rick,
      I have thought of doing what you suggest. Let's see how I accommodate the space to clearly reflect the results of all the strategies without too much clutter. I mentioned the possibility of not tracking the dividend investments because there will be life changing events this year in my life (Move to the US). I won't be allowed to work for a few months possibly 6, but maybe a year. For that reason I am accumulating money but will inevitably have to refrain from investing this year as I will need liquidity. I imagined, that tracking investments in an account with zero activity the entire year offers no value. At least not until I resume activity there.
      What do you think?
      LT

      Delete
  3. Great articles, thanks very much for posting. Regards,

    ReplyDelete
    Replies
    1. Thanks Chris,
      Glad you liked it. Doing my best over here to get us all out of poverty.
      LT

      Delete
  4. This is one of the reasons I appreciate sites like yours LT; "curation". There is so much noise regarding self directed investing, it's great to find a place to pick up new, "non-gimmick" related information. Thanks for your efforts in providing this blog.

    Take care LT

    ReplyDelete
    Replies
    1. Thanks frankblack5.
      Agreed. Too much hype and too many scams out there.
      I'm always looking to finding new no BS, valuable sites. So, if you see anything interesting out there, I'd appreciate your letting me know.
      Thanks for dropping by!
      LT

      Delete
  5. Great work. What would be your bottom line, for portfolio with enough money, where commissions are not an issue?
    1. What is the final list of ETF's?
    2. How many positions to hold?
    3. Special rules for entry/exit?
    4. Shorting under-performing ETF's (or buying inverse ones)?
    Thanks.

    ReplyDelete
    Replies
    1. Thanks for the question.
      I don't trade ETF Rotation systems. At least not right not. I would say a portfolio of 20K at least where my projected commissions per year are less than 1% of the portfolio. That would be a start.

      1. No final list of ETFs. I shared 5 simple systems,it's up to the reader to come up with his own.
      2. Hold 2 or 3 positions. Never 1.
      3. Special rules for entry and exits? No. Just the mathematical formula discussed in chapter 1.
      4. No shorting nor buying inverse ETFs

      Cheers,
      LT

      Delete
    2. I didn't really investigate this topic thoroughly, however I would like to refer to two papers:
      1. Profitable Momentum Trading Strategies for Individual Investors by Bryan Foltice and Thomas Langer.
      2. Are Momentum Strategies Profitable? Evidence from Singapore by Vikash Ramiaha, Tony Naughtonb and Madhu Veeraraghavanc.
      Although these papers deal with stocks and not ETFs, they address the issues I mentioned (and the ones I didn't mention) in my previous post, such as: number of positions, trade frequency, portfolio size, volatility, commissions, shorting. You can google them online for free.
      My ideas from the papers (but not only):
      1. Portfolio size of 5 to 8 - discussed in paper 1.
      2. No need to include volatility in the formula to pick the ETFs. If you use not highly correlated ETFs to build the portfolio (with at least 5 ETFs), the volatility takes care of itself (just a feeling, I don't have proof of that). Also, from paper 1, trading frequency addresses volatility.
      3. Could be useful to include shorting (or buying inverse). Relevant data is shown in paper 2 (dollar neutral approach gives higher returns).
      Would be nice to hear your thoughts on that.
      Regards.

      Delete
  6. Great articles! Couple of questions:

    1. Do you use a Stop Loss on any of the positions that you are invested in for the month?
    2. Do you move a position you are invested in mid month if it falls below it's month moving average during the month?
    3. Any benefit that you have heard of to evaluating on a more frequent basis?

    Thanks!

    Ryan

    ReplyDelete
    Replies
    1. Great questions Ryan,

      1. Stop Losses only tend to degrade the performance of ETF Rotation systems. Keep in mind that even if SPY has corrected more than 50% in the past, it doesn't do it in a single month. The system would be out of it much earlier. Also, the system invests in 3 ETFs at the same time. So, if one of them falls 10%, it would only be a 3.33% fall for the overall portfolio. Plus the portfolio is invested in non-correlated assets, so chances are, some other ETF will mitigate that. The worst month in the back test was -8.14% which happened on April 2004 while invested in EFA, ICF and TLT. No single month was worse than -8.14%.

      2. No. Positions are held for an entire month.

      3. I haven't heard of benefits. I know one drawback is that you will spend more money in commissions. Unfortunately ETFReplay doesn't allow to go below one month cycles. At least as of this writing.

      Cheers,
      LT

      Delete
  7. I would like to know how you do your back testing. Is there a service that does that for a fee?
    -Jay Alagia

    ReplyDelete
    Replies
    1. Hi Ray,
      I did it all through etfreplay.com which cost $35 per month.
      Cheers,
      LT

      Delete
  8. Thanks, LT. More questions:: Do you have year by year comparison of the backtested strategy with SPY?
    I want to know the odds of it winning in any year and odds of it winning any 3, 5 and 10 years.Also have you considered backtesting Fidelity select sector funds? They go back 20 years.
    -Jayantilal

    ReplyDelete
    Replies
    1. I do. shoot me a private at traderlazy at gmail dot com and I'll send you the data. It would be an unformatted mess if I tried to add it to this comments box.
      LT

      Delete
  9. P.S. I see that you have posted annual results.
    -J

    ReplyDelete
  10. LT,
    I sent you e-mail at private but did not get response.
    Can you provide us with ETFs of your choice every month for a fee?
    If they do not change, we will pay for the no change advice as well.
    -Jayantilal

    ReplyDelete
  11. Hi Jayantilal,

    Sorry, I didn't see your email I just checked again and didn't find anything.
    As for providing ETF of my choice for a fee, I wouldn't feel like it is fair. 1- Because I don't invest this way yet due to lack of capital. 2- because it is simple for people to do it themselves by purchasing an ETFReplay subscription for $35 a month or you can follow thetatrend.com for example who does it for free. He recently talked about making it a subscription based service for $14.95 per month which is good value.

    Hope those things help.
    Regards,
    Henrik


    ReplyDelete
    Replies
    1. Dear Henrik,
      Thanks.
      Another question :: Why EEM? Why not small caps international. SCZ for example? If we use IWM, for domestic,Doesn't SCZ make sense for International?
      I have registered with my e-mail jayalagia@gmail.com You can send me added research you have done if you wish.
      -Jayantilal

      Delete
    2. The beauty is that you can select any ETF you want. Of course you want to prioritize those with low management fees, and those that pay a higher dividend.
      In this case, among other things I went with because of its popularity and because it has a past history that allows me to go back several years and perform meaningful back tests.

      Cheers,
      LT

      Delete
  12. Excellent series Henrik. I'm interested in setting it up in TOS (paperTrade) and I'm stuck at setting up the formulas. Do you know how the original author set up the formulas(s) and what tools he used? Like I mentioned, I use TOS so any guidance on how to "scan for" or set a watchlist to lay out these parameters would be greatly appreciated. "Sorry for the n00b question".

    Thanks Henrik!

    ReplyDelete
    Replies
    1. Absolutely man.
      Shoot me a private at traderlazy@gmail.com so I can send you the explanations. I avoided them in these articles as it would have made them too lengthy.

      Cheers,
      LT

      Delete
  13. Hi LT, have you tried backtesting the system from 2010 (not 2004)?
    By the way, TD Ameritrade has over 100 commission free ETFs.

    ReplyDelete
    Replies
    1. Absolutely. Here are the returns 2010 - 2014:

      2010: +26.3%
      2011: +35.5%
      2012: +8.5%
      2013: +8.7%
      2014: +7.9%

      Summing up, it outperformed the SPX in 2010 and 2011.
      It underperformed the market in 2012, 2013 and 2014.

      Overall, in the five year period, it out performed the market

      Cheers,
      LT

      Delete