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

Wednesday, January 7, 2015

ETF Rotation Systems to beat the Market - American Equities

Because I'm usually hungry for trading and investing related content, it's unavoidable for me to spend two or three solid hours during the weekends on the web looking to learn something new. Most times, I find way too many things that lack real value or that do not offer real actionable ideas. Long gone are the days when I used to waste my life reading random gurus' opinions and forecasts. There is tons of information on the markets out there. A very small portion of it, is truly worth something. However, from time to time you find that little gem and you're allowed to experience that Eureka moment once again.

One trading technique that has really picked my interest in the last couple of months is ETF Rotation systems. I already knew about them but never paid much attention. It was via fellow blogger/trader Dan from ThetaTrend.com that I became really interested in this concept.

The basic goal
The idea is to basically try to beat the markets by being invested in specific sectors that have some "recent" momentum going in their favor, discarding others that are performing poorly. By doing this, ETF Rotation Systems aim to deliver better returns than the market and with less volatility (smoother equity curve + smaller and shorter draw-downs).

Rotation vs being long SPY
We all know that most retail traders lose money in the long run. We also know that every year, 75% to 90%+ of the professional money managers fail to beat the S&P500 index. It's because of that, that being an index follower makes sense if you have a very long term horizon in mind. Simply buying shares of SPY, collecting the quarterly dividends and riding it year in and year out will guarantee that you will beat most traders, including professionals. You will spend way less money in management fees and trading commissions and hopefully, given enough time, you will be able to grow your capital at a compound annual growth rate of around 7%.

However, being simply long the market via an index following approach is a bumpy road and you are destined to suffer gut wrenching draw-downs that may take years to recover. At the same time, being long an index is not suitable for everybody. For example a period like 1929-1932 where the market lost 89% of its value would have been devastating right in front of your retirement. Or more recently, a 55% draw-down on the SPX between 2008 and 2009 which was indeed life changing for many and not in a good way. Yes, you would have gotten out of the draw-down years later had you remained invested, but you would have had to delay your retirement because of the market crash. 4 or 5 more years working for "the man" just because of the freaking market crash. Not funny.

So, instead of simply being long say SPY, an ETF Rotation system would attempt to only be long a portion of the SPY. That is, a particular sector inside the index, discarding the rest of the stocks that are doing bad. A current example as of this writing would be long the market except for the energy sector, which has been taking it on the chin lately. Every month, the system calculates the recent performance of its instruments of choice and puts the capital to work on the best of them, discarding the rest until next month when everything is re-evaluated. If the best symbols this month are the same symbols the portfolio is already invested in, then no trades, simply enjoy the party until next month.

The rotation system of your choice doesn't have to be constrained to equities. In fact it is healthy to mix up different asset classes like US Stocks, International Stocks, Commodities, Real Estate, Bonds, etc.

Will an ETF Rotation system always beat the market?
Of course not. With ETF Rotation systems you are always chasing recent performance. You may end up moving into an ETF that has performed well in the last three months, only to see it decay going forward. That's why your evaluation function is crucial as well as back-testing your ideas. However, depending on the instruments of choice and your tests you will find that most of the time it is way better to invest in recently strong ETFs, than to try and catch falling knifes on the worst performing ones.

Your ETF Rotation system may not beat the market every single year, but as long as it beats the market by a good margin over the long run, with fewer heart attacks, it is worth it.


The ETF Rotation Systems to beat the Market series
After this lengthy introduction it is time to say Welcome! to the The ETF Rotation Systems to beat the Market series: the series of articles where I explore and back-test different rotation systems designed with the goal of beating the markets while minimizing draw-downs and volatility in the portfolio.


American Equities Rotation
For the first portfolio of the series I'm going to simply focus on a single asset class: Equities. This is a simple portfolio that chooses the best three sectors every month from the pool below:

XLB: U.S. Materials Sector SPDR
XLE: U.S. Energy Sector SPDR
XLF: U.S. Financial Sector SPDR
XLI: U.S. Industrials Sector SPDR
XLK: U.S. Technology Sector SPDR
XLP: U.S. Consumer Staples Sector SPDR
XLU: U.S. Utilities Sector SPDR
XLV: U.S. Health Care Sector SPDR
XLY: U.S. Consumer Discretionary Sector SPDR


How are the three best ETFs selected every month?

There are three ingredients involved in the formula:

- The 3 months return (40% weight in the final score)

- The 20 day return (30% weight in the final score)

- The 20 days volatility (30% weight in the final score). Volatility is the annualized standard deviation of daily returns. So, the 20-day Volatility is the standard deviation of the past 20 1-day returns multiplied by sqrt(252) (annualized). The idea is to penalize the instruments that are having large variations in their daily returns. Those that are quiet and consistent are favored.

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


Here is the result from 2003 to 2014.
Definitely better than the S&P500 and less volatile:
CAGR: 11.4% vs 9.5% SPY (better yearly returns)
Volatility: 16.6% vs 19.5% SPY (less volatile)
Worst draw-down: -43.3% vs -55.2% SPY (smaller draw-down)

Not bad, but with lots of room for improvement.
Let's apply a filter rule in order to minimize draw-downs: If a selected ETF is below its 10-month moving average, then the third of the portfolio that would correspond to it will be invested in SHY instead (1-3 year Treasury bonds fund). If two of the three best ETFs are trading below their 10-month average, then 2 thirds of the portfolio would go to SHY. If all three of the best ETFs are below their 10-month moving average, the entire portfolio goes to SHY. This is in order to avoid major draw-downs, as it is obvious that if the three best sectors are below such a long term moving average, then what can we expect from the rest? Something pretty bad must be happening out there, better to move the capital to a safe instrument.

Here are the results:
CAGR: 10.4% vs 9.5% SPY (better yearly returns)
Volatility: 13.2% vs 19.5% SPY (way less volatile)
Worst draw-down: -25% vs -55.2% SPY (much smaller draw-down)

The 2008 collapse would have only been a 25% draw-down for this portfolio. However, by moving the capital to a safe instrument (SHY) from time to time we inevitably had to sacrifice some of the performance, as the yearly return went from +11.4% in the first case to +10.4% when the filter rule was applied.

This is a simple ETF Rotation system entirely based on equities. Not spectacular by any means but apparently effective and better than simply long SPY (which the vast majority under-perform anyways). In the upcoming chapters I will share some portfolios with very interesting results, some of them truly remarkable. Stay tuned.

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

12 comments:

  1. Great work, I'm enjoying your series. For the 20 day volatility calculation, just adding volatility to the final score would favor rather than penalize large variations. How are you inversing the volatility in your calculation? A simple inverse, or something more fancy?

    ReplyDelete
    Replies
    1. That's a good question. Volatility is not added to the calculation. The instruments are simply ranked. For example:
      For the 3 month return the ETFs are ranked from 1 to 13 in this case.
      For the 20 day return the ETFs are ranked from 1 to 13.
      For the volatility the systems are ranked in reversed putting the lowest value in first place in the ranking .

      Now with all the ranks assigned for all ETFs in all three categories, for each ETFs you use the formula (3monthReturnRank * 0.4) + (20 dayReturnRank * 0.3) + (20 dayVolatilityRank * 0.3) which gives you a global weighted rank. The ETFs with the smallest weighted rank are chosen.

      Regards,
      LT

      Delete
  2. Results after taxes?

    ReplyDelete
    Replies
    1. Futile exercise. Depends on your personal situation.
      LT

      Delete
  3. > How are the three best ETFs selected every month?

    Very interesting set of articles; thanks for doing it.

    From this I gather you reranked monthly? Beginning of the month, or every 30 days after you started? (Do you think it matters?)

    ReplyDelete
  4. Hi Michael,
    The ranking is calculated the last day of the month so you rotate to the instruments the first day of the next month at market close.
    Obviously, this matters. It could be a totally different story if you rotated every three months or every 15 days.
    Thanks for reading.
    LT

    ReplyDelete
    Replies
    1. This comment has been removed by the author.

      Delete
    2. Sorry, I realized I was unclear in my "does it matter"... I meant does it matter WHEN in the month you recalculate and rotate? End of month, middle of month, somewhere in the third week... I'm guessing not, as long as you're consistent.

      Delete
    3. Got it. Shouldn't matter. I think it should be as you say as long as you are consistent.
      LT

      Delete
  5. When you pick the top 3 to rotate to, do you put an equal amount of money into each one?

    ReplyDelete
  6. hi lt i love this etf rotation idea for my pension as its tax free ! (until the end)
    how can i contact you for help building a spreadsheet to help rank them for me each month.
    regards dom

    ReplyDelete