Sunday, May 1, 2011

GB9A Relative Strength Portfolio

Here is a relative strength portfolio that I just started trading. I'll explain what it is and the rationale for using it.

The portfolio uses 9 asset classes. DBC (Commodities), EEM (Emerging Markets), EFA (EAFE index), IAU (Gold), SH (Short S&P), SPY (S&P 500), TIP (Inflation protected treasuries), TLT (15 year treasury), VNQ (real estate).

Why so few etfs? I've found that packing too many etfs into a relative strength portfolio doesn't add performance. It actually hinders it from rotating to the best performing sector. It's best to keep it simple and not over complicate it by adding similar asset classes with slightly different shades. Some people like to create separate portfolios for bonds, international stocks, and US stocks. That is typical asset allocation. However, I would like to point out that in a bull market your bond allocation will be in high yield bonds which don't provide you the traditional protection that treasury bonds add. Your asset allocation won't protect you in the same way a fixed asset allocation will. So in this light I'd rather go with one portfolio and let it rotate to the best asset class.

How many to hold? I'm selecting the top 2 asset classes from the list. Why? The top 2 provides a little more stability to the returns while still allowing the portfolio to completely exit an asset class. Using 3 etfs in such a small portfolio means you have to hold an under performing asset class. To hold more asset classes you need a larger portfolio.

How is the relative strength calculated? It is calculated using the 3 month price performance weighting this calculation at 70 percent. It also uses the 20 say volatility weighting this calculation at 30 percent.  I've back tested 6 month price performance the the draw downs are much higher. The 3 month catches the big trends but adapts fast enough to get you out of an asset class as it begins to lose favor. A smaller time frame can help you get out even sooner but you will get whipped around more and it doesn't seem to help overall returns. The volatility filter helps reduce draw downs and provides more stable returns. As the volatility of an etf increases it's rank will be pushed lower. Volatility eats into returns and is also a sign that an asset class is beginning to fail. It will hurt overall returns but will reduce risk and draw downs. I think I've found a good balance with this weighting.

How often do you update the portfolio? The portfolio is updated once per month. Updating more often doesn't increase returns, nor does it reduce risk. You only whipsaw yourself. No stops are used but I will start calculating an equity curve and see if that helps reduce draw down and risk.

What about risk management? That's up to you. You can use stops, and equity curve, puts, nothing or a combo. I'm going to let it run without stops and trust that the portfolio to rotate as it is designed into treasuries, or an S&P short if the market dives. I like using cheap out of the money puts to protect against major catastrophes in excess of 30 percent losses. Less than that I'll let the portfolio run and rotate me out into safety at the beginning of each month.

Important Notes About the Back Test
The max draw down on the performance graphs is calculated monthly. It is not your true max draw down. During the month the portfolio could draw down much greater than 15 percent. All the performance test is showing you is that from month to month it never exceeded 15 percent, but it could have drawn down 30 percent or more during the month. A couple of the etfs on the list did not exist before 2006 so the performance before 2006 would be impacted by this.

So in a nutshell here are the rules.

1) Buy the top 2 etfs in the portfolio according to the above ranking rules
2) Update the holdings once per month. If the etfs being held are no longer the top 2 performers they are sold and replaced by the new top 2 performers. If nothing has changed in the rankings then nothing is bought or sold that month.
3) Take the good and the bad. Sit tight and let the portfolio work. It will out perform at times, and also correct sooner or larger than the overall market at times.

I'm trading this portfolio and I'll keep updates on the 1st of the month on it's progress. I'll also post some more info on this portfolio and additional tests as I go along.

No comments:

Post a Comment