Over the past few months, I’ve been employing a momentum strategy to invest within my registered accounts. Outside of my registered accounts, I’m investing for income. I set up a model that I described previously, and so far it has worked extremely well. It doesn’t outperform the market each week, but it tends to help me get in front of the fast money, benefiting the rising prices caused by increased demand. My model included only ETFs traded on the TSX that represent bonds, gold, REITs, TSX 60 stocks, small cap stocks, and emerging market stocks. Being able to pick the asset class or sector most likely to outperform has translated into very good performance. For example, I owned gold from July to November of 2011. During that time, gold rose 15.4% and the TSX fell 13.5%. Since then, I’ve switched to real estate, then emerging markets, then back to real estate. Those further trades (up to June 1, 2012) earned me an additional 4.9%, whereas the market fell 4.2%. Although these results are extremely encouraging, I expect far more average performance, perhaps with increased volatility, during periods of rising stock markets.
Even though my strategy has been successful, I started to feel that it was somewhat restricted. It included gold, real estate, stocks and bonds as asset classes, and differentiated only between large cap, small cap and emerging market stocks. I felt there should be a place for the US market, the European market and certain other countries: Brazil, Japan and Hong Kong (as a proxy for China). Part of the source of that feeling came from reading economic news that showed Canada leading the pack out of the recession of early 2009, but then faltering in 2011. When our economy slowed and our market faltered in spring of last year, the US and, to a lesser extent, Hong Kong have outpaced the Canadian market. (Europe has done about the same, with Brazil performing far worse.)
For this reason, I added EWH, EWZ, SPY, IWM and XIN to my asset rotation model. The result was surprising. The model begins on May 24, 2002 (because that’s when enough data was available). Over the past 10 years, the stock market (represented by XIU) grew from $10,000 to $15,296 (as of May 25, 2012). The old model represented growth from $10,000 to a value of $46,366. The new model, by contrast, shows growth to $73,163. That makes sense to me, because with more assets to choose between, the model is able to select more profitable positions. The old model recommended 135 trades over the period, whereas the new model recommends 140 trades, which works out to around one switch each month. (Trading costs are ignored in the calculation; they can be significant in small accounts.) I notice that there are periods of uncertainty where the model switches every week over a period of 8-10 weeks, so I’d like to find a way to smooth the results and reduce trading frequency. I’ll share it if I find one.