I think I’ve established that asset rotation works. I’ve done a little further reading, and it seems that moving between asset classes isn’t the only stage for this strategy. It can also be successful to move between GIC sectors. This surprised me, because those sectors all represent stocks in a single market. But they must have varied enough cyclical characteristics that they present opportunities to move capital profitably.

Here’s how it would work in four steps. Remember that this won’t avoid sudden, serious market downturns. However, it will help a person to stay informed in a way that is efficient for the amount of time and effort required.

The first step is asset-class selection. According to studies, this step is most important. For example, I’ve been reviewing momentum variance between stocks, bonds and cash for the last year or so. Doing a backtest showed that this would have been profitable over the past 10 years of volatility. However, it produced only one signal each to switch from stocks to bonds and from bonds back to stocks, during the summer of 2010. Because it was the summer doldrums and the differential was small, I ignored it (successfully). Adding in a larger number of asset classes over the period of the backtest improved the results greatly. Here’s what I’m using right now: cash, gold (IGT), bonds (XBB), small-cap stocks (XCS), emerging markets stocks (XEM), large-cap stocks (XIU) and real estate stocks (XRE). I would have preferred to use something unrelated to the equity market for real estate, but that’s unrealistic.

In order to complete a reasonable backtest, I had to substitute some American ETFs that had a longer history. Based on the formula below, here are the results in graphical format. The beginning value is $10,000 on 24-May-02. It is interesting to note that the formula stayed in cash for almost a year, to April-03.

You’ll notice that while the loss of value wasn’t as significant as with buy & hold, it happened earlier and it still hasn’t broken the previous high.

Second is the selection of the asset class to own. Each asset class is evaluated on a weekly basis and compared for outperformance. Because short-term momentum tends to persist, we look for the asset class that has the best momentum over the past 1 – 12 months. In fact, we will average the performance over 1, 3, 6 and 12 months ((1m + 3m + 6m + 12m)/4). Even in a down market, the performance of the asset class with the best momentum is not too bad.

Third is to invest in the preferred asset class. These days, it’s easy to buy an ETF that represents that sector or market. There also may be inexpensive mutual funds that do the same thing. Or it may be possible to buy a diversified group of individual equities within the sector or market.

Each week, complete the market review and find the asset class with the best momentum. If a new asset class takes the lead over your existing holding, sell your investment and buy the new one. In practice, this should lead to approximately 4-6 trades per year.

Decision Moose is a website that shows one financial advisor’s output from a system similar to this. I understand that he include “bonus points” for macroeconomic indicators. He has also found a way to smooth the results so that a new asset class doesn’t take over for only a single week.

Next week, I’ll look at whether or not it makes sense to combine this speculative approach with a fundamental investment plan.

Asset Rotation Mechanics

2 thoughts on “Asset Rotation Mechanics

  • June 6, 2012 at 7:40 pm
    Permalink

    Robert, you’re making my head hurt. This seems so complex! I’ll have to buy you a coffee or lunch one day and bring a pad and pen. 😉

    Reply
    • June 6, 2012 at 7:46 pm
      Permalink

      I’d be happy to meet anytime you’re in Calgary. Having said that, in my next Wednesday post I’ll try to clarify what I’m doing.

      As a side note, trading commissions make a huge difference, especially when starting with a small(ish) amount. 14 trades per year (on average) x 2 trades per time (sell & buy) x $25 (as an example) = $700. As a percentage of $10,000, that’s 7% per year (far higher than any MER).

      Reply

Leave a Reply

Your email address will not be published. Required fields are marked *