Monday, March 21, 2016

Market Timing Using a Spreadsheet

I’m no market timer, but my investing spreadsheet looks like one. I’ve coded just about all of my investment decisions into one spreadsheet. Whenever I add new savings, the spreadsheet tells me what to buy to bring my portfolio back to its target asset allocation percentages. However, if we look at the rebalancing decisions in isolation, they look like brilliant market timing.

It’s no secret that the Canadian dollar has been extremely volatile compared to the U.S. dollar over the past year. I had no idea this would happen. I didn’t make any predictions. I don’t know what will happen in the future. But my rebalancing spreadsheet looks like it made good predictions.

Before the Canadian dollar began dropping, I was adding new money to U.S. stocks to get my portfolio back to its target percentages. After the Canadian dollar dropped, I was buying Canadian stocks. The recent run-up in the Canadian dollar has me back to buying U.S. stocks. I’ve consistently bought low. How can a simple spreadsheet be so smart?

The answer is that rebalancing looks brilliant whenever the spread between assets grows and then shrinks again. Rebalancing is a loser if the spread between assets just keeps growing. So, as long as the Canadian dollar keeps trading within a given range and Canadian stocks keep up with U.S. stocks over the long run, rebalancing will look a lot like being able to see the future.

16 comments:

  1. Do you rebalance if you do not add new investments?
    If so, do you rebalance regularly or when spread reaches certain level in %% or $$?

    ReplyDelete
    Replies
    1. @AnatoliN: I do occasionally rebalance without adding new money if my allocations are off by enough. Basing the thresholds on percentages can make them too narrow for small portfolios and too wide for large portfolios. Similarly, basing thresholds on dollar amounts can make them too wide for small portfolios and too narrow for large portfolios. I use a method described in the following post:

      http://www.michaeljamesonmoney.com/2012/03/portfolio-rebalancing-based-on-expected.html

      It may seem complex at first, but now that I have it coded into my portfolio spreadsheet, I don't have to think about it. My spreadsheet sends me an email if I have to do anything. But this happens rarely.

      Delete
    2. I use TD's webbroker and my trading cost is zero. Any rule of thumb for me?

      Delete
    3. @AnatoliN: I assume you mean that you don't pay trading commissions. But you do pay spreads. The formulas work the same with commissions set to zero.

      Delete
    4. you are right, twice. misread the formula.
      Should I count both spreads, of the source and target securities? Webbroker does not show spreads for index funds. Do they exist? where can I find them?

      Delete
    5. @AnatoliN: I don't know much about Webbroker, but with any trading, there is a bid price and an ask price. The gap between these prices is the spread which represents how much more you have to pay to buy a security than you will get from selling it. I treat the midpoint of the spread as the fair price. So, you lose half the spread with every buy or sell.

      Delete
  2. This phenomenom makes me think that Canadians with an asset allocation like yours (and mine) are kind of advantaged with this rebalancing plan.

    Investing about 1/3 of your portfolio in Canada to get the benefit of home bias and the other 2/3 in US and International, expose us to US and foreign currencies with minimal risk. Most of the time, the falling Loonie protect us when the world's economy is slowing down (my 2015 return is mostly because of the USD strength vs CAD). Then, when the TSX goes up faster than other stock markets, you can buy them with a stronger Loonie (better purchasing power with low prices).

    I think it's the counterpart of having our stock market loaded with energy and comodities. Best of both world?

    ReplyDelete
    Replies
    1. @Le Barbu: Comparing a world-diversified portfolio to an all-Canadian portfolio, things look best for the diversified portfolio when Canadian stocks (or the dollar) are falling. But the shine comes off when Canada booms. I think diversifying will win out in the end, but how good it looks ebbs and flows.

      Delete
    2. Sure, it's not like comparing apples to apples but this kind of diversification is a lot better than staying 100% canadian with 30-40% bonds to lower the volatility.

      As you explained, the gap between stock markets return should be narrower on the long run than the one between stocks and bonds.

      The first is a win/nil game, the second a nil/lose game.

      Another benefit is you almost never have to sell to rebalance, new money is enough most of the years.

      Delete
  3. Buying stressed/out of favour assets pays off over the long term.

    A bit surprised your spreadsheet told you to buy US assets before CAD dropped. That was just after a very fast rise in American stocks.

    ReplyDelete
    Replies
    1. @BHCh: It was a few months before the Canadian dollar dropped when I had new money to add and the spreadsheet said to buy U.S. stocks. In part, this is because my group RRSP at work is set up to buy only Canadian stocks, and I stay in balance with the rest of my portfolio.

      Delete
    2. Yep, that would explain it.

      Delete
  4. Nice post, your spreadsheet is a good trailing indicator.

    ReplyDelete
    Replies
    1. @Mark: Actually, for the period of time I discussed, my spreadsheet was also a good leading indicator. However, I would never count on it for making active bets.

      Delete
  5. i get the concept of spread. for this formula, should I use two halves of spreads associated with two securities?

    ReplyDelete
    Replies
    1. @AnatoliN: The decision to rebalance is individual for each ETF. It's only when two or more ETFs give their independent rebalance indications that the formula says to rebalance. For each individual ETF, the variable s is the entire bid-ask spread of just that ETF divided by the share price.

      Delete