Monday, 7 May 2007

Portfolio Magic ... 3+1 = 5!

If someone told you that you should buy for your investment portfolio a holding with low annual returns and high volatility of returns (wide swings from year to year, i.e. the "sleep at night" factor) would you not think that this is against common sense, that this would lower your overall returns and make the swings even worse? For example, why add an investment that increased on average 10.3% annually and with volatility of 24.3% (this is the standard deviation, the usual measure of volatility) to a portfolio that has a higher return of 13.7% annually with lower volatility of only 14.7%? Well, contrary to common sense, the marvels of diversification can make this worthwhile. Certain precise, though quite simple and understandable, circumstances that go to the heart of diversification, make this seemingly odd result come about. How does it work you ask?

The key is to combine investment assets that move, in the best case, in contrary directions – when one is going up the other is going down and vice versa. Beneficial, but less so, is the case where the assets move randomly or independently – sometimes together, sometimes in opposite directions. There is still benefit when the assets tend to move together, but not all the time, in a weak positive relationship. Most of the investment assets that can easily be bought by the average investor tend to fall into the last category so it isn't possible to build a high return portfolio with zero volatility but the improvement can be considerable.

Let's continue with the example, which comes from Roger Gibson's fine book, Asset Allocation. The mystery low return/high volatility asset is the Goldman Sachs Commodity Index and the portfolio is a combination of the S&P 500, the MSCI Europe, Australia and Far East Index and the (US) National Association of Real Estate Investment Trusts Equity Index. Gibson shows in chapter 8 that over the period 1972 to 1998, the combination of the four holdings would have increased returns a little, from 13.7% to 13.8%, and reduced volatility a lot, from 14.7% to 10.6%. The reason again for this remarkable result is again, that when commodities go up the others go down and vice versa, even though over the long term, all of them go up. To cap it all, each asset on its own, underperformed the combination of all four. That's the power of diversification.

An investment category that moves differently from others is termed an asset class. So what are the asset classes that can/should be combined in a portfolio that maximizes returns while minimizing volatility? The main ones identified by finance researchers are:

  • cash / Treasury bills

  • bonds, with further useful (i.e. with diversification benefits) subdivisions into corporate, government and international

  • equities, with useful subdivisions into large cap, small cap, value, international Europe, EAFE and emerging markets

  • real estate, purchasable primarily in the form of REITs, though one's home can be considered an investment to a degree

  • commodities - energy, minerals, foodstuffs, gold, basic materials (steel, aluminium)

For examples of how these can be turned into actual portfolios with the purchase of specific ETFs or low-cost mutual funds (the latter being primarily available in the US alas), check out the Index Fund Advisors portfolios. The excellent book All About Asset Allocation by Richard Ferri that I previously reviewed also has example portfolios. The Efficient Frontier shows basic international equity portfolios with ETFs. Bylo Selhi has a handy list of the various low-cost funds available in Canada, some of which represent the above asset classes. For Canadians, since US mutual funds cannot be purchased, US-traded ETFs are the only route. Most (if not all) can be listed and searched for at Index Universe or Morningstar. In the UK, the iShares ETFs are the only low cost option for anything other than UK stocks or bond indices, or European stock indices, which are available from any number of fund companies.

A word of caution is that, as Ferri shows with numerous graphs on every asset class, the advantageous correlation relationships can be absent or reversed from year to year or for periods of many years at a time. The relationships work over long periods when the variations smooth out. While 3+1 is no longer always equal to 4, patience is still a virtue in investing.


Mike said...

For Canadians, since US mutual funds cannot be purchased, US-traded ETFs are the only route.
You can buy Canadian traded ETFs as well -

Outroupistache said...

Oops. Yes, that's quite right, I own some Canadian ETFs myself. Bad phrasing by me. What I meant to say is that for some elements of the diversification strategy like US small caps or value funds which are not available in Canadian ETFs or low cost (<0.75% MER) mutual funds, US ETFs are the only route since US mutual funds cannot be bought by Canadians (and that, so I am told by my broker BMO Investorline, is because our regulators like the OSC won't allow non-Ontario/Canadian licensed funds to sell to Canadians. In my view that's part of the reason mutual funds fees are so high here - no competition. As I said before the only way I know to counter that is buy a piece of the banks, be an exploiter instead of an exploitee.

Mike said...

Very true, I'm in the process of setting up an ETF portfolio and it's about 70% in US ETFs.

Regarding your second point, I'm also about to buy part of one of the banks.

Patrick said...

Is there rebalancing going on here? Otherwise, how is it mathematically possible for a combination of assets to outperform every individual constituent?

CanadianInvestor said...

Yes, there was annual rebalancing that took advantage of the non-correlation so that over a period of years the portfolio does better than the individual parts held all along. In any particular year, as you note, the returns cannot be other than the sum of the individual holdings.

Wikinvest Wire

Economic Calendar

 Powered by Forex Pros - The Forex Trading Portal.