The Benefits and Imperfections of Asset Class Investing

What's Wrong with Judging Investment Performance with an Index
One of my pet peeves is articles about investment performance based on the price variation of indexes such as the TSX, the Dow or the S&P 500. Unfortunately they do not reflect real world individual investor experience. Though it is possible to buy ETFs or mutual funds whose objective is to track an index, such things as trading costs / commissions, tracking error, bid-ask spreads and distributions can cause actual results to vary from the index.

Another thing I find annoying is that people often limit themselves to indexes for only two asset classes - stocks and bonds. We've all seen the classic 40% bonds, 60% stocks. There are a lot more asset classes out there with which to diversify, like real estate (REITs), real return bonds, foreign developed or emerging market equities (which introduce the issue of currency hedging), small and value cap tilts, commodities. Current theory says we should take advantage to maximize diversification and the ETFs are there to allow the average investor to do that, so why not model it?

Another issue of note is whether and when to rebalance a portfolio. Would a policy of reviewing the portfolio every December, and rebalancing if too much out of sync with target allocations, have done better than simply buying and holding?

Finally, the recent (on-going?) financial crisis and market crash provides a real high stress period in which to see how various realistic portfolios fared.

Assumptions: So, I've done some calculations in as realistic a way as possible. I started with $100,000 in May 2007 before the troubles really began and took it to the close last Friday, September 11, 2009. The portfolio is split 70% equity, 30% fixed income, with finer sub-divisions of both for the 4- and 16-asset portfolios. I used passive index ETFs available on US exchanges and the TSX.

The USD-CAD exchange rate I've used is based on the mid-market closing rate, which I've adjusted for the initial purchases by adding 1% to approximate the foreign exchange fee embedded in Canadian broker rates. At the Dec.17, 2008 rebalancing date, I have not adjusted for FX since almost all of the FX fee could be avoided by doing wash trades offered by most brokers and/or keeping the USD distributions in a USD account when received, as all brokers allow for non-registered accounts and some do for registered accounts. Also I have not deducted any US withholding tax from the US ETF distributions, which is ok for ETFs held in registered accounts but not in a TFSA or a non-reg account. So that assumption might slightly overstate returns depending on the account in which a portfolio is held.

As to rebalancing, I modeled none in Decmeber 2007 because the ETFs had not strayed far from their May target percentage allocations. Ths cash distributions were merely accumulated in the account. I ignored interest on the cash since it would have been too little to matter. ... all this stuff about my assumptions shows why so many people don't like taking the trouble to do realistic calculations - it's painstaking!

Results:

• big surprise, the simplest portfolio, consisting of only the iShares S&P/TSX 60 Index (XIU) and the iShares ScotiaCapital DEX Bond Index (XBB) fared best in every way!! It dropped the least to the review point of Dec.17, 2008 and has recovered almost fully (less than 1% below) to the starting value. That's why I've named it the "KISS Me Quick" portfolio - It has treated you well - who doesn't like a kiss!? It's KISS = Keep It Simple Stupid and it sure is quick to implement.
• big surprise again, the more diversified the portfolio, the worse the results! Huh, I though diversification was supposed to help, but whether or not the strategy was buy-and-hold or rebalancing, the fancy 16 asset portfolio did worst: it dropped the most and has receovered the least. That's why it's called the "Diversification Guru" - we all know what gurus are really worth.
• wow, rebalancing really worked well. In the short time since last December, all three rebalanced portfolios have outdone the buy-and-hold approach by anywhere from 7% to 9%.
• no surprise, diversification by holding fixed income is a lot better than just equities; if only XIU had been in the portfolio, there would have been a 31% drop in value, even including distributions. That's much worse than the 22% fall of even the worst portfolio, the 16-asset version. And XIU as of Sept.11th was still 14% below its initial value of May, 2007. The strong recovery of XIU has not made up the ground lost up to December. The reason is that no rebalancing occurred, as it could not with a single asset.

Why More Diversification Didn't Work

• real estate and foreign markets - some of the extra asset classes fell harder than Canada's; the UK's banks made up a bigger chunk of the FTSE index and they had just as much trouble as US banks
  
• currency shifts - up to last December, the CAD's big drop relative to USD as the flight to safety occurred cushioned some of the blow of drastically falling stock markets but since then the strength of CAD (check all the blue appreciation of the last 3 months at RatesFX) has limited the upside.

Further Thoughts:

• the future may not be like the past - this time and in this relatively short period, it was bonds, particularly government bonds, that provided the critical diversification benefit. Safety of principal was the issue. That may not be the case if inflation for instance, is the next big threat. In an uncertain world, different assets for different threats is still my best guess at what will allow me to survive if not thrive quite as much as the strategy which has worked best in retrospect.