Battling fund myths
Beware of number twisting

A recent release by Mackenzie Financial Corporation has sparked all kinds of public and private debate. It rebutted the media's persistent pro-indexing message with performance comparisons aimed to showcase active managers' superiority - which was reported in a recent National Post column. Rebuttals proliferated and it became clear that not only is Mackenzie's comparison flawed, but that even rebuttals had holes.

Problems with the pitch

There are five glaring problems with Mackenzie's performance comparisons. In an effort to illustrate the superiority of active managers, they highlighted the ten largest global and Canadian stock funds as of June 30, 2004 - along with the annualized returns for each over the past five years, compared to either an index or exchange-traded fund (ETF).

First, five years is a short enough time frame to make any comparison overly sensitive to the measurement period's "end-date". It catches about a year's worth of a bull market, about three years of a bear, and a few final months of recovery. In other words, most of their chosen period was a bear market.

Second, the selection of the largest funds is biased. Assets are influenced by two main factors - net sales and performance. We also know that the latter influences the former. Hence, choosing the largest funds naturally results in the selection of the better recent performers. It would have been a little better is they had chosen the biggest funds at the beginning of the five-year period - not at the end of the period.

Third, the benchmark used for the Canadian equity fund comparison is wholly unsuitable for two reasons. The i60s ETF is used as the benchmark for this category, but it has no limit on how much it can hold in a single stock. Mutual funds, on the other hand, have long been limited to investing 10 percent in a single holding, based on cost. While some funds exceeded the 10 percent threshold based on market value - i.e. Altamira and Legg Mason Canada - most did not and could not once Nortel began its historic ascent.

(Note: Nortel Networks, at its peak, accounted for more than 35% of the S&P/TSX Composite Index and 45% of the S&P/TSX 60 Index.)

Mackenzie chose the uncapped i60s as the benchmark for its comparison in a period dominated by a bear market. During Nortel's bull market ascent, however, not only were fund companies lobbying the OSC to lift the 10% rule (to be able to keep pace with the index), but many also changed their target benchmark from the standard (uncapped) index to a capped version of the same. The word 'bull' seems appropriate for its other meaning at the moment.

The other benchmark issue relates to the fact that most Canadian stock funds - particularly the largest - typically hold no more than 55 to 70 percent in Canadian stocks thanks to big foreign content allocations and cash reserves. So, comparing against an all-Canadian benchmark seems a bit silly since it may not yield any meaningful insight. (See Benchmarking Problems for more on this topic.)

Fourth, Mackenzie didn't highlight U.S. or overseas equities. That's probably because it doesn't have any funds that are among the largest. In fact, no Mackenzie U.S. stock fund is even in the top 40 by assets. Its overseas fund ranks fifteenth by assets. But its funds in both of these omitted categories are average performers at best.

Finally, if this same method were done for the previous five-year period, just four Canadian and one global stock fund would have beaten their respective benchmarks. I guess we know why this illustration wasn't released before this year.

The rebuttals

Rebuttals showed up in the Post, the Toronto Star, and on the web. Some good arguments surfaced (i.e. the Post), while some others (i.e. Toronto Star) printed much weaker arguments. The latest hat tossed into this ring belongs to ETF and indexing specialists Barclays Global Investors Canada (BGI). BGI published a detailed rebuttal on its website. But despite its best effort, it too does a bit of a disservice to readers.

BGI boasts of its "apples-to-apples" comparison, which extends the time frame out ten years and uses the i60C (i.e. the capped i60s) as the appropriate benchmark for Canadian stock funds. I agree with them on both counts, except that an all-Canadian benchmark is only part of a more suitable benchmark that should include cash and global stocks. As a result, BGI's rebuttal makes the index look better than reality.

You see, most funds - especially the biggest ones - have a policy of maximizing foreign content. Not just allowing this amount, but a policy of taking advantage of the current 30 percent limit. For years, this worked in favour of such funds as foreign stock exposure juiced performance by 3 to 4 percentage points per year through the late 1990s. But over the past five years, it's been a significant handicap.

The performance of the S&P/TSX Capped 60 index generated a return of more than 13 percent per year, for the ten years ending June 30, 2004. The MSCI World Index turned in a performance of just over 7 percent per year over the same period. Barclays' method unfairly punishes funds that, by policy, maximize foreign content.

In fairness, however, even these adjustments don't change the final outcome. Rather, they just narrow the gap of active managers' underperformance. However, the problem of a biased and limited selection of active managers remains. Barclays stuck with this because it's the method used by Mackenzie. However, Mackenzie's method was clearly inappropriate, so continuing with this method yields no meaningful insight other than to judge a small sample of funds.

Also bear in mind that the S&P/TSX indices have only been around for about five years. Data prior to that (i.e. half of Barclays' illustration) is based on simulations.

More often than not, problems with such comparisons cannot be fully resolved. But knowing the weaknesses in such presentations will allow you to better judge whether they have any merit, or whether it's just another example of creative number twisting.