The following table compares the compounded annual returns (CARs) for the last 15 years between portfolios comprised of the median (half did better, half did worse) actively managed funds versus similarly weighted portfolios of index funds.

Globe & Mail Active Fund Data	1	2	3	4	5	6	7	8
Asset Class	Index	Active	I-A	Margin	I-MER	I-MER-A	Margin	Cum Diff
	no MER	with MER	diff	diff/Active	0.5	diff	diff/Active	over 15 yrs
TSX Comp TR - Canadian Equity	11.3	10.6	0.7	6.6%	10.8	0.2	1.9%	12.5%
SCM Bond Universe - Can Bonds	8.2	6.6	1.6	24.2%	7.7	1.1	16.7%	43.4%
S&P 500 - US Equity	10.7	7.3	3.4	46.6%	10.2	2.9	39.7%	141.5%
MSCI EAFE - International Equity	8.3	6.8	1.5	22.1%	7.8	1.0	14.7%	40.3%
increasing Canadian Equity
20/40/20/20	9.3	7.6	1.8	23.2%	8.8	1.3	16.6%	57.1%
30/40/15/15	9.5	7.9	1.6	20.0%	9.0	1.1	13.7%	50.9%
40/40/10/10	9.7	8.3	1.4	17.0%	9.2	0.9	11.0%	44.2%
increasing US Equity
20/40/20/20	9.3	7.6	1.8	23.2%	8.8	1.3	16.6%	57.1%
15/40/30/15	9.4	7.4	2.0	26.7%	8.9	1.5	20.0%	67.3%
10/40/40/10	9.5	7.3	2.2	30.4%	9.0	1.7	23.6%	77.5%
increasing International Equity
20/40/20/20	9.3	7.6	1.8	23.2%	8.8	1.3	16.6%	57.1%
15/40/15/30	9.1	7.4	1.7	23.2%	8.6	1.2	16.4%	52.9%
10/40/10/40	8.8	7.2	1.7	23.1%	8.3	1.2	16.1%	48.9%

Returns data, as of 31Dec06, comes from Globe & Mail (16Feb07 online) 15-year mutual fund reviews as well as Libra Investment Management. Benchmark returns are in CA$ and include reinvested dividends. Raw data is summarised here.

Notes for the 2006 edition: 1. There was no 2005 edition due to lack of data. 2. Some of the previous calendar-year data for actively-managed funds from G&M sources has changed slightly. It's not clear why this has occurred.

Column 1 shows pure index returns, i.e. without allowing for management expense ratios (MERs.) This is a theoretical rate of return that can be achieved only if all costs are eliminated.

Column 2 shows the performance of the median actively-managed fund or fund portfolio.

Column 3 indicates by how many percentage points each year a portfolio of indexes outperformed an equivalent actively managed portfolio, i.e. Col3 = Col1 - Col2.

Column 4 expresses this annual outperformance as a percentage (margin) over an actively managed portfolio, i.e. Col4 = (Col1 - Col2)/Col2. A value of 13.2% means that the margin by which the indexed portfolio outperformed represents 13.2% of the actively-managed portfolio's annual return.

Column 5 shows the annual performance of a portfolio implemented with index funds after allowing for an average MER of 0.50%.

Column 6 indicates by how many percentage points each year an index fund portfolio outperformed an equivalent actively managed portfolio, i.e. Col6 = Col1 - MER - Col2.

Column 7 expresses this annual outperformance as a percentage increase (margin) over an actively managed portfolio, i.e. Col7 = (Col1 - MER - Col2)/Col2. This is similar to Column 4 except that it accounts for MERs.

Column 8 shows the total cumulative difference in returns over the entire 15 year period between an indexed portfolio and an actively-managed portfolio.

Following the data for the individual asset classes is data for typical portfolios. The aa/bb/cc/dd indicates the percentage weightings respectively in Canadian equities, Canadian bonds, US equities and international equities. The typical Canadian balanced fund (median 15-yr CAR ~8%) has a 40/30/15/15 composition.

One should be hesitant to draw any strong conclusions from just 15 years of data since that's a relatively short period to study. But that's all the data that's generally available for Canadian mutual funds. This hesitation is especially important considering that the past 15 years has been dominated by (a) a sustained reduction in inflation and interest rates, and (b) the greatest bull market of the century in the US.

That said, the ~1 percentage point annual outperformance of indexed portfolios is striking. This represents an ~50% cumulative outperformance over actively managed funds. This also demonstrates the powerful effects of compounding over long time periods, even a small annual outperformance becomes substantial over time.

Why these numbers are biased against indexed portfolios:

1. The 0.50% MER "handicap" imposed on the index portfolios is at the high end of the range. Most investors can reduce the effective MER using a hybrid of Exchange-Traded Funds (ETFs like XIUs and SPYs), low-MER index funds and in the case of fixed income, bond ladders. Those with more than $150K can reduce the total MER to 0.30% or less using CIBC index products, either exclusively or in combination with ETFs.
2. Actively-managed Canadian equity funds, perhaps in an attempt to skirt the 30% maximum foreign content restriction [eliminated in the 2005 budget] on RRSPs, often invest a substantial amount of their assets in foreign equities. If for no other reason than the secular decline in the value of the Canadian dollar until 2003, this has given them a substantial edge that's not available to Canadian index funds. (From 1971 to 2002, the CA$ has depreciated by an average of 1.3% annually against the US$ and by almost 5% annually against the German Mark and Japanese Yen.)
3. Actively-managed equity funds are usually "impure" in that their mandate allows them to own significant amounts of other asset classes. For instance, a fund that's categorized Canadian Equity could hold 10% in cash and 25% in US and overseas equities without violating its mandate. Since presumably this gives the fund manager the flexibility to mitigate the effects of bear markets in their designated asset classes, no effort is made in this comparison to avoid actively-managed that are significantly "impure." If fund managers possess the skill to predict when bear markets will begin and end then this decision should work in their favour compared to index funds whose "hands are tied" by the index's composition rules.
4. The 15-year numbers for actively-managed funds are inflated due to survivorship bias. Many of the poorest performing funds no longer exist so their sub-par performance no longer drags down the averages. It's difficult to assess how much an effect this has, however, academic studies have found that survivorship bias can inflate the compounded annual returns of surviving funds by several percentage points annually.
5. In some asset classes the number of actively-managed funds that have existed for a full 15 years is quite small.
6. In taxable accounts the indexed portfolios will do much better on an after-tax basis due to their generally lower turnover. Again it's hard to assess how much of an effect this has. (In RRSPs this is a lesser issue because the tax effects -- but not the transaction costs -- can be ignored.)

1. Generally speaking, low-MER actively-managed funds outperform similar high-MER funds. How much would the results differ if we excluded the high-MER funds from the fund averages?
2. Studies have shown that one derives a "rebalancing bonus" by periodically selling the better performing asset classes and using the proceeds to buy more of the laggards. This bonus can add an annual performance improvement of 1% or more. The comparison makes no provision for rebalancing. (The benefits of rebalancing can be better realized inside tax-sheltered retirement accounts like RRSPs.)
3. Returns are only one dimension of portfolio behaviour. An investor's ability to sleep well during the inevitable market downturns is also important, however the comparison does not address the volatilities (standard deviations) of the various portfolios.

1. Vanguard's John Bogle says, "costs matter." The above numbers provide strong evidence that this is true.
2. Small differences add up over time. A 1% improvement in a portfolio's annual return adds ~16% in total return after 15 years. And thanks to the power of compounding, that advantage grows to 22% after 20 years, 35% after 30 years and 49% after 40 years! (A 2% annual advantage grows from 34% after 15 years to a whopping 120% after 40 years.)
3. For "typical" asset mixes one could achieve ~10% compounded annual returns over the past 15 years. (40% in bonds may not be "typical" for everyone but it is typical of many balanced funds.)

15 years ending 31Dec06
15 years ending 31Dec05 data not available
15 years ending 31Dec04
15 years ending 31Dec03
15 years ending 31Dec02
15 years ending 31Dec01
15 years ending 31Dec00
15 years ending 30Jun00
15 years ending 31Dec99
15 years ending 30Jun99
15 years ending 31Dec98