Indexing Works (Even In Down Markets)
By: Gerry Rocchi
The myth that active managers will typically beat the index persists. Gerry Rocchi, of Barclays Global Investors Canada Limited, believes there are three main reasons why the myth dies hard.
Conventional wisdom has it that indexing doesn’t work in down markets, that only active management typically adds value in a bear market. But numbers recently released by Globefund have dealt a body blow to this myth.
The Globefund data (see Table 1) for 2002 show that for all of the major mutual fund asset classes (Canadian equity, U.S. equity, international equity, Canadian bonds) the median active funds in Canada did not beat their respective benchmark indexes – even in last year’s punishing markets.
By definition of the median, 50 per cent of funds in each asset class did better than the median fund in 2002 and 50 per cent performed worse. Therefore, what these results also tell us is that Canadian mutual fund investors in these categories had less than a 50 per cent chance of earning more than the index return in this, supposedly, ideal year for active management.
The weak performance of median active funds in 2002 is not surprising when you consider what Economics Nobel Laureate William F. Sharpe has called “the arithmetic of active management.”
In an elegant article published in 19911, Sharpe noted that the sum of all active investors’ portfolio holdings is, by definition, the market. Therefore, across all active funds in a given asset class, one fund’s over-weighting of stock ABC is matched, somewhere, by another fund’s under-weighting of ABC. As a result, the average return on those portfolios, before costs, must be equal to the return on the market. And this market return is simply the index return – that is, the return that you would achieve if you invested in the entire market without over- or under-weighting individual stocks.
But if average returns before costs are equal to the index, or market, return, it follows that, after costs, the return on the average actively managed dollar will be less than the return on the index, since the index itself does not pay fees.
The Globefund facts in Table 1 are, therefore, quite consistent with what Sharpe’s theoretical reasoning on indexed investment would lead one to expect.
When people talk about a ‘stockpicker’s market,’ they usually mean a bear environment where active managers – ‘stock-pickers’ – have a good chance to beat the market. Sharpe’s paper, however, is a sobering reminder that all the phrase ‘a stockpicker’s market’ can ever really mean is that the market (up, down, sideways) is always just the sum of everyone’s stock picks – with every winning pick versus the market matched, necessarily, by a losing pick.
The Globefund results for 2002 are not unique. Nevertheless, the myth that active managers will typically beat the index persists. Why is this?
We believe there are three main reasons why the myth dies hard.
First, in any given year, some funds do beat the index. The managers of those funds naturally publicize these returns widely and investors tend to focus on them. The problem, however, is that it is hard to pick winning funds in advance and it is difficult for active funds to beat the index consistently, year after year. This year’s winners are often next year’s losers and vice versa.
Once again, the Globefund data are revealing. Out of the 31 mutual funds in the Canadian equity category that achieved first quartile2 performance in 1997, only 14 were first quartile in 1998 and none were first quartile by 2002. If we start the comparison in 1999, Globefund shows that out of the 61 first quartile funds in that year, only 10 retained that position in 2000 and, by 2002, only two of the original 61 were first quartile.
Second, ‘style drift’ can bias results. In some cases, a fund will achieve superior returns with assets that are not part of the benchmark index for that fund’s asset category. For example, Canadian equity funds can hold up to 30 per cent U.S. equities or other foreign property. Or a fund might invest, in part, in a portfolio of small cap stocks whose risk/return profile is very different from that of the larger cap stocks that dominate the index for the fund category.
Clearly, ‘style drift’ is not ‘cheating.’ Who can object if, for example, a fund in the Canadian equity asset category earns a better return by investing some of its assets in the U.S. in a year when Canada has under-performed that market? At the same time, it is fair to ask:
- How did that fund perform against the blended index returns for the asset classes in which it invested?
- Would the investors in that fund have been even better off by investing in some combination of the S&P 500 Index (the usual benchmark for the U.S. equity asset category) and the S&P/TSX Composite (the benchmark for Canadian equity)?
When the assets move, the goalposts should follow.
Measurement is not the only problem with style drift. Returning to our example above, what if the investor already had exposure to the U.S. market through other funds – with the Canadian equity fund serving as the Canadian piece of a carefully crafted asset allocation plan? Now that investor finds him or herself with an unplanned, over-exposure to the U.S. This is fine if the U.S. market performs as the active manager expects, but could be disastrous if it does not.
Finally, in the Canadian equity asset category – probably the one with the highest profile among Canadian investors – there is the Nortel factor. In recent years, the weight of Nortel in the S&P/TSX Composite Index, the benchmark for the Canadian equity category, was quite high. But regulations did not allow active managers to match this weight. In hindsight, this turned out to be lucky for them and, as a result, more than the usual percentage of active managers was able to beat the Composite Index in recent years.
But what if we compare the median fund in the Canadian equity category with a more appropriate benchmark, one that shared the constraint these funds faced with respect to Nortel? The S&P/TSX Capped Composite Index (an index that limits the weight of any one stock to no more than 10 per cent) is just such a benchmark.
The results are very interesting. When compared with the capped index benchmark, data from Globefund and Standard & Poor’s show that the median fund trails the index over the last year (-13.5 per cent versus –12.4 per cent for the capped index), three years (-2.2 per cent versus -1.5 per cent), five years (two per cent versus 4.2 per cent), and 10 years (8.8 per cent versus 10.6 per cent). To sum up:
- Funds are exposed to the market, just like indexes.
- The median performing fund typically trails the index.
- Some funds will beat the index, but it is hard to identify those funds in advance.
- In the fund world, this year’s winners can be next year’s losers.
- Most important of all, these conclusions hold in both down and up markets.
Gerry Rocchi is chief executive officer of Barclays Global Investors Canada Limited in Toronto.
1. William F. Sharpe, The Arithmetic of Active Management, The Financial Analysts’ Journal, Vol.47, No. 1, January/February 1991. pp. 7-9.
2. A first quartile fund for a given period has performance that is better than 75 per cent or more of all of the funds in its group.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -