Should you be an Active Investor?

A second fundamental reason that we started The Index Investor was because we believed that too many people were wasting their time trying to beat the market. We believed that many people could save themselves a lot of time, energy, and heartache, while increasing their chances of achieving their financial goals, if they would use indexed investment products. Recently, we looked at a new set of data to see if we needed to change our beliefs.

The data set in question is publicly available: it comes from Morningstar, and was published in Money Magazine's February, 2002 issue. It consists of 493 U.S. mutual funds that have at least ten years of historical returns data (through 2001), and which invest in either U.S. or non-U.S. equities. Forty seven funds invested in non-U.S. equities; 446 funds invested in U.S. equities. Of the latter, 253 were large cap funds (including growth, balanced, and value funds), 81 were midcap funds, 52 were small cap funds, and 60 were sector specific funds. These funds collectively had assets of $1.6 trillion at year end 2001, or about 45 percent of the total amount of assets invested in mutual funds that invest primarily in equities.

Our first question was simple: how many of these funds beat the comparable index fund over ten years, after expenses? We compared the U.S. large and midcap funds to the Vanguard S&P 500 Index Fund. The Vanguard MidCap fund hasn't been in existence for ten years, and, on top of that, the Morningstar definition of "midcap" seemed to include more than a few companies that are actually in the 500. Of the 334 funds in our comparison group, 78 beat the index over ten years after expenses, or 23.4% of the funds in our sample.

We compared the small cap funds to the Vanguard Small Cap Index Fund. Over the ten year period, 21 of them beat the index after expenses, or 40.4% of the funds in our sample.

Finally, we compared the ten European funds in our sample to the Vanguard European Index Fund. After expenses, only two actively managed funds (20% of the sample) beat the index.

To put this analysis into perspective, we need to make two further points. First, we compared the index funds to an active management all star team. The funds in our database were those that were good enough to survive for at least ten years, which is no mean feat in a business that routinely merges less successful funds into more successful ones. Second, we did not have enough data to take tax consequences into account. If we did, even fewer actively managed funds would have beaten their respective index fund competitors, because the former trade more actively, and therefor generate more taxable ordinary income and capital gains distributions for their shareholders. So our results show the actively managed funds in their most favorable light.

Our second question was about the relative importance of asset allocation versus size and style tilts. In our way of thinking, asset allocation refers to the percentage of your portfolio that is invested in different asset classes, such as U.S. and international equities. Size and style tilts refer to the way you allocate your funds within an asset class (e.g., U.S. equities) between funds that invest in large, mid, and small cap companies, and/or funds that invest in growth stocks, value stocks, or a blend of both. (Readers with good memories will remember how much we dislike that term "growth stocks" -- see our December, 2001 discussion of momentum investing for more on this).

Among the 493 actively managed funds, the average compound annual return over the ten year period was 6.5% for the funds investing in non-U.S. shares, and 12.6% for funds investing in U.S. shares. (Note that these are not weighted by the assets of each fund, because fund assets were not proportional over the ten year period). The difference of 6.1% represents the impact of asset allocation.

The table below shows the average ten year compound annual returns for funds with different combinations of size and style tilts:

What we found most interesting about this table was how little most size and style tilts added compared to the overall average for all actively managed funds of 12.6% (not to mention the 13.5% return for the Vanguard S&P 500 Fund and the 12.7% return for the Vanguard Small Cap Index Fund). At the level of aggregate size and style, the best improvements came from tilting toward small caps (.6%) and value (.7%). At the level of the nine possible size/style combinations, it was midcap balanced which delivered the best performance improvement, at 2.3% above the overall average, while large cap growth delivered the worst, at (1.5%) below the average. Relative to the asset allocation decision, size and style tilts were clearly less important.

Sector tilts, however, were another story. The compound annual ten year returns delivered by these actively managed funds were as follows: Technology, 20.0%; Health/Biotech, 14.4%; Financial Services, 18.5%; Natural Resources, 10.9%, and Utilities, 9.0%. At 7.4%, the difference between the return on the average Technology Sector fund and the overall return on actively managed U.S. funds was larger than the 6.1% difference between the latter and non-U.S. funds (although the difference between all the other sector fund averages and the U.S. average was less than 6.1%). These data further confirm what we wrote last year: sectors tilts have the potential to deliver greater return increments than size or style tilts (of course, because they also represent more concentration than either size or style tilts, they also carry with them higher risk, as any heavy investor in a technology sector fund can tell you these days). That being said, you also have to remember that today you can also invest in sector index vehicles in addition to actively managed funds (unfortunately, sector index funds haven't been around for ten years, so we can't compare their performance to the active funds). Given the relatively high expenses charged by active sector funds, this is an important consideration.

Okay, so what we've found so far is that a few actively managed funds have the potential to deliver higher returns, after expenses (but maybe not after taxes) than index funds. The next logical question is how do you go about identifying future index beaters?

This is where the real problems arise. In a nutshell, we found no way to use historical data to identify, ahead of time, actively managed funds that later beat their respective index fund competitors.

We started by testing the (somewhat reasonable, we thought) assumption that a fund that charged higher expenses (to pay all those extra bright people, no doubt) should deliver higher returns. Unfortunately, what we found was that the correlation between an actively managed fund's ten year compound annual return and its expense charges was (.17). That's right, it was weakly negative, implying that higher expenses resulted in relatively worse performance. Imagine that…Hold that double mocha latte! A recent study by John Chalmers et al ("Fund Returns and Trading Expenses") drives home this point. Its premise was that the relationship between fund returns and their trading expenses (as opposed to their overall expenses) provides a more powerful test of the value of active fund management. Chalmers found that the average fund incurred trading expenses equal to 0.75% of assets, and that the level of these expenses was negatively correlated with returns. In short, "the level of [additional] returns generated by [active] fund managers' trading activity fell short of the expenses they incurred [in making those trades]."

Our next step was to see if a fund's relative performance over the first five years of our data (that is, between 1992-1996) could be used to predict its relative performance over the next five years (1997-2001). What we found was very, very discouraging (if you're an active investor). We found that only one fund whose performance was in the top ten percent of all actively managed U.S. funds over the first five years was also in the top ten percent after the second five years. One: The Fidelity Select Electronics Fund. Only one fund that was in the first decile in the first five years was in the second decile in the second five years: the Seligman Communications and Information Fund. And only two funds that were in the top ten percent of funds during the first five years were in the third decile in the second five years: the Fidelity Select Computers and Fidelity Select Banking Funds. That's only four out of 446 funds -- less than one percent -- that managed to end up in the top 30% of funds in the second five years after having been in the top ten percent after the first five years. In short, they're not kidding when they tell you that past performance is no guarantee of future performance.

We're not the only ones who have reached this conclusion. The Financial Services Authority in the U.K. last year published a report ("Past Imperfect: The Performance of UK Equity Managed Funds") that concluded that "retail investors could not usefully exploit information on past performance." Others have reached the conclusion that, at best, while good performance does not persist, bad performance does, and can be used as a signal to sell a fund. In no case has any study reached the conclusion that past good performance can be successfully used as a way to pick future winners.

Not only does good performance by an actively managed fund in one period not predict good performance in the next one, you cannot even be sure what it means. As Stan Beckers described in his Journal of Portfolio Management article ("Manager Skill and Investor Performance"), luck alone plays a great, and usually unacknowledged role (what a surprise!). He performed a simulation experiment similar to the one we described in last month's issue. What he found was that, strictly because of luck, "some [active] managers could easily stay in business for ten years without having any skill whatsoever…The natural conclusion is that manager selection on the basis of returns data (and league rankings) is extremely hazardous…This has very sobering consequences for the mutual fund industry, where typically the only information in the public domain is the returns data track record."

In summary, what emerges (yet again) from this analysis is confirmation of the basic beliefs that have guided us since we launched The Index Investor:

• Most investors do not spend enough time thinking about how to allocate their investments across different asset classes, despite the fact that this decision has by far the largest impact on the long term returns they will earn on their portfolios.
  
  
• Most investors are overconfident. We believe that active investment management -- trying to "beat the benchmark" through superior manager selection, market timing and/or security selection -- is exceptionally difficult, and very, very few people consistently do it profitably year after year.
  
  
• Most investors would be better off if they used low cost, tax efficient index investment vehicles (mutual funds or exchange traded funds) to implement their asset allocation decisions.

| Is the Average Investor a Good Asset Allocator? | Should you be an Active Investor? | Model Portfolio Update |