Standard & Poor's (S&P) publishes a quarterly report comparing the returns of actively managed mutual funds to returns calculated for our indexes. The results should come as no surprise to readers of the Journal of Indexes-three out of five mutual funds underperform the indexes in any one-, three- or five-year time period. When people go through the data with us, especially people from the active fund side of the debate, we work through the usual questions such as fund fees and survivorship bias-both of which are accounted for, and neither of which explain the active funds' underperformance-and then come to the question of real numbers vs. calculated numbers.
In comparing indexes and active funds, these questions focus on the treatment of operating expenses, the costs of trading and the revenues derived from securities lending. All of these figure into the costs and returns of a real mutual fund, and none of these are included in the calculations of index returns.
S&P's calculations of index total returns assume that trades done to affect index changes are made at closing prices, that dividends are paid on ex-dates and that there are no cash inflows or outflows to be handled. While these assumptions all make sense and are virtually universal among index providers, they do exclude some real costs from the results. At the same time, some common activities that generate real gains to funds are also ignored. Securities lending is a simple example. More complex (and certainly not without risk) are strategies that involve trading index changes earlier or later than the official date, or optimization approaches that hold only some of the stocks in an index.
Are all these assumptions and factors minor adjustments, or do they matter? We can look at some data. The table on the next page gives figures on funds tracking the S&P 500, S&P MidCap 400, S&P SmallCap 600 and Russell 2000 indexes for the year ending March 31, 2005. The data come from S&P's mutual fund database and include funds that both identify themselves as index funds and identify the index they track; exchange-traded funds are included. (The data may not include all relevant public mutual funds.)
The conclusions are similar across the four indexes shown, so we will focus on the first column covering the S&P 500. One thing we see is that even counting funds isn't that simple. For the S&P 500, the data include 59 independent funds, but because many of these funds have several share classes, there are a total of 173 "funds." Since fees vary from share class to share class, returns vary across share classes, too: In a couple of the fund groups in the S&P 500 list, returns between one share class and another differ by as much as 93 basis points.
It is the overall range of returns, and the small number of funds that beat the index, that is most surprising. Clearly, which index fund one buys makes a difference. For the year ended March 2005, the S&P 500 returned 6.69 percent. Of the 173 funds tracking the S&P 500, the best returned 7.21 percent, providing 52 basis points of outperformance, while the worst lagged the market by 187 basis points, returning just 4.82 percent. On average the funds returned 6.1 percent- a significant difference from the index itself.
Interestingly, the average return weighted by fund size was 6.55 percent, getting close to the index. Apparently, large index funds do better: Either there are efficiencies of scale or less need to demand large fees at the larger funds. Only seven funds managed to outperform the index. The numbers here cover one year and neither time nor space permits examining longer time periods.
These figures confirm two things. For one, Jack Bogle's Cost Matters hypothesis clearly does matter. And two, real numbers make a real difference.
Comparing indexes to funds tracking the indexes isn't the only place where real numbers matter. Numerous people spend a lot of time trying to develop investment strategies that will consistently beat the indexes. If simply tracking- and matching performance with-the indexes must be studied carefully in the bright light of real numbers, imagine how dark some comparisons of new strategies and indexes could be. The CFA Institute publishes guidelines on how investment managers should report their performance track records. These guidelines specify that only real portfolios with real money that were really managed can be reported. Theoretical results may be interesting, but they are not a track record.
Index providers argue that running an index isn't as simple as it may look. Certainly those managing index funds will argue that tracking an index is harder than it looks. Data comparing indexes and active managers confirm that beating the indexes is also difficult-difficult enough that very few manage to do it with any consistency. All of which is a reminder that real numbers count.
Interested readers can find the index vs. active funds comparisons on www.indexes.standardandpoors.com; look for the SPIVA report-SPIVA stands for S&P Index vs. Active.