Measuring aggregate equity market performance is problematic. Prominent equity indexes provide perspectives that are either distorted by the effects of market-cap weighting or by measuring only selective slices of the stock market. There is not a single accurate measure of the U.S. equity market. If not used in context, equity market benchmark indexes potentially obscure more than they reveal.
Equity market indexes provide a perspective, but not a complete one. The S&P 500, for example, is often used as a surrogate for the U.S equity market. Clearly, however, it is an incomplete picture of the entire equity market. In recent years the value of the S&P 500 Index as a useful equity market barometer increasingly has been questioned (Blake, 2002). 1 'Is Time Running Out for the S&P 500?' Institutional Investor , May 2002, pages 52-64.The shortcomings of the S&P 500 Index are not unique. Most equity indexes measure only a certain segment of the total market, and thus they "report" only their particular perspective. Indexes that attempt to measure the total U.S. equity market (e.g. Wilshire 5000) typically do so by weighting securities according to their market capitalization. Market capitalization (i.e. market "cap") is calculated by multiplying the number of a firm's outstanding shares by the current market price per share. Thus the returns of firms with the largest market cap are weighted proportionally to their percentage piece of the total equity market-cap pie. In fact, the vast majority of equity indexes utilize a market-cap weighting protocol. (It should be noted that Wilshire Associates also provides an equal-weighted 5,000 index which has a distinctly different return pattern compared to the market-cap-weighted 5,000). The logic for market-cap-weighting an aggregate equity index is both rational and defensible. The performance of widely held stocks, such as General Electric, ExxonMobil, WalMart, etc., certainly impact a greater number of investors relative to thinly held securities. Nevertheless, even though market-cap weighting makes sense when attempting to measure equity market performance, it has essentially no value to most individual investors.
For example, an investor who purchased $2,000 worth of General Electric stock and $2,000 worth of Culligan stock will have no interest in a market-cap-weighting procedure to determine how his/her portfolio is behaving. Regardless of the market capitalization of each particular stock held in the portfolio, each holding represents 50% of the original invested value of his/her portfolio.
Ultimately, individual investors are only concerned about how their own portfolio is performing over any given period of time. Thus, it's entirely possible that any number of individual investors may be experiencing good results while the "market" ¾ as measured by any number of equity indexes ¾ appears to be performing poorly, or vice versa. It is, of course, impossible to provide equity market indexes that apply specifically to individual investors. As a result, certain equity indexes have evolved into benchmark barometers of the U.S. equity market, including the Dow Jones Industrial Average, Standard & Poor's 500, Russell 2000, Wilshire 5000 and others. In recent years, Morgan Stanley (MSCI) indexes have gained popularity. Vanguard, for instance, recently changed the benchmark index for its Midcap Index Fund from the S&P Midcap 400 to the MSCI U.S. Midcap 450. Likewise, Vanguard Small Cap Index changed its bogy from the Russell 2000 to the MSCI U.S. Small Cap 1750 Index.
This article examines the potential gaps that can exist between prominent equity benchmark indexes, the equity market they are attempting to measure and the aggregate performance of U.S. equity mutual funds. The intent of this analysis is not to disparage equity indexes, but rather to illuminate the challenges inherent in "measuring the market" via indexes.
The data in Table 1 present snapshots of the U.S. equity market at five annual intervals beginning in December 1999 and ending in December 2003. Each column of data represents results for the preceding 12-month period.
Different indexes tell dramatically different stories. As of December 31, 1999 , the mean return of 6,242 U.S.-based stocks (i.e. companies) was 42.7%. The median return was -3.9%, which tells us that 50% of the 6,242 stocks had a return lower than -3.9% in 1999 ¾ a very different picture than that painted by a mean return of 42.7%. The share-weighted return 2 in 1999 was 41.7%. The similarity of the mean return and the share-weighted return suggests that the mean return was heavily influenced by a relatively small number of large-cap stocks in 1999. This interpretation is supported by the 21.0% return posted by the Vanguard 500 Index and an even lower return (15.3%) from the Vanguard Midcap Index. The Vanguard Small Cap Index had a one-year return of 22.6% in 1999, a surprising result given the fact that 54.2% of all stocks had a negative one-year return in 1999.
The share-weighted return (as calculated by the author) assigns a greater weight to stocks that have more shares outstanding and less weight to thinly held stocks. For instance, General Electric had the largest amount of shares outstanding within the Morningstar database as of December 31, 2002. In fact, of the 465 billion total shares outstanding, GE had 9.95 billion, or 2.14% of the total. Therefore, the 2002 return for GE (-37.7%) was multiplied by .0214. This same calculation procedure was applied to all 6,004 stocks, resulting in a share-weighted average return of -26.0%.
The mean return of U.S.-based equity mutual funds in 1999 was 28.4% and the median return was 19.6%. The asset-weighted return of 29.2% (same concept as share-weighted return of individual stocks, but applied to mutual funds) was higher than the mean return, suggesting that equity performance in 1999 was better among larger funds. It is the case that larger funds (in terms of net assets) are funds that invest in large-cap stocks. Hence, equity returns in 1999 were primarily driven by large-cap stocks.
For the 12-month period ending on December 31, 2000 , the mean return of the 6,090 U.S. stocks in existence for the entire period was -5.1%. The median return was -13.8%. The share-weighted return (a technique similar to the market-cap weighted return ) was -1.5%. Thus, using three different aggregate metrics, the equity market performance ranges from -1.5% to -13.8%.
Addressing the question, "How is the market doing?" is complicated further by an appeal to equity indexes. At year-end 2000 the Vanguard 500 Index had a return of -9.1%, the Midcap Index posted a gain of 17.6%, the Small Cap Index lost 3.8% and the Total Stock Index fell 10.6%.
Using overall equity mutual fund performance as a barometer of the aggregate unweighted equity market is unreliable. At year-end 2001, for example, the mean return for 6,197 U.S. stocks was 18.1% and the median return was 3.5%. The mean return of 2,317 U.S. equity funds was -10.9% and the median return was even worse at -11.9%. The performance of equity mutual funds, as a group, was closely approximated by the S&P 500 in 2001 ¾ a reminder of the large-cap bias among equity mutual funds.
Of particular interest is the relatively consistent level of median return compared to mean and share-weighted returns. The latter two measures vary widely over the five years, whereas the median return exhibits far less volatility. During bull markets, mean and share-weighted returns tended to be higher than median return. Enter the bear market in mid-year 2000; by June 2001 both mean and share-weighted returns were hovering closer to median return. This convergence generally persisted through December 2002.
Bull markets, with their inherent excesses, more easily distort both mean and share-weighted returns, but median return remains largely unaffected. Major equity indexes (and their myriad clone funds) appear to be affected by the same forces which affect mean and share-weighted returns. This is not surprising inasmuch as share-weighting is very similar to market-cap weighting, and nearly all equity indexes utilize market-cap weighting in calculating their return.
In December 1999 the U.S. equity market was still enjoying, or at least riding on the tailwind of, the bull market of the 1990s. At least that is the picture painted by a prior one-year return of 21.0% for the Vanguard Index 500 (the most prominent S&P 500 Index clone fund). And yet at that same time, more than half of all U.S. stocks had a 12-month return of -3.9% or worse (half of the stocks have a return above the median and half below). Moreover, the Vanguard Total Stock Index (a Wilshire 5000 clone) reported a one-year return of 23.8%, despite the fact that 54.2% of all U.S. stocks had a negative return during the prior year. This is a clear example of the potential distortions that can be present in equity indexes that calculate portfolio return using a market-cap-weighting procedure.
Two periods that present an interesting contrast are the 12-month periods ending on December 31, 1999 , and December 31, 2001 . As already noted, more than 54% of all stocks had experienced a negative one-year return as of year-end 1999. The returns of major indexes (S&P 500, Midcap Index, Small Cap Index and Total Stock Index) were reporting strong, positive returns. Moreover, the mean return for all 6,242 stocks was a lofty 42.7%, and the share-weighted return was also strong (41.7%). The only sign of weakness was a median return of -3.9%.
By December 31, 2001 , a fundamental shift in equity performance had taken place. Three of the four major equity indexes were reporting negative one-year returns and the share-weighted aggregate stock return was -3.1%. Equity mutual funds had been hammered by year-end 2001.
Ironically, amidst all the "meltdown," the mean return for all 6,197 stocks over the prior one-year period was 18.1% and the median return was a positive 3.5% (the only time period in which the median return was a positive number other than 2003). Furthermore, the percentage of stocks with negative one-year returns was lower (46.9%) in the "bear" 2001 than in the supposed bull-market year of 1999 (54.2% of all stocks had a negative one-year return).
The revealing statistic behind this enigma is the 2001 share-weighted return of negative 3.1%. Despite a relatively robust one-year period in the broad equity market (as of December 31, 2001 ), the largest firms had not performed well. As a result, while equity "barometers" ¾ which are market-cap weighted ¾ signaled weakness, the equity market, on average, was having one of its stronger periods as measured by mean stock return (18.1%) and median stock return (3.5%).
Clearly, the equity market is too diverse to be measured accurately by one measure only. When assessing the performance of equities, it is vitally important to know "by what measure" the market is being evaluated. This research suggests that relying solely upon prominent equity indexes can provide a potentially distorted view of U.S. equity market performance.
|One-year Period Ending
||Dec 2003 |
|Mean Stock Return (%)
|Median Stock Return (%)
Stock Return (%)
|Number of Stocks which
Survived the Entire Period
|Percentage of Stock with
Negative One-year Return
U.S. Equity Index Funds
|Vanguard 500 Index
|Vanguard Mid Cap
Small Cap Index**
|Vanguard Total Stock Index
U.S. Equity Mutual Funds
Funds Return (%)
Fund Return (%)
Fund Return (%)
|Number of Funds which
Survived the Entire Period
|Percentage of Funds with
Negative One-year Return
The following are observations drawn from the results in Table 1.
- The differential between mean, median and share-weighted stock returns can be very large. The same is true of prominent U.S. equity index funds (i.e. equity market indexes).
- The impact of mid-cap and small-cap stock performance tends to have a relatively minor impact on indexes that attempt to measure the aggregate U.S. equity market (e.g. Vanguard Total Stock Index). This is due to market-cap weighting.
- The percentage of U.S. stocks that experienced negative one-year rolling returns was surprisingly consistent from 1999 through 2002, despite dramatic swings in performance as measured by mean return, median return, share-weighted return and the returns of prominent index funds.
1 Blake, Rich, 'Is Time Running Out for the S&P 500?" Institutional Investor, May 2002, pages 52-64.
2 The share-weighted return (as calculated by the author) assigns a greater weight to stocks that have more shares outstanding and less weight to thinly held stocks. For instance, General Electric had the largest amount of shares outstanding within the Morningstar database as of December 31, 2002. In fact, of the 465 billion total shares outstanding, GE had 9.95 billion, or 2.14% of the total. Therefore, the 2002 return for GE (-37.7%) was multiplied by .0214.
This same calculation procedure was applied to all 6,004 stocks, resulting in a
share-weighted average return of -26.0%.