| |
 May 2011 Not All Benchmarks are the Same But Does It Really Matter?
But is it possible to over-think this process? Are major benchmarks really that different? Investment professionals tend to think so. Most struggle to get the public to measure them by their benchmark of choice. Use a different one and they'll quickly point out your "mistake', calling into question any conclusions you may have drawn. This, too, can go to extremes. Case in point: Recently a mid-sized money management firm decided against purchasing equity benchmarking software that relied on the Russell series rather than the S&P counterparts. What made this interesting was the fact they were looking for index products to represent different capitalizations and styles, not actively managed portfolios. They felt the Russell series would render significantly different results than the comparable S&P. Are they right? On the one hand it wouldn't seem so. After all, the Russell 1000 and the S&P 500 are measures of large cap equity performance. Both are broadly based and have long track records. However, on the other hand, there are some notable differences between the two, so the only real way to compare them is to actually look at the numbers.
Construction Differences Both benchmarks are capitalization-weighted. This means the larger the stock, the greater its weight in the index. A stock's size or capitalization is simply the number of shares outstanding multiplied by its current market price. As a stock's share price goes up, so does its capitalization. Because not all stocks go up at the same rate, those that are rising fastest become more dominant in the index. To some investors, this is a weakness of capitalization weighted-indexes, essentially turning them into momentum benchmarks. Those who feel this way often advocate equal-weighted indexes which keep roughly the same percentage in each component. The benefit there is that an X% move in any one stock is equivalent to an X% move in any other. The downside is that in order to maintain this equality, equal-weighted indexes have to be constantly rebalanced to offset the effects of uneven trading. Capitalization-weighted indexes don't have this problem.
However even capitalization-weighted indexes are periodically rebalanced, and this is where the Russell 1000 and S&P 500 differ significantly. The Russell 1000, like its sister Russell 2000 and Russell 3000, is reconstituted once a year in late June. Additions and deletions are decided in the month or so preceding that date and then the index begins the coming year with the 1000 selections. So far so good, but then it gets interesting. Throughout the ensuing year, companies merge, sell out, or yes, even go bankrupt. When that happens, Russell removes them from the index but they don't' replace them until the next rebalancing in June. As a result, by the time June rolls around, the Russell 1000 may have substantially less than 1000 stocks. As of the end of February 2011 with roughly three months to the next rebalancing, the Russell 1000 was only 978 stocks. While this may seem odd it does makes sense in that this is truly an "unmanaged" index. It ends the year with the stocks that were chosen at the beginning (or less) without any changes in the interim. This is in stark contrast to the S&P 500. Like the other major S&P indexes, its constituents are determined by a committee that is charged with their construction. As corporate actions occur, stocks being acquired move out of the index and new ones replace them. Often when a stock leaves the S&P 500, the committee will tap its replacement from the S&P Mid Cap 400, and a replacement for the latter from the S&P Small Cap 600. In essence, a change in one can spark a ripple effect throughout all three indexes. It's not always moving up, either. Sometimes companies split themselves into two or more entities or will simply drop in market value. When this occurs, the committee can move them down to the S&P 400 or S&P 600 as appropriate. Not only does the S&P index committee replace or remove stocks throughout the year, they also have the ability to tilt the index between sectors. S&P divides domestic stocks into ten sectors so at any point in time each S&P index has a weighting in each sector. However if, for example, an industrial company is purchased by a rival and removed from the index, there is no requirement that it be replaced by another industrial. In fact, in recent years, its very likely not to be. This was particularly noticeable in the late 1990s when tech stocks were leading the market and quickly becoming large caps. Analysts criticized the S&P index committee for managing a supposedly unmanaged index by adding hot tech and telecommunication stocks at the expense of old line industrials and basic materials producers. With tech stocks leading the market, this definitely had an impact on the index's overall performance. It also had a major effect when techs imploded a few years later, too.
Then there's the issue of the actual stocks themselves. Obviously the Russell 1000 has many more -- in fact twice as many -- components as the S&P 500. (That's at least true in June, see above.) One might immediately conclude from this that each stock in the S&P 500 has more of an impact on the index than in the Russell 1000, but it's not that easy. Because capitalization plays a role, the largest stocks in each index will dominate returns, so it's more of a question of the mix of stocks rather than their number. The mix between the Russell 1000 and the S&P 500 is interesting, too. Russell's index for the broad domestic equity market is the Russell 3000 which is composed of the Russell 2000 a measure of small caps and the Russell 1000 a measure of large -- and mid caps. That's right, stocks considered mid caps (generally with market caps of $2 billion to $7.5 billion) are also part of the Russell 1000. S&P has a separate mid cap index, the S&P 400, for mid caps. When it's combined with the S&P 500 and the Small Cap 600, you end up with the Composite 1500, S&P's index for the broad U.S. equity market. Although the S&P 500 sounds like a purer large cap index than the Russell 1000, it's smallest component (as of February 28, 2011) has a market cap of just $2.08 billion which is actually smaller than the Russell 1000's smallest of $2.31 billion. And the Russell's midcaps have more limited effect than you might at first think. Its average weighted market cap isn't that much different from the S&P's ($83 billion vs. $93 billion, respectively) despite the fact the Russell's median ($5.8 billion) is firmly in the mid cap range while the S&P's is roughly two times greater ($11.4 billion). In short, the two indexes have some striking differences which would lead one to think they would make significantly different benchmarks. How could they not? Well, let's take a look.
Separated at Birth Over the common period of January 1978 through February 28, 2011, a $100 investment in the Russell 1000 would have grown to $3,579 while a similar investment in the S&P 500 would be worth $3,526. That's right, a whopping $53 dollar difference! What's up with that? One way this could occur would be if it was simply a odd coincidence. The two indexes could wildly diverge for most of the 33 years only to cross one another's path in 2011. Well they could, but they didn't. Chart 1 shows not only the endpoint for the period, but also how the two indexes tracked one another on the way. Although it may be difficult to see, there really are two lines on Chart 1. The green is the Russell 1000 while the red is the S&P 500. Correlation over the entire 33 years was a remarkably high 0.9964. Correlations run from mirror images of -1.0 to perfect lockstep of 1.0. The correlation between the Russell 1000 and S&P 500 is about as close to lockstep as you can get and still fall short. And it's not like the correlations have been lower in the past. Chart 2 shows the rolling 5-year correlation between the Russell 1000 and the S&P 500. As you'll notice, the line is basically horizontal around 100% (or 1.0 in the definition above). Again, that shows a high correlation remaining constant over time. What about various time periods? Well, there's not much difference there, either. Chart 3 has eight different time periods (all ending February 28, 2011) as well as the overall cumulative returns. As you'd expect from such highly correlated indexes, the results are virtually identical. The longer the time period, the more similar they become.
The fundamentals are comparable, too. The S&P is a little pricier based on earnings (P/E 17.1x vs. 16.3) while the Russell 1000 looks more expensive from a Price-to-Book approach (2.31 vs. 2.08). In both cases, the ratios are quite similar. Risk, as measured by standard deviation (17.39 Russell vs. 17.30 S&P), semi-standard deviation to zero (11.98 vs. 11.87) and Sharpe Ratio (.2056% vs. .2068%) also show little difference. So any way you look at it (with the exception of their actual composition), these two indexes are essentially identical. Not only that, they have been for the past 33 years. That's certainly not what you would have expected.
Under the Cap
According to data from Russell Investments, at the end of 2010, the Russell 1000 captured 92% of the U.S. market capitalization while the S&P 500 represented 76%. That's pretty impressive, but what's more so is how top-heavy the capitalizations are. The Russell Top 200 captures 67% of the market cap while the Russell Top 50 covers 42%. Because the top 50 stocks in both the Russell 1000 and the S&P 500 are essentially identical, both represent over 40% of the overall market and it's these stocks that make the benchmarks behave similarly. These largest of the large caps don't go bankrupt (well, if you ignore GM, Worldcom and Enron) or merge themselves out of existence like the smaller stocks that are periodically eliminated from the Russell 1000 or are replaced in the the S&P 500. Their values don't fluctuate nearly as much as smaller company shares and the membership in the top 50 is relatively stable. No wonder the two indexes are so highly correlated. There are some important lessons to take away from this. First, broad-based indexes like the Russell 3000 and the S&P Super Composite 1500 may not be a good a measure of the overall market as they first appear. Sure, they have representation from mid and small caps as well as large caps, but they're still so dominated by the largest of the large caps, the mid and small stocks really don't have an effect. To truly benchmark their behavior, investors either need to use capitalization-specific indexes (e.g. the Russell 2000 or the S&P 400), or equally-weighted versions of the broader indexes. Secondly, the large cap indexes aren't even the best benchmarks for the universe of large cap stocks. If the top 50 stocks of the S&P 500 control the benchmark, the other 450 are left out in the cold. Yes, a few may fall into the mid cap or small cap range, but the vast majority of those other 450 do have market caps over $7.5 billion. Again an equally-weighted index is probably the solution here. And the final lesson, perhaps the most important one for investors, is not to simply select a benchmark by its name. Clearly the construction methods of the Russell and S&P indexes are different, so obviously you'd think the results would be. As you've seen, however, they aren't. Managers who insist on being compared to "the right" index are more often than not, just blowing smoke. In reality, any cap weighted large cap index will probably yield similar results. The Russell 1000 and the S&P 500 have some distinctive differences, but in the long run it's all the same. Search this site! Just enter you key word or words:
Get current quotes or follow your own custom portfolio,
courtesy of E-Line Financials:
|
||||||||||||||||||||||||||||
ONEY MANAGERS WANT TO BE measured by an appropriate benchmark. Peer groups are constantly changing and are subject to survivorship bias. Riskier portfolios may often have an advantage over more conservative ones, so risk-adjusted evaluations may be more appropriate. Different asset classes exhibit differing characteristics, so performance yardsticks should be compatible with the subject portfolio. All of these are legitimate issues and are frequently considered when choosing the most appropriate benchmark.
Source: Ibbotson Associates
Source: Ibbotson Associates
Source: Ibbotson Associates
Data Source: Russell Investments/S&P
