This month’s KnowRisk topic may, at first, seem very simple. After all, even the most inexperienced investor knows what an index is. Checking to see how “the markets are doing” based on the movements of one of the popular stock indexes is water-cooler conversation in offices around the world.
But while all investors are familiar with indexes as a concept, how many truly know what is being measured by the various indexes that they consult for a snapshot of the overall markets? Is the index that investors use as a convenient measurement against which to judge their own portfolios’ performance the best available gauge of how those investments are doing? How do various indexes that are presumed to describe similar market segments really compare to each other, or to the segments they purport to measure?
This article seeks to answer these, and many other basic questions. We will discuss how indexes are constructed, the use and misuse of various indexes, the ways in which various indexes can overlap in unexpected ways, and other topics that will help you to understand and properly apply these important measurement tools.
Kicking the Tires
If you were asked to explain the concept of a car to someone who had never seen one, your task might seem straightforward. After all, there are certain key features that all cars share: a passenger compartment, wheels, axles, an engine, transmission, etcetera. With this description, your friend should be able to determine whether or not the next vehicle he sees is a car without too much trouble. But nothing about your description could help your friend determine whether the car in question is a huge SUV, like a Cadillac Escalade, or a tiny hybrid, like a Toyota Prius. In other words, your description will help your friend know that a car is not a bicycle, but it won’t help him to figure out what kind of car he should buy.
Similarly, we could probably all, if pressed, provide a general definition of what an index is. Something like: indexes are a statistical measurement derived by aggregating the change in value of a selected group of securities over time. And we could all probably name a few indexes off the tops of our heads: the Dow Jones Industrial Average, the S&P 500, the Russell 1000, the NASDAQ, and foreign indexes like Britain’s FTSE, Japan’s Nikkei, and Germany’s DAX.
However, just as was the case with cars, it is a mistake to think that defining the idea of indexes is the same thing as understanding how each is constructed, or that all of them are equally useful for any given application.
What Did the Market Do Today?
When talk around the aforementioned water coolers turns to how “the market” is doing, it’s a good bet that everyone understands this “market” to mean one of two indexes: The Dow Jones Industrial Average (“the Dow”), or the S&P 500. These are, by far, the most widely reported figures related to the financial markets, with their movements discussed endlessly in print and on television. But what, exactly, do these two indexes measure?
According to its parent company, Dow Jones, the Dow Jones Industrial Average includes 30 (though the real figure is currently 35) stocks selected by the editors of The Wall Street Journal, with the aim of “represent[ing] large and well-known U.S. companies.” Recent upheavals in the markets necessitated the removal and replacement of two of the Dow’s components, when Citibank and General Motors were replaced with Travelers Companies, and Cisco Systems, respectively. Normally, though, the lineup of stocks included in the Dow is very stable, and represents many of the most familiar companies in the U.S.
The other measurement people often refer to as the “market” is the S&P 500. While the Dow measures a select set of large-cap equities, the S&P includes 500 (actually, it can include as many as 507 — as of this writing, there are 504) stocks at all levels of capitalization, and therefore offers a considerably broader view of the market than does the Dow. In addition, the S&P is, unlike the Dow, a capitalization-weighted index. This means that the top 25 stocks by capitalization in the index are responsible for more than 50% of the overall index’s return.
While these two indexes tend to move in the same direction (their 3-year correlation is 0.98, their 5-year correlation is 0.97, and their 10-year correlation is 0.94), that correlation does not mean that the indexes’ movements have the same duration or magnitude. In addition, their composition is clearly very different. Those differences are only rendered more starkly when we compare the relative capitalizations and the style of their component companies.
In the charts above, we can clearly see that in addition to the huge difference in number of components that make up each of these indexes, there are significant differences in terms of the size and types of stocks they include. The most obvious difference stems from the fact that, while the Dow includes no mid- or small-cap stocks, fully 14% of the S&P at the time of this measurement was made up of stocks in these capitalizations. Moreover, while the S&P’s components are roughly equally distributed among value, core, and growth stocks, growth stocks represent less than 4% of the Dow, while core equities make up nearly two-thirds of the index.
Turning to the relative weightings that the Dow and the S&P have to each economic sector, we find a number of other significant disparities. Simply consider the 40.1% weighting that the S&P allots to the Service Economy, and compare it to the 28.3% weighting that the Dow assigns to the same sector. Digging deeper, we see that within the Manufacturing Economy, the Dow’s weighting in the Industrial Materials sector is nearly twice that of the S&P. Finally, the Dow includes no stocks in the utilities sub-sector at all, while they make up nearly 4% of the S&P.
Clearly, there are significant differences in the composition of these two indexes, each of which is used as shorthand for how “the market” is doing. But on the other hand, in spite of the public’s tendency to think of the movements of the Dow and the S&P as two versions of the same idea, these indexes were never intended by their designers to represent the same snapshot of the markets. What happens when we consider two indexes that are focused on the same market segments?
But Surely “Value” Indexes Are Basically the Same?
At first, it seems reasonable to assume that the various capitalization and style-specific indexes put out by different companies would be essentially the same. These more specific indexes include many that investors and advisors depend upon as benchmarks against which to judge the performance of specific managers. For example: if an investor is considering an allocation to a large-cap value manager, then he and his advisor would want to compare the performance and characteristics of the managers they scrutinize against those of a large-cap value index. But which large-cap value index?
Surely, as you move from broad-market indexes toward more specifically targeted indexes that focus on a particular capitalization and style, there should be increasing agreement about how best to represent those market segments in an index, right?
Here again, our chart demonstrates that there are significant differences between two of the most popular large-cap value indexes, the Russell 1000 Value and the Dow Jones Large-Cap Value. Perhaps the most obvious and surprising difference is between the two indexes’ capitalizations. The Dow Jones Large-Cap Value index is almost entirely composed of large-cap, with a small portion allocated to mid-cap stocks. This mid-cap allocation is to be expected, as some of the stocks in the index will rise into large-cap status, or fall into mid-cap status, as their business improves or falters. More interestingly, though, we can see that the Russell 1000 Value has a small but significant allocation in the small-cap range, which may surprise investors who are looking to this index as a large-cap value benchmark.
This comparison is not intended to characterize one benchmark or the other as “superior.” However, for benchmarking purposes, it is entirely reasonable to think that they are unlikely to be equally applicable to any given manager’s or investor’s portfolio. For example — if you are considering a manager whose portfolio includes 45 stocks, and you are benchmarking him against the 500+ stocks in the S&P 500, how reasonable is that comparison? Are the sector weightings of the benchmark truly a good match for the weightings in an investor’s portfolio? Many investors would be surprised to learn that the benchmark against which they are measuring their returns allows weightings that would violate their portfolio’s investment policy.
Building Custom, “Blended” Indexes
Often, it is the case that there is no single index that provides a good benchmark for an investor’s or institution’s composite portfolio (that is, the diversified portfolio that results after the asset allocation and portfolio construction process have been completed). After all, most sufficiently diversified portfolios include not simply equities of all capitalization levels, but also fixed income investments and alternative investments of various kinds. Clearly, measuring such a portfolio against an equity-only index will produce a skewed view of its relative performance. For this reason, in investment consulting, we often create customized “blended” indexes that provide as close to an “apples-to-apples” comparison with the portfolio in question as possible. These combine the characteristics of several indexes that are each the best available match for a given portfolio segment, using relative percentages that mirror the appropriate allocations in the investment portfolio.
Below, you will find two charts that help to illustrate the construction of blended indexes. The chart on the left illustrates the asset allocation of a sample investment portfolio. In the absence of an established benchmark index that includes all of these asset classes in a similar ratio to that of the portfolio, a consultant might construct a benchmark index similar to the one pictured on the right. Note that, for each segment of the portfolio, the benchmark index includes a corresponding percentage of a comparable index.
With this blended index, we seek to determine, as closely as possible, how well the overall portfolio is performing in comparison with an index that closely replicates its characteristics.
But Are You Truly Blending?
Earlier, we touched on the fact that some indexes include some surprising components — recall that the Russell 1000 Value, which is ostensibly a large-cap index, includes some small-cap stocks. Knowing, then, that an index’s name may not always provide a complete description of its components and construction, it may not be as simple as we might hope to create a truly diversified blended index. If you choose two indexes that would seem to cover completely different capitalization levels, for example, you may find a surprising amount of overlap between the stocks included in each. And if two or more of the various components of your blended index cover the same securities, then you may not receive the benefits that you were seeking by blending these indexes in the first place.
Equitas conducted an in-depth study of the Russell family of indexes, to determine exactly how much diversification could be achieved by combining them in various ways. The results were both surprising and illuminating. The graph below, which expands the traditional 9-box graphs we used earlier in this piece to include extremes of style (Deep-Value, High-Growth) and capitalization (Giant, Micro), is designed to illustrate the complete range of purchasable equities available in the U.S. market. Over this matrix, we have illustrated all of the stocks included in a given index, as plotted by their combination of style and size, at the time of measurement (in this case, 2006). Below, we see that the blue circle represents the Russell 1000, which is considered a core large-cap index. The tan circle represents the Russell Mid-cap, which is also a core index.
As you can see, the Russell Mid-cap index ranges up into the large-cap space, while the Russell 1000 dips deeply into the mid-cap space. Overall, these indexes feature a 28.9% overlap. In other words, nearly a third of all stocks included in one index are also included in the other. Anyone who constructs a blended index that uses both of these indexes is experiencing considerably less diversification among his large and mid-cap benchmarks than he might otherwise assume.
What if we focus on indexes that cover the same capitalization level, but different styles of stocks within those capitalizations? Specifically, let’s look at three Russell Mid-Cap indexes: the core-focused Mid-Cap Index we saw in the previous graph, alongside the Russell Mid-Cap Growth and the Russell Mid-Cap Value. What we might expect to see, based on these indexes’ titles, is a series of circles, each centered on the midpoint of the appropriate style column. What we see instead, though, is a series of ovals that extend far into the adjoining style columns, and produce a surprising degree of overlap with each neighboring index.
Consider, based on these figures, how difficult it might be to achieve a truly useful blended benchmark for a portfolio that features both a core and a growth allocation in the mid-cap space using only Russell mid-cap benchmarks? After all, a huge percentage of the performance of each of these indexes is accounted for by precisely the same stocks. Obviously, knowing how much overlap there is between the indexes used to construct a blended benchmark is vital.
Exchange Traded Funds
Exchange Traded Funds, or ETFs, have become increasingly popular in recent years as vehicles through which investors can access the diversification of various indexes without incurring high management fees. Of course, this also means that investors who put their money into ETFs cannot depend on the experience and acumen of a manager with a talent for picking stocks — they are entirely subject to the movements of the components of the index. Moreover, while many investors assume that investing in an ETF means that they have invested in 100% of an index, the usual figure is closer to 80%. There are a number of reasons for this discrepancy, including liquidity issues: given the enormous sums that some ETFs have available to invest (in certain cases, in excess of $1 trillion), it can be impossible for them to invest in some of the smallest stocks in many indexes without artificially impacting the price of the stock in question.
Even with these caveats, though, ETFs can be an excellent tool for investors who seek exposure to alternative investments that might otherwise be hard to access. For example, what if an institution wished to have exposure to a natural resource, like timber, in its portfolio? Firstly, such an institution might consider a Real Estate Investment Trust that is focused on timber. But there’s also the option of investing in an ETF that buys timber-associated investments (paper companies, for example). These ETFs can offer exposure to an attractive sector for a comparatively low price.
Just as with any index, investors who are considering ETFs must ensure that they are fully aware of what, exactly, they are buying. If you are interested in gaining exposure to the growing “smartphone” market, you must keep in mind that these devices are considered “telecommunications”, and not “technology”, in spite of the fact that they sport many of the same features as computers. On the other hand, a “technology” ETF is likely to include software and peripherals manufacturers, in addition to chip and computer makers. Here again, truly understanding the index is vital to wise investing.
Fixed Income Indexes
When an investor purchases a bond, the principal underpinning the transaction is fairly simple: the government, corporation, or other entity that issues the bond promises to repay the money that the investor has loaned out by a specific date, and in the interim, the issuer will pay interest, usually twice per year. In essence, all that the investor in this scenario must concern himself with is whether or not the issuer will remain solvent, and be able to pay both the interest and the principal amount on schedule.
Of course, most fixed-income investors do not have such a simple relationship with their bond investments. Here, as with stocks, most institutional and individual investors achieve exposure to fixed income through either index funds or active managers. In situations like these, investors look to the characteristics of the overall portfolio (credit quality, maturity range, duration, etc.), in comparison with one of the standard fixed income indexes.
Again, just as with equity indexes, it is vital to know whether the composition of the portfolio in question is truly comparable to the benchmark index. For example, it is entirely possible to invest in an intermediate fixed income portfolio that has heavy weightings in bond classes that are simply not included in the benchmark. For example, a manager might decide to shift 66% of his portfolio into mortgage-backed securities, while the index against which he is measuring his performance contains little to no exposure to this sector. To draw a rough analogy, it’s as if an equity portfolio manager suddenly shifted 2/3rds of his portfolio into the tech sector. While this sort of tech sector overweighting might be possible in a small-cap growth portfolio, the distinction between styles of fixed income investing are not always so obvious to many investors.
Conclusion
Just like any tool, indexes can be used properly or improperly. Proper use of indexes as a benchmark against which to measure an investment’s characteristics and returns can help to determine which among a number of similar portfolios is the best choice for your investment goals. Choosing the wrong index, or the wrong combination of indexes, can obscure potential risks, or deliver apples-to-oranges comparisons that only make the process of evaluating investments more difficult. This is why it is vital to discuss your benchmarking options, along with all other facets of your investment strategy, with your advisor.
