The Russell Pure Style Indexes: Tools for the next level of style investing
- The Russell Pure Style Indexes are designed to provide focused exposure to the value and growth segments of the U.S. equity market.
- The Pure Style Indexes methodology extends that of Russell’s standard Value and Growth Indexes.
- While aligned with the standard style indexes, the Pure Style Indexes have increased exposures to value and growth style factors, which resulted in greater distinction in returns, in historical simulations.
On April 7, 2015, Russell Indexes launched the Russell Pure Style Index Series, which builds on Russell Indexes’ long history of index innovation. The index series has Russell’s standard style methodology at its core and is further enhanced to offer a more concentrated expression of the growth and value investment styles. Covering all capitalization components of the Russell 3000® Index, the series comprises:
Russell Top 200® Pure Growth Index
Russell Top 200® Pure Value Index
Russell Midcap® Pure Growth Index
Russell Midcap® Pure Value Index
Russell 2000® Pure Growth Index
Russell 2000® Pure Value Index
Russell 1000® Pure Growth Index
Russell 1000® Pure Value Index
While the standard Russell Value and Growth style indexes can be used as performance benchmarks or as asset class proxies, the Pure Style Index Series provides tools that allow more precise refinement of equity exposures.
In the first part of this paper, we review the uses of style indexes and Russell’s motivation for creating these new Pure Style Indexes. We also briefly touch on the methodology behind the Pure Style Indexes. In Part 2, we examine the characteristics and factor capture of the new indexes, contrasting them with standard style indexes. We also examine the behavior of the Pure Style Indexes over style cycles.
Style index uses
Russell style indexes have been the primary benchmarks for use by style investors ever since the launch of Russell’s first value and growth indexes in 1987. As of year-end 2014, more than half of all U.S. institutional equity products are benchmarked to a style index, and of those products, 99 percent are benchmarked to a Russell style index.1
From their inception, style indexes have mainly been used in two ways. The first use is to serve as benchmarks for two types of active managers, who are fishing in distinctly different waters. Roughly speaking, one type of active manager searches for “bargain” stocks, where perceived potential value is most often defined by low price-to-book or price-to-earnings ratios. Another type of active manager looks for stocks with high expected earnings growth and sales trends. The different approaches to stock picking are known as value and growth styles, respectively, and the groups of stocks associated with these styles can reasonably be distinguished by stock characteristics. Accurate measurement of styles with style indexes has both led to increased awareness and validation of styles and has fueled the rise of style investing.
The second main use of style indexes is to support institutional investors’ strategic asset allocation processes. After a pension plan sponsor, for example, decides on an overall percentage allocation to equities, there is usually another step: deciding on the segmentation of the equity allocation into “core” investments (a category with no overall style or cap bias), and then into value, growth, small-cap, mid-cap and large-cap investments. Indexes are used as proxies for the market segments in asset allocation analysis. In that context, value and growth indexes are typically constructed to be modular and to split the overall market into complementary components that neatly roll up to the overall broad market index.
A simple “knife-edge” dividing line between the market’s value and growth segments would result in some stocks near the dividing line being arbitrarily classified as (for example) value, even though their characteristics were very similar to those of stocks classified as growth, just on the other side of the dividing line. Stocks near the line might actually be held by both value and growth managers. As we discuss in more detail in our methodology overview, this problem is elegantly solved by Russell’s nonlinear probability (NLP) algorithm, which allows stocks with neither strong value nor strong growth characteristics to be included, with diminished weight, in both indexes. This scoring system successfully accomplishes the goals of modular and all-inclusive style indexes. Russell style indexes continue to be the primary benchmarks for active value and growth managers in the U.S. institutional market place, and they continue to attract index users seeking exposure to broad segments of the market with style tilts.
The next level of style investing
The modular approach of Russell’s standard style methodology leads to an intentional overlap of roughly 30 percent of the total market value weight in the two indexes (Figure 1), which results in a reduction in the strength of the style exposure. This may not suit the needs of all investors.
While standard style indexes are useful for benchmarking and strategic asset allocation, some investors want to have even greater control over style exposures within their portfolios. They seek to more precisely construct portfolios by striking a balance between growth and value segments with large-, mid- and small-cap exposures. Dynamic and tactical tilting may also be part of their strategy. They want indexes to support their decisions with exposures that are sharp, focused and non-overlapping.
With its creation of the Russell Pure Style Indexes Series, Russell Indexes has responded to this demand. These new indexes include only stocks with purely growth and purely value characteristics. There is no overlapping (Figure 2), and stocks are weighted according to their relative style attractiveness.
Table 1 highlights the differences in stock selection rules between Russell’s standard style and Pure Style indexes.
The Pure Style indexes have been further differentiated by cap size. The well-known Russell 1000® Index comprises mega- and mid-cap stocks. To ensure that there is no overlapping between indexes in the new series, we employ the Russell Top 200® Index universe for our Top 200 Pure Style indexes, and the Russell Midcap® Index universe for our Midcap Pure Style indexes. The Top 200 universe had about 68% of the market value of the Russell 1000, while the Midcap universe had about 32% of the market value of the Russell 1000 as of the 2014 reconstitution. The Russell 2000® Index universe is employed for the small cap Pure Style indexes. Taken together, as shown in Table 2, there are six Pure Style indexes drawn from universes without any gaps or overlaps (two styles by three cap tiers).
Two additional Pure Style indexes are based on the Russell 1000 universe, which, again, encompasses the Top 200 and Midcap universes. Thus there are eight Pure Style indexes in all.
While we have touched on some of the key aspects of both the standard and the Pure Style methodologies, in the next section we expand on them, so the reader may get a deeper understanding of Russell’s index methodology.
The overall objective of the Pure Style methodology was to produce a series of concentrated style indexes, where the constituent companies of each index are either fully growth or fully value. How is this done? To explain, we first present an overview of the style methodology used in the construction of Russell’s standard style indexes. From there, we look at how the Pure Style methodology builds on this – and then extends it – to create a comprehensive series of more concentrated, “pure” style indexes. While the standard style indexes are the best representation of a given stock universe, the Pure Style methodology narrows that universe and delivers a tailored exposure to only those stocks considered to be either fully value or fully growth.
Starting from the basics: An overview of Russell’s standard style methodology3
As mentioned, value and growth managers tend to construct their portfolios from different initial groups of stocks. These groupings can be distinguished by various stock characteristics, such as valuation ratios or growth metrics. As of 20154, the stock characteristics used by Russell Indexes for measuring value and growth styles are:
- Book-to-price ratio (value variable)
- I/B/E/S forecast medium-term EPS growth (growth variable)
- Historical sales per share growth (growth variable)
For each variable, stocks of the parent index are sorted on the variable and assigned a score ranging from zero to 1, representing the strength of value exposure indicated by the variable. For example, a book-to-price score of zero represents the greatest growth (or least value) exposure within the group, while a score of 1 represents the greatest value exposure within the group.
Given the scores for each variable, the methodology combines the scores in a weighted average to form a composite value score (CVS), which also ranges from zero to 1. Again, a CVS of zero represents the greatest combined growth exposure among stocks of the parent index, while a CVS of 1 represents the greatest combined value exposure within the group. This sorting and scoring is done separately for stocks of the Russell 1000 and Russell 2000 indexes. CVS scores from the Russell 1000 are used for the Russell Top 200 and Midcap Pure Style indexes.
The CVS measures the exposure of a stock to value or growth. The final step of the standard style methodology is to divide the parent index’s constituents into equal market value halves, after sorting on the CVS. Instead of simply separating the constituents into two parts, above and below a breakpoint, the Russell methodology employs a nonlinear function to assign a probability to each stock, where a probability of zero indicates membership wholly in the growth index, while a probability of 1 indicates membership wholly in the value index. The probability, in turn, is multiplied by the number of shares (and thereby the market value) of the stock in the parent index, so that each stock is fully apportioned between the value and growth indexes. An example of this function is shown in Figure 3. The function is calibrated so that roughly 35% of the market value of the parent index is wholly in the growth index, 35% is wholly in the value index, and about 30% is divided between the growth and value indexes. This middle range represents stocks that might reasonably appear in both value managers’ and growth managers’ portfolios. Conversely, value managers’ portfolios are not likely to hold stocks where the probability is zero, and growth managers’ portfolios are not likely to hold stocks where the probability is 1.
Creating a “pure” style index: Extending the traditional style methodology
The Russell Top 200, Russell Midcap, Russell 1000 and Russell 2000 indexes are the parent indexes from which candidate Pure Style constituents are selected. Except where noted, the methodology described below is applied to all four market capitalization segments.5 The Pure Style methodology addresses issues of liquidity, pure style exposure, constituent style weighting and index-level sector concentrations.
Pure Style indexes are more concentrated than their standard style index counterparts, and stock liquidity is a concern. To address this, Russell utilizes a four-week average daily dollar trading volume (ADDTV) to measure liquidity. An ADDTV cutoff is established such that ADDTV must be greater than or equal to the bottom 10% of the Russell 2000 Index when sorted by ADDTV. All candidate constituents, for any of the Pure Style indexes, must meet this minimum ADDTV requirement.
Pure style exposure
Once a pool of liquid candidates is created from the parent index, the style probability assignments from the standard style indexes are used to further narrow the list. Only those stocks having a value probability of 1 are considered for inclusion in the Pure Value Index, and only those stocks having a growth probability of 1 (equivalent to a value probability of zero) are considered for inclusion in the Pure Growth Index. The group of stocks with a value probability of 1, as shown above in Figure 3, corresponds to roughly 35% of the candidates in the parent index. The candidate list is then further reduced by half, retaining the half with the highest CVS scores for the Pure Value Index, or the lowest CVS scores for the Pure Growth Index.6
Assigning style weights
In order to further focus the exposure of the index on value or growth, the selected stocks for the Pure Value Index are weighted in proportion to their CVS, and the selected stocks for the Pure Growth Index are weighted inversely to CVS. It is easiest to grasp the weighting scheme in a picture. The weights for Russell 1000 Pure Value and Pure Growth Indexes, as of the 2014 reconstitution, are depicted in Figure 4. This illustrates that the Russell 1000 Pure Value Index constituent weights have a clear, direct relationship to the constituent CVS scores. For comparison, the weights of the Russell 1000 Value Index are shown. Clearly, the Pure Value Index concentrates the exposure on those stocks with the highest value exposure, as measured by the CVS score. Similarly, the Pure Growth Index concentrates on the highest growth exposure.
Managing sector exposures
Alternatively weighted indexes may have notable sector deviations from the parent benchmark. To ensure that index characteristics are driven by intended exposures, rather than by sector biases, the Russell Pure Style Indexes methodology incorporates sector constraints. Simply put, a Pure Style Index may not have a sector deviation greater than 10% from its parent style index’s sector weight. Any weight in excess of 10% is redistributed pro-rata across the constituents of the unconstrained sectors.
The net result of the above methodology is to create a more focused exposure, while maintaining adequate liquidity and diversity in a tradable index. The minimum liquidity requirement ensures that the least-liquid stocks of the Russell 3000 Index universe do not enter these concentrated portfolios, while the sector caps restrict overconcentration in sectors.
The inherent reduction in diversification due to concentration is also partially mitigated by the weighting scheme. All Pure Style Indexes have far fewer actual stocks than their standard style counterparts, in order to maintain highly concentrated style exposure. All else equal, that would make Pure Style Indexes far less diversified than their standard style counterparts. However, all is not equal, due to differences in weighting schemes. The standard style weights are based on capitalization levels, which tend to be top-heavy, resulting in the largest stocks often dominating the returns. The Pure Style weights are based on style scores, and by the index construction, only those stocks with high scores are included. The resulting indexes are less top-heavy compared to capitalization-weighted, which partially (or in some cases, completely) offsets the reduction in diversification from having fewer actual stocks. See Appendix A for an analysis of this effect.
We’ve described the standard Russell Style methodology and its role in the Pure Style methodology. In Part 2 we review the resultant style factor capture, performance and portfolio characteristics that these indexes would have exhibited over time.
Concentration of the value and growth investment styles is a primary objective of the Pure Style Indexes. By construction, these indexes have greater exposure to value and growth as measured by the composite value scores. We would expect to confirm greater exposures to value and growth factors by using factor models and other metrics commonly employed in the industry. We demonstrate that this is the case when using a standard set of portfolio characteristics and two different factor models.
— Download the complete paper —