Russell High Efficiency Factor Index Series Providing investors with efficient exposure to return drivers James Barber, CFA i ; Scott Bennett ii ; Mark Paris, CFA 1iii Recognizing the need is the primary condition for design. Charles Eames Introduction In 2014, Russell Indexes celebrates 30 years of smarter beta. In the decades since the 1984 launch of the world s first small capitalization index the Russell 2000 Russell has continued to be a leader in the industry with the development of innovative indexes that better define markets and their relevant sub-components. Russell has pioneered style benchmarking with the Russell Value and Growth indexes and, more recently, the Russell Defensive and Dynamic indexes. We are now pleased to announce the launch of the Russell High Efficiency Factor Index (HEFI) series, which builds on our heritage of innovation. The Russell HEFI series combines over 30 years of insights into style- and factor-based investing to give investors a comprehensive set of tools and building blocks to manage their total portfolio outcome. The Russell HEFI series uses a consistent, factor-based weighting methodology to provide exposure to commonly identified and utilized factors: Low Volatility, Momentum, Quality and Value. Our proprietary methodology provides strong factor capture via indexes that are active risk aware, investable, and low turnover. The HEFI series is offered within six Russell large cap universes, namely, Global, Developed, Developed ex-u.s., U.S., Developed Europe and Emerging Markets. The consistent methodology utilized across the Russell HEFI series offers a unique advantage to investors who are looking to combine exposures. In this paper we provide a detailed overview of the Russell HEFI series, outlining the benefits and insights of our innovative and consistently applied methodology and detailing the four factors that make up the series. Finally, in the appendix, we describe the investment beliefs underlying the exposures to each of the factors and provide descriptive information for each of the indexes. The Russell HEFI series combines over 30 years of knowledge regarding style and factor based investing. 1 The authors would like to acknowledge the significant contributions of the following Russell associates in the development of the Russell High Efficiency Index series: Nicole Bahr, Guillermo Cano, David Carino, Mary Fjelstad, Evgenia Gvozdeva, Bryson Hirai-Hadley, Sarah Mars, Sean Smith and Pradeep Velvadapu. Russell Investments // Russell High Efficiency Factor Index Series APRIL 2014
Russell High Efficiency Factor Indexes: Overview The Russell HEFI series steps away from traditional capitalization-based weighting, as used in benchmarks, to factor-based weighting which starts with each stock s benchmark weight and adds an active weight based on the stock s factor score. The Russell HEFI series utilizes Russell s market-tested, non-linear probability (NLP) algorithm 2 in a fresh way to deliver a robust factor-based weighting methodology. High efficiency references the ability of the Russell HEFI series to give investors: Significant exposure to the underlying factor; Active risk awareness; Low turnover; High capacity; Low levels of stock-specific risk; Moderate tracking error; Meaningful active share levels; Full transparency, and The ability to combine exposures In developing the Russell HEFI series, we focused on identifying equity market factors that were relevant, comprehensive, universally robust, persistent and implementable. In determining factor specifications, we relied on our extensive capital market insights and drew on our heritage in researching active managers, constructing multifactor portfolios and designing market-leading indexes. Further, we ensured that all of our factor specifications were consistent with academic research findings and empirically relevant using industrystandard risk models. 3 Table 1 details the factors and the underlying variables used. 4 A detailed discussion of all four factors can be found in the appendix. Table 1: HEFI Series Factor definitions INDEX Russell High Efficiency Quality Index (HEQI) Russell High Efficiency Low Volatility Index (HELVI) Russell High Efficiency Momentum Index (HEMI) Russell High Efficiency Value Index (HEVI) UNDERLYING VARIABLES Return on assets Debt to equity 5-year earnings variability 52-week total return volatility 60-month total return volatility 11-month total return, lagged 1 month Book/price ratio Earnings/price ratio In developing the Russell HEFI series we focused on identifying equity market factors that were relevant and implementable. 2 For further details on the NLP, see Chapter 26 in Portfolio Performance Measurement and Benchmarking, Christophersen, Cariño, Ferson, New York: McGraw-Hill, 2009. 3 Axioma s U.S. and Global ex-u.s. Medium Horizon Fundamental Risk Models. 4 The reader will note that the underlying variables used in the High Efficiency Quality Index (HEQI) and High Efficiency Low Volatility index (HELVI) are the same variables used in Russell s market cap weighted Stability Indexes series. For more information on the Stability Indexes series, refer to Russell Stability Indexes Construction and Methodology (November 2012). Russell Investments // Russell High Efficiency Factor Index Series 2
Presented below is a high-level summary of the steps involved in generating the underlying stock weights in the Russell HEFI series. This summary details the key aspects of the methodology; a more comprehensive overview is contained in the Russell High Efficiency Factor Indexes Construction and Methodology document. Select the parent index The first step in constructing the Russell High Efficiency Series is to select a parent index (e.g., Russell 1000 Index, Russell Global Large Cap, etc.). Every constituent of the parent index is eligible for inclusion in the HEFI index for the respective region. Generate the factor scores For each underlying variable, a score is assigned to each stock by using the non-linear probability algorithm, such that each stock is scored from zero to 1 for each variable (for example, book/price for value). A composite factor score for each stock is calculated by taking a simple average of the individual variable scores. The final composite factor score for each stock is then re-scaled from -1 to 1. Convert the factor scores to active weights A maximum active weight, known as the Weight Adjustment Factor (WAF), is set at 1% for all securities. An active breakpoint (X B ), above which stocks are overweighted and below which stocks are underweighted, is chosen. Each stock factor score is then converted to an Unconstrained Active Weight (UAW) by using the non-linear probability algorithm, the WAF and the X B. The UAW is unconstrained in that it allows short positions to be held. Impose a long-only constraint To apply the long-only constraint, the negative active weights are limited to prevent short positions. The positive active weights are then adjusted so that the resulting underweights and overweights sum to zero. 5 This results in a Constrained Active Weight (CAW) for each stock. Stocks with a benchmark weight less than or equal to the CAW will not be included in the final index Final weight in the High Efficiency Index The final stock weight in the HEFI is equal to the weight in the parent index plus the Constrained Active Weight (CAW). Given this high-level description of the index construction, we turn now to a more detailed consideration of the use of the NLP in factor scoring and the subsequent active weighting of the constituents in the HEFI series. In building factorbased indexes, there are two major considerations that determine the effectiveness of the strategy: the factor scoring of each stock and the determination of its weight in the factor index. Russell s factor-scoring approach In building factor-based indexes, there are two major considerations that determine the effectiveness of the strategy: the factor scoring of each stock and the determination of its weight in the factor index. These two components are essential to ensuring that the portfolio exposures are relevant, effective and efficient. The first step in scoring securities is to standardize the raw values of each variable that define a factor. Standardization transforms the raw values of the different variables of each factor to the same scale and allows for the comparison of different variables. If this had not been done, we could not have compared the Return on Assets value to the Debt to Equity value variables used in the Russell HEQI Index. This becomes important when combining singlevariable values into a composite value. 5 Leibowitz, M., S. Emrich and A. Bova, 2009. Modern Portfolio Management: Active Long/Short 130/30 Equity Strategies, New Jersey: Wiley Finance. Russell Investments // Russell High Efficiency Factor Index Series 3
One standardization method commonly used within the industry is the Z-score, which expresses the raw value of a variable relative to the mean and the standard deviation of its distribution. Unfortunately, the distributions for most investment factors are not normal and can have significant skews, which are preserved in the Z-score distribution. Further, Z-scoring does not control for outliers, resulting in securities that have large outlier values. If these large outlier values left unmanaged, they can have a large effect on the resulting portfolio. Percentile ranking is another traditional scoring approach that aims to overcome some of the issues with standard Z-scores. This approach solves the skew issue and neutralizes outliers. However, transforming raw scores to ordinal (percentile rank) scores discards any information about the shape of the distribution and makes all securities equidistant from each other. For example, if you have three stocks with momentum returns of, respectively, 100%, 90% and 20%, you would lose the information contained within this set that the third stock had a much lower momentum return than the top two. We believe that preserving such information about the distribution enhances the scoring model. For this reason, we have utilized Russell s NLP algorithm in the standardization process. This method effectively re-scales outliers but preserves, to an extent, the key distributional characteristics. When implementing the NLP scoring approach, we first calculate percentile ranks for each of the variables and retrieve the raw score that corresponds to the 90th percentile (X u ), the 10th percentile (X L ) and the 50th percentile (X M ) of the distribution. We then calculate a score (Y) based on the raw values of each variable (X) by using Equation 1. Equation 1: Non-linear probability algorithm Where: Y = Non-linear score X M = 50th percentile breakpoint X U = 90th percentile breakpoint X L = 10th percentile breakpoint The non-linear weighting algorithm allows for a monotonic relationship between a stock s factor score and its active weight. After calculating a score for each of the variables, we calculate a composite factor score by simply taking the weighted average scores of the variables for each factor. Finally, in preparation for the active weighting step, we re-scale the composite score such that it ranges between -1 and +1. Russell s weighting approach Once a composite score is calculated for each factor, we translate these scores into active weights by using the NLP algorithm. The non-linear weighting algorithm allows for a monotonic relationship between a stock s factor score and its active weight. This ensures that a higher factor score results in a higher exposure to a stock within the factor index, but only marginally so at the extremes of the distribution. That is, as the factor score of the stock increases, the active weight should increase as well, but at a rate that is decreasing. The Russell NLP weighting algorithm allows for this benefit. It does not crowd the majority of active share 6 into a handful of stocks; instead, it spreads the active share across the full spectrum of stocks that represent the factor. Similar to the scoring approach explained in the section above, here we first calculate the percentile rank for the composite factor scores and retrieve the score that corresponds to the 90th percentile (X u ), the 10th percentile (X L ) and the 50th percentile (X B ) 7 of the distribution. In the weighting algorithm, the (X B ) value is the active breakpoint that determines the number of overweight and underweight positions held in the index. Higher values of (X B ) will result in more concentrated active positions, while lower values will result in more diversified active positions. 6 For a definition of active share, see How Active Is Your Fund Manager? A New Measure That Predicts Performance, M. Cremers and A. Petajisto, Review of Financial Studies 22, 3329-3365 (March 2009). 7 The exception is the Russell High Efficiency Low Volatility Index series, where the active breakpoint is set at 70%, which leads to the index holding 30% of the names as an overweight relative the parent index. Russell Investments // Russell High Efficiency Factor Index Series 4
We then calculate an Unconstrained Active Weight (UAW) for each stock by using a variation 8 of Equation 1 and the Weight Adjustment Factor (WAF) to scale these active weights to a targeted level. The underweights are then constrained to be no greater than the respective benchmark weight, ensuring that the index holds no short positions. Finally, the sum of all the active underweights is distributed across the active overweights to enforce the long-only constraint that all index weights sum to 100%. 9 This results in the final HEFI active weight which we term the Constrained Active Weight (CAW) for each stock. Stocks with a benchmark weight less than or equal to the CAW will not be included in the final index. 8 In the variation, X B is substituted for X M. 9 Leibowitz, M., S. Emrich and A. Bova, 2009. Modern Portfolio Management: Active Long/Short 130/30 Equity Strategies, New Jersey: Wiley Finance. Russell Investments // Russell High Efficiency Factor Index Series 5
Active Weight (Relative to parent Russell Index) Underweight Overweight In Figure 1 we provide a graphical representation of the conversion of the factor scores to non-linear active weights described above. On the horizontal axis we have plotted sample factor scores for all securities in the parent Russell index, and on the vertical axis we have plotted the resulting CAW in a hypothetical HEFI index. In the diagram we see a strong relationship between the factor score and the CAW; higher factor scores are associated with higher CAWs, and vice versa. Preserving this relationship results in a high factor-/active weight correlation and results in stronger factor exposure. Figure 1: Constrained active weight vs. underlying sample factor score Negative Factor Score Positive (Note: The blue dots denote securities held in the Russell High Efficiency Index. The grey dots in the chart denote securities that are held in the parent index, but not in the Russell High Efficiency Index, because their benchmark weight was equal to or below the CAW.) Although we can see a strong relationship between the factor score and the CAW, it is clearly not linear. The disproportional relationship is intentional and reflects our view that stock returns associated with factor exposures are not strictly linear. As an example, using momentum, on average we expect high-momentum stocks to outperform low-momentum stocks, but we don t necessarily expect the subsequent return to be linearly related to a stock s momentum exposure. We typically find that there is little differentiation in returns across stocks found within the highest quintile of a factor, or across those within the lowest quintile of a factor. In other words, there is a limit to how much more return can be expected as a result of the factor exposure increasing. Recent academic research is also supportive of this notion. 10 The benefits of the Russell HEFI approach In this section we explore in detail the key benefits of the Russell HEFI methodology and highlight the robustness of the methodology to consistently provide institutional investability. The key benefits that we explore include: Consistent factor capture; Active risk awareness; Low turnover; and Modularity. 10 Working paper: Robustness and Monotonicity of Asset Pricing Anomalies, D. Maslov and O. Rytchkov (2013). Russell Investments // Russell High Efficiency Factor Index Series 6
Russell HEFI: Consistent factor exposure Earlier we discussed the four factors that make up the Russell HEFI series and the benefits they bring to investors by helping them target a desired total portfolio outcome. In order to allow investors to take full advantage of these exposures, the Russell HEFI series needs to be able to consistently deliver those exposures. As described, the Russell HEFI methodology moves away from capitalization weighting and puts factor exposure at the heart of the index construction: a stock s factor exposure is the sole determinant of the resulting active position. The stronger the exposure is to a particular factor, the larger the resulting active weight in the index relative to the parent index. This is a defining characteristic of the Russell HEFI methodology. In Table 2, below, we compare the correlation of the factor exposure and resulting active position of the Russell HEFI methodology and three alternative methodologies. In the table we see that in the U.S. large cap market, the Russell HEFI Value methodology has the highest factor exposure/active weight correlation of the methodologies The stronger the exposure is to a particular factor, the larger the resulting active weight in the index relative to the parent index Table 2: Factor exposure/active weight correlation 11 of HEFI active factor exposures vs. other factor-weighting methodologies STRATEGY Russell HEFI Capitalization weighted Score x market capitalization Factor weighted METHODOLOGY SUMMARY Active weights are derived by using a non-linear factor score Select the top third of the parent index constituents based on the factor score and then create a capitalizationweighted portfolio Constituent weights are derived by multiplying each constituent s factor score by its market-capitalization weight Constituent weights are derived by dividing each constituent s factor score by the sum of all factor scores FACTOR EXPOSURE/ ACTIVE WEIGHT CORRELATION 0.66 0.34 0.38 0.32 We have also looked at our factor capture through both returns-based and holdings-based analyses utilizing the Axioma Fundamental Medium Horizon risk models. The Axioma risk model has factor proxies for value, momentum and volatility. As there is no direct proxy for quality in the Axioma risk model, we have proxied quality with Axioma s leverage factor. In Table 3 we show the returns-based results using a multivariate regression of the Russell U.S. and Global HEFI series excess returns 12 over the parent index against the Axioma factor returns for value, momentum, volatility and leverage to estimate the factor exposures (regression coefficients) of the HEFI underlying indexes. We report the factor exposures relevant to each index. We also show the t-statistics for each exposure; generally, where the t-statistic is greater than 2, there is a strong relationship between the factors. For the U.S. and Global HEFI series, we see high exposures to the Axioma factors with very strong relationships (t-statistics), highlighting the efficacy of the construction methodology in delivering the intended exposure. 11 The correlation has been calculated by using the Value Score as the factor and is based on the Russell 1000 Index as at December 31, 2013. 12 In this paper, the phrase excess return refers to the return of the HEFI index minus the return of the parent index. Russell Investments // Russell High Efficiency Factor Index Series 7
Jan-99 Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 Jan-12 Jan-13 Axioma Factor Exposure Table 3: Multivariate regression coefficients of U.S. and Global HEFI excess returns against Axioma factor returns (January 1999 December 2013) Russell High Efficiency Value Index U.S. GLOBAL Exposure to Axioma Value 0.72 1.05 t-statistic 3.39 5.705 Russell High Efficiency Momentum Index Exposure to Axioma Momentum 0.79 0.71 t-statistic 10.54 9.529 Russell High Efficiency Low Volatility Index Exposure to Axioma Volatility* -0.55-0.74 t-statistic -9.76-8.64 Russell High Efficiency Quality Index Exposure to Axioma Leverage* -0.29-0.48 t-statistic -4.84-2.93 *Axioma calculates high-volatility and high-leverage exposures, so the intended exposure to these factors should be negative. In Figure 2 we show holdings-based factor estimates utilizing the Axioma risk model for the Russell HEFI U.S. series. The chart shows that our exposure to the corresponding Axioma factors has historically been persistent, highlighting the consistency of exposures that we are able to achieve with the Russell HEFI methodology. Figure 2: Holdings-based exposure to Axioma factors, U.S. HEFI series (1999 2013) 13 0.80 0.60 0.40 0.20 -- -0.20-0.40-0.60 Medium-Term Momentum Value Leverage Volatility 13 The exposure shows the HEFI s exposure to the corresponding Axioma factor. (e.g., High Efficiency Quality represents the holdings-based exposure to the Axioma Leverage factor. High Efficiency Momentum represents the holdings-based exposure to the Axioma Momentum factor. High Efficiency Value represents the holdings-based exposure to the Axioma Value factor. High Efficiency Low Volatility represents the holdingsbased exposure to the Axioma Volatility factor). The analysis uses the Axioma U.S. Fundamental Medium Horizon risk model. Russell Investments // Russell High Efficiency Factor Index Series 8
Russell HEFI: Active risk aware The Russell HEFI series has been designed to be an efficient and effective tool for harvesting factor returns, while also managing the risks that come with those exposures. For the Russell HEFI series, we have focused on building diversified exposures to help ensure that the active risk in each HEFI index is driven by factor exposures (systematic risk) and not by any individual stock (idiosyncratic risk). There are two ways the Russell HEFI methodology explicitly controls for active risk. The first is through defining the active breakpoint, which determines the number of stocks that will be held overweight relative to the parent Index. For the Russell HEFI series, the active breakpoint is set at 50% and results in the indexes owning at least 50% of the stocks in the parent index. 14 The second risk-control parameter is the weight-adjustment factor (WAF), which is set at 1%. Together, these two parameters ensure diversified exposures that target a large number of small active stock positions, as opposed to a small number of very large active stock positions. They help ensure that the index exposure is not heavily dominated by any particular sector and/or country, a common problem with existing factor-based indexes. Figure 3, below, shows that while there were sector differences between the Russell U.S. HELVI and Russell 1000, the HELVI was not dominated by any one sector or sectors. The result was that the Russell HEFI series delivered consistently high active share levels, with moderate levels of tracking error, as highlighted in Table 4. Figure 3: Comparative sector exposures for Russell U.S. HELVI and Russell 1000, December 31, 2013 Russell U.S. HELVI R1000 Index 11% 14% 5% 10% 13% 11% 13% 6% 17% 5% 16% 12% 4% 12% 15% 18% 8% 10% Consumer Discretionary Consumer Staples Energy Financial Services Health Care Materials & Processing Producer Durables Technology Utilities Table 4: Active risk characteristics of Russell U.S. HEFI Indexes, December 31, 2013* MOMENTUM VALUE QUALITY LOW VOLATILITY Tracking error 5.10% 5.96% 2.73% 7.32% Maximum active position 1.15% 1.05% 0.90% 0.89% Sum of Top 10 positions 13.03% 16.75% 16.08% 15.68% Sum of Top 10 active weights 1.83% 2.22% 1.86% 2.79% Active share 40.51% 39.64% 33.07% 45.27% *Tracking error was calculated versus the Russell 1000 for the July 1999 December 2013 time period. The maximum active position is the absolute value of the largest underweight/overweight. Maximum active positions are slightly outside the maximum of 1%, due to market movements. 14 The exception is the Russell High Efficiency Low Volatility Index series, where the mid-breakpoint is set at 70%, which leads to the Index holding approximately 30% of the names as an overweight relative the parent index. Russell Investments // Russell High Efficiency Factor Index Series 9
Russell HEFI: Turnover Turnover can be a material drag on the net performance an investment achieves via factor indexes, due to transaction costs and potential tax liabilities. Further, indexes that have high levels of turnover, and that reconstitute frequently, can be hard to replicate. And yet, in order for factor indexes to generate the necessary exposures, they typically require more frequent reconstitution and experience higher levels of turnover than traditional capitalization-weighted indexes. The Russell HEFI series keeps turnover at moderate levels without sacrificing intended exposures. One of the biggest drivers of turnover is the frequency of index reconstitution: typically, more frequent reconstitutions lead to higher turnover, albeit with better factor exposure. In developing the Russell HEFI series, we tested the impact of different reconstitution frequencies on both turnover and factor exposure. We found marginal decreases in factor exposures between quarterly and semiannual reconstitutions; however we consistently saw materially lower levels of turnover for semiannual reconstitutions. Although annual reconstitutions resulted in the lowest turnover levels this came with materially lower factor exposures. The Russell HEFI series keeps turnover at moderate levels without sacrificing intended exposures. In order to deliver the desired factor exposure while keeping turnover at a reasonable level, Russell uses a semiannual reconstitution cycle, with reconstitution occurring at the end of June and end of December each year. We believe the semiannual reconstitution frequency provides for the best trade-off between exposure to the factor and turnover. Banding To further minimize turnover, the Russell HEFI series applies a banding logic to minimize trades that have an immaterial impact on the portfolio exposure. Additions and deletions are not affected by the banding logic; additions are added to the index at the full target weight, and deletions are fully removed. Only securities that were members of the index prior to t reconstitution and are also current members are subject to the banding logic. The banding works in the following manner: First, at reconstitution, a band of plus or minus 10 basis points is computed around the new target weight of each stock. If a stock s current weight (weight in the Russell HEFI index prior to rebalance/reconstitution) is within the band, no action is taken. If a stock s current weight is outside of the band, the weight is adjusted toward the boundary of the band. Second, the band is adjusted so as not to enable a target overweight to become an underweight, or vice versa. That is, the lower boundary of the band for a target overweight is not less than the benchmark weight. Similarly, the upper boundary of the band for a target underweight is not greater than the benchmark weight. For the Russell HEFI series, we saw, on average, a 20% decrease in the turnover of each index as a result of the banding logic with a de minimis impact on the returns. The final turnover result of both the reconstitution frequency and the banding logic is displayed in Table 5. Table 5: Average annualized one-way turnover (December 1996 December 2013) MOMENTUM VALUE QUALITY LOW VOLATILITY U.S. 60.2% 30.4% 14.8% 20.7% Developed 63.4% 34.2% 19.4% 25.4% Developed ex U.S. 61.4% 34.6% 18.7% 28.0% Global 64.4% 35.2% 20.5% 25.2% Developed Europe 58.9% 30.0% 19.0% 29.6% Emerging Markets 66.6% 40.8% 28.8% 36.3% Modularity and diversification One of the benefits of factor-based investments is that factors are not perfectly correlated. This can present opportunities to enhance returns and/or minimize risks. The standard factorbased indexes currently in the market utilize vastly different selection and construction methodologies. These competing methodologies can result in highly contradictory portfolio Russell Investments // Russell High Efficiency Factor Index Series 10
Dec-96 Dec-97 Dec-98 Dec-99 Dec-00 Dec-01 Dec-02 Dec-03 Dec-04 Dec-05 Dec-06 Dec-07 Dec-08 Dec-09 Dec-10 Dec-11 Dec-12 Dec-13 structures, with each methodology bringing its own unique ideology. This can make it very difficult to ensure that the intended exposure is realized. Further, the factor exposure achieved from different construction methods can differ substantially, which can severely reduce the effectiveness of the combined strategy. All indexes in the Russell HEFI series have been specifically designed to be able to complement one another. This helps investors to more easily and effectively exploit the benefits of multifactor investing and better control the total portfolio outcome. Using a consistent portfolio construction process allows the Russell HEFI series to generate similar levels of active share across the different factors and also similar factor exposure levels. For example, this means that the average active stock position in the Russell High Efficiency Momentum Index is the same as the average active stock position in the Russell High Efficiency Value Index. Further, the factor exposures in both indexes are of a similar magnitude, resulting in consistent exposures that are highly complementary. The active returns of the Russell U.S. HEFI are not highly correlated with each other (see Table 6), and this presents an exploitable opportunity to reduce active risk. In Table 6 we also see that some factors tend to be more correlated with each other; for example, Value/Volatility and Momentum/Quality are the most-correlated combinations over the period, although we do see large variations in correlations over shorter time horizons. Figure 4, which charts calendar year excess returns from 1996 to 2013, also shows that, historically, the Russell U.S. HEFI indexes had varying periods of outperformance and underperformance. Table 6: Correlation of monthly excess returns: Russell U.S. HEFI indexes (July 1996 December 2013) R1000 HEMI R1000 HEQI R1000 HEVI R1000 HELVI R1000 HEMI 1.00 0.47-0.35-0.26 R1000 HEQI 0.47 1.00-0.31-0.12 R1000 HEVI -0.35-0.31 1.00 0.56 R1000 HELVI -0.26-0.12 0.56 1.00 Figure 4: Russell 1000 HEFI, calendar-year excess returns (December 1996 December 2013) 6 4 2 Excess return (%) 0-2 -4-6 -8 Russell 1000 High Efficiency Momentum Index Russell 1000 High Efficiency Value Index Russell 1000 High Efficiency Quality Index Russell 1000 High Efficiency Low Volatility Index Russell Investments // Russell High Efficiency Factor Index Series 11
Bringing it all together At Russell we believe that investment factors are significant drivers of equity returns. The excess returns associated with Value-, Momentum-, Quality- and Low Volatility based investment strategies have persisted across markets and through time. The Russell HEFI series is designed to provide investors with efficient exposure to these return drivers, and it draws on our more than 30 years experience in delivering targeted market exposures. The HEFI indexes can be used to manage strategic and dynamic exposures within a total portfolio and they can be easily combined. Strategic exposure: The Russell HEFI series can be used to provide systematic exposure to factors that align with investment philosophies. The consistency of the exposure can help investors ensure that their desired long-term investment exposures are reflected in the total portfolio across different markets. Dynamic exposure: The Russell HEFI series may allow investors to take advantage of shortterm mispricing of a long-term rewarded factor, due to market inefficiencies and behavioral biases. This short-term cyclicality may offer rewarding tactical investment opportunities. The Russell HEFI series is designed to provide investors with efficient exposure to these return drivers and draws on over 30 years experience in delivering targeted market exposure. The complementary nature of the Russell HEFI series and the low correlation of active returns across the factors enable investors to build robust multifactor portfolios. The ability to effectively combine factors within a portfolio has historically been limited to asset management firms possessing sophisticated quantitative capabilities. The Russell HEFI series brings many of these quantitative techniques and insights to investors in a modular framework which is easy to implement and manage. Investors seeking to achieve the best total portfolio outcomes need reliable tools to access these factors. In a world where precision matters, the Russell HEFI series provides exposures that are targeted, consistent, investable and complementary. i James Barber is Chief Investment Officer, Equities at Russell Investments. ii Scott Bennett is Director, Equity Strategy & Research at Russell Investments. iii Mark Paris is a Senior Research Analyst at Russell Indexes. Russell Investments // Russell High Efficiency Factor Index Series 12
Appendix: Factor summary Russell High Efficiency Value Index (HEVI) A value investment strategy involves identifying those stocks that are trading at a discount to some measure of fair value. The theory of value investing was pioneered by Graham and Dodd 15 and has been a key focus for both investors and academic researchers ever since. The value premium is one of the best-documented investing anomalies, and value investing is generally accepted across the investment industry as a persistent excess return generating strategy. Much of the related research and discussion focuses on the justification for the value return premium. Historically, value returns were seen to be associated with a compensation for taking on extra risk. More recently, though, there is increasing evidence to support the view that the value premium is the result of behavioral biases, as originally proposed by Lakonishok, Shleifer and Vishny 16. The behavioral research suggests that investors mistakenly expect the high growth rate of growth firms and the low growth rate of value firms to persist in the future, and they price the stocks accordingly. When growth rates mean-revert, and when earnings expectations are not met, growth firms are penalized more severely than value firms. The value effect might exist at least partly because many investors lack the patience to wait for mean-reversion. The dramatic shortening of investment horizons in recent years, as shown in average mutual fund turnover and holding period statistics, suggests that this value investing opportunity persists. In Table A1, below, we highlight that the excess returns to value as represented by HEVI have been pervasive across all regions and of similar magnitude. Exhibit A1 shows, however, that clearly there is not always a positive reward to value.. Table A1: High Efficiency Value Index (HEVI) return summary (July 1996 December 2013) U.S. GLOBAL DEV. DEV. EX-US DEV. EUROPE Annualized return 10.8% 10.9% 10.7% 9.8% 9.8% 8.8% Parent index 8.2% 7.5% 7.5% 6.5% 8.2% 6.8% Annualized standard deviation 16.6% 17.5% 16.8% 17.9% 20.4% 25.1% Parent index 16.1% 16.6% 16.3% 17.5% 18.9% 25.1% Sharpe ratio 0.54 0.53 0.53 0.46 0.43 0.35 Annualized excess return 2.6% 3.5% 3.2% 3.3% 1.6% 2.0% Tracking error 6.0% 5.6% 5.6% 4.7% 4.5% 4.8% Information ratio 0.44 0.62 0.57 0.71 0.35 0.42 Turnover 30.4% 35.2% 34.2% 34.6% 30.0% 40.8% EM 15 Graham, B. and D. Dodd, 1934, Security Analysis, New York: McGraw-Hill 16 Lakonishok, J., A. Shleifer, and R. Vishny, 1994, Contrarian investment, extrapolation, and risk, Journal of Finance, Vol. 49, No. 5, 1541-1578 Russell Investments // Russell High Efficiency Factor Index Series 13
Jul-96 - Jun-97 Jan-97 - Dec-97 Jul-97 - Jun-98 Jan-98 - Dec-98 Jul-98 - Jun-99 Jan-99 - Dec-99 Jul-99 - Jun-00 Jan-00 - Dec-00 Jul-00 - Jun-01 Jan-01 - Dec-01 Jul-01 - Jun-02 Jan-02 - Dec-02 Jul-02 - Jun-03 Jan-03 - Dec-03 Jul-03 - Jun-04 Jan-04 - Dec-04 Jul-04 - Jun-05 Jan-05 - Dec-05 Jul-05 - Jun-06 Jan-06 - Dec-06 Jul-06 - Jun-07 Jan-07 - Dec-07 Jul-07 - Jun-08 Jan-08 - Dec-08 Jul-08 - Jun-09 Jan-09 - Dec-09 Jul-09 - Jun-10 Jan-10 - Dec-10 Jul-10 - Jun-11 Jan-11 - Dec-11 Jul-11 - Jun-12 Jan-12 - Dec-12 Jul-12 - Jun-13 Jan-13 - Dec-13 Excess Return (%) Figure A1: High Efficiency Value Index rolling 12-month excess return (July 1996 December 2013) 40 30 20 10 0-10 -20-30 Developed LC ex-us High Efficiency Value Index Developed LC High Efficiency Value Index Russell 1000 High Efficiency Value Index Emerging LC High Efficiency Value Index Global LC High Efficiency Value Index Developed Europe LC High Efficiency Value Index Russell High Efficiency Momentum Index (HEMI) Momentum-based investment strategies focus on identifying those stocks with strong performance over the most recent 12-month period (absolute or relative), with the expectation that this strong performance will continue. The positive excess return associated with momentum strategies was first highlighted by Jegadeesh and Titman 17. The proposed explanations for price momentum generally fall into rational and behavioral categories, with the latter being the modern consensus. Momentum may occur due to investors underreacting to new information and information being slowly incorporated into prices. Or momentum can arise due to investors overreacting to private information and causing prices to be pushed away from fundamentals. Momentum returns have persisted through time; however, they are prone to shorter-term cyclicality. Momentum strategies have tended to lag the market during transitions in the market cycle (e.g., moving from a bear market to a bull market). The persistence of past winners has tended to occur over a horizon of three to 12 months, so momentum strategies tend to be very high turnover and can incur large amounts of transaction costs. It is important to find an optimal way to reduce turnover while still allowing a passive strategy to achieve the momentum exposure. Table A2 highlights that the excess returns to momentum strategies have been pervasive across regions and most effective in markets outside the U.S. The cyclicality of the excess return can be observed in Exhibit A2. 17 Jegadeesh, N. and S. Titman, 1993, Returns to buying winners and selling losers: Implications for stock market efficiency, Journal of Finance, Vol. 48, 65-91. Russell Investments // Russell High Efficiency Factor Index Series 14
Jul-96 - Jun-97 Jan-97 - Dec-97 Jul-97 - Jun-98 Jan-98 - Dec-98 Jul-98 - Jun-99 Jan-99 - Dec-99 Jul-99 - Jun-00 Jan-00 - Dec-00 Jul-00 - Jun-01 Jan-01 - Dec-01 Jul-01 - Jun-02 Jan-02 - Dec-02 Jul-02 - Jun-03 Jan-03 - Dec-03 Jul-03 - Jun-04 Jan-04 - Dec-04 Jul-04 - Jun-05 Jan-05 - Dec-05 Jul-05 - Jun-06 Jan-06 - Dec-06 Jul-06 - Jun-07 Jan-07 - Dec-07 Jul-07 - Jun-08 Jan-08 - Dec-08 Jul-08 - Jun-09 Jan-09 - Dec-09 Jul-09 - Jun-10 Jan-10 - Dec-10 Jul-10 - Jun-11 Jan-11 - Dec-11 Jul-11 - Jun-12 Jan-12 - Dec-12 Jul-12 - Jun-13 Jan-13 - Dec-13 Excess Return (%) Table A2: High Efficiency Momentum Index (HEMI) return summary (July 1996 December 2013) U.S. GLOBAL DEV. DEV.EX-U.S. DEV. EUROPE Annualized return 9.8% 10.0% 9.6% 8.8% 10.7% 8.5% Parent index 8.2% 7.5% 7.5% 6.5% 8.2% 6.8% Annualized standard deviation 17.2% 17.5% 16.8% 17.4% 18.5% 24.9% Parent index 16.1% 16.6% 16.3% 17.5% 18.9% 25.1% Sharpe ratio 0.48 0.48 0.47 0.42 0.50 0.35 Annualized excess return 1.6% 2.5% 2.1% 2.3% 2.6% 1.7% Tracking error 5.1% 4.9% 4.7% 4.5% 4.9% 4.9% Information ratio 0.32 0.51 0.46 0.51 0.53 0.35 Turnover 60.2% 64.4% 63.4% 61.4% 58.9% 66.6% EM Figure A2: High Efficiency Momentum Index (HEMI) rolling 12-month excess return (July 1996 December 2013) 30 20 10 0-10 -20-30 Developed LC ex-us High Efficiency Momentum Index Developed LC High Efficiency Momentum Index Russell 1000 High Efficiency Momentum Index Emerging LC High Efficiency Momentum Index Global LC High Efficiency Momentum Index Developed Europe LC High Efficiency Momentum Index Russell Investments // Russell High Efficiency Factor Index Series 15
Russell High Efficiency Quality Index (HEQI) Quality-based investment strategies are focused on identifying companies that have greater ability to deliver sustainable returns to shareholders. These companies are typically characterized by high profitability, low leverage and low earnings volatility. The term quality was first used in this context, again by Graham and Dodd, and reflected firm-specific characteristics such as size, reputation, financial position and prospects. Recent research by Campbell, Hilscher and Szilagyi 18, 19 Novy-Marx 20 and Asness et al 21. confirms excess returns associated with quality-based investment strategies. The recognition of quality as a factor has led to its being a major component of Russell s Stability Indexes series. Unlike other factors, such as value, momentum and low volatility, the characteristics used to define quality companies are highly subjective, and as a result, quality-based investment strategies can be difficult to compare. That said, the most commonly utilized attributes are leverage, profitability and earnings stability. Russell believes that the excess returns to quality exist due to behavioral reasons: the tendency of investors to favor more volatile (high-leverage) stocks with more explosive potential upside in the short term, which can lead to higher-quality companies being mispriced in the shorter term, and to attractive longer-term returns. Table A3 highlights the performance of quality, and while the average outperformance of quality strategies has not been as large as that we have seen for other factors, the stable returns that have been offered by high-quality stocks may benefit the investor over the longer term. As Exhibit A3 shows, exposure to quality has been rewarded during many time periods, although it has not delivered excess returns in more recent years. Table A3: High Efficiency Quality Index (HEQI) return summary (July 1996 December 2013) U.S. GLOBAL DEV. DEV.EX-U.S. DEV. EUROPE Annualized return 9.7% 9.1% 9.0% 8.0% 9.6% 7.3% Parent index 8.2% 7.5% 7.5% 6.5% 8.2% 6.8% Annualized standard deviation 15.7% 16.5% 15.9% 16.2% 17.2% 23.4% Parent index 16.1% 16.6% 16.3% 17.5% 18.9% 25.1% Sharpe ratio 0.50 0.45 0.45 0.39 0.46 0.30 Annualized excess return 1.5% 1.6% 1.6% 1.5% 1.4% 0.5% Tracking error 2.7% 2.9% 2.8% 3.3% 3.8% 3.7% Information ratio 0.56 0.56 0.56 0.44 0.36 0.13 Turnover 14.8% 20.5% 19.4% 18.7% 19.0% 28.8% EM 18 Campbell, J., J. Hilscher, and J. Szilagyi. 2008, In search of distress risk, Journal of Finance, Vol. 63 (6), 2899-2939 19 Campbell, J., J. Hilscher, and J. Szilagyi. 2011, Predicting financial distress and the performance of distressed stocks, Journal of Investment Management, Vol. 9 (2), 14-34 20 Novy-Marx, R., 2013, The other side of value: the gross profitability premium, Journal of Financial Economics, Vol. 108(1), 1-28 21 Asness, C., A. Frazzini, and L. H. Pedersen, 2013, Quality minus Junk, Working paper, AQR Capital Management, New York University Russell Investments // Russell High Efficiency Factor Index Series 16
Jul-96 - Jun-97 Jan-97 - Dec-97 Jul-97 - Jun-98 Jan-98 - Dec-98 Jul-98 - Jun-99 Jan-99 - Dec-99 Jul-99 - Jun-00 Jan-00 - Dec-00 Jul-00 - Jun-01 Jan-01 - Dec-01 Jul-01 - Jun-02 Jan-02 - Dec-02 Jul-02 - Jun-03 Jan-03 - Dec-03 Jul-03 - Jun-04 Jan-04 - Dec-04 Jul-04 - Jun-05 Jan-05 - Dec-05 Jul-05 - Jun-06 Jan-06 - Dec-06 Jul-06 - Jun-07 Jan-07 - Dec-07 Jul-07 - Jun-08 Jan-08 - Dec-08 Jul-08 - Jun-09 Jan-09 - Dec-09 Jul-09 - Jun-10 Jan-10 - Dec-10 Jul-10 - Jun-11 Jan-11 - Dec-11 Jul-11 - Jun-12 Jan-12 - Dec-12 Jul-12 - Jun-13 Jan-13 - Dec-13 Excess Return (%) Figure A3: High Efficiency Quality Index (HEQI) rolling 12-month excess return (July 1996 December 2013) 20 10 0-10 -20 Developed LC ex-us High Efficiency Quality Index Developed LC High Efficiency Quality Index Russell 1000 High Efficiency Quality Index Emerging LC High Efficiency Quality Index Global LC High Efficiency Quality Index Developed Europe LC High Efficiency Quality Index Russell High Efficiency Low Volatility Index (HELVI) Low volatility based investment strategies are focused on identifying companies that have had more stable return patterns than the broader market. The higher returns associated with low-volatility strategies can run counter to the adage that higher volatility (risk) is generally associated with higher returns. The benefits of low-volatility investing were first documented by Haugen and Heins 22 and have recently gained much more interest, following the 2008 financial crisis. Research by Clark, Silva and Thorley 23 and Blitz and van Vliet 24 (2007) show that low-volatility strategies tend to benefit from their avoidance of the most volatile stocks in the market, which typically deliver lower average returns. Blitz and van Vliet propose that investors are willing to pay higher prices for more volatile stocks in the short term, due to a preference for leverage (high beta) and lottery-style payoffs. Over the long term, the more stable return pattern of lowvolatility stocks stands to benefit investors, because avoiding severe downturns can have powerful effects on compounding. 22 Haugen, R., and A. Heins, 1975, Risk and the Rate of Return on Financial Assets: Some Old Wine in New Bottles, Journal of Financial and Quantitative Analysis, Vol. 10, No. 5 23 Clarke, R., H. Silva, and S. Thorley, 2006, Minimum-Variance Portfolios in the U.S. Equity Market, The Journal of Portfolio Management, Vol. 33, No. 1: pp. 10-24 24 Blitz, D. and P. Vliet, 2007, The Volatility Effect, The Journal of Portfolio Management, Vol. 34, No. 1: pp. 102-113 Russell Investments // Russell High Efficiency Factor Index Series 17
Jul-96 - Jun-97 Jan-97 - Dec-97 Jul-97 - Jun-98 Jan-98 - Dec-98 Jul-98 - Jun-99 Jan-99 - Dec-99 Jul-99 - Jun-00 Jan-00 - Dec-00 Jul-00 - Jun-01 Jan-01 - Dec-01 Jul-01 - Jun-02 Jan-02 - Dec-02 Jul-02 - Jun-03 Jan-03 - Dec-03 Jul-03 - Jun-04 Jan-04 - Dec-04 Jul-04 - Jun-05 Jan-05 - Dec-05 Jul-05 - Jun-06 Jan-06 - Dec-06 Jul-06 - Jun-07 Jan-07 - Dec-07 Jul-07 - Jun-08 Jan-08 - Dec-08 Jul-08 - Jun-09 Jan-09 - Dec-09 Jul-09 - Jun-10 Jan-10 - Dec-10 Jul-10 - Jun-11 Jan-11 - Dec-11 Jul-11 - Jun-12 Jan-12 - Dec-12 Jul-12 - Jun-13 Jan-13 - Dec-13 Excess Return (%) There are different approaches to gaining low-volatility exposures, and investors should be mindful of the sector concentration and turnover associated with some low-volatility strategies. Table A4 highlights the performance characteristics of the Russell High Efficiency Low Volatility Index. Here we see that although the absolute risk of low-volatility strategies has been consistently lower than that of the broader market, the tracking errors have been typically quite high. As Figure A4 reveals, the reward to low volatility has been very cyclical, with several periods of both underperformance and outperformance. Table A4: High Efficiency Low Volatility Index (HELVI) return summary (July 1996 December 2013) US GLOBAL DEV. DEV.EX-U.S. DEV. EUROPE Annualized return 9.4% 9.7% 9.6% 9.6% 9.7% 8.7% Parent index 8.2% 7.5% 7.5% 6.5% 8.2% 6.8% Annualized standard deviation 12.5% 12.6% 12.3% 13.4% 14.8% 20.6% Parent index 16.1% 16.6% 16.3% 17.5% 18.9% 25.1% Sharpe ratio 0.57 0.58 0.59 0.55 0.52 0.38 Annualized excess return 1.2% 2.2% 2.1% 3.1% 1.5% 1.9% Tracking error 7.3% 7.2% 7.4% 7.1% 6.7% 7.4% Information ratio 0.17 0.31 0.29 0.43 0.22 0.26 Turnover 20.7% 25.2% 25.4% 28.0% 28.6% 36.3% EM Figure A4: High Efficiency Low Volatility Index (HEQI) rolling 12-month excess return (July 1996 December 2013) 40 30 20 10 0-10 -20-30 -40-50 Developed LC ex-us High Efficiency Low Volatility Developed LC High Efficiency Low Volatility Index Russell 1000 High Efficiency Low Volatility Index Emerging LC High Efficiency Low Volatility Index Global LC High Efficiency Low Volatility Index Developed Europe LC High Efficiency Low Volatility Russell Investments // Russell High Efficiency Factor Index Series 18
ABOUT RUSSELL INDEXES Russell s indexes business, which began in 1984, accurately measures U.S. market segments and tracks investment manager behavior for Russell s investment management and consulting businesses. Today, our series of U.S. and global equity indexes reflect distinct investment universes asset class, geographic region, capitalization and style with no gaps or overlaps. Russell Indexes offers more than three dozen product families and calculates more than 700,000 benchmarks daily, covering 98% of the investable market globally, 80 countries and more than 10,000 securities. Approximately $5.2 trillion in assets are benchmarked to the Russell Indexes. For more information about Russell Indexes call us or visit www.russell.com/indexes. Americas: +1-877-503-6437; APAC: +65-6880-5003; EMEA: +44-0-20-7024-6600 Disclosures Russell Investments is a trade name and registered trademark of Frank Russell Company, a Washington USA corporation, which operates through subsidiaries worldwide and is part of London Stock Exchange Group. Russell Investments is the owner of the trademarks, service marks and copyrights related to its respective indexes. Westpeak Global Advisors, LLC and Goldman Sachs Asset Management, L.P. are developers of technologies used in the Russell HEFI Indexes. Russell Indexes has independently developed intellectual property that is used to construct and maintain the Russell HEFI Indexes. Indexes are unmanaged and cannot be invested in directly. The returns provided for each Russell Index may include data for periods prior to when the index was in live production. Historical returns for these Russell Indexes prior to the live production date are calculated using the same Russell methodology; however, application to the performance calculation may vary due to data sources, corporate actions, and the availability of historical data with respect to certain securities. Please contact the Russell Index Client Service Team for further detail. Unless otherwise noted, the source for the data in this report is Russell Investments. This material is proprietary and may not be reproduced, transferred or distributed in any form without prior written permission from Russell Investments. It is delivered on an as is basis without warranty. This is not an offer, solicitation or recommendation to purchase any security or the services of any organization. Copyright Russell Investments 2015. All rights reserved. First use: April 2014. Revised February 2015. CORP-9443-04-2016E Russell Investments // Russell High Efficiency Factor Index Series 19