THE PROBLEM

The idea for ActiveBeta Indexes, as a conceptual framework and an investment product, developed out of conversations that took place at the start of 2008 between Westpeak Global Advisors and a large pension fund. The investment problem being confronted by this fund was as follows. The equity portion of the fund was equally split between passive index replication and active management. The large size of the fund required an investment in a substantial number of active managers in order to diversify individual manager investment process and business risk. For the three and five years ending 2007, the active managers employed by the fund delivered only benchmark returns, as a group, but the fund paid alpha fees for this outcome. Disappointed by the overall performance of their active managers, the fund’s Investment Committee decided to move to a 70 percent allocation to passive index replication. A smaller allocation to active management would allow the fund to concentrate investments in the most skilled active managers, reduce the noise that a large group of active managers inherently creates, and allocate active management fees and risk budgets more efficiently.

Yet, some senior officials at the fund did not necessarily view a higher allocation to passive management as an optimal decision. They believed that common and persistent sources of active returns do exist in the marketplace but that the implementation choices available to capture these sources of active returns were either incomplete and perhaps even misleading (such as value and growth style benchmarks), or inefficient, opaque, and unduly expensive (such as traditional active management). In their view, a clear need existed for an investment solution that better defined the investment styles of active managers (i.e., common sources of active returns) and provided an efficient, transparent, and cost-effective passive capture of a significant portion of traditional active management returns. These senior officials challenged us to develop such an investment solution for the industry.

ActiveBeta Indexes is our attempt at providing this solution.

THE SOLUTION

Three basic questions became the focus of our research efforts.

1. What gives rise to common sources of active returns (investment styles)?

2. How can such active returns persist in highly efficient and adaptive markets?

3. What is the best way to capture the common sources of active returns in an index structure? That is, what is the best way to passively replicate a significant portion of active management returns?

The answers to these questions led to the development of the ActiveBeta Framework and the associated ActiveBeta Indexes.

Current Paradigm of Style Investing

Investment styles have basically emerged out of academic research that focused on discovering stock market anomalies and/or explaining the cross-section of average stock returns. Decades of research and empirical evidence have led to the current industry practice of defining investment styles in terms of value and growth. Index providers and fund consultants have also developed performance benchmarks and fund classification schemes based on the value-growth categorization. However, the current empirical research-driven framework has some limitations. In this framework, so-called market anomalies are first researched, through extensive data crunching, and then an attempt is made to rationalize their existence and persistence. As a result, there is little disagreement on the existence of active returns relating to some strategies, such as value, as these active returns have been well documented. But, because of the lack of a coherent conceptual and theoretical framework, there is little or no agreement on what causes such active returns to exist and persist over time. Similarly, there is no explanation for why growth investing, as it is currently defined, constitutes an investment style of its own.

Furthermore, in the prevailing style framework, the core investment style is defined as the market. That is, the selection universe and performance benchmark for core managers is the market index. Value managers have a value index as a benchmark, rather than the market index, because they are viewed as having a value bias relative to the market. Similarly, growth managers are evaluated against a growth index, not the market index. So, the fact that the core managers’ benchmark is the market index implies that these managers have no systematic biases relative to the market. Practitioners know that this is not the case. Core managers do have systematic tilts.

The current paradigm has emerged from the incorrectly defined growth investment style. In its prevailing definition, growth is better classified as non-value, not an independent style of its own. Therefore, the combination of value and growth simply produces the market, which then is used as the performance yardstick for core managers. In the current style framework, both growth managers and core managers are evaluated against inappropriate performance benchmarks.

We approach investment styles from a different conceptual perspective. In our view, an investment style, by definition, represents a source of return common across a large number of active managers. If a common source of return persists over time, despite its widespread exploitation, then it must emanate from fundamental influences that drive stock returns. So, to identify appropriate investment styles, or common and persistent sources of active returns, we have to identify, and study the behavior of, the fundamental drivers of stock returns.

Active Betas: Systematic Sources of Active Equity Returns

Stock prices are fundamentally driven by earnings expectations and interest rates used to discount future earnings. Stock price changes are driven by changes in expectations relating to these variables. If changes in expectations depict a systematic behavior, which is not discounted by current prices, then some portion of future returns would also become systematic or predictable. Our research findings, presented in Chapters 3 through 6, indicate that short-term and long-term earnings expectations and discount rates do behave in a systematic or predictable manner over time. Short-term earnings expectations depict positive serial correlation, or an average tendency to trend in the short run. Long-term earnings expectations and discount rates depict negative serial correlation, or an average tendency to mean revert in the long run.

This systematic behavior arises out of the fact that in an open and competitive system, a large group of stocks cannot sustain above-average earnings growth indefinitely. In Chapter 6, we discuss in more detail the influences that explain the systematic behavior of earnings expectations and why it is difficult to fully incorporate this behavior in the formation of current expectations and market prices. Our research further establishes that the payoff to momentum strategies arises from the systematic tendency of short-term earnings expectations to trend in the short run. On the other hand, the payoff to value strategies arises from the systematic tendency of long-term growth expectations to mean revert in the long run. Consequently, we refer to the active returns of momentum and value strategies that are driven partly by the systematic behavior of earnings expectations and discount rates as systematic sources of active equity returns, or Active Betas.

Redefining Investment Styles

Based on the ActiveBeta Framework and the associated research, we argue that persistent investment styles arise out of systematic sources of active equity returns that active managers attempt to capture. As such, momentum and value better represent the investment styles of active managers, compared to value and growth.

An analysis of the active returns of the so-called growth managers, presented in Chapter 9, clearly highlights that the so-called growth managers are, in fact, momentum players. That is, they attempt to capture the systematic active return sources associated with the average tendency of short-term expected earnings growth to trend, or to be positively serially correlated. It is not clear what fundamental or systematic source of return the growth investment style, as currently defined by the academic literature and the various growth indexes, attempts to capture. Growth indexes force investors to invest in securities that have high current long-term earnings growth expectations (and, hence, high valuations). But, the systematic behavior of long-term expected earnings growth is to mean revert. Buying securities that are currently classified as high-growth is prone to failure, as these securities subsequently experience decelerating growth expectations (mean reversion), causing their price-earnings (P/E) ratios to contract and future returns to be significantly lower than average. The opposite happens when investors buy value securities. These securities are currently classified in the low-growth category (e.g., low P/E multiples, on average, imply low growth). Going forward, they experience significant expansions in P/E multiples and higher than average returns, as mean reversion of long-term earnings expectations leads to upward revisions in the growth prospects of these companies. It should, therefore, come as no surprise that value indexes systematically outperform and growth indexes systematically underperform the market indexes in the long run in all markets and market segments. The evidence presented in Chapter 9 on the relative performance of value versus growth across a large number of market universes raises serious doubts about the validity of growth as an independent investment style that a large number of active managers follow. With billions of dollars tracking or linked to growth indexes, the research raises concerns regarding the current theory and practice of investment styles.

Further, the active return decomposition analysis presented in Chapter 9 highlights that the so-called core managers are not style-neutral. They have significant combined momentum and value tilts. A large proportion of their portfolio holdings are in stocks that have both high momentum and high value attributes. Given this evidence, how can the market index be the appropriate opportunity set (neutral portfolio) for core managers? Core is a separate investment style with simultaneous biases toward momentum and value.

We need to rethink and revisit the design and structure of investment style indexes.

ActiveBeta Indexes: Capturing Systematic Sources of Active Equity Returns

In order to provide an efficient, transparent, and cost-effective capture of Active Betas, and an alternative to current style indexes, we created an internally consistent family of ActiveBeta Indexes. For any given selection universe, for example, the FTSE All-Share U.K. Index, this family of indexes comprises independent market capitalization-weighted ActiveBeta Momentum and ActiveBeta Value Indexes. The two indexes include high momentum and high value stocks, respectively, and independently target about a 50 percent market capitalization coverage of the selection universe. The two independent indexes are then combined, in equal proportions, to create the ActiveBeta Momentum and Value Index (MVI).

The motivation for creating a combined ActiveBeta MVI is to develop a more representative benchmark for the so-called core managers. The ActiveBeta MVI provides roughly 75 percent of the market capitalization coverage of the underlying selection universe. It does not include stocks that have both low-momentum and low-value attributes. These stocks have a negative exposure to the systematic sources of active returns, and they underperform the market on a consistent basis. This group of stocks creates a tremendous performance drag on the market index. These are also the stocks that core managers do not consider for investment. By excluding these stocks, the ActiveBeta MVI better reflects the neutral portfolio of core managers.

The ActiveBeta MVIs outperform the underlying market index or universe on which they are based, in each and every market and market segment we have studied, net of implementation costs. For Developed Markets universes, the after-cost information ratios generated by the ActiveBeta MVIs range from 0.4 to 0.8. The outperformance of the ActiveBeta MVIs has been remarkably consistent, on a year-by-year basis and in each market, during the 1992 through 2008 time period that we have analyzed. The historical performance of the ActiveBeta Indexes is presented in Chapter 8.

The ActiveBeta MVIs outperform because the relative performance of value and momentum is linked to the risk aversion of investors. When risk aversion is high (e.g., in down markets), momentum outperforms value. When risk aversion is low (e.g., in up markets) value outperforms momentum. The link of momentum and value with the risk aversion of investors causes the active returns of the two strategies to become negatively correlated. The negative correlation of momentum and value is observed in all markets and is fairly stable over time. Combining two systematic sources of returns (e.g., momentum and value) that independently provide positive active returns but are negatively correlated offers investors an opportunity to realize significant diversification gains without necessarily sacrificing returns. Since value and growth are not independent sources of active returns, and growth, as currently defined, systematically delivers negative active returns in the long run, investors do not have this diversification opportunity in the current value/growth style framework.

ACTIVEBETA INDEX APPLICATIONS

The existence and efficient capture of Active Betas has many investment process applications. Some of these are summarized next.

Decomposing Active Returns

Since ActiveBeta Indexes capture systematic sources of active equity returns, they more accurately represent the generic investment processes of active managers, or investment styles. As such, ActiveBeta Indexes can be used to provide a more accurate decomposition of an active portfolio’s returns into a systematic source (Active Beta) component and a skill (pure alpha) component. This analysis, when applied to a large universe of core, value, and growth managers, leads to the conclusion that a significant portion of traditional active management returns comes from the systematic sources of active equity returns, or Active Betas (see Chapter 9).

Revisiting the Alpha-Beta Portfolio Structure

Currently, equity portfolios are structured in terms of a market beta (passive) component and an alpha (active) component. However, if a significant portion of the traditional alpha component comes from Active Betas, then the alpha-beta return separation could be misleading. Therefore, investors should consider moving away from the current alpha-beta return separation portfolio structure to a new structure that comprises three components, namely market beta, Active Betas, and pure alpha.

With ActiveBeta Indexes, investors now have an efficient vehicle for capturing Active Betas, without taking undue investment process risks or paying alpha fees for the delivery of additional forms of beta.

Benchmarking of Active Manager Performance

As discussed, momentum and value better represent the investment styles and processes of active style managers than do value and growth. Similarly, a combination of momentum and value better represents the generic investment process of core managers than does a market index. As such, we argue that the ActiveBeta Momentum Index, ActiveBeta Value Index, and ActiveBeta MVI constitute more appropriate benchmarks for evaluating the performance of growth, value, and core active managers, respectively.

A Fairer Deal on Management Fees

Asset owners have long held the view that active management fees are too high relative to the value-added generated by traditional active management. However, asset owners have lacked the tools to make their argument credible. With ActiveBeta Indexes, investors now have a framework and a tool, through a return decomposition analysis, to argue that the Active Beta component of a manager’s alpha should not be compensated at active management fee levels. On the other hand, asset owners should be willing to pay higher management fees to managers who deliver pure alpha that is in addition to and uncorrelated with Active Betas. We realize these suggestions may be viewed as provocative by many, but we believe they represent a step in the right direction and will be beneficial for both asset owners and asset managers in the long run.

Creating Investment Vehicles

An internally consistent family of ActiveBeta Indexes for global equity markets provides a vehicle for investors to capture the systematic sources of active equity returns, either independently through the ActiveBeta Momentum and ActiveBeta Value Indexes or in combination through the ActiveBeta MVI. However, we would also advance the argument that an independent passive replication of an investment style, such as value, may not constitute the optimal long-term active strategy. Investors commonly believe that value is the most powerful and consistent source of active return. This belief is based on the fact that growth, the other (mis-specified) investment style, typically delivers negative active returns. Momentum is currently not viewed as an investment style, partly because the academic literature has yet to agree on a reasonable rationale for the existence of momentum active returns.

The ActiveBeta Framework, in Chapter 6, provides a reasonable rationale for (1) the existence and persistence of momentum and value active returns and (2) the countercyclical (diversifying) nature of momentum and value active returns. In addition, there is no fundamental reason to believe that either momentum or value will provide higher active returns than the other in the long run. In fact, based on the Fama-French Factor Returns, momentum and value have provided similar after-cost risk-adjusted excess returns since 1927 in the United States. Therefore, if no excess return superiority can reasonably be established, then investors might be better served to pursue a combined capture of momentum and value, through the internally consistent ActiveBeta MVI or other such vehicles, to better diversify risk. Stated differently, value investing is characterized by significant downside risk and drawdown relative to the underlying market index. But, this risk can be substantially reduced by combining value with momentum, without sacrificing the active return potential in the long run. The risk-reducing characteristics, and the resulting superior risk-adjusted performance, of the ActiveBeta MVI are clearly evident in all the market universes we have studied. This analysis is presented in Chapter 8.

Customizable Solutions

The ActiveBeta Indexes have been created to demonstrate the working and practical applicability of the ActiveBeta Framework. These indexes provide nearly unlimited capacity and possess other characteristics that we hope will make them widely used as true benchmarks.

However, in addition to the ActiveBeta Index Methodology, we have developed other innovative and patent-pending methods and techniques that provide a more efficient and customizable capture of Active Betas. In particular, we discuss a new portfolio construction technique to create informationally efficient active portfolios. This technique allows for a more efficient capture of the information embedded in a signal, at an active risk level selected by the investor, but without the use of optimizers in order to provide complete transparency in portfolio construction and performance attribution.

We apply this innovative technique to create ActiveBeta Portfolios. An informationally efficient capture of Active Betas results in information ratios that are, on average, 50 percent higher than those achieved by the ActiveBeta Indexes. The historical performance of ActiveBeta Portfolios is presented in Chapter 10.

Summary

In summary, the ActiveBeta Momentum Index, ActiveBeta Value Index, and ActiveBeta MVI more accurately reflect the investment processes of active managers and, as such, constitute more appropriate performance benchmarks for active style and core managers. These indexes also provide a vehicle for an efficient, transparent, and cost-effective passive replication of a significant portion of traditional active management returns, thus allowing asset owners to streamline their portfolio structures and allocate their management fees and active risk budgets more efficiently.

OUR CONTRIBUTION

Through this work, we hope to contribute to the literature and practice of investment management in several ways. First, we present new research and broaden the scope of existing research on the behavior of short-term and long-term earnings expectations and the relationship of momentum and value with earnings expectations by studying and documenting new relationships, not just in the U.S. Large Cap universe but also in other market segments and geographical regions. This comprehensive coverage of markets and market segments, representing many different market environments over the time period studied, makes it possible to ascertain which relationships represent true fundamental influences, as opposed to potentially spurious links specific to a given market and/or a given time period.

Second, we study the various relationships mentioned earlier over a more recent time period, thus providing an out-of-sample confirmation of or challenge to previously published research.

Third, we link independent and narrowly focused pieces of research, both proprietary and published, to create a coherent and new ActiveBeta Framework. This framework establishes the existence of systematic sources of active equity returns and provides a fresh perspective on the rationale for the existence and persistence of active returns associated with broadly diversified momentum and value strategies, something that has been the subject of intense debate and discussion in the world of investments.

Fourth, we design and create an efficient and transparent vehicle, the ActiveBeta Indexes, for investors to implement the capture of systematic sources of active equity returns in actual portfolios. An internally consistent family of market capitalization-weighted momentum indexes, as well as combinations of momentum and value indexes, for various global equity markets and market segments does not currently exist in the marketplace. The ActiveBeta Indexes provide an effective capture of each systematic source, independently as well as in combination, to increase efficiency of capture and ease of implementation. ActiveBeta Indexes make it possible for investors to better understand the nature of active equity management returns and to structure equity portfolios more efficiently by directly capturing the systematic sources of active returns. Active Betas also challenge the current paradigm on style investing and offer a new framework for style replication and performance benchmarking of active style managers.

Finally, in addition to the ActiveBeta Framework and Index Methodology, we present and discuss other proprietary and patent-pending tools and techniques in Chapter 10. We introduce a new methodology for creating ActiveBeta style indexes at varying levels of market coverage and active risk to better suit the investment needs and portfolio structures of asset owners. This customization capability is not available or offered in the design of current style indexes. We also present an innovative portfolio construction technique to implement an informationally efficient capture of systematic sources of active equity returns.

STRUCTURE OF THE BOOK

ActiveBeta Indexes is structured in four main sections, outlined here, to help you navigate the book:

Section One: Background (Chapters 1 and 2). This section describes the basis for and evolution of market indexes. The discussion is then extended to style indexes, exploring the development and limitations of current style indexes.

Section Two: ActiveBeta Conceptual Framework (Chapters 3 through 6). This section details the theoretical framework behind the ActiveBeta Indexes, as well as the research supporting the framework. The concept of Active Betas is introduced, followed by our research into the nature and relationships of the systematic sources of active equity returns. Lastly, this section delves into the pricing and persistence of the systematic sources of active returns.

Section Three: ActiveBeta Indexes (Chapters 7 through 9). This section discusses the objectives of a style index and illustrates the methodology employed to create the ActiveBeta Indexes. The discussion then proceeds to a detailed analysis of the historical performance of the ActiveBeta Indexes. This section also demonstrates the various applications of the ActiveBeta Indexes, such as in style investing, performance attribution, portfolio structuring, and asset allocation.

Section Four: ActiveBeta Customizable Solutions (Chapters 10 and 11). This section offers a variety of alternative solutions for capturing the systematic sources of active equity returns. Several innovative techniques are detailed, including the patent-pending informationally efficient ActiveBeta Portfolios.

ACKNOWLEDGMENTS

The development of the ActiveBeta Framework has involved the contributions of numerous individuals, whom we wish to acknowledge. Their assistance to the authors has been invaluable.

Specifically, we wish to thank Jerry Chafkin and Andrew Lo at AlphaSimplex Group. Jerry and Andrew have offered their guidance to us over the past few years, helping to advance our thinking with regard to the theoretical development and practical application of the concepts in this book.

Our partners at FTSE also merit special consideration, particularly Mark Makepeace and Paul Walton. Mark, Paul, and their team provided valuable support and guidance to help with the launch of the FTSE ActiveBeta Indexes.

The ActiveBeta Framework, and this book in particular, could not have been accomplished without the tireless aid of our coworkers at Westpeak Global Advisors. We wish to especially thank Barbara Carlough for dealing with our numerous changes, Anders Fridberg for offering his research insight, Michelle Niemann for making this book more readable with her talent for editing and structuring, and Rachael Sax for producing much of the data and converting it into an understandable format.

Last, but certainly not least, Ingrid Hanson has been instrumental to completing this book. Ingrid kept a group of neophyte authors on track throughout the writing process, making sure the end result matched the original objective. Her attention to detail and organizational skills allowed us to produce the book you hold today.

SECTION One

Background

CHAPTER 1

The Evolution of Market Indexes and Index Funds