Beyond Traditional Beta 032015

For Institutional Investors Only Beyond Traditional Beta | March 2015 1

March 2015 Marketing Material For Professional Investors Only

Beyond Traditional Beta Deutsche Bank AG Deutsche Asset & Wealth Management db X-trackers ETF Team Winchester House 1 Great Winchester Street London EC2N 2DB United Kingdom Hotline: +44 (20) 7547 1747 [email protected] www.etf.deutscheawm.com Content 1. Introduction 1 2. Combining Active and Passive

Investment Strategies 4 3. Factors Explored 9 4. Beyond Traditional Beta 19

1. Introduction As a growing number of financial indices are developed to offer exposure to investment strategies in a systematic way, investors are confronted with new challenges. How should they evaluate these non-traditional indices? What distinguishes a valid strategy index from one based on a random back-test? Where does the dividing line between index-based and discretionary fund management now lie? Does an asset-based or factor-based approach to portfolio allocation make more sense? In this paper, authored by the passive asset management team at Deutsche Asset and Wealth Management (Deutsche AWM1

), we aim to answer these questions by putting the evolving role of financial indices into a historical context, by providing a framework to evaluate strategy indices and by suggesting how active and passive approaches to investment can continue to coexist.

The principal and long-standing view of a financial index is as a benchmark for the broad equity (or bond) market. In other words, the index is seen as a way of measuring the systematic risk and the associated return from the market as a whole. In traditional asset management theory, the sensitivity of a portfolio to market risk can be calculated and expressed as a coefficient, called beta. A broad market index has a beta of one. A portfolio of stocks with above-average sensitivity to market movements has a beta of more than one, and a defensive portfolio has a beta of below one. Positive alpha, the target of all active managers, means an above-zero residual return once beta has been accounted for. These conceptual shortcuts - beta for the passive, index-based part of an investment portfolio and alpha for the actively managed element that aims to add value, whether in absolute or relative terms - remain largely in place amongst practitioners. But this theoretical framework is inadequate in a world where non-traditional indices offer exposure to different segments of the market, blurring the lines between long-held concepts of passive and active management.

1 On p.18 we outline Deutsche AWMs experience in managing passive investment mandates and the role of Deutsche Banks quantitative equity research team.

Beyond Traditional Beta

For Professional Investors Only Beyond Traditional Beta | March 2015 2

Investment strategies based on a systematic screening of the market portfolio are not new. The principles of value-focused investing, for example, were set out by Graham and Dodd before the Second World War (Graham, Dodd, 1934) and have since been successfully adopted by many investors. But during the last two decades empirical evidence has accumulated of the existence of multiple factors that provide a more granular view of the behaviour of markets. Within an equity portfolio, for example, the evidence suggests that long-term performance can be explained not just by traditional beta (market risk) and alpha (the value added by active management) but also by exposure to factors such as value, size, momentum, volatility and quality. And factor-based research has now extended to the fixed income, commodity and currency markets. More recently, new equity factor indices have also enabled investors to capture exposure to factors in a systematic and transparent way. So the traditional concept of a single market beta has been replaced by the idea of coexisting, multiple market betas. From a portfolio perspective these betas can compete or work in a complementary fashion. Prompted by the adoption of factor strategies by some of the largest institutional investors in the world, factor-based asset allocation is a research area attracting great interest. In the remainder of this paper, following a brief summary of the historical evolution of financial indices, we examine equity factor index approaches in detail. We describe how asset allocation practice is changing to reflect the evolution of investment theory and review how passive and active approaches to asset management are likely to coexist in future. We set out a framework for evaluating the key requirements for a valid strategy index before examining in detail the risk/reward characteristics of common equity factor approaches. As markets move beyond the traditional concept of a single market beta, our objective is twofold: to provide investors with a conceptual framework to assess and evaluate newer forms of index in a consistent fashion; and to help investors combine factor approaches in their portfolios.



Source: Deutsche Bank AG

Portfolio Theory Risk/return

modelled at portfolio level

1950s

Capital Asset Pricing Model (CAPM)

Portfolio beta (systematic risk) distinguished from alpha (residual)

1960s

First commercial index funds developed (Samsonite, Vanguard Index Trust)

1970s

Popularisation of index funds

New indices segment markets by size, geography, development status

1980s

Fama-French/Carhart publish studies of size, value/momentum equity market factors

First ETFs launched 1990s

Research into factors, behavioural finance extended

Rapidly increasing usage of ETFs, index funds

2000s

Popularisation of factor-based investment strategies

Indexing overlaps with quantitative investment management

2010s

The Evolution of Indexing 1 The Evolution of Indexing: Theory and Practice? The use of indices in asset management has evolved over time, mirroring developments in academic theory (see Figure 1). Most of the index funds launched in the 1970s and 1980s aimed to replicate the broad market portfolio, an approach consistent with Portfolio Theory (Markowitz, 1952) and the Capital Asset Pricing Model (CAPM- Sharpe, 1964), theoretical market models set out in the preceding decades. Since the 1990s, increasing interest in factor-based approaches to investment have prompted widespread interest in non-traditional beta. Advances in computing power and extensive data histories have helped researchers to develop quantitative market models, while falls in trading costs have made real-time transactions in index-based portfolios (for example, via the ETF market) more feasible for a broad range of investors.



2. Combining Active and Passive Investment Strategies

Strategic Asset Allocation - the Traditional Approach The traditional approach to long-term (strategic) asset allocation is set out in Figure 2. Investment committees charged with setting asset allocation policy have typically followed a three-step process. The first step in the traditional process is to set target portfolio weightings (or weighting ranges) for individual asset classes, such as bonds, equities and alternatives. The key inputs in determining these target weightings are long-term estimates of individual asset classes returns and volatility, together with estimates of intra-asset class correlations. Allocations to individual equity and bond markets (or market regions) are typically also made at this first stage. The second step in the traditional process is to decide on the passive/active split within each asset class. Part of the target allocation to equity and bond markets is then run in a passive (i.e., index-tracking) manner, typically using a broad market index as the benchmark. Responsibility for the remainder of the portfolio is then assigned to active managers. The third step in the process is to select the fund managers. A traditional asset allocation process usually starts with the most important decision, i.e., the strategic asset class targets. Overall, the equity/bond split can be seen as the key decision from a return perspective: given equities greater historical volatility and long-term return premium over bonds, portfolios with greater risk appetite and higher return expectations target a higher equity weighting, while those with more conservative risk appetite target higher bond weightings. A 60/40 target equity/bond weighting split has been called the classic asset allocation recipe (Mesomeris, Wang, Salvini, Assetand-Fenoel, 2012). According to Christopher Brightman of Research Affiliates, a 60/40 portfolio of US equities and US bonds generated a nominal return of 7.6% per annum and a real (post-inflation) return of 5.4% per annum over a period of more than a century (1871-2010: Brightman, 2012). Many US pension funds currently have long-term return targets that are consistent with a 60/40 target asset allocation and with the expected continuation of its historical returns: according to Boston Colleges Centre for Retirement Research, the mean long-term nominal return assumption of US pension funds in 2012 was 7.75% (Munnell, Aubry, Hurwitz, 2013).2

2 In Europe pension funds return targets are somewhat lower. According to a survey of 190 European

pension funds with combined assets of 1.9 trillion, published in November 2014 by Create Research, the median long-term nominal return expectation (net of fees) is 5% a year (Rajan, 2014).




Unstable Risk Contributions and Correlations The principal drawback of this long-standing approach to asset allocation is that certain assumptions implicit in the model may not hold in practice. In particular, an asset class-based approach makes sense if the individual portfolio building blocks can be relied upon to provide adequate levels of diversification. This requires that the volatilities of individual asset classes, though different, remain largely stable over time; and that the correlations between asset classes are also relatively stable. As is evident from Figures 3 and 4, these assumptions cannot be relied upon.

Source: Deutsche Bank AG, Bloomberg Financial LP, 31/12/1991-31/12/2014, based on five-year rolling returns and first five-year period ending 31/12/1996.

Asset Selection

bonds/ equities/ alternatives

Passive/ Active Split Within Asset Class

broad market exposure for passive component

Manager Selection

active/ passive managers chosen within asset classes

93%

7%

S&P 500 US 10 Year Govt Bond

Risk Contribution to 60/40 Portfolio by Asset Class (January 1992 December 2014)

The traditional approach to asset allocation

3

2

2



Source: Deutsche Bank AG, Bloomberg Financial LP, 31/12/1991-31/12/2014, based on five-year rolling returns and first five-year period ending 31/12/1996.

Figure 3 (see previous page) shows that equity risk dominates the overall risk of a 60/40 portfolio consisting of US equities and 10-year US government bonds (equity risk contributed 93% of the portfolio risk during 1992-2014). And, based on a five-year rolling view of risk contribution, equities contributed 90% or more of risk at certain points during this historical window. Using one-year historical correlations and volatilities, Bruder and Roncalli showed that the equity component of a fund with a target 50/50 asset allocation split between global equities and global bonds contributed up to 100% of the portfolio risk on two occasions between 2000-2012, both during periods of market stress (Bruder, Roncalli, 2012). A policy of relying on alternative asset classes (such as hedge funds, property and commodities) to provide extra portfolio diversification has also been questioned since the financial crisis. Figure 5 illustrates the three-year rolling average pairwise correlation between asset classes in a portfolio consisting of equities, bonds and four popular alternative assets (hedge funds, real estate investment trusts, commodities and private equities since 1999). The 2008 liquidity crisis caused an unwelcome surge in pairwise correlations just at a time when investors were relying on the extra diversification potential of alternative assets.

Source: Source: Deutsche Bank AG, Bloomberg Financial LP, 01/01/1991-31/12/2014, based on three-year rolling returns and first three-year period ending 01/01/1994.

0% 10% 20% 30% 40% 50% 60% 70% 80% 90%

100%

1996 2000 2004 2008 2012 US 10 Year Govt Bond S&P 500

0%

20%

40%

60%

80%

100%

Jan 94 Jan 96 Jan 98 Jan 00 Jan 02 Jan 04 Jan 06 Jan 08 Jan 10 Jan 12 Jan 14 Average Pairwise Correlation

3-Year Rolling Average Pairwise Correlations between Equities, Bonds and Four Alternative Asset Classes

5-Year Rolling View of Risk Contribution to 60/40 Portfolio by Asset Class 4

5



Allocation by Factors - an Alternative Approach The period since the financial crisis has witnessed increasing interest in alternative approaches to asset allocation, based on the recognition that asset class volatilities and correlations can become particularly unstable in times of market stress. The behaviour of markets in 2008/09 may reflect the fact that notionally uncorrelated asset classes in fact share exposure to common underlying drivers of returns: it therefore makes sense to try to identify these common drivers or factors. In an influential review of the active management of the Norwegian Government Pension Fund portfolio, published in 2009, Professors Ang, Goetzman and Schaefer concluded that a significant component of [the portfolios] performance is explained by exposure to systematic factors (Ang, Goetzman, Schaefer, 2009). While conceding that the factors fared poorly during the financial crisis, the authors argued that exposure to such factors is actually appropriate for a longterm investor since the factors earn risk premiums over the long run. Most importantly, Ang, Goetzman and Schaefer recommended that the factor exposures be made part of the funds benchmark: in other words, that either passive (index-tracking) or active managers be given the explicit responsibility of generating the factor returns. Several recent academic papers have focused on a factor-based approach to asset allocation. Bender et al. showed that, based on historical evidence, a portfolio of factor risk premia produced annual returns similar to those of an asset-based portfolio, but with lower levels of correlation between factors (Bender, Briand, Nielsen, Stefek, 2010). More recently, Asness, Moskowitz and Pedersen have published evidence of the existence of factor-type returns across non-equity asset classes (Asness, Moskowitz, Pedersen, 2012). A comprehensive review of factor-based asset allocation across asset classes is beyond the scope of this paper. Nevertheless, we suggest how such a factor-based approach might work in Figure 6. In the next section, we focus in particular on equity factors and how to combine them.


Risk/ return budgeting

long-term return goals and risk

Factor Selection

Screening and selection of factors across asset classes

Factor weightings set

Manager selection

Manager appointed for index-tracking and factor mandates

A Factor-Based Asset Allocation Model 6



Active Management - a Continuing Role Moving from an asset class-based to a factor-based asset allocation does not necessarily imply a downsized role for active portfolio management. Although part of the long-term portfolio returns traditionally assumed to accrue from alpha may now be seen as the result of exposure to systematic factor risk premia (see Figure 7), the decision on whether to implement an actively or systematically rebalanced factor allocation is the focus of an increasingly intense debate. The optimal allocation to factors is indeed a complex topic. While several managers aim to generate active returns by switching between factors across time (i.e., implementing so-called factor timing strategies) most of the investors looking at factor investing are currently contemplating a passively rebalanced portfolio of factors. The main reason for this is probably that the risks induced by factor timing (namely missing a rally of an under-weighted strategy while in the meantime suffering from the drawdown of an over-weighted strategy) are ultimately considered as higher than the potential gains brought by such an approach.


A Changing View of the Sources of Portfolio Returns 7

Fund Governance under Factor-Based Allocation Moving from a traditional to a factor-based asset allocation may require a rethink of fund governance arrangements. This is because under a factor-based allocation scheme explicit responsibility for fund performance moves up the decision-making chain, away from active portfolio managers (in the traditional model) and towards those setting the factor allocation. It is therefore important that those responsible for fund governance (e.g., pension fund trustees) set out a transparent framework for reporting and performance evaluation. Trustees should be well acquainted with the risk-return characteristics of different factors and with the possibility that certain factors may underperform broad market indices for extended periods in particular market environments. Factor risk characteristics are explored in further detail in the next section.



3. Factors Explored As outlined in the previous section, a shift has taken place amongst both academics and practitioners regarding the sources of equity portfolio returns. From viewing returns as a simple combination of beta (systematic, broad-based market risk) and alpha (the result of active manager skill), investors now increasingly regard systematic beta - the return derived from exposure to identifiable equity factors - as a third, distinct component of performance. Systematic beta is also often referred to as strategic or smart beta. We prefer to avoid the latter term: traditional broad indices are likely to remain the principal type of market benchmark, so should not be seen as any less valid or useful than newer forms of index. How a Factor Model Works: an Example The way in which financial theory has evolved from the single-factor Capital Asset Pricing Model (CAPM) of the 1960s can be understood by means of an example: the three-factor model outlined by Eugene Fama and Kenneth French in 1992 (Fama, French, 1992). Fama and French found that, based on 50 years of equity market data, there was only a weak relation between beta (as understood in the CAPM framework) and average stock returns. Instead, they found that a multi-factor model using market beta plus two additional factors (size and value) had much better explanatory power. The Fama-French three-factor model is commonly stated as

where:

= the portfolios expected rate of return = the risk-free rate of return = the return of the market portfolio

= the size factor (Small Minus Big) = the value factor (High book value to market value ratio Minus

Low") and = coefficients for the market, size and value factors

= residual (alpha), the result of active management decisions. Since Fama and French introduced their three-factor model in 1992, researchers have produced evidence for the existence of additional factors, all with the ability to contribute to our understanding of equity market returns. These factors include quality (Sloan, 1996), momentum (Jegadeesh and Titman, 1993) and low beta (Baker, Bradley and Wurgler, 2011).



Screening Factors Below we explore in further detail the risk/return characteristics of these factors. However, since all factors are identified by means of the study of empirical data on securities returns, a question immediately arises: what distinguishes a true factor from one that may be the coincidental result of the data and period used for the analysis? In other words, how confident can investors be that the factor characteristics identified in sample will persist out of sample? Deutsche AWM aims to address such concerns by requiring the factors chosen in its equity factor product range to undergo a screening process, outlined in Figure 8.


To pass the screening process, equity factors should be: explainable (they should have a strong basis for existence and be

uniquely identifiable) established (they should have a well-established history in academic

literature) attractive (they should offer potential for attractive long-term

risk/return) persistent (there should be a rationale for the persistence of the

factor risk premium) accessible (they should be accessible at a level of cost that avoids

the dilution of the factor premium)

Equity Factors

Explainable

Established

Attactive Persistent

Accessible

The Factor Screening Process 8



Factor Characteristics: Value The risk premium associated with the value factor is well-established in academic and practitioner literature. Indeed, value-focused investment strategies predate academic theories such as CAPM and the subsequent multi-factor models of market behaviour. For example, in the 1930s Graham and Dodd described their preferred approach as the discovery of undervalued individual common stocks, which presumably are available even when the general market is not particularly low, and whose margin of safety resides in the discount at which the stock is selling below its minimum intrinsic value. Six decades later Fama and French described value (and size) as two of the three factors (together with market risk) with significant ability to explain stocks long-term returns. Value can be measured in different ways, for example via stocks price-to-book value ratios (the method used by Fama and French in their 1992 paper), their price-to-earnings ratios, their dividend yield or by a ratio of standardised earnings to enterprise value. The DB Equity Value Factor Index relies on two equally weighted measures of value to determine a companys composite value factor score: 12-month trailing dividend yield and the ratio of earnings before interest, tax, depreciation and amortisation (EBITDA) to enterprise value (EV). These two components can be seen as complementary: yield is a more defensive measure, while the EBITDA/EV ratio (earnings power) is a more cyclical measure of value. In the DB Equity Value Factor Index, companies yield and earnings power are compared to their sector averages and weightings are then neutralised across sectors. This is done with the aim of ensuring that structural differences in value scores between different types of company do not result in imbalanced sector weightings at the index level.


Explanations for the existence of factor risk premia commonly fall into two categories: risk-based and behavioural. The risk-based (or economic) explanation is more consistent with concepts of market efficiency, since the return associated with a factor risk premium is seen as compensation to an investor for bearing a type of systematic risk. Behavioural explanations for factor risk premia tend to focus more on investor irrationality or on market inefficiencies. Risk-based explanations for the value factor premium include: the higher potential exposure of value stocks (given their low valuation) to financial distress or default; higher cash flow risk; and value stocks greater sensitivity (by comparison with growth stocks) to economic downturns.

The Factor Screening Process 9



Behavioural explanations for the value premium focus more on investors perceived irrationality in mispricing growth stocks. According to this theory, in aggregate investors tend to extrapolate fast-growing companies growth rates too far into the future, overpaying for growth stocks as a result. Value companies, to which no such growth premium is attached, then tend to outperform the broader market over the long term. The attractiveness of a value factor investment strategy has been demonstrated by the widely documented long-term outperformance of broad market indices by value strategies (see, for example, Asness, Moskowitz, Pedersen, 2012). The likely persistence of a value strategys returns should be evaluated in the context of the explanation for the sources of the value factor premium: if the source of the premium is risk-based, it is unlikely that it will disappear; if the source of the premium reflects market inefficiencies, such as the tendency of institutional investors to benchmark their performance against broad market indices (which, by their nature, may be seen as having a natural tilt towards growth stocks), or the inability of many investors to sell short overpriced stocks, then these inefficiencies also seem likely to persist. Value can be seen as one of the most accessible factor investment strategies: despite the long-standing popularity of the strategy, there is little evidence that there has been dilution of the factor risk premium, nor of any significant investment capacity constraints in portfolios replicating value factor indices. By comparison with some other non-traditional indices, a value factor index does not rely, for example, on substantial weightings in smaller-capitalisation stocks.



Factor Characteristics: Quality In an influential 1996 paper, Richard Sloan of the University of Pennsylvania argued that investors appear unable to distinguish between the cash flow and accruals components of corporate earnings, even though empirical evidence shows that the relative magnitude of the cash and accruals components has an impact on the durability of current earnings. Fundamentally, accruals can be seen as less reliable earnings than cash flow, since they involve subjective judgments regarding the period in which revenues and expenses are recognised. Sloans paper followed another influential study of the potential for discrepancies between reported earnings and earnings based on cash flow, authored by stock analyst Terry Smith (Smith, 1992), then head of UK company research at UBS Phillips and Drew. Smith said he wished to investigate why several large and ostensibly profitable UK quoted companies, such as Polly Peck and British & Commonwealth, had gone bankrupt in the early 1990s, and concluded that investors needed to pay more attention to companies cash flow. The DB Equity Quality Factor index uses two accounting metrics to determine a companys quality score: return on invested capital (ROIC), a measure of profitability, is combined with the year-on-year change in accruals based on net operating earnings. The higher the ROIC score and the lower the accruals score, the higher the quality score. In the DB Equity Quality Factor index, ROIC and change in accruals are compared to their sector averages and weightings are then neutralised across sectors. Financial stocks are excluded from the index as a result of their special accounting characteristics.


The quality anomaly is well-established in academic and practitioner literature. Smiths and Sloans studies from the 1990s have been followed by a number of other papers examining the accruals effect (Chan, Chan, Jegadeesh and Lakonishok, 2001; Fairfield, Whisenant, Yohn, 2003). Explanations for the existence of the quality factor premium focus on behavioural inefficiencies. Simply stated, the quality premium may reflect investors lack of attention to the different components of reported earnings. This inattention then causes stocks with relatively high or low accruals components to be mispriced, an effect that can be exploited systematically. Several studies have shown that stocks selected according by the underlying components of the quality factor scoring process exhibit high attractiveness from a risk/return perspective. For example, a 2011 paper showed that high accruals are negatively related to future earnings changes and stock returns (Allen, Larson, Sloan, 2011).

Determining the Quality Factor Score 10



The likely persistence of the accruals anomaly has been widely studied. While some academics suggest that its effect has weakened over time (for example, Green, Hand, Soliman, 2009), others argue that it will persist as a reflection of behavioural inefficiencies among investors (Hirshleifer, Hou, Teoh, 2012). Additionally, Deutsche Banks policy of combining accruals and ROIC to measure a companys quality score means that low accruals must be complemented by a real economic return. The accruals anomaly has been documented across markets, industries and in larger and smaller companies, suggesting that accessibility to a quality strategy is relatively unrestricted.



Factor Characteristics: Momentum The momentum factor aims to exploit the tendency for trends in stock price movements to persist over time. Jegadeesh and Titman showed in a paper published in 1993 that a strategy which simultaneously buys past winners and sells past losers generates significant abnormal returns over holding periods of between three months and a year (Jegadeesh, Titman, 1993). These abnormal profits appear to be independent of market, size or value factors and can be observed in market data from different countries and regions, and across asset classes. The DB Equity Momentum Factor Index uses a two-step process to determine a companys momentum score (see Figure 11): first, the 11-month momentum of a stock is calculated by dividing its price one month earlier by the price 12 months earlier; second, the momentum is neutralised by adjusting for each companys specific risk.


The abnormal returns generated by a strategy that is long past winners and short past losers have been well-documented, ensuring that the momentum factor is one of the best-established risk premia in academic literature. Most explanations for the existence of the momentum factor focus on behavioural anomalies and structural market inefficiencies. Behavioural economists posit that investors initially underreact, then overreact to information affecting share prices. Institutional explanations for the momentum premium suggest that, by focusing on broad market indices for performance measurement purposes and by seeking to minimise the risk of underperforming such benchmarks, many investors reinforce stock price trends. The attractiveness and persistence of a momentum strategy have been established in multiple academic studies, across markets and over extended time periods. However, momentum tends to suffer sharp reversals during periods of changing market risk appetite. In other words, a pure momentum strategy generates high Sharpe ratios and positive alphas, but also negatively skewed return streams (see Daniel, Moskowitz, 2013). One way of mitigating this potential negative skewness in the returns of a momentum strategy is to aim to separate the momentum factor from a stocks embedded volatility exposure (see Alvarez, Kassam and Mesomeris, 2010): this is the methodology employed by the Deutsche Bank Equity Quantitative Strategy Group to determine a stocks momentum score. Momentum is one of the more accessible and high-capacity factor investment strategies: for example, its returns do not depend on the overweighting of less liquid segments of the market, such as small-capitalisation stocks.

Determining the Momentum Factor Score 11



Factor Characteristics: Low Beta The low beta (also known as the low volatility) anomaly may be regarded as one of the most puzzling findings in finance. In a recent, 40-year study of stock returns, low-beta portfolios were demonstrated to offer both higher returns and lower drawdowns than the broad market index (see Baker, Bradley, Wurgler, 2011).This establishes the attractiveness of an investment strategy focused on low-beta stocks. This outcome runs counter to the fundamental principles of portfolio theory and CAPM, under which a higher-beta portfolio should offer higher expected returns. In another, similar study a strategy of betting against beta was shown to have a Sharpe Ratio twice as high as a value strategy and 40% higher than a momentum strategy (see Frazzini, Pedersen, 2011). The low beta anomaly is not a recent discovery: it was pointed out over four decades ago (see Black, Jensen, Scholes, 1972) and confirmed by Haugen and Baker in 1991 (Haugen, Baker, 1991). In other words, this factor risk premium is one of the most well-established in academic literature. The DB Equity Low Beta Index uses a straightforward process to determine a companys low beta score (see Figure 12): the covariance of a stocks returns to the returns of the overall market is divided by the market variance, with all measurements taken over a five-year period.


The most popular explanations for the existence of the low-beta anomaly are behavioural. According to such theories, many investors overpay for perceived lottery stocks - i.e., those that promise high rates of return - leaving less glamorous low-beta stocks relatively underpriced and with the ability to outperform. In chasing such stock market winners, investors may be suffering from the behavioural trait of overconfidence (most respondents to surveys seem to think they are better-than-average drivers, or that they can outperform the stock market). The persistence of the returns from low-beta investing has been well-documented: Frazzini and Pedersen (2011) found that low-beta stocks had outperformed in 18 of 19 global equity markets they studied. The authors also documented a low-beta effect in government, corporate bond and futures markets. As the low-beta effect has been documented across markets and in different capitalisation segments, the accessibility of this factor for a systematic investment strategy can be ranked as relatively high.

Determining the Low Beta Factor Score 12



Combining Factors in a Portfolio Traditionally considered as difficult to forecast, the relative performance of the four factors described above depends on the macroeconomic environment. For example, during inflationary growth periods quality does relatively well, while momentum and value tend to outperform low beta. In a stagflationary period low beta does well, while momentum and quality outperform value. The value factor performs best in reflationary and disinflationary periods (see Figure 13, where the performance of the capitalisation-weighted MSCI World index is also shown for reference).

Source: Deutsche Bank AG. The returns shown for the value, quality, momentum and low beta factors are based on the returns of the Deutsche Bank Equity Factor Indices of the same name over the period 31.12.2000-31.12.2014. The Deutsche Bank Equity Factor Indices have no prior operating history and the returns illustrated are based on the retroactive application of the index methodology. Performance is calculated in total return USD and shown gross of dividend withholding tax, rebalancing and index costs. Past performance, actual or simulated, is not a reliable indicator of future returns.

Detailed performance metrics for the factor indices included in Figure 13 are shown in Appendix 1. The tendency of the four equity factors to perform differently in different macroeconomic environments suggests that investors may gain significant diversification benefits by combining the factors in a portfolio. While there are many approaches to combining equity factors, investors often consider the following weighting methods: Equal Weighting; Risk Parity Weighting (given the similar historical volatilities of the

four factors, this allocation is broadly similar to equal-weighting); Momentum-based allocation (e.g., selecting the two of the four

factors with the highest 11-month momentum and allocating 70% and 30%, respectively, to the factors with the highest and second-highest momentum, respectively).

A simulated performance record of the equal-weighted and momentum-based factor portfolios over the period 2000-2014 is shown in Figure 14 (see next page), with the MSCI World index included for comparison.

Simulated Factor Performance in Different Macro Environments 13



Source: Deutsche Bank, Bloomberg, 31.10.2001 31.12.2014. The performance data in this section is based on monthly data calculated in USD and shown gross of rebalancings and index costs. All rebalancing in the hypothetical portfolios are done at zero costs. The Equity Factor Indices have no prior operating history. All performance data is simulated and calculated by means of a retrospective application of the index methodology.The simulated returns of the hypothetical returns are based on the backtested performance of the Equity Factor Indices before the launch date (Sept. 14). Past performance, actual or simulated, is not a reliable indicator of future results.

3

3 The DB Quantitative Strategy Group was ranked #1 in both the 2011 & 2012 All-Europe Institutional

Investor Research Survey and the 2011 & 2012 US Research Institutional Investor Survey. Both the US and Europe teams were ranked in the top 3 in the Greenwich Survey for 2011.

Simulated Performance of Factor-Based Equity Portfolios

14

Choosing a Fund Structure The passive asset management team at Deutsche AWM offers a full range of capabilities, from the development of investment strategies, fund and ETF structuring to the physical replication of large, index-based portfolios. These portfolios track a variety of indices, from popular benchmarks to sophisticated strategy indices. The firm is already a prominent global issuer of ETFs (according to Lan et al., 2015, Deutsche AWM was the second largest ETF issuer in Europe and the fifth largest globally as at end 2014), as well as passive funds and segregated mandates. Now, Deutsche AWM has designated passive asset management as a key growth area and made significant investments in people, platforms and product development to achieve this goal. Through its strategic beta product range, Deutsche AWMs passive asset management team provides, cost-efficient, alternatively weighted strategies which aim to achieve optimized risk-adjusted returns when compared with their benchmarks. To deliver this objective, whether in flagship products like ETFs or in customised segregated mandates, Deutsche AWM Passive Asset Management may rely on internal and external research, index providers and its own portfolio construction expertise. Over the years Deutsche Bank has developed significant expertise in equity factors, as evidenced by the numerous publications from the banks Equity Quantitative Strategy Group. This groups research in factor identification and portfolio construction is considered to be market-leading by surveys of large institutional clients.3 This research expertise, combined with extensive experience in portfolio construction, optimised replication and efficient execution, puts Deutsche AWMs Passive Asset Management team in a strong position to implement a large range of equity factor portfolios.

14



4. Beyond Traditional Beta In this paper we have described how a broad range of transparent, systematic and standardised equity investment strategies, based upon developments in academic theory, is becoming available to investors. A rapidly increasing interest in strategies focused on factor risk premia is evident across a broad range of investor categories, including some of the worlds largest institutional investors. This trend poses challenges for traditional models of portfolio asset allocation and fund governance. We have outlined how alternative approaches in both these areas might be structured. We presented the screening process used by the DB Equity Factor Indices to identify factors with long-term viability and investment capacity, before reviewing in detail the risk/return characteristics of four popular equity factor strategies: value, quality, momentum and low beta. Finally, we discussed the relative performance of factor strategies in different market environments and examined potential ways of combining equity factors in a portfolio. Broad market indices remain central in investment management practice, both as the basis for the measurement of active managers performance and as the target for many index-tracking portfolios. However, the evolution of investment theory towards a world involving multiple, systematic sources of risk and return is now well-entrenched. As a result, non-traditional beta is only likely to grow in importance over the coming decades.

Source: Deutsche Bank AG, Bloomberg LP, 31.10.2001 31.12.2014. The DB Equity Factor Indices have no prior operating history before September 2014. All performance data is simulated and calculated by means of a retrospective application of the index methodology before the launch date . Performance is calculated in total return USD and shown gross of dividend withholding tax, rebalancing and index costs. Past performance, actual or simulated, is not a reliable indicator of future results. vs MSCI World over previous one year period. 2 vs MSCI World over the observation window starting on 31.10.2000 and ending on 31.12.2014.

Simulated Performance of DB Equity Factor Indices (2000-2014) Appendix 1



References

Allen, E., Larson, C., Sloan, R., Accrual Reversals, Earnings and Stock Returns, 2011

Alvarez, M-A., Kassam, A., Mesomeris, S., Factor Neutralization and Beyond, 2010

Ang, A., Goetzmann, W., Schaefer, S., Evaluation of Active Management of the Norwegian Government Pension Fund Global, 2009

Asness, C., Moskowitz, T., Pedersen, L., Value and Momentum Everywhere, 2012

Baker, M., Bradley, B., Wurgler, J., Benchmarks as Limits to Arbitrage: Understanding the Low-Volatility Anomaly, 2011

Bender, J., Briand, R., Nielsen, F., Stefek, F., A New Approach to Diversification, 2010

Black, F., Jensen, M., Scholes, M., The Capital Asset Pricing Model: Some Empirical Tests, 1972

Brightman, C., Expected Return, 2012

Bruder, B., Roncalli, T., Managing Risk Exposures Using the Risk Budgeting Approach, 2012

Chan, K., Chan, L., Jegadeesh, N., Lakonishok, J., Earnings Quality and Stock Returns, 2001

Daniel, K., Moskowitz, T., Momentum Crashes, 2013

Fairfield, P., Whisenant, S., Yohn, T., Accrued Earnings and Growth: Implications for Future Profitability and Market Mispricing, 2003

Fama, E., French, K., The Cross-Section of Expected Stock Returns, 1992 Frazzini, A., Pedersen, L., Betting against Beta, 2011

Graham, B. and Dodd, D., Security Analysis, 1934

Green, J., Hand, J., Soliman, M., Going, Going, Gone? The Demise of the Accruals Anomaly, 2009

Haugen, R., Baker, N., The Efficient Market Inefficiency of Capitalization-Weighted Stock Portfolios, 1991

Hirshleifer, D., Hou, K., Teoh, S., The Accrual Anomaly: Risk or Mispricing?, 2012

Jegadeesh, N., Titman, S., Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency, 1993

Lan, Mercado, Rajendra, Gademsetty, Deutsche Bank ETF Annual Review & Outlook, 2015

Markowitz, H., Portfolio Selection, 1952

Mesomeris, S., Wang, Y., Salvini, M., Avettand-Fenoel, J-R., A New Asset Allocation Paradigm, 2012

Munnell, A., Aubry, J-P., Hurwitz, J., How Sensitive is Public Pension Funding to Investment Returns?, 2013

Rajan, A., The Alpha behind Alpha: Rebooting the Pension Business Models, 2014.

Sharpe, W., Capital Asset Prices: A Theory of Market Equilibrium under Conditions of Risk, 1964

Sloan, R., Do Stock Prices Fully Reflect Information in Accruals and Cash Flows about Future Earnings?, 1996



Risk Factors db X-trackers UCITS ETFs Investors should note that the db X-trackers UCITS ETFs are not capital protected or

guaranteed and investors in each db X-trackers UCITS ETF should be prepared and able to sustain losses of the capital invested up to a total loss.

The value of an investment in a db X-trackers UCITS ETF may go down as well as up and past performance is not a guide to the future.

Investment in db X-trackers UCITS ETFs involve numerous risks including among others, general market risks relating to the relevant index, credit risks on the provider of index swaps utilised in the db X-trackers UCITS ETFs, exchange rate risks, interest rate risks, inflationary risks, liquidity risks and legal and regulatory risks.

Not all db X-trackers UCITS ETFs may be suitable for all investors so please consult your financial advisor before you invest in a db X-trackers UCITS ETF

db X-trackers UCITS ETFs following a direct replication investment policy, may engage in securities lending. In these instances the db X-trackers UCITS ETFs face the risk of the borrower not returning the securities lent by the db X-trackers UCITS ETF due to e.g. a default situation and the risk that collateral received by the db X-trackers UCITS ETFs may be liquidated at a value lower than the value of the securities lent out by the db X-trackers UCITS ETFs.

db X-trackers UCITS ETFs employing an indirect investment policy will use OTC derivative transactions. There are appropriate arrangements in place to reduce the exposure of the db X-trackers UCITS ETF to the counterparty, in some cases up to 100%, but there is no guarantee that such arrangements will be perfect and the counterparty may lose up to 100% of its investment if the counterparty defaults.

db X-trackers UCITS ETFs may be unable to replicate precisely the performance of an index.

An investment in a db X-trackers UCITS ETFs is dependent on the performance of the underlying index less costs, but an investment is not expected to match that performance precisely. There may be a tracking difference between the performance of the db X-trackers UCITS ETFs and the underlying index e.g. due to the impact of fund management fees and administrative costs among other things. The returns on the db X-trackers UCITS ETFs may not be directly comparable to the returns achieved by direct investment in the underlying assets of the db X-trackers UCITS ETFs or the underlying index. Investors' income is not fixed and may fluctuate.

db X-trackers UCITS ETFs shares may be denominated in a currency different to that of the traded currency on the stock exchange in which case exchange rate fluctuations may have a negative effect on the returns of the fund.

The value of any investment involving exposure to foreign currencies can be affected by exchange rate movements.

Tax treatment of the db X-trackers UCITS ETFs depends on the individual circumstances of each investor. The levels and bases of, and any applicable relief from, taxation can change.

DB Affiliates significant holdings: Investors should be aware that Deutsche Bank or its affiliates (DB Affiliates) may from time to time own interests in any individual db X-trackers UCITS ETF which may represent a significant amount or proportion of the overall investor holdings in the relevant db X-trackers UCITS ETF. Investors should consider what possible impact such holdings by DB Affiliates may have on them. For example, DB Affiliates may like any other Shareholder ask for the redemption of all or part of their Shares of any Class of the relevant db X-trackers UCITS ETF in accordance with the provisions of this Prospectus which could result in (a) a reduction in the Net Asset Value of the relevant db X-trackers UCITS ETF to below the Minimum Net Asset Value which might result in the Board of Directors deciding to close the db X-trackers UCITS ETF and compulsorily redeem all the Shares relating to the db X-trackers UCITS ETF or (b) an increase in the holding proportion of the other Shareholders in the db X-trackers UCITS ETF beyond those allowed by laws or internal guidelines applicable to such Shareholder.

db X-trackers shares purchased on the secondary market cannot usually be sold directly back to the db X-trackers ETFs. Investors must buy and sell shares on a secondary market with the assistance of an intermediary (e.g. a stockbroker) and may incur fees for doing so. In addition, investors may pay more than the current net asset value when buying shares and may receive less than the current net asset value when selling them.

Full disclosure on the composition of the db X-trackers UCITS ETFs portfolio and information on the Index constituents, as well as the indicative Net Asset Value, is available free of charge at www.etf.deutscheawm.com . For further information regarding risk factors, please refer to the risk factors section of the prospectus, or the Key Investor Information Document.



Disclaimer This marketing communication is intended for professional clients / qualified investors only. The information contained in this document does not constitute investment advice. Complete information on the sub-funds including risks can be found in the prospectus of Concept Fund Solutions plc ("CFS") as well as the relevant supplements in their prevailing version. These and the relevant key investor information documents constitute the only binding sales documents for the sub-funds. Germany: Investors can obtain these documents along with copies of the articles of association and the latest published annual and semi-annual reports from the Paying and Information Agent, (Deutsche Bank AG, Institutional Cash & Securities Services, Issuer Services, Post IPO Services, Taunusanlage 12, 60325 Frankfurt am Main (Germany)) in German in printed form free of charge, or download them from www.etf.db.com. Austria: Investors can obtain these documents for sub-funds that are admitted for distribution in Austria, along with copies of the articles of association and the latest published annual and semi-annual reports from the Austrian Paying Agent, Deutsche Bank sterreich AG, Stock im Eisen-Platz 3, A-1010 Vienna, in German in printed form free of charge, or download them from www.etf.db.com. Switzerland: Investors can obtain these documents along with copies of the articles of association and the latest published annual and semi-annual reports from the Swiss Representative in German in printed form free of charge, or download them from www.etf.db.com. The Representative and Paying Agent in Switzerland for the sub-funds is Deutsche Bank (Suisse) S.A., Place des Bergues 3, 1201 Geneva and its branch offices in Zurich and Lugano. All statements of opinion reflect the current assessment of Deutsche Bank AG and are subject to change without notice.

All statements of opinion reflect the current assessment of Deutsche Bank AG and are subject to change without notice.

As explained in the prospectus of CFS plc, distribution of the aforementioned sub-funds is subject to restrictions in certain jurisdictions. For this reason, the sub-funds mentioned herein may neither be offered nor sold in the USA, nor to, or for the account of, US persons or persons residing in the USA. This document and the information contained herein may only be distributed and published in jurisdictions in which such distribution and publication is permissible in accordance with applicable law in those jurisdictions. Direct or indirect distribution of this document is prohibited in the USA as well as to or for the account of US persons and persons residing in the USA. All prices shown here are provided for informational purposes only and do not serve as an indicator of trading prices.

The calculation of performance uses the BVI (Bundesverband Investment und Asset Management) method and therefore does not take the Upfront Sales Charge into account. Individual costs such as fees and other charges, which would have a negative impact on the performance, have not been taken into account. The information contained in this document does not constitute a financial analysis but qualifies as marketing communication. This marketing communication is neither subject to all legal provisions ensuring the impartiality of financial analysis nor to any prohibition on trading prior to the publication of financial analyses.



db X-trackers_1700

This document is intended for discussion purposes only and does not create any legally binding obligations on the part of Deutsche Bank AG and/or its affiliates (DB).This material was not produced, reviewed or edited by the Research Department, except where specific documents produced by the Research Department have been referenced and reproduced above. Without limitation, this document does not constitute an offer, an invitation to offer or a recommendation to enter into any transaction. When making an investment decision, you should rely solely on the final documentation (including the most recent key investor information document, which is available in English and certain other relevant languages on www.etf.db.com) relating to the transaction and not the summary contained herein. These documents are available free of charge from Deutsche Bank, London Branch. DB is not acting as your financial adviser or in any other fiduciary capacity with respect to this proposed transaction. The transaction(s) or products(s) mentioned herein may not be appropriate for all investors and before entering into any transaction you should take steps to ensure that you fully understand the transaction and have made an independent assessment of the appropriateness of the transaction in the light of your own objectives and circumstances, including the possible risks and benefits of entering into such transaction. For general information regarding the nature and risks of the proposed transaction and types of financial instruments please go to www.globalmarkets.db.com/riskdisclosures.You should also consider seeking advice from your own advisers in making this assessment. If you decide to enter into a transaction with DB, you do so in reliance on your own judgment. The information contained in this document is based on material we believe to be reliable; however, we do not represent that it is accurate, current, complete, or error free. Assumptions, estimates and opinions contained in this document constitute our judgment as of the date of the document and are subject to change without notice. Any projections are based on a number of assumptions as to market conditions and there can be no guarantee that any projected results will be achieved. Past performance is not a guarantee of future results. The DB Factor Indices have no prior operating history. All performance data is simulated and calculated by means of a retrospective application of the index methodology. Performance is calculated in total return USD and shown gross of dividend withholding tax, rebalancing and index costs. Past performance, actual or simulated, is not a reliable indicator of future results. Any opinions expressed herein may differ from the opinions expressed by other DB departments including the Research Department. DB may engage in transactions in a manner inconsistent with the views discussed herein. DB trades or may trade as principal in the instruments (or related derivatives), and may have proprietary positions in the instruments (or related derivatives) discussed herein. DB may make a market in the instruments (or related derivatives) discussed herein. Investors should be aware that DB may from time to time own interests in any individual sub-fund of CFS which may represent a significant investment or proportion of the overall investor holdings in the relevant sub-fund. Investors should consider what possible impact such holdings or any disposal thereof by DB may have on them. The distribution of this document and availability of these products and services in certain jurisdictions may be restricted by law. You may not distribute this document, in whole or in part, without our express written permission. DB SPECIFICALLY DISCLAIMS ALL LIABILITY FOR ANY DIRECT, INDIRECT, CONSEQUENTIAL OR OTHER LOSSES OR DAMAGES INCLUDING LOSS OF PROFITS INCURRED BY YOU OR ANY THIRD PARTY THAT MAY ARISE FROM ANY RELIANCE ON THIS DOCUMENT OR FOR THE RELIABILITY, ACCURACY, COMPLETENESS OR TIMELINESS THEREOF. Any individual intending to invest in any investment described in this document should consult his/her professional adviser and ensure that he/she fully understands the risks associated with making such an investment and has sufficient financial resources to sustain any loss that may arise from it. CFS is an investment company with variable capital incorporated on 17 November 2004 and authorised in Ireland as an undertaking for collective investment in transferable securities pursuant to the European Communities (Undertakings for Collective Investment in Transferable Securities) Regulations, 2011 (S.I. No. 352 of 2011) as amended. The registered office of CFS (Reg. No.: 393802), a company registered in Ireland, is 78 Sir John Rogersons Quay, Dublin 2, Ireland. DB is authorised under German Banking Law (competent authority: BaFin Federal Financial Supervising Authority) and DB AG London Branch is regulated by the Financial Conduct Authority for the conduct of UK business. Deutsche Bank AG 2015. All rights reserved.

Deutsche Asset & Wealth Management represents the asset management and wealth management activities conducted by Deutsche Bank AG or any of its subsidiaries. Clients will be provided Deutsche Asset & Wealth Management products or services by one or more legal entities that will be identified to clients pursuant to the contracts, agreements, offering materials or other documentation relevant to such products or services. 2015 Deutsche Asset & Wealth Management. All rights reserved.

Documents

Beyond Traditional Beta 032015