EDHEC-Risk Publication Risk Allocation Factor Investing Smart Beta PDF

An EDHEC-Risk Institute Publication
Risk Allocation, Factor Investing

and Smart Beta: Reconciling
Innovations in Equity Portfolio
Construction
July 2014
with the support of
Institute
With the support of Amundi ETF & Indexing. Research conducted as part of the Amundi ETF & Indexing Research Chair “ETFs and
Core Satellite Portfolio Management". Any remaining errors or omissions are the sole responsibility of the authors.
2 Printed in France, July 2014. Copyright EDHEC 2014.
The opinions expressed in this survey are those of the authors and do not necessarily reflect those of EDHEC Business School
and Amundi ETF & Indexing.
Risk Allocation, Factor Investing and Smart Beta: Reconciling Innovations in Equity Portfolio Construction — July 2014
Table of Contents
Executive Summary.................................................................................................. 7
Introduction................................................................................................................15
1. Selecting Desired Factor Exposures................................................................19
2. Designing Efficient and Investable Proxies for Risk Premia.................... 25
3. Risk Allocation with Smart Factor Indices.....................................................39
Conclusion................................................................................................................63
Appendix...................................................................................................................65
References................................................................................................................77
About Amundi ETF & Indexing.............................................................................83
About EDHEC-Risk Institute.................................................................................84
EDHEC-Risk Institute Publications and Position Papers (2011-2014).........89
An EDHEC-Risk Institute Publication 3

About the Authors
Noël Amenc is a professor of finance at EDHEC Business School, director of

EDHEC-Risk Institute, and chief executive officer of ERI Scientific Beta. He has
conducted active research in the fields of quantitative equity management,
portfolio performance analysis, and active asset allocation, resulting in
numerous academic and practitioner articles and books. He is on the editorial
board of the Journal of Portfolio Management and serves as associate editor
of the Journal of Alternative Investments and the Journal of Index Investing.
He is a member of the Monetary Authority of Singapore Finance Research Council
and the Consultative Working Group of the European Securities and Markets
Authority Financial Innovation Standing Committee. He co-heads EDHEC-Risk
Institute’s research on the regulation of investment management. He holds a
master’s in economics and a PhD in finance from the University of Nice.
Romain Deguest is a senior research engineer at EDHEC-Risk Institute.

His research on portfolio selection problems and continuous-time asset-
pricing models has been published in leading academic journals and presented
at numerous seminars and conferences in Europe and North America. He
holds masters degrees in Engineering (ENSTA) and Financial Mathematics
(Paris VI University), as well as a PhD in Operations Research from Columbia
University and Ecole Polytechnique.
Felix Goltz is Head of Applied Research at EDHEC-Risk Institute. He carries

out research in empirical finance and asset allocation, with a focus on
alternative investments and indexing strategies. His work has appeared in
various international academic and practitioner journals and handbooks.
He obtained a PhD in finance from the University of Nice Sophia-Antipolis
after studying economics and business administration at the University of
Bayreuth and EDHEC Business School.
Ashish Lodh is Senior Quantitative Analyst. He does research in empirical

finance, focusing on equity indexing strategies and risk management. He
has a master’s in management with a major in finance from ESCP Europe. He
also has a bachelor’s degree in chemical engineering from Indian Institute
of Technology.
Lionel Martellini is professor of finance at EDHEC Business School and

scientific director of EDHEC-Risk Institute. He has graduate degrees in
economics, statistics, and mathematics, as well as a PhD in finance from
the University of California at Berkeley. Lionel is a member of the editorial
board of the Journal of Portfolio Management and the Journal of Alternative
Investments. An expert in quantitative asset management and derivatives
valuation, his work has been widely published in academic and practitioner
journals and has co-authored textbooks on alternative investment strategies
and fixed-income securities.
4 An EDHEC-Risk Institute Publication
About the Authors
Eric Shirbini is Global Product Specialist with ERI Scientific Beta. Prior to
joining EDHEC-Risk Institute, Eric was a quantitative analyst at UBS, BNP Paribas
and Nomura International. During this time he worked on a diverse range of
topics including multi-factor models, fundamental stock valuation, equity
market indices, portfolio construction and portfolio trading. At BNP Paribas
Eric managed a team of analysts who were responsible for the Global Equity
Research Database. He holds a BSc and PhD from University College London
and an MBA from CASS Business School.

About the Authors

Executive Summary

Executive Summary
Introduction: From Cap-Weighted Designing Efficient and Investable

Indices to Smart Factor Indices Proxies for Risk Premia
This publication argues that current smart In this study we focus on four well-known
beta investment approaches only provide rewarded factors – the Size and Value
a partial answer to the main shortcomings factors (Fama and French (1993)1), the
of capitalisation-weighted (cap-weighted) Momentum factor (Carhart (1997)2) and
indices, and develops a new approach to the Low Volatility factor (Ang et al. (2006,
equity investing referred to as smart factor 2009)3). For each rewarded factor, we
investing. It provides an assessment of the introduce a corresponding smart factor
benefits of simultaneously addressing the index, which can be regarded as an efficient
two main shortcomings of cap-weighted investable proxy for a given risk premium. In
indices, namely their undesirable factor a nutshell, a risk premium can be thought
exposures and their heavy concentration, of as a combination of a risk (exposure)
by constructing factor indices that explicitly and a premium (to be earned from the
seek exposures to rewarded risk factors risk exposure). Smart factor indices have
while diversifying away unrewarded risks. been precisely engineered to achieve a
The results we obtain suggest that such pronounced factor tilt emanating from
1 - Fama, E. and K. French.
1993. Common Risk Factors
smart factor indices lead to considerable the stock selection procedure (relevant
in the Returns on Stocks and improvements in risk-adjusted performance. risk exposure), as well as high Sharpe ratio
Bonds. Journal of Financial
Economics 33(1): 3-56. For long-term US data, smart factor emanating from the efficient diversification
2 - Carhart, M.M. 1997. On
Persistence in Mutual Fund
indices for a range of different factor tilts of unrewarded risks related to individual
Performance. Journal of roughly double the Sharpe ratio of the stocks (fair reward for the risk exposure).
Finance 52(1): 57-82.
3 - Ang, A., R. Hodrick, Y. broad cap-weighted index. Outperformance The access to the fair reward for the
Xing and X. Zhang. 2009. The
Cross-Section of Volatility
of such indices persists at levels ranging given risk exposure is obtained through
and Expected Returns. The from 2.92% to 4.46%, even when assuming a well-diversified, also known as smart-
Journal of Finance 61(1):
259-299. Ang, A., Hodrick, unrealistically high transaction costs. weighted, portfolio, as opposed to a
Y. Xing and X. Zhang.
2009. High Idiosyncratic
Moreover, by providing explicit tilts to concentrated cap-weighted portfolio, of
Volatility and Low Returns: consensual factors, such indices improve the selected stocks so as to ensure that the
International and Further U.S.
Evidence. Journal of Financial upon many current smart beta offerings largest possible fraction of individual stocks'
Economics 91: 1-23.
where, more often than not, factor tilts unrewarded risks is eliminated.
result as unintended consequences of ad
hoc methodologies. In fact, this publication The results in Exhibit A confirm that the
shows that by using consensual results combination of relevant security selection
from asset pricing theory concerning and appropriate weighting schemes in
both the existence of factor premia and a two-step process leads to substantial
the importance of diversification, it is improvements in risk-adjusted performance
possible to go beyond existing smart with respect to the use of a standard
beta approaches which provide partial cap-weighted index, which typically implies
solutions by only addressing one of these an inefficient set of factor exposures and
issues. an excess of unrewarded risk.
On the one hand, starting with a focus

on the systematic risk exposure, we find

Executive Summary
that a higher Sharpe ratio can be achieved that the reward harvested through the
with the same weighting scheme, here a factor exposure is a compensation for a
cap-weighting scheme, for stocks selected corresponding increase in risk. In contrast,
on the basis of their loadings on the value, we note that high momentum and low
size, momentum and low volatility factors, volatility selections lead to lower levels
compared to the case where the full universe of max Drawdown compared to the no
is held in the form of a cap-weighted selection case, suggesting that the excess
portfolio. performance earned on these two factors
has at best a behavioural explanation, and
The results we obtain, reported in Exhibit is not necessarily related to an increased
A, show that while the Sharpe ratio of the riskiness.
cap-weighted index is 0.24 on the sample
period, it reaches values as high as 0.39 for On the other hand, shifting to the
a mid cap stock selection, 0.30 for a high management of specific risk exposures,
momentum stock selection, 0.29 for a low we find that even higher levels of Sharpe
volatility stock selection or 0.35 for a value ratio can be achieved for each selected
stock selection. These results suggest that factor exposure through the use of a
4 - Diversified Multi-Strategy
weighting is an equal-
a systematic attempt to harvest equity risk well-diversified weighting scheme, which we
weighted combination of premia above and beyond broad market take to be an equally-weighted combination
the following five weighting
schemes - Maximum exposure leads to additional risk-adjusted of 5 popular smart weighting schemes.4
Deconcentration, Diversified
Risk Weighted, Maximum
performance. It should be noted at this Thus, the Sharpe ratio of the so-called
Decorrelation, Efficient stage that substantially higher levels diversified multi-strategy combination
Minimum Volatility and
Efficient Maximum Sharpe of max Drawdown are incurred for the reaches 0.52 for mid cap stocks, 0.48
Ratio (see www.scientificbeta.
com for more details).
mid cap and value selections, confirming for high momentum stocks, 0.50 for low
Exhibit A: Performance comparison of USA Cap Weighted Factor Indices and USA Multi-Strategy Factor Indices. The exhibit shows
the absolute performance, relative performance, and risk indicators for Cap Weighted (CW) Factor Indices and Multi-Strategy
Factor Indices for four factor tilts – Mid Cap, High Momentum, Low Volatility, and Value. The complete stock universe consists of
the 500 largest stocks in the USA. The benchmark is the cap-weighted portfolio of the full universe. The yield on secondary market
US Treasury Bills (3M) is the risk-free rate. The return-based analysis is based on daily total returns from 31/12/1972 to 31/12/2012
(40 years). All weight based statistics are average values across 160 quarters (40 years) from 31/12/1972 to 31/12/2012.
Broad Mid Cap High Momentum Low Volatility Value
CW
CW Diversified CW Diversified CW Diversified CW Diversified
Multi Multi Multi Multi
Strategy Strategy Strategy Strategy
Ann. Returns 9.74% 12.54% 14.19% 10.85% 13.30% 10.09% 12.64% 11.78% 14.44%
Ann. Volatility 17.47% 17.83% 16.73% 17.60% 16.30% 15.89% 14.39% 18.02% 16.55%
Sharpe Ratio 0.24 0.39 0.52 0.30 0.48 0.29 0.50 0.35 0.54
Historical Daily 1.59% 1.60% 1.50% 1.64% 1.50% 1.42% 1.28% 1.59% 1.47%
5% VaR
Max Drawdown 54.53% 60.13% 58.11% 48.91% 49.00% 50.50% 50.13% 61.20% 58.41%
Ann. Excess - 2.80% 4.45% 1.10% 3.56% 0.35% 2.90% 2.04% 4.70%
Returns
Ann. Tracking Error - 5.99% 6.80% 3.50% 4.88% 4.44% 6.17% 4.74% 5.82%
95% Tracking Error - 9.39% 11.56% 6.84% 8.58% 9.20% 11.53% 8.72% 10.14%
Information Ratio - 0.47 0.66 0.32 0.73 0.08 0.47 0.43 0.81
Source: scientificbeta.com.

Executive Summary
volatility stocks and 0.54 for value stocks. in consideration when designing a
These results suggest that multi-strategy sophisticated allocation methodology (see
factor-tilted indices obtain the desired Exhibit B).
factor tilts without undue concentration,
which provides an explanation for their The first, and arguably most important,
superior risk-adjusted performance with dimension relates to whether risk is defined
respect to the cap-weighted combination by the investor from an absolute perspective
of the same selection of stocks. in the absence of a benchmark, or whether
it is instead defined in relative terms with
Overall, it appears that the combined effects respect to an existing benchmark, which
of a rewarded factor exposure ensured by a is more often than not a cap-weighted
dedicated proper security selection process index. In the former situation, one would
and an efficient harvesting of the associated use volatility as a relevant risk measure,
premium through improved portfolio while tracking error with respect to the
diversification leads to a Sharpe ratio cap-weighted index would instead be used
improvement of around 100% compared in the latter case.
to the broad cap-weighted index.
The second dimension concerns whether
one would like to incorporate views
Risk Allocation with Smart Factor regarding factor returns in the optimisation
Indices process. While additional benefits can be
Once a series of smart factor indices have obtained from the introduction of views
been developed for various regions of on factor returns at various points of the
the equity universe, they can be used as business cycle, we focus in what follows
attractive building blocks in the design of only on approaches that are solely based
an efficient allocation to these multiple on risk parameters, which are notoriously
risk premia. easier to estimate with a sufficient degree
of robustness and accuracy (Merton
In an attempt to identify, and analyse (1980)). The third dimension is related to
the benefits of, the possible approaches the objective of the allocation procedure.
to efficient risk allocation across the Indeed, there are several possible targets for
various smart factor indices, we identify the design of a well-diversified portfolio of
four main dimensions that can be taken factor exposure, depending upon whether
Exhibit B – The Various Dimensions of Allocation Methodologies across Assets or Risk Factors.

Executive Summary
one would like to use naive approaches parity constraints – that is we restrict our
(equal dollar allocation or equal risk analysis to portfolios such that each one of
allocation) or scientific approaches based the four factors has the same contribution
on minimising portfolio risk (volatility in to the portfolio volatility. More precisely
the absolute return context or tracking we consider a global minimum variance
error in the relative return context). The portfolio subject to factor risk parity
fourth and last dimension related to the constraints (denoted by Multi Beta GMV
presence of various forms of constraints - Fact. Allocation), as well as a maximum
such as minimum/maximum weight deconcentration portfolio subject to factor
constraints, turnover constraints, or factor risk parity constraints (denoted by Multi
exposure constraints, which are obviously Beta MDecon - Fact. Allocation), a portfolio
highly relevant in the context of risk factor which can be regarded as the closest
allocation. approximation to an equally-weighted
portfolio that satisfies the factor risk parity
To illustrate the benefits of an efficient constraints.5
allocation to smart factor indices, we
consider a second dataset over the 10-year We note that the GMV allocation process
5 - So as to avoid introducing
overly strong biases in
period from 31-Dec-2003 to 31-Dec leads to the lowest volatility, as expected.
country exposures, we also 2013 using five sub-regions of the global When analysing the performances in
introduce a set of constraints
dedicated to ensure that developed universe, namely, US, UK, Dev. terms of bull versus bear market regimes
each one of the five regions
is not too strongly under- or
Europe Ex UK, Japan, Dev. Asia Pacific (defined as positive versus negative returns
over-represented with respect Ex Japan. Using the four smart multi- for the cap-weighted index), we observe
to its market capitalisation
in the cap-weighted global strategy indices as proxies for the value, that the addition of risk parity constraints
developed index.
size, momentum and volatility rewarded to the GMV allocation tends to stabilise
tilts in each region, we obtain a total of the returns across market conditions. For
(5x4) 20 constituents. example, in the absence of a factor risk
parity constraint, the GMV allocation leads
Absolute Return Perspective to a massive outperformance of 11.94%
We start from the absolute return with respect to the cap-weighted index in
perspective and consider in Exhibit bear markets, which is due to the almost
C five allocation strategies to the 20 exclusive domination of the low volatility
aforementioned smart factor indices – an factor, with a defensive bias that proves
equal dollar contribution portfolio (denoted extremely useful in such market conditions.
by Multi Beta EW Allocation), an equal risk On the other hand, the relative return in
contribution portfolio (denoted by Multi bull markets is negative at -3.90% due
Beta ERC Allocation), and then a global to the performance drag associated with
minimum variance portfolio (denoted by exclusively holding defensive equity
Multi Beta GMV Allocation). Given that exposure in bull market conditions. In
these allocation strategies lead in general to this context, one key advantage of the
concentrated factor exposures (for example introduction of factor risk parity constraints
the minimum variance portfolio heavily is that it leads to a much more balanced
loads on the low volatility factor indices in return profile across market conditions
each region), we also introduce factor risk with positive outperformance in both

Executive Summary
Exhibit C – Multi Beta Allocations across Smart Factor Indices (Developed Universe). The exhibit shows the allocations of the EW,
ERC, GMV under geographical constraints, and both the Max-Deconcentration and GMV Diversified Multi-Strategy indices under
geographical and risk parity constraints, invested in the 20 Diversified Multi-Strategy indices with stock selection based on mid cap,
momentum, low volatility, and value in the five sub-regions – US, UK, Dev. Europe Ex UK, Japan and Dev. Asia Pacific Ex Japan. The
period goes from 31-December-2003 to 31-December-2013.
Developed Diversified Multi-Strategy
(2004-2013)
CW (All Multi Beta EW Multi Beta ERC Multi Multi Beta Multi Beta
Stocks) Allocation Allocation Beta GMV MDecon-Fact GMV-Fact
Allocation Allocation Allocation
Ann. Returns 7.80% 11.37% 11.07% 10.57% 11.17% 10.88%
Ann. Volatility 17.09% 15.32% 14.33% 12.84% 17.23% 17.21%
Sharpe Ratio 0.36 0.64 0.66 0.70 0.56 0.54
Max Drawdown 57.13% 54.40% 51.82% 45.07% 55.22% 55.32%
Excess Returns - 3.56% 3.27% 2.76% 3.37% 3.07%
Tracking Error (TE) - 6.75% 7.51% 6.36% 3.08% 3.34%
95% TE - 13.84% 14.89% 11.85% 5.19% 5.55%
Information Ratio - 0.53 0.44 0.43 1.09 0.92
Outperf Prob (3Y) - 98.36% 89.34% 89.07% 100.00% 100.00%
Max Rel. Drawdown - 6.35% 9.54% 13.10% 4.03% 5.47%
Ann. Rel. Returns - 2.50% 0.48% -3.90% 3.81% 3.18%
Ann. Rel. Ret. Bear - 4.65% 6.74% 11.94% 2.50% 2.66%
bear and bull markets (at 2.66% and 3.18% In this context, a multi-smart beta solution
respectively). can be regarded as a reliable cost-efficient
substitute to expensive active managers,
We also find that the introduction of and the most relevant perspective is not an
factor risk parity constraints has led to a absolute return perspective, but a relative
substantial improvement in information return perspective, with respect to the
ratios with an information ratio above 1 cap-weighted index.
for the Max-Deconcentration allocation
under risk parity constraints. Interestingly In what follows, we focus on two approaches,
we note that the introduction of factor a naive diversification approach leading
risk parity constraints leads to 100% to a relative equal risk allocation (R-ERC)
out-performance probabilities over a portfolio, which focuses on equalising the
three-year horizon. Overall, all tested contribution of the smart factor-tilted
strategies lead to extremely substantial indices to the portfolio tracking error, and a
levels of outperformance with respect to scientific diversification approach leading to
the cap-weighted index, with excess returns a relative global minimum variance (R-GMV)
ranging between 276 and 356 basis points portfolio, also known as minimum tracking
per annum. error portfolio, which focuses on minimising
the variance of the portfolio relative returns
Relative Return Perspective with respect to the cap-weighted index.
It is often the case that investors maintain
the cap-weighted index as a benchmark, From the results reported in Exhibit D, we
which has the merit of macro-consistency note that the focus on relative return leads
and is well-understood by all stakeholders. to lower tracking error levels compared to

Executive Summary
Exhibit D – Relative ERC and GMV Allocation across Smart Factor Indices (Developed Universe). The table compares performance
and risk of Scientific Beta Diversified Multi-Strategy indices converted in US dollars. We look at relative ERC and relative GMV
allocations invested in the 20 Diversified Multi-Strategy indices with stock selection based on mid cap, momentum, low volatility,
and value in the five sub-regions – US, UK, Dev. Europe Ex UK, Japan and Dev. Asia Pacific Ex Japan. The period goes from
31-December-2003 to 31-December-2013.
Developed CW Diversified Multi-Strategy
(2004-2013)
Multi Beta Relative ERC Multi Beta Relative GMV
(All Stocks) (All Stocks)
Allocation Allocation
Ann. Returns 7.80% 10.92% 9.96%
Ann. Volatility 17.09% 16.10% 16.64%
Sharpe Ratio 0.36 0.58 0.50
Max Drawdown 57.13% 54.14% 55.50%
Excess Returns - 3.12% 2.15%
Tracking Error (TE) - 2.56% 2.43%
95% TE - 4.70% 4.27%
Information Ratio - 1.22 0.88
Outperf Prob (3Y) - 100.00% 89.34%
Max Rel. Drawdown - 5.10% 4.95%
Ann. Returns Bull - 31.38% 31.02%
Ann. Returns Bear - -25.25% -26.93%
the portfolios that had an absolute return First-generation smart beta investment
focus. For example, the ex-post tracking approaches only provide a partial answer
error is around 2.50% for these two to the main shortcomings of cap-weighted
portfolios (2.43% for the relative minimum indices. Multi-Strategy factor indices, which
variance portfolio and 2.56% for the relative diversify away unrewarded risks and seek
equal risk contribution portfolio). Such exposure to rewarded risk factors, address
low tracking error levels, associated with the two main problems of cap-weighted
substantial outperformance (more than indices (their undesirable factor
300 basis points per annum for the R-ERC exposures and their heavy concentration)
portfolio), eventually leads to exceedingly simultaneously.
high information ratios. In particular, the
relative ERC has an information ratio of The results suggest that such Multi-
1.22, which is the highest level among all Strategy factor indices lead to considerable
portfolio strategies tested so far, with an improvements in risk-adjusted performance.
outperformance probability of 100% over For long-term US data, smart factor
any given three-year investment horizon indices for a range of different factor
during the same period. tilts roughly double the Sharpe ratio of
the broad cap-weighted index. Moreover,
outperformance of such indices persists
Conclusion: From Cap-Weighted at levels ranging from 2.92% to 4.46%,
Indices to Smart Factor Indices even when assuming unrealistically high
We find that well-rewarded factor-tilted transaction costs. The outperformance
indices constitute attractive building blocks of Multi-Strategy factor indices over
for the design of an improved equity portfolio. cap-weighted factor indices is observed for

Executive Summary
other developed stock markets as well. By

providing explicit tilts to consensual factors,
such indices improve upon many current
smart beta offerings where, more often
than not, factor tilts result as unintended
consequences of ad hoc methodologies.
Moreover, additional value can be added at

the allocation stage, where the investor can
control for the dollar and risk contributions
of various constituents or factors to
the absolute (volatility) or relative risk
(tracking error) of the portfolio. As a result,
extremely substantial levels of risk-adjusted
outperformance (information ratios) can be
achieved even on the absence of views on
factor returns. The portfolio strategies we
have presented in this publication can be
regarded as robust attempts at generating
an efficient strategic factor allocation
benchmark in the equity space. Obviously,
active portfolio managers may generate
additional value on top of this efficient
benchmark by incorporating forecasts of
factor returns at various points of the
business cycle in the context of tactical
factor allocation decisions.

Introduction

Introduction
Alternative equity indices or smart beta theory concerning both the existence
strategies are seen to provide tremendous of factor premia and the importance
growth potential. A recent survey (the of diversification, it is possible to go
EDHEC European ETF Survey 2013 by beyond existing smart beta approaches
Ducoulombier et al. (2014)) reveals which provide partial solutions by only
that while only 30% of investment addressing one of these issues.
professionals already use products
tracking smart beta indices, more than Asset pricing theory in fact suggests that
one third of respondents are considering there are two main challenges involved in
investing in such products in the near a sound approach to equity investing. The
future.1 This publication argues that first challenge is the efficient diversification
current smart beta investment approaches of unrewarded risks, where "diversification"
1 - Estimates of the total assets
managed in smart beta funds or only provide a partial answer to the main means "reduction" or "cancellation" (as in
mandates are notoriously hard
to get with sufficient reliability
shortcomings of capitalisation-weighted "diversify away"). Indeed, unrewarded risks
as data on dedicated mandates (cap-weighted) indices, and develops a are by definition not attractive for investors
which likely play a considerable
role for smart beta adoption new approach to equity investing referred who are inherently risk-averse and
by sophisticated institutional
clients is not available publicly,
to as smart factor investing. It then therefore only willing to take risks if there
and identification of the smart provides an assessment of the benefits of is an associated reward to be expected in
beta category is difficult.
However, some observers put addressing the two main problems of cap- exchange for such risk taking, as shown by
the total assets at USD 200bn
as of 2013. This number is cited
weighted indices (their undesirable factor Harry Markowitz (1952) in his seminal work
in <http://www.top1000funds. exposures and their heavy concentration) on portfolio diversification.2 The second
com/analysis/2013/05/29/
pushing-smart-beta-further/>. simultaneously by constructing factor challenge is the efficient diversification
The Economist, in July 2013,
estimates the assets managed
indices that explicitly seek exposures to of rewarded risks. Here the goal is not to
in smart beta funds to be USD rewarded risk factors, while diversifying diversify away rewarded risk exposures so as
142bn (“The rise of Smart Beta”,
The Economist, July 2013). CNBC away unrewarded risks. The results to eventually eliminate or at least minimise
(Smart beta: Beating the market
with an index fund”, CNBC,
suggest that such smart factor indices them, since this would imply giving up
November 7th, 2013, <http:// lead to considerable improvements in on the risk premia; the goal is instead to
www.cnbc.com/id/101149598>
) reported that about seven risk-adjusted performance. For long- efficiently allocate to rewarded risk factors
percent of ETF assets are linked
to smart beta indices and such
term US data, smart factor indices for so as to achieve the highest reward per unit
ETFs have seen a 43 percent a range of different factor tilts roughly of risk. In William Sharpe's (1964) CAPM,
growth over 2013 compared
to 16 percent growth of the double the Sharpe ratio of the broad cap- there is a single rewarded risk factor so the
overall ETF market. Investment
consultancy firm Towers Watson
weighted index. Outperformance of these second challenge is non-existent, and the
has stated that its institutional indices ranges from 2.92% to 4.46% per only focus should be on holding a well-
clients have allocated about USD
20bn to smart beta strategies annum on this long period and persists diversified proxy for the market portfolio.
at the end of 2012, an increase
of 33% over levels seen one
even when assuming unrealistically high In a multi-factor world, where the equity
year earlier (See <http://www. transaction costs. Moreover, by providing risk premium is multi-dimensional
next-finance.net/Smart-Beta-
strategies-continue-to>). explicit tilts to consensual factors, such (including not only market risk, but also size,
2 - Unrewarded risks can be risks
specific to a particular company
indices improve upon many current smart book-to-market (B/M) ratio, momentum,
or systematic risk exposure for beta offerings where, more often than volatility, etc.), an important component
which no reward is expected. It
can be shown that for a factor not, factor tilts result as unintended of an investor's equity investment process
model with the assumption of
zero alpha and replicable factors,
consequences of ad hoc methodologies. In is the determination of the appropriate
the specific risk of the true fact, this publication shows that by using (e.g. Sharpe ratio maximising) allocation to
(long-short) maximum sharpe
ratio (MSR) portfolio is zero. consensual results from asset pricing these rewarded risk exposures.

Introduction
This analysis of the dual challenges to (equal-dollar contribution or equal risk

rational equity investing is enlightening contribution indices).3 One problem with
with respect to a proper understanding these smart beta indices, however, is that
of the intrinsic shortcomings of cap- they fail to address the second problem,
weighted (CW) indices that are typically namely the explicit control of rewarded
used as default investment benchmarks risk exposures. Hence, by switching from
by asset owners and asset managers. On a CW index to an equally-weighted (EW)
the one hand, CW indices are ill-suited or global minimum variance (GMV) or
investment benchmarks because they tend GMV index for example, the investor is
to be concentrated portfolios that contain switching from one arbitrary bundle of
an excessive amount of unrewarded risk. factor exposures to another arbitrary
On the other hand, CW indices implicitly bundle of factor exposures, which may or
embed a bundle of factor exposures that may not be consistent with the investor's
are highly unlikely to be optimal for any needs and beliefs. On the other hand, index
investor, if only because they have not providers have also launched so-called
been explicitly controlled for. For example, factor indices, which focus on addressing
3 - In fact, scientific
CW indices show by construction a large the second shortcoming of CW indices,
and naive approaches to cap bias and a growth bias, while the namely their lack of controlled factor
diversification are not
competing approaches; academic literature has instead shown exposure.4 Such factor indices are meant
in particular, introducing
some form of shrinkage of
that small cap and value where the to be investable long-only or long-short
the scientifically diversified positively rewarded risk exposures. proxies for some of the rewarded factors
portfolio towards a naively
diversified (equal-weight or that have been analysed in academic
equal risk parity) portfolio has
been shown to improve the
This analysis also allows light to be shed on literature, such as the value factor, the
out-of-sample risk-adjusted the benefits and shortcomings of existing size factor, the momentum factor or the
performance.
4 - Fundamental indices and alternatives to CW indices. Broadly low volatility factor.5 One problem with
other indices that weight
stocks according to some
speaking, there have been two main these factor indices, however, is that they
fundamental measure of innovations in recent years. On the one fail to address the first problem, namely
economic size (Arnott et
al. (2005)) do not explicitly hand, a number of index providers have the excessive concentration problem
try and improve the
concentration problem nor do
launched so-called smart indices or smart leading to the presence of unrewarded
they explicitly aim to address beta indices, which focus on addressing risk. This is because the weighting scheme
the problem of inefficient
factor exposure. This is the the first shortcoming of CW indices, used in the design of factor indices is
reason why we do not include
these indices in the afore-
namely their excessive concentration either CW (leading to an excessive degree
mentioned list of recent that leads to an excessive presence of of concentration) or factor exposure
innovations. Such approaches
can be regarded as ad-hoc unrewarded risk. Such smart beta indices maximising (also leading to a lack of
attempts at constructing an
index based on a measure of
include various approaches that are based diversification).
company size that is different either on scientific diversification (e.g.
from market cap.
5 - The low volatility factor indices aiming to implement a minimum In a nutshell, CW indices suffer from two
is in fact an anomaly; since
it stipulates that the most
variance or MSR allocation to selected main problems, namely the presence of
risky stocks underperform, as stocks subject to a number of constraints excessive concentration and the presence
opposed to outperform the
least risky stocks (Ang et al. either on weights or on parameter of an underlying arbitrary set of factor
(2006), Baker, Bradley, and
Wurgler (2011), Bali, Cakici,
estimates that are meant to improve the exposures, and existing alternatives
and Whitelaw (2011)). robustness of the portfolio construction (namely smart indices or factor indices)
methodology) or naive diversification are reasonably successful attempts

Introduction
at addressing one of these problems, smart factor indices, once they have
but they do so while leaving the other been carefully constructed. This portfolio
dimension unattended. In the end, risk construction process distinguishes from
factors are like vectors; they are defined the unconditional approach, where the
through the direction they point to, but investor seeks the optimal exposure to risk
also their magnitude Having access to a factors that are rewarded in the long term
good proxy for a factor is hardly relevant by utilising a sophisticated risk allocation
if the investable proxy only gives access to framework. A sound approach to smart
a fraction of the fair reward per unit of risk factor index allocation requires the proper
to be expected from the factor exposure execution of three different steps:
because of the presence of unrewarded • Choice of factors that are rewarded in
risk due to excessive concentration. The the long term;
first contribution of this publication is • Designing factor-tilted portfolios that
to demonstrate that it is in fact feasible capture the fair risk-adjusted reward
to address two problems simultaneously associated with exposure to the factor;
through the use of smart factor indices, • Choice of a methodology for deriving
which are smart (meaning well-diversified) the optimal multi-factor exposures.
6 - This careful distinction
lies at the heart of Smart
indices with selected factor exposures that
Beta 2.0 approach (Amenc naturally combine the benefits of smart This paper discusses investment choices
and Goltz (2013)).
indices and the benefits of factor indices. at each of these steps different steps in
In brief, smart factor indices are meant detail, and provides an analysis of readily
to be the outcome of a process carefully implementable investment solutions
distinguishing the security selection stage that draw on allocations across publicly
from the portfolio construction process.6 available factor indices. The remainder of
The security selection stage is meant to the paper is organised as follows. In the
ensure that the right factor-tilt will be next section, we describe the selection of
associated to each index. For example, appropriate factors. Then we describe the
one would select a set of value stocks to design of well-diversified factor indices
construct a proxy for a value factor or a (referred to as smart factor indices) and
set of low volatility stocks to construct a compare them with the conventional
proxy for the low volatility factor. On the approach to factor indices. The third
other hand, the portfolio construction section illustrates suitable allocation
phase is meant to seek to diversify away decisions across smart factor indices. The
unrewarded risk as much as possible by last section provides conclusions.
using some naive or scientific approach
to diversification. As such the factor index
is made "smart", that is better diversified,
and the investor can hope to gain a larger
fraction of the reward (Sharpe ratio)
associated with these factors. The second
contribution of this publication is to
introduce a formal framework that can be
used by investors to allocate to the various

1. Selecting Desired Factor
Exposures

1. Selecting Desired Factor Exposures
In this section, we review the empirical asset their investment process is, however, that
pricing literature to identify the factors there is a clear economic intuition as to
that are most likely to bear a long-term why the exposure to this factor constitutes
reward. Both equilibrium models such as a systematic risk that requires a reward and
Merton’s (1973) inter-temporal capital asset is likely to continue producing a positive
pricing model and no arbitrage models such risk premium.7
as Ross’s (1976) Arbitrage Pricing Theory
allow for the existence of multiple priced In this study we focus on four well known
7 - It should be noted risk factors. The economic intuition for rewarded factors – Size, Value, Momentum
that not all investors are
necessarily interested in the existence of a reward for a given risk and Low Volatility. Fama and French have
harvesting every available
risk premium. In fact some
factor is that exposure to such a factor is identified that value (book-to-market) and
investors prefer to pay the undesirable for the average investor because size (market cap) explain average asset
premium to avoid exposure
to a certain risk factor e.g. it leads to losses in bad times when marginal returns, as a complement to the market beta
in the case of the illiquidity
premium. However, some
utility is high (Cochrane (2001)). While asset (Fama and French (1993)). Carhart (1997)
investors could decide to pricing theory provides a sound rationale empirically proved the existence of another
try to capture the reward
associated with a risk for the existence of multiple factors, theory priced factor – the momentum factor. The
premium, even if this reward
is related to taking on
provides little guidance on which factors low volatility factor, which qualifies as an
additional risk. For example, should be expected to be rewarded. anomaly rather than a risk factor, is the
while the reward may occur
in equilibrium due to a result of the famous ‘volatility puzzle,’ which
factor paying off poorly in
bad times (when marginal
The first order necessary condition for states that low-volatility stocks tend to
utility of consumption is a factor to be deemed important is the outperform high-volatility stocks in the
high), investors who have a
particularly long time horizon existence of empirical research which shows long run (Ang et al. (2006)).8
may be less sensible to such
risks. Long horizon investors
that the identified factor has a significant
may thus be particularly impact on the cross section of stock returns
inclined to seek exposure to
such rewarded factors, from in US and international equity markets. Empirical Illustration
which short term investors
may shy away due to the
Several systematically rewarded risk factors Exhibit 1.1 shows the returns of signal
associated risk. have been documented in literature; Harvey weighted quintile portfolios that represent
8 - It is beyond the scope
of this paper to provide an et al. (2013) document a total of 314 of portfolios with varying degree of exposure
exhaustive analysis of all
rewarded risk factors. For
such factors. The practice of identifying to each factor. Signal weighting is done
example, we do not include empirical factors is referred to as ‘factor by weighting the stocks in proportion to
a ‘quality’ factor which is
another recently documented fishing’. Therefore, a key requirement of their rank by relevant sorting characteristic
risk factor. This factor
could be based on a simple
investors to accept factors as relevant in following Asness et al. (2013). For example,
company attribute like gross
profitability (Novy-Marx Exhibit 1.1: Performance of quintile portfolios sorted by factors and weighted by rank – The exhibit shows mean annualised returns
(2013)) or more complex
of quintile portfolios. For each factor, the quintiles are constructed on related stock characteristics – market cap for size, B/M ratio
composite measures like a
combination of profitability,
for value, past 1 year minus 1 month returns for momentum, and past 2-year volatility for low volatility. Stocks in each quintile
growth, safety, and dividend are rank weighted. All portfolios are rebalanced quarterly and the analysis is based on daily total returns from 31/12/1972 to
payout (Asness et al. (2013)). 31/12/2012 (40 years).
We have made the choice Market Cap B/M Ratio Momentum Volatility
to not include this factor
as empirical evidence is High 9.74% 19.67% 15.19% 10.21%
much more recent and the
Quintile 2 10.92% 13.78% 13.43% 12.31%
definition less consensual
than for the factors included Quintile 3 13.23% 12.07% 14.01% 12.17%
in this paper. However, the
conceptual arguments we
Quintile 4 12.14% 9.99% 11.99% 12.97%
make in this paper carry Low 14.71% 8.05% 9.64% 12.44%
through to any rewarded risk
factors. High-Low -4.97% 11.62% 5.56% -2.23%

in any value quintile consisting of 100 indistinguishable from zero even for
stocks, the stock with highest B/M ratio relatively long sample periods. However,
will have 100 times more weight than the one may reasonably expect that stocks
stock with lowest B/M ratio. Same rank have higher rewards than bonds because
based weighting is followed in all quintiles investors are reluctant to hold too much
for all factors. The difference between Value equity due to its risks. For other equity risk
and Growth quintiles is 11.62% and that factors, such as value, momentum, low risk
between Mid Cap and Large Cap quintiles and size, similar explanations that interpret
is 4.97%. the factor premia as compensation for risk
have been put forth in the literature.
The debate about the existence of positive
premia for these factors is far from closed. It is worth noting that the existence of the
While positive premia for these factors are factor premia could also be explained by
documented in an extensive literature, some investors making systematic errors due to
9 - As an example, one may authors question the robustness or the behavioural biases such as over-reaction
consider the ongoing debate
on the low risk premium.
persistence of the reward associated with or under-reaction to news on a stock.
Early empirical evidence these factors.9 In fact, one can argue that However, whether such behavioural
suggests that the relation
between systematic risk empirical evidence will not be sufficient to biases can persistently affect asset prices
(stock beta) and return is
flatter than predicted by
draw a clear conclusion as to which set of in the presence of some smart investors
the CAPM (Black, Jensen, factors are acceptable for a given investors. who do not suffer from these biases is a
and Scholes (1972)). More
recently, Ang, Hodrick, Xing, Empirical results always carry a risk of point of contention. In fact, even if the
and Zhang (2006, 2009)
find that stocks with high
data-mining, i.e. strong and statistically average investor makes systematic errors
idiosyncratic volatility have significant factor premia may be a result due to behavioural “biases”, it could still
had low returns. Other
papers have documented of many researchers searching through the be possible that some rational investors
a flat or negative relation
between total volatility and
same dataset to find publishable results who are not subject to such biases exploit
expected return. However, a (see e.g. Harvey et al. (2013)). Therefore, the any small opportunity resulting from
number of recent papers have
questioned the robustness of choice of relevant factors should consider the irrationality of the average investor.
such results and show that
the findings are not robust
the economic rationale behind the reward The trading activity of such smart investors
to changes to portfolio for a given factor (see Kogan and Tian may then make the return opportunities
formation (Bali and Cakici
(2008)) or to adjusting for (2013)). The following subsection explains disappear. Therefore, behavioural
short-term return reversals
(Huang et al. (2010)). More
why investors should expect a reward for explanations of persistent factor premia
generally, McLean and Pontiff the four main risk factors discussed in this often introduce so-called “limits to
(2013) assess empirically
if risk premia for a range paper. arbitrage,” which prevent smart investors
of factor have remained
significant after the effect
from fully exploiting the opportunities
has been widely publicised. Moreover, simple, straightforward factor arising from the irrational behaviour of
10 - It has been argued that
value-tilted indices, which definitions may be useful to avoid the risk other investors. The most commonly-
draw on proprietary and ad
hoc definitions of composite
of data-mining of complex and unproven mentioned limits to arbitrage are short-
scores, such as commercially factor definitions.10 sales constraints and funding-liquidity
available fundamentally
weighted indices, are constraints.
highly sensitive to the
methodological choices made
in the index construction Economic Rationale The table below summarises the main
process (see e.g. Blitz and
Swinkels (2008) and Amenc Given the wide fluctuation in equity returns, economic explanations for common factor
(2011)).
the equity risk premium can be statistically premia.

Risk-Based Explanation Behavioural Explanation

Value Costly reversibility of assets in place leads to high Overreaction to bad news and extrapolation of the
sensitivity to economic shocks in bad times recent past leads to subsequent return reversal
Momentum High expected growth firms are more sensitive to Investor overconfidence and self-attribution bias
shocks to expected growth leads to returns continuation in the short term
Low Risk Liquidity-constrained investors hold leveraged positions Disagreement of investors about high-risk stocks
in low-risk assets which they may have to sell in bad leads to overpricing in the presence of short sales
times when liquidity constraints become binding constraints.
Size Low profitability leads to high distress risk and N.A.
downside risk. Low liquidity and high cost of investment
needs to be compensated by higher returns.
Value Zhang (2008) provide empirical evidence

Zhang (2005) provides a rationale for the that past winners have temporarily higher
value premium based on costly reversibility loadings on the growth rate of industrial
of investments. The stock price of value production. This higher sensitivity of
firms is mainly made up of tangible assets firms with higher expected growth rates
which are hard to reduce while growth is a natural result of firm valuation and is
firms’ stock price is mainly driven by growth similar to the higher interest rate sensitivity
options. Therefore value firms are much (duration) of bonds at high levels of interest
more affected by bad times. Choi (2013) rate (see Johnson (2002)). Low momentum
shows that value firms have increasing betas stocks on the other hand have low expected
in down markets (due to rising asset betas growth and are less sensitive to changes in
and rising leverage) while growth firms have expected growth.
more stable betas. The value premium can
thus be interpreted as compensation for the Behavioural explanations for momentum
risk of suffering from losses in bad times. profits focus on the short-term
In an influential paper, Lakonishok, over-reaction of investors. Daniel et al.
Shleifer and Vishny (1994) argue that (1998) show that two cognitive biases,
“value strategies exploit the suboptimal overconfidence and self-attribution, can
behaviour of the typical investor”. Their generate momentum effects. In particular,
explanation of the value premium focuses they show that investors will attribute the
on the psychological tendency of investors recent performance of the winning stocks
to extrapolate recent developments into they have selected to their stock picking
the future and to ignore evidence that skill and thus further bid up the prices for
is contrary to the extrapolation. Glamour these stocks, thus generating a momentum
firms with high recent growth thus tend to effect in the short term, with stock prices
obtain valuations that correspond to overly only reverting to their fundamental values
optimistic forecasts while distressed firms at longer horizons.
obtain stock market valuations which are
overly pessimistic. Low Risk
Frazzini and Pedersen (2014) provide a
Momentum model in which liquidity-constrained
Momentum stocks are exposed to investors are able to invest in leveraged
macroeconomic risk. In particular, Liu and positions of low-beta assets, but are forced

to liquidate these assets in bad times when

their liquidity constraints mean they can no
longer sustain the leverage. Thus low-risk
assets are exposed to a risk of liquidity
shocks and investors are compensated for
this risk when holding low-beta assets.
High-beta assets, on the other hand, expose
investors to less liquidity risk and rational
investors may thus require less expected
return from these stocks than what would
be in line with their higher market beta.
Behavioural explanations for the low-risk

premium argue that high-risk stocks tend
to have low returns because irrational
investors bid up prices beyond their rational
value. For example, Hong and Sraer (2012)
show that when there is disagreement
among investors on the future cash flow
of firms, short sales constraints will lead
to overpricing of stocks where investor
disagreement is high. As disagreement
increases with a stock’s beta, high-beta
stocks are more likely to be overpriced.
Size
Small stocks tend to have lower profitability
(in terms of return on equity) and greater
uncertainty of earnings (see Fama and
French (1995)), even when adjusting for
book-to-market effects. Therefore, such
stocks are more sensitive to economic
shocks, such as recessions. It has also been
argued that stocks of small firms are less
liquid and expected returns of smaller firms
have to be large in order to compensate for
their low liquidity (Amihud and Mendelson
(1986)). It has also been argued that smaller
stocks have higher downside risk (Chan,
Chen and Hsieh (1985)).


2. Designing Efficient and
Investable Proxies for Risk
Premia

2. Designing Efficient and Investable Proxies

for Risk Premia
2.1. Conventional Approaches to index, simply because the integration of

Factor Indices the attributes characterising the value
Factor indices fall into two major categories. exposure into the weighting does not take
The first involves selecting stocks that are the correlations between these stocks into
most exposed to the desired risk factor account.
and the application of a cap-weighting
scheme to this selection. While this Exhibit 2.1 shows that although
approach responds to one limitation of Fundamentals weighted strategy
cap-weighted indices, namely the choice outperforms the broad cap-weighted index
of exposure to a good factor, the problem and delivers a Sharpe ratio of 0.34, the Value
of poor diversification arising from high Multi-Strategy Index exhibits even better
concentration in a small number of stocks Sharpe ratio (of 0.54) as it takes into account
remains unanswered. The second method the correlation between the value stocks.
involves maximising the exposure to a This is confirmed by a lower Goetzmann-
factor, either by weighting the whole of Li-Rouwenhorst (GLR) measure (19.51%)
the universe on the basis of the exposure than the Fundamentals based strategy. The
to this factor (score/rank weighting), or by index also remains de-concentrated with
selecting and weighting by the exposure high Effective number of stocks (ENS).
score of the stock to that factor. Here again,
the maximisation of the factor exposure Moreover, the value tilt is an implicit result
does not guarantee that the indices are of the weighting methodology and it is
well-diversified. questionable whether an investor seeking
a value tilt would wish to hold any weight
To overcome these difficulties, index in growth stocks which will be present in a
providers that generally offer factor indices fundamentally-weighted index. Exhibit 2.2
on the basis of the first two approaches shows that the fundamentally-weighted
have recently sought to take advantage of strategy holds, albeit in relatively lower
the development of smart beta indices to amount, growth stocks, which is not in
offer investors a new framework for factor line with the objective of a value exposure
investing (Bender et al. (2013)). In fact, seeking investor. On the other hand, the
index providers have recognised that the Value Multi-Strategy index, by construction,
traditional factor indices they previously invests a larger percentage of the portfolio
offered are not good investable proxies of in value stocks.
the relevant risk factors due to their poor
diversification, and that the smart beta Similarly, seeking exposure to the size factor
indices aiming at improved diversification through equal weighting of a broad universe
have implicit risk exposures. As a result, is certainly less effective than selecting the
providers are proposing to select and smallest size stocks in the universe and then
combine indices according to their implicit diversifying them, including with an equal-
factor exposures. For example, one could weighted weighting scheme. Furthermore,
seek exposure to the value factor through a minimum volatility portfolio on a broad
a fundamental-weighted index. This, universe does not guarantee either the
however, will not produce a well-diversified highest exposure to low volatility stocks


for Risk Premia
Exhibit 2.1: Performance of the USA Fundamentals Based Strategy and the USA Value Multi-Strategy Index – The exhibit shows the
absolute performance, relative performance, and diversification indicators for the Fundamentals Based Index and the USA Value
Multi-Strategy Index. The Fundamentals Based Strategy is constructed by selecting the largest 500 stocks in the USA universe by the
average of their book value, trailing five-year dividend, trailing five-year cash flow and trailing five-year sales and then weighting
the selected stocks by the same measure of size. The GLR measure is defined as the ratio of the portfolio variance to the weighted
variance of its constituents. The effective number of stocks (ENS) is defined as the reciprocal of the Herfindahl Index, which in turn
is defined as the sum of squared weights of portfolio constituents. The complete stock universe consists of 500 largest stocks in
USA. The benchmark is the cap-weighted portfolio of the full universe. The yield on secondary market US Treasury Bills (3M) is the
risk-free rate. The return based analysis is based on daily total returns from 31/12/1972 to 31/12/2012 (40 years). All weight based
statistics are average values across 160 quarters (40 years) from 31/12/1972 to 31/12/2012.
Broad CW Fundamentals Based Strategy Value Diversified Multi
Strategy
Ann. Returns 9.74% 11.20% 14.44%
Ann. Volatility 17.47% 16.94% 16.55%
Sharpe Ratio 0.24 0.34 0.54
GLR 26.51% 25.82% 19.51%
Effective Number of Stocks 113 108 190
Ann. Alpha 0.0% -0.08% 2.26%
Market Beta 1.00 0.98 0.91
Small Minus Big (SMB) - 0.00 0.16
High Minus Low (HML) - 0.34 0.31
Momentum (MOM) Beta - -0.04 0.03
R-squared 100% 98.0% 95.0%
Exhibit 2.2: Distribution across Value Quintiles of the USA Fundamentals Based Index and the USA Value Multi-Strategy Index
– The exhibit shows the distribution of portfolio weights across value quintiles for the Fundamentals Based Index and the USA
Value Multi-Strategy Index. The analysis based statistics are average values across 160 quarters (40 years) from 31/12/1972 to
31/12/2012.
or the best diversification of this low factor choice, and which we refer to as
volatility portfolio. As the examples show, Smart Beta 1.0, are not satisfactory neither
the drawback of this approach is that it in terms of control of rewarded risk nor in
maximises neither factor exposure nor terms of diversification of the unrewarded
diversification of the indices. Such factor risks.
indices which are not based on explicit


for Risk Premia
2.2. Smart Beta 2.0 Approach to to unrewarded strategy specific risks.

Smart Factor Indices For example, Minimum Volatility portfolios
An important challenge in factor index are often exposed to significant sector
construction is to design well-diversified biases. Similarly, in spite of all the attention
factor indices that capture rewarded risks paid to the quality of model selection
while avoiding unrewarded risks. The Smart and the implementation methods for
Beta 2.0 approach allows investors to explore these models, the specific operational
different Smart Beta index construction risk remains present to certain extent.
methods in order to construct a benchmark For example, robustness of Maximum
that corresponds to their own choice of Sharpe Ratio scheme depends on a good
factor tilt and diversification method. It estimation of the covariance matrix and
allows investors to manage the exposure expected returns. The parameter estimation
to systematic risk factors and diminish the errors of optimised portfolio strategies
exposure to unrewarded strategy specific are not perfectly correlated and therefore
risks (see Amenc, Goltz and Lodh (2012), have a potential to be diversified away
and Amenc and Goltz (2013)). (Kan and Zhou (2007), Amenc et al. (2012)).
11 - The Diversified A Diversified Multi-Strategy approach,11
Multi-Strategy weighting Stock selection, the first step in Smart Beta which combines the 5 different weighting
is an equal weighted
combination of the following 2.0, allows investors to choose the right schemes in equal proportions, enables the
five weighting schemes:
Maximum Deconcentration (rewarded) risk factors to which they want non-rewarded risks associated with each
Diversified Risk Weighted, to be exposed. When it is performed upon a of the weighting schemes to be diversified
Maximum Decorrelation,
Efficient Minimum Volatility particular stock-based characteristic linked away.
and Efficient Maximum
Sharpe Ratio. to stocks’ specific exposure to a common
12 - The following selection factor, such as size, stock selection allows The Smart Beta 2.0 framework thus
rules are applied to select
stocks for each tilt: Mid this specific factor exposure to be shifted, allows the full benefits of smart beta to
Cap: bottom 50% free float
adjusted market cap stocks. regardless of the weights that will be be harnessed, where the stock selection
Value: top 50% stocks applied to portfolio individual components. defines exposure to the right (rewarded)
by book-to-market (B/M)
ratio defined as the ratio A well-diversified weighting scheme allows risk factors and the smart weighting scheme
of available book value
of shareholders’ equity to for the reduction of unrewarded or specific allows unrewarded risks to be reduced.
company market cap; High
Momentum: top 50% stocks
risks. Stock specific risk (such as management
by returns over the past 52 decisions, product success, etc.) is reduced We now turn to an empirical analysis on US
weeks excluding the last
4 weeks; Low Volatility: through the use of a suitable diversification long-term data of a set of Multi-Strategy
bottom 50% stocks by the strategy. However, due to imperfections in factor indices constructed for the four main
standard deviation of weekly
stock returns over the past the model there remain residual exposures factors introduced above.12 All indices are
104 weeks. This score based
selection is done twice a
year (June and December) Exhibit 2.3: Smart Beta 2.0 framework – The exhibit shows the two key index construction steps which allow factor indices to be
for Momentum and once designed within a Smart Beta 2.0 framework.
a year (June) for the other
three factors. Multi-Strategy
factor indices are constructed
by applying the Diversified
Multi-Strategy weighting
scheme and CW factor indices
are constructed by applying
float adjusted market cap
weighting on each stock
selection.


for Risk Premia
rebalanced quarterly and dividends are Similarly, the low volatility smart factor
reinvested in the index. The analytics on index has a low market beta (0.78), as low
US indices in subsequent sections use 40 market beta stocks are usually also low
years daily total returns. Stock-level data volatility stocks. Cap weighted indices, by
for portfolio construction and portfolio construction, load heavily on large cap
valuation is obtained from CRSP. We first stocks. Therefore any alternative to cap
assess the achievement of the desired weighting, especially diversification-based
factor tilts and then assess risk-adjusted weighting schemes which aim to be more
performance. deconcentrated, will induce the exposure
to the small cap factor. As a result, smart
factor indices have small size exposure as
2.3. Achieving the Objective of well. However, it is important to note that
Factor Exposure the magnitude of the small size beta is
First we examine how well these Multi- largest for the smart factor index that is
Strategy factor indices fulfill their first explicitly exposed to small size (i.e. the Mid
objective (i.e. to provide exposure to the Cap Diversified Multi-Strategy index (0.32)),
desired risk factor). Exhibit 2.4 shows the while the average small size beta for the
Carhart 4-factor regression statistics for other three smart factor indices is 0.12.
the Multi-Strategy factor indices and Similarly the Momentum Diversified Multi-
for the cap-weighted (poorly diversified) Strategy index has a momentum beta of
factor indices. The Mid Cap Multi-Strategy 0.17 compared to 0.01 on average for the
index has a size beta of 0.32, the High other smart factor indices; and the Value
Momentum Multi-Strategy index has a Diversified Multi-Strategy index has a value
momentum beta of 0.17, and the value beta of 0.31 compared to 0.13 on average
tilted index has a value beta of 0.31. for the other smart factor indices.
Exhibit 2.4: Exposure of USA Cap Weighted Factor Indices and USA Multi-Strategy Factor Indices to Equity Risk Factors – The exhibit
shows 4-factor regression analysis indicators for Cap Weighted Factor Indices and Multi-Strategy Factor Indices for four factor
tilts – Mid Cap, High Momentum, Low Volatility, and Value. The Market factor is constructed on the basis of the daily returns of a
cap-weighted index of all stocks net of the risk-free rate. The Small size factor is long the CW portfolio of market cap deciles 6 to
8 (NYSE, Nasdaq, AMEX) and short the CW portfolio of the largest 30% of stocks. The value factor is long the CW portfolio of the
highest 30% and short the CW portfolio of the lowest 30% of stocks ranked according to B/M ratio. The Momentum factor is long
the CW portfolio of the highest 30% and short the CW portfolio of the lowest 30% stocks ranked by past returns over 52 weeks
(excluding the most recent 4 weeks). The regression coefficients (betas and alphas) which are statistically significant at the 95%
level are highlighted in bold. The complete stock universe consists of the 500 largest stocks in the USA. The yield on secondary
market US Treasury Bills (3M) is the risk-free rate. All statistics are annualised. The analysis is based on daily total returns from
31/12/1972 to 31/12/2012 (40 years).
Mid Cap High Momentum Low Volatility Value
CW Diversified CW Diversified CW Diversified CW Diversified
Multi Multi Multi Multi
Strategy Strategy Strategy Strategy
Ann. Alpha 0.88% 2.59% 0.07% 1.73% 0.92% 2.52% -0.58% 2.26%
Market Beta 1.01 0.93 1.01 0.94 0.86 0.78 0.99 0.91
Small Size Beta 0.31 0.32 0.01 0.16 -0.14 0.02 0.00 0.16
Value Beta 0.15 0.16 0.00 0.09 0.05 0.14 0.42 0.31
Momentum Beta 0.02 0.00 0.20 0.17 -0.01 0.00 0.05 0.03
R-squared 94.3% 92.0% 98.6% 95.5% 94.7% 90.2% 98.3% 95.0%


for Risk Premia
The exercise shows that the simple stock factor indices consistently post superior
selection process (prior to any optimisation) Sharpe ratio than CW factor indices. The
results in portfolios which have the desired historical Daily 5% Value-at-Risk and
exposure ex-post. If one wants to have Maximum Drawdown of Multi-Strategy
a strong factor tilt, using stock selection factor indices and CW factor indices are
is the most transparent and simple way similar. The maximum Drawdown of Multi-
to achieve it. In other words, a careful Strategy Factor indices is not different from
distinction between security selection and that of CW factor indices. It shows that the
weighting scheme allows investors to turn increase in performance and the reduction
risk into “a choice rather than a fate”, to in portfolio risk do not come at the cost of
paraphrase an insightful comment by the extreme risk.
late Peter Bernstein (1996).
In addition to the non-parametric methods
to compute extreme risk (historical VaR), we
2.4. Avoiding Non Rewarded Risks also report VaR based on a semi-parametric
through Diversification approach based on Extreme Value Theory
Exhibit 2.5 presents the performance (EVT). This approach avoids the disadvantages
13 - A GARCH-EVT model
is used to estimate VaR
summary of the four Multi-Strategy of non-parametric methods such as the
in which EVT is applied factor indices and their Cap-Weighted limited number of data points for extreme
through the Peak-over-
Threshold (POT) method. counterparts. As broad cap-weighted indices events and the fact that the information
In particular, a GARCH(1,1)
model is estimated and
remain the widely accepted reference, we on skewness, kurtosis, and higher-order
then the residuals are use the broad cap-weighted index of the moments contained in the data points from
extracted from the estimated
model. True generalised 500 largest stocks as the benchmark. All the body of the distribution becomes less
Pareto Distribution (GPD)
function is then fitted to
factor-tilted portfolios, irrespective of the relevant for the lower quintiles in the left
the exceedances of the weighting scheme used, outperform the tail. At the same time, unlike full parametric
threshold using the maximum
likelihood method. The 1% broad cap-weighted index. It verifies that approach, it is not exposed to model risk.13
VaR is calculated through the
fitted GPD and a forecast of
the four chosen risk factors do earn, on The results show that Multi-Strategy indices
volatility generated through average, a positive risk premium in the are way superior to the CW factor indices
the GARCH(1,1) model. For
additional details about the long run. in terms of both EVT 1%VaR (extreme risk)
model and an empirical study,
see Loh and Stoyanov (2013).
and Return to EVT 1%VaR ratio (extreme
For each factor tilt, the Multi-Strategy risk-adjusted returns).
factor index earns higher returns than
the CW factor index for the same tilt. Both Multi-Strategy and Cap Weighted
Value and Mid Cap have been the most factor indices are exposed to systematic risk
rewarding factors in the last 40 years in the factors which are quite different from those
US market. The Mid Cap and Value smart of the broad CW index. Reward to these risk
factor indices earn a premium of 4.45% and factors varies over time and they experience
4.70% annual respectively. Low Volatility periods of underperformance relative to
and High Momentum are comparatively the broad market. Consequently all factor
less rewarded; however, their smart factor indices are exposed to relative risk i.e. risk of
indices have earned 2.90% and 3.56% excess underperforming the broad CW benchmark
returns respectively. If one looks at the in the short term which is shown by the
risk-adjusted performance, Multi-Strategy ‘maximum relative Drawdown’ numbers.


for Risk Premia
Each factor index, be it a CW factor index Going a step further and taking correlations
or a Multi-Strategy factor index, undergoes into account, we also report the ratio of
periods of relative Drawdown which can be portfolio variance to the weighted variance
triggered by events specific to stock markets. of its constituents (based on the GLR ratio16)
The occurrence of these Drawdowns does as a measure of diversification. A weighting
not mean that the factor premium is not scheme which exploits correlations to bring
robust across different economic conditions. down portfolio volatility will have a low
An analysis conditional on business market GLR ratio.
cycles, presented in the following sections,
shows that all Multi-Strategy factor indices The results in Exhibit 2.5 show that Multi-
earn positive premia in both the contraction Strategy factor indices are in fact better
and expansion phases of the USA business diversified as they have considerably higher
14 - Within the framework cycles. ENS and lower GLR ratio than their CW
of Smart Beta 2.0, one could
choose to put tracking error
counterparts. On the other hand, CW factor
constraints in smart factor Also, one must not forget that Multi- indices display high GLR ratios and - with
indices. Details on relative
risk control can be found at Strategy factor indices have a limited set the exception of Mid Cap factor17 - a low
Goltz and Gonzalez (2013).
However, it is not desired
of securities to diversify across as they are effective number of stocks, suggesting that
because tracking error constructed on 50% of the stock universe. while they may improve the exposure to
constraints increase the
correlation among smart This induces considerable tracking error rewarded risk factors compared to the
factor indices and thus
reduce the diversification
relative to the broad CW index. However, broad cap-weighted index, they actually
benefits upon their this tracking error is not a drawback if the aggravate the concentration problem. In
combination. A more practical
approach to manage tracking associated outperformance is high enough, contrast, multi-strategy factor-tilted indices
error risk would be to put
constraint on smart factor
which is the case with Multi-Strategy factor obtain the desired factor tilts without
allocation rather than putting indices, suggesting that they harvest the undue concentration, which provides an
it on each smart factor index.
15 - The effective number relevant factor premia in an efficient way. In explanation for their superior risk-adjusted
of stocks (ENS) is defined
as the reciprocal of the
fact, the results show that the information performance.
Herfindahl Index, which in ratios of Multi-Strategy factor indices range
turn is defined as the sum
of squared weights across from 0.47 for Low Volatility to 0.81 for
portfolio constituents.
where
Value.14 2.5. Robustness of Performance
N is the total number of Benefits
stocks in the portfolio and Wi
is the weight of i-th stock. This outperformance of smart factor
16 - Denoting RP the daily
return series of an index,
indices over traditional factor indices is not 2.5.1. Probability of Outperformance
Ri is the daily return series surprising. In fact, a lack of diversification Of particular interest is the information on
of the i-th stock, and Wi
the weight of i-th stock, has been identified as a major drawback of the probability of outperformance which
the GLR ratio is defined as
CW indices. When it comes to factor-tilted is defined as the historical probability of
17 - The Mid Cap selection indices, Multi-Strategy factor indices show outperforming the cap-weighted reference
picks up the bottom 250
market cap stocks in the considerable improvement both over the index over a given investment horizon.
broad USA universe. The
weight profile of these stocks
broad cap-weighted index and over the This measure is reported for investment
is flatter meaning that the CW factor indices. We report the effective horizons of 5 years by using a rolling
difference in market cap
of largest and small stock number of stocks (ENS)15 which can be used window analysis with one-week step size.
is not very high. Therefore
the Mid Cap CW index
as a measure of deconcentration. Given a We compute the frequency of obtaining
does not necessarily suffer fixed number of constituent stocks, an index positive excess returns if one invests in the
from the problem of high
concentration. with balanced weights will have a high ENS. strategy for a period of three or five years


for Risk Premia
Exhibit 2.5: Performance Comparison of USA Cap Weighted Factor Indices and USA Multi-Strategy Factor Indices – The exhibit
shows the absolute performance, relative performance, and diversification indicators for Cap Weighted Factor Indices and Multi-
Strategy Factor Indices for four factor tilts – Mid Cap, High Momentum, Low Volatility, and Value. Maximum relative Drawdown is
the maximum drawdown of the long-short index whose return is given by the fractional change in the ratio of the strategy index to
the benchmark index. The GLR measure is defined as the ratio of the portfolio variance to the weighted variance of its constituents.
The effective number of stocks (ENS) is defined as the reciprocal of the Herfindahl Index, which in turn is defined as the sum of
squared weights of portfolio constituents. The complete stock universe consists of the 500 largest stocks in USA. The benchmark is
the cap-weighted portfolio of the full universe. The yield on secondary market US Treasury Bills (3M) is the risk-free rate. The return
based analysis is based on daily total returns from 31/12/1972 to 31/12/2012 (40 years). All weight based statistics are average
values across 160 quarters (40 years) from 31/12/1972 to 31/12/2012.
Broad
Diversified
Diversified
Diversified
Diversified
Strategy
Strategy
Strategy
Strategy
Multi
Multi
Multi
Multi
CW CW CW CW CW
Ann. Returns 9.74% 12.54% 14.19% 10.85% 13.30% 10.09% 12.64% 11.78% 14.44%
Ann. Volatility 17.47% 17.83% 16.73% 17.60% 16.30% 15.89% 14.39% 18.02% 16.55%
Sharpe Ratio 0.24 0.39 0.52 0.30 0.48 0.29 0.50 0.35 0.54
Historical Daily 5% VaR 1.59% 1.60% 1.50% 1.64% 1.50% 1.42% 1.28% 1.59% 1.47%
Max Drawdown 54.53% 60.13% 58.11% 48.91% 49.00% 50.50% 50.13% 61.20% 58.41%
EVT 1% VaR 2.39% 2.28% 2.11% 2.43% 2.16% 2.19% 1.90% 2.39% 2.12%
Return to EVT 1% VaR 0.11 0.19 0.26 0.14 0.22 0.13 0.23 0.16 0.26
Monthly EVT 1% VaR 11.05% 11.46% 10.68% 11.19% 10.25% 9.99% 8.70% 12.22% 11.12%
Ann. Excess Returns - 2.80% 4.45% 1.10% 3.56% 0.35% 2.90% 2.04% 4.70%
95% Tracking Error - 9.39% 11.56% 6.84% 8.58% 9.20% 11.53% 8.72% 10.14%
Max Relative Drawdown - 35.94% 42.06% 14.44% 17.28% 33.82% 43.46% 20.31% 32.68%
GLR 26.51% 19.12% 16.72% 28.52% 21.08% 29.60% 22.20% 26.46% 19.51%
Effective Number of Stocks 113 181 191 65 199 64 201 69 190
and is computed using a rolling window outperformance of Multi-Strategy factor

analysis with 1 week step size. It is an indices (over CW factor indices) is well
intuitive measure to show how often the spread out over time and therefore is
strategy has managed to outperform the consistent. The outperformance cannot
cap-weighted reference index in the past be explained merely by some specific event
irrespective of the entry point. or some specific period, but by diversifying
away the bulk of the idiosyncratic risk at the
Exhibit 2.6 shows that all Multi-Strategy constituent level as well as the weighting-
factor indices have better chances of scheme specific risk.
outperforming the CW benchmark relative
to their respective CW factor indices. For 2.5.2. Transaction Costs, Liquidity
a 5-year horizon, the Low Volatility CW and Capacity
index beats the benchmark with a 54.27% Smart beta strategies, in their unaltered
probability while the Low Volatility Multi- form often incur large turnover and are
Strategy index beats it with 84.96% exposed to liquidity risk – the risk of
probability. It shows that the long-term investing substantial amount in illiquid


for Risk Premia
Exhibit 2.6: Probability of Outperformance of USA Cap Weighted Factor Indices and USA Multi-Strategy Factor Indices – The exhibit
shows the outperformance probability for Cap Weighted Factor Indices and Multi-Strategy Factor Indices for four factor tilts –
Mid Cap, High Momentum, Low Volatility, and Value. The probability of outperformance is the historical empirical probability of
outperforming the benchmark over a typical investment horizon of 3 or 5 years irrespective of the entry point in time. The analysis
is based on daily total returns from 31/12/1972 to 31/12/2012 (40 years).
18 - For Multi-Strategy factor Mid Cap High Momentum Low Volatility Value
indices, turnover is managed
Broad
Diversified
Diversified
Diversified
Diversified
through optimal control of
Strategy
Strategy
Strategy
Strategy
Multi
Multi
Multi
Multi
rebalancing of the indices - a CW CW CW CW CW
technique based on rebalancing
thresholds (see Leland (1999),
Martellini and Priaulet (2002)). At
Outperformance Probability - 70.08% 74.12% 78.21% 84.42% 50.10% 76.35% 69.82% 78.83%
each quarterly rebalancing, the new
optimised weights are implemented
(3Y)
only if the resulting overall weight Outperformance Probability - 75.33% 78.88% 86.76% 91.25% 54.27% 84.96% 72.05% 88.35%
change exceeds the threshold. The (5Y)
threshold is calibrated using the
past data, and it is fixed at the level
that would have resulted in not
more than a 30% annual one-way
turnover historically. The idea behind stocks. Both these limitations could Since Multi-Strategy factor indices aim to
this rule is to avoid rebalancing
when deviations of new optimal
result in high transaction costs and other replace active style investing, the impact of
weights from the current weights are operational hurdles such as large trading turnover on the performance, and not the
relatively small. This technique brings
down transaction costs by a large lags in the implementation of the strategy. absolute turnover, is the matter of concern.
extent without having a big impact
on the strategy's performance. In
The multi-strategy index performance A transaction cost of 20 bps per 100%
the case of Diversified Multi- reported here relates to portfolios which one-way turnover represents the worst case
Strategy weighting scheme, the
turnover control is applied to the have been subjected to turnover control18 observed historically and 100 bps represents
five constituent strategies before
combining them.
and capacity adjustments19 which ensure an 80% reduction in market liquidity. The
19 - The following capacity rules are easy implementation of these strategies. excess returns net of unrealistically high
applied to limit liquidity issues that
may arise upon investing and upon transaction costs, even for high momentum
rebalancing. 1. Holding Capacity
Rule - the weight of each stock is
Exhibit 2.7 compares the turnover incurred indices, remain quite significantly high.
capped to avoid large investment by rank weighted (and hence concentrated)
in the smallest stocks. 2. Trading
Capacity Rule - the change in factor indices constructed using 500, 250, Another wide-spread criticism of smart beta
weight of each stock is capped to
avoid large trading in small illiquid
100, and 50 stocks with the turnover of strategies is their limited capacity compared
stocks at the rebalancing. Formally, Multi-Strategy factor indices (on 250 stocks). to the CW benchmark which by definition
we adjust weights so that Wi,I ≤10.
Wi,CW ∀ i ∈ [1,N] and ∆Wi,I ≤ Wi,CW It is clear that as one concentrates in a lesser invests very small amounts in smaller and
∀ i ∈[1,N], where Wi,I is the weight
of i-th stock in the Multi-Strategy
number of stocks in an attempt to select less liquid stocks. Exhibit 2.8 shows that the
factor index and Wi,CW is the weight the most value (or least volatile stocks), weighted average market capitalisation of
if same stock in a cap-weighted
index that comprises of the same the associated turnover increases to very factor indices ranges from $2.73 billion for
stocks as the Multi-Strategy factor
index in question.
high levels. In a nutshell the concentrated the Mid Cap Multi-Strategy index to $13.67
20 - Momentum strategies factor indices not only suffer from poor billion for the Low Volatility Multi-Strategy
typically result in high turnover
(Chan et al. (1999)). Momentum diversification, but also exhibit quite high index compared to $44.9 billion for the
chasing strategies have short
horizons because persistence in
levels of turnover. Interestingly, on the other broad CW index. Another way to assess
price movement is a short-term hand, Multi-Strategy indices on 250 stocks the impact of holding less liquid securities
phenomenon and mean-reversion
is observed over longer horizons. result in lower turnover than that incurred is to have an estimation of trading days
Therefore to extract the momentum
premium, momentum score
by the rank weighting of 500 stocks. required to enter (or exit) the investment.
assignment is done semi-annually ‘Days to Trade’ is average number of days
which results in higher turnovers.
21 - Even if one assumes that only Indeed, with the exception of the momentum required to trade the total stock position
about 10% of average daily traded
volume can be traded, one would
tilt, all smart factor indices have one-way in the portfolio of $1 billion, assuming
still get a very reasonable ‘Days to annual turnover in the range of 22%-25%, that 100% of the ‘Average Daily Traded
Trade’ number for the smart factor
indices. which is well below the threshold of 30%.20 Volume (ADTV)’ can be traded every day.21


for Risk Premia
Exhibit 2.7: Turnover of USA Rank Weighted and Multi-Strategy Factor Indices - The exhibit shows one-way annualised turnover
of Rank Weighted and Multi-Strategy Factor Indices for four factor tilts – Mid Cap, High Momentum, Low Volatility, and Value. The
complete stock universe consists of 500 largest stocks in USA. All statistics are average values across 160 quarters (40 years) from
31/12/1972 to 31/12/2012.
22 - The measure is
computed for all stocks at
each rebalancing in the last
10 years (40 quarters) and the
We report the 95% percentile of this We show above that Multi-Strategy factor
95th percentile is reported. statistic across all stocks and across all indices in the USA universe, which are based
23 - Amenc, et al. (2012),
‘EDHEC-Risk North American rebalancing dates22 to get an estimate of on the 500 largest stocks, do not show
Index Survey 2011’ reveals
that 58.6% respondents who
extremely difficult trades. The results show any significant illiquidity that could hinder
use, who are going to use or that all Multi-Strategy factor indices have smooth implementation of the strategy.
who consider using smart
beta strategies consider these extreme trades which can be implemented However, it is interesting to assess whether
strategies as a complement
to cap-weighted indices;
within about 1/4th of a trading day. liquidity can be further improved. We thus
27.6% of them consider construct high liquidity versions of the
them as a replacement of
active managers; and only Recent surveys show that smart beta same portfolios by selecting the top 60%
23% consider them as a
replacement of cap-weighted
is perceived as an alternative to active of stocks by liquidity among the stocks
indices. management and/or as a complement to included in the factor-tilted portfolios.
the existing cap-weighted indices rather Exhibit 2.9 displays performance and
than as a replacement of the cap-weighted risk characteristics of the resulting High
benchmark.23 Therefore the assessment of Liquidity Multi-Strategy factor indices. As
the capacity effect (liquidity) and turnover expected, weighted average market cap and
of smart beta indices, including smart factor ‘Days to Trade’ numbers show significant
indices, must be done in a way similar to improvement. Furthermore, the indices
that of active managers. In that regard, both maintain most of the outperformance
the turnover and weighted average market of the original portfolios even though
cap of Multi-Strategy factor indices stay outperformance is reduced by a few basis
at quite manageable levels. Furthermore, points which can be explained by a potential
the systematic rules-based methodology illiquidity premium (Xiong et al. (2012)).
of Multi-Strategy factor indices guarantee
increased transparency that active managers
cannot provide.


for Risk Premia
Exhibit 2.8: Implementation Costs of USA Multi-Strategy Factor Indices - The exhibit shows weighted average market cap, turnover,
and outperformance net of transaction costs of Multi-Strategy Factor Indices for four factor tilts – Mid Cap, High Momentum,
Low Volatility, and Value. The complete stock universe consists of the 500 largest stocks in the USA. All statistics are average values
across 160 quarters (40 years) from 31/12/1972 to 31/12/2012.
USA Diversified Multi-Strategy
USA Broad CW
Ann. One-Way Turnover 2.65% 23.73% 63.46% 25.75% 23.83%
Relative Returns - 4.45% 3.56% 2.90% 4.70%
Relative Returns net of 20
- 4.41% 3.43% 2.85% 4.65%
bps transaction costs
- 4.22% 2.92% 2.65% 4.46%
Weighted Avg Mkt Cap ($m) 44 959 2 734 12 786 13 666 8 326
Days to Trade $1 bn
0.03 0.24 0.18 0.20 0.19
Investment (95% quintile)
Exhibit 2.9: Implementation Costs of USA High Liquidity Multi-Strategy Factor Indices – The exhibit shows weighted average market
cap, turnover, and outperformance net of transaction costs of High Liquidity Multi-Strategy Factor Indices for four factor tilts –
Mid Cap, High Momentum, Low Volatility, and Value. The complete stock universe consists of the 500 largest stocks in the USA. All
24 - The National Bureau statistics are average values across 160 quarters (40 years) from 31/12/1972 to 31/12/2012.
of Economic Research USA High Liquidity Diversified Multi-Strategy
(NBER) Business Cycle USA Broad CW
Dating Committee publishes Mid Cap High Momentum Low Volatility Value
business cycle reference dates
Ann. One-Way Turnover 2.65% 30.87% 66.37% 27.30% 27.46%
which can be obtained at
http://www.nber.org/cycles/ Relative Returns - 4.28% 2.80% 2.09% 4.10%
cyclesmain.html. Contractions
start at the peak of business
- 4.22% 2.67% 2.03% 4.04%
cycle and end at the trough. bps transaction costs
Expansions start at the Relative Returns net of 100
trough and end at peak. - 3.97% 2.14% 1.81% 3.82%
Weighted Avg Mkt Cap ($m) 44 959 3 293 18 513 19 691 11 683
Days to Trade $1 bn
0.03 0.24 0.12 0.15 0.13
Investment (95% quintile)
2.5.3. Sensitivity of Performance economic cycles, equity market conditions

to Economic Cycles and Market such as bullish or bearish markets may
Conditions have a considerable impact on how
As discussed before, the rewarded factors different portfolio strategies perform.
yield premiums in the long term in exchange For example, Amenc et al. (2012) show
of risks that can lead to considerable considerable variation in the performance
underperformance or relative drawdowns of some popular smart beta strategies in
in smaller periods. Therefore it is important different sub-periods, revealing the pitfalls
to analyse the time varying performance of of aggregate performance analysis based
Multi-Strategy factor indices in an attempt on long periods. Moreover, separating
to identify and characterise the nature bull and bear market periods to evaluate
of the risk premiums. One approach is to performance has been proposed by various
use the NBER definition of business cycle authors such as Levy (1974), Turner, Starz
to breakdown the analysis period into and Nelson (1989) and Faber (2007). Ferson
alternating sub-periods of ‘contraction’ and Qian (2004) note that an unconditional
and ‘expansion’ phases.24 In addition to evaluation made for example during bearish


for Risk Premia
markets will not be a meaningful estimation by a larger margin in expansion phases

of forward performance if the next period while the Low Volatility Multi-Strategy
were to be bullish. It is therefore important index was favoured by contraction phases.
to assess the robustness of performance This difference in sensitivities to market
with respect to such conditions. conditions suggests room for improvement
through allocating across multiple Multi-
Exhibit 2.10 shows annualised excess returns Strategy factor indices, an issue we turn to
of the four Multi-Strategy factor indices over in the next subsection.
broad CW index in different business cycles
and different equity market conditions. 2.5.4. Performance of Regional Indices
Exhibit 2.10 shows that the performance Having shown the robustness of Multi-
of Multi-Strategy factor indices depends Strategy factor indices using US long-term
on market conditions. For example, the track records, we test the consistency
Mid Cap Multi-Strategy index post much of their performance across different
higher outperformance in bull markets developed stock markets. Due to limited
(+5.37%) than in bear markets (+3.02%). availability of reliable data for non-US
The converse is true for the Low Volatility markets the time period of analysis is 31
Multi-Strategy index which underperforms December 2003 to 31 December 2013 (10
by 0.81% in bull markets and outperforms years). The Multi-Strategy factor indices
by 7.33% in bear markets. Similarly, the Mid and CW indices are governed by the same
Cap Multi-Strategy index has outperformed methodology as described for the USA data,
Exhibit 2.10: Conditional Performance of USA Multi-Strategy Factor Indices - The exhibit shows relative performance of Multi-
Strategy Factor Indices for four factor tilts – Mid Cap, High Momentum, Low Volatility, and Value in two distinct market conditions
– bull markets and bear markets and in contraction and expansion phases of US economy (NBER). Calendar quarters with positive
market index returns comprise bull markets and the rest constitute bear markets. The complete stock universe consists of the 500
largest stocks in the USA. The benchmark is the cap-weighted portfolio of the full universe. All statistics are annualised. The analysis
is based on daily total returns from 31/12/1972 to 31/12/2012 (40 years).
USA High Liquidity Diversified Multi-Strategy
Bull Markets
Ann. Relative Returns 5.37% 3.44% -0.81% 3.89%
Ann. Tracking Error 5.86% 4.10% 5.17% 5.08%
Information Ratio 0.92 0.84 -0.16 0.77
Bear Markets
Ann. Relative Returns 3.02% 3.43% 7.33% 5.33%
Information Ratio 0.36 0.55 0.93 0.75
NBER Contraction Phases
NBER Expansion Phases


for Risk Premia
and the only difference across regions is

the number of stocks. Stock universe sizes
for developed regions are: 500 (USA), 300
(Eurozone), 100 (UK), 500 (Japan), and 400
(Asia Pacific ex Japan).
Exhibit 2.11 shows that all Multi-Strategy

factor indices exhibit superior Sharpe ratios
relative to the broad CW index and their
respective CW factor indices. Information
ratios of the four Multi-Strategy factor
indices are usually higher than those of CW
factor indices and often reach impressive
levels such as 0.84 for USA Value and 0.69
for UK Momentum.
Since the analysis period is very short,

certain CW factor indices in certain regions
do not necessarily outperform the broad
CW index despite being tilted towards
the long-term rewarded factors – the
problem of sample time dependency. For
example, Japan High Momentum CW and
UK Value CW indices have excess returns
of -0.45% and -2.27% respectively in the
10-year period. The benefit from using a
well-diversified weighting scheme is more
visible in these cases as their corresponding
Multi-Strategy factor indices outperform
by 1.22% and 1.77%.


for Risk Premia
Exhibit 2.11: Performance of Multi-Strategy Factor Indices in developed markets – The exhibit shows the absolute and relative
performance of Multi-Strategy Factor Indices in 5 developed regions for four factor tilts – Mid Cap, High Momentum, Low Volatility,
and Value. Developed universes and their respective stock universe sizes are: USA (500), Eurozone (300), UK (100), Japan (500),
and Developed Asia Pacific ex-Japan (400). The benchmark is the Cap-Weighted index on the full universe for each region. The
risk-free rates used for these regions are Secondary Market US T-bill (3M), Euribor (3M), UK T-bill (3M), Japan Gensaki T-bill (1M)
and Secondary Market US T-bill (3M) respectively. All statistics are annualised. The analysis is based on daily total returns from
31/12/2003 to 31/12/2013 (10 years).
Broad
Diversified
Diversified
Diversified
Diversified
Strategy
Strategy
Strategy
Strategy
Multi
Multi
Multi
Multi
CW CW CW CW CW
USA
Ann. Returns 7.68% 10.41% 10.80% 8.64% 9.40% 7.94% 10.08% 7.56% 10.54%
Ann. Volatility 20.23% 22.33% 20.29% 20.38% 20.07% 17.82% 16.99% 22.46% 20.63%
Sharpe Ratio 0.30 0.40 0.45 0.35 0.39 0.36 0.50 0.27 0.43
Ann. Relative Returns - 2.73% 3.12% 0.96% 1.72% 0.26% 2.40% -0.12% 2.86%
Information Ratio - 0.54 0.74 0.22 0.34 0.06 0.47 -0.03 0.84
Eurozone
Ann. Returns 6.35% 7.99% 8.41% 9.09% 10.60% 8.39% 9.19% 6.09% 7.68%
Ann. Volatility 20.58% 18.63% 16.69% 19.74% 16.66% 18.35% 14.96% 22.81% 20.26%
Sharpe Ratio 0.21 0.32 0.38 0.35 0.51 0.34 0.47 0.18 0.28
Ann. Relative Returns - 1.64% 2.05% 2.73% 4.25% 2.04% 2.84% -0.26% 1.33%
Information Ratio - 0.26 0.29 0.57 0.60 0.45 0.39 -0.07 0.29
UK
Ann. Returns 8.32% 11.76% 11.10% 9.46% 12.71% 8.19% 11.86% 6.04% 10.09%
Ann. Volatility 19.18% 19.67% 17.95% 20.57% 17.99% 16.57% 15.33% 21.35% 19.43%
Sharpe Ratio 0.30 0.46 0.47 0.33 0.56 0.34 0.60 0.16 0.38
Ann. Relative Returns - 3.44% 2.78% 1.14% 4.39% -0.13% 3.54% -2.27% 1.77%
Information Ratio - 0.48 0.38 0.19 0.69 -0.02 0.47 -0.46 0.31
Japan
Ann. Returns 4.09% 4.97% 5.72% 3.64% 5.31% 5.34% 7.15% 5.65% 6.86%
Ann. Volatility 22.62% 21.21% 19.26% 22.39% 19.95% 19.50% 17.42% 22.60% 20.15%
Sharpe Ratio 0.17 0.23 0.29 0.15 0.26 0.26 0.40 0.24 0.33
Ann. Relative Returns - 0.89% 1.64% -0.45% 1.22% 1.26% 3.06% 1.56% 2.77%
Information Ratio - 0.13 0.21 -0.09 0.16 0.21 0.35 0.41 0.45
Developed Asia Pacific ex Japan
Ann. Returns 12.91% 15.31% 15.91% 16.12% 18.01% 13.86% 14.24% 14.92% 16.82%
Ann. Volatility 23.93% 23.08% 20.72% 25.45% 22.13% 22.85% 17.74% 24.36% 21.93%
Sharpe Ratio 0.47 0.60 0.69 0.57 0.74 0.54 0.71 0.55 0.70
Ann. Relative Returns - 2.40% 2.99% 3.21% 5.10% 0.94% 1.33% 2.01% 3.91%

3. Risk Allocation with
Smart Factor Indices

3. Risk Allocation with Smart Factor

Indices
In Section 2, we introduced the concept of momentum and low volatility risk premia.
a smart factor index, which can be regarded The risk and return characteristics of this
as an efficient investable proxy for a given equally-weighted portfolio of the four
risk premium. In a nutshell, a risk premium selected smart factor indices for the US
can be thought of as a combination of region are summarised in Exhibit 3.1.
a risk (exposure) and a premium (to be
earned from the risk exposure). Smart factor The results we obtain suggest a massive
indices have been precisely engineered to outperformance (398 basis points per
achieve a pronounced factor tilt emanating annum over the sample period), with an
from the stock selection procedure (right outperformance probability exceeding 80%
risk exposure), as well as high Sharpe ratio over all 3-year periods and with a volatility
emanating from the efficient diversification lower than the CW index volatility. Overall,
of unrewarded risks related to individual the Sharpe ratio is more than doubled, going
stocks (fair reward for the risk exposure). from 0.24 for the CW reference index to
The access to the fair reward for the 0.52 for the EW multi-smart beta portfolio.
given risk exposure is obtained through a In a nutshell, this outperformance comes
well-diversified smart-weighted portfolio from an efficient response to the two
(as opposed to a concentrated cap-weighted main shortcomings of CW indices, namely
portfolio) of the selected stocks so as to their inefficient factor exposures and their
ensure that the largest possible fraction insufficient reward for each given factor
of individual stocks' unrewarded risks is exposure.
eliminated.
While an equally-weighted scheme is
the simplest approach one can use, it is
3.1. The Many Dimensions of likely that the use of more sophisticated
Efficient Risk Allocation weighting schemes could provide additional
Section 3 is dedicated to showing that value, in particular when it comes to the
such smart factor indices can be used as management of the risks relative to the
attractive building blocks in the design CW benchmark. Clearly, a unique allocation
of an efficient allocation to the multiple framework that generates optimal portfolios
risk premia to be harvested in the equity for all possible investors does not exist; the
universe. allocation methodology should instead be
tailored to accommodate the preferences
3.1.1. From Equally-Weighted to and constraints of each investor.
Efficient Risk Allocation to Smart
Factor Indices In an attempt to identify, and analyse
In the context of generating a “smart” the benefits of, the possible approaches
(meaning efficient) allocation to the smart to efficient risk allocation across the
factor indices, a natural first, albeit naive, various smart factor indices, we identify
approach, consists in forming an equally- four main dimensions that can be taken
weighted portfolio of the selected smart into consideration when designing a
factor indices, in this case the indices that sophisticated allocation methodology
serve as proxies for the value, mid cap, (see Exhibit 3.2). These dimensions


Indices
define the allocation framework and respect to the CW index would instead
the choices implemented in each one of be used in the latter case (restricting the
these dimensions allow asset managers analysis to the second moment of the
to perform the risk allocation exercise so portfolio excess return distribution with
as to best accommodate investors' needs. respect to the CW index).
The first dimension relates to whether The second dimension concerns whether
risk is defined by the investor from an one would like to incorporate views
absolute perspective in the absence of a regarding factor returns in the optimisation
benchmark, or whether it is instead defined process. While additional benefits can be
in relative terms with respect to an existing obtained from the introduction of views
benchmark, which is more often than not on factor returns at various points of the
a CW index. In the former situation, one business cycle, we focus only on approaches
would use volatility as a relevant risk that are solely based on risk parameters,
measure (restricting the analysis to the which are notoriously easier to estimate
second moment of the portfolio return with a sufficient degree of robustness and
distribution), while tracking error with accuracy (Merton (1980)).
Exhibit 3.1 – Implementation of EW Allocation across Smart Factor Indices. The table compares performance and risk of Scientific
Beta Diversified Multi-Strategy indices. The Multi Beta Diversified Multi-Strategy index is the equal combination of the four Diversified
Multi-Strategy indices with stock selection based on mid cap, momentum, low volatility, and value respectively. All statistics are
annualised and daily total returns from 31-December-1972 to 31-December-2012 are used for the analysis. The CRSP S&P500 index is
used as the cap-weighted benchmark. The yield on Secondary US Treasury Bills (3M) is used as a proxy for the risk-free rate.
US Long Term Diversified Multi-Strategy
(Dec 1972 - Dec
CW Index Mid Cap Momentum Low Vol Value Multi Beta EW
2012)
Allocation
Ann. Returns 9.74% 14.19% 13.30% 12.64% 14.44% 13.72%
Ann. Volatility 17.47% 16.73% 16.30% 14.39% 16.55% 15.75%
Sharpe Ratio 0.24 0.52 0.48 0.50 0.54 0.52
Max Drawdown 54.53% 58.11% 49.00% 50.13% 58.41% 53.86%
Ann. Excess Returns 4.45% 3.56% 2.90% 4.70% 3.98%
Ann. Tracking Error 6.80% 4.88% 6.17% 5.82% 5.23%
95% Tracking Error 11.55% 8.58% 11.53% 10.14% 8.95%
Information Ratio 0.66 0.73 0.47 0.81 0.76
Outperf. prob (3Y) 74.12% 84.42% 76.35% 78.83% 80.38%
Exhibit 3.2 – The Various Dimensions of Allocation Methodologies across Assets or Risk Factors.


Indices
The third dimension is related to the argue that measuring and controlling factor
objective of the allocation procedure. exposures is obviously highly relevant in the
Indeed, there are several possible targets for context of risk factor allocation.
the design of a well-diversified portfolio of
factor exposures, depending upon whether 3.1.2. On the Relevance of Measuring
one would like to use naive approaches and Controlling Factor Exposures
(equal dollar allocation or equal risk One natural question that arises with
allocation) or scientific approaches based smart factor indices is to analyse what
on minimising portfolio risk (volatility in the their exposure to common equity factors
absolute return context or tracking error is, including precisely the market factor,
in the relative return context). If views on value factor, size factor, momentum factor
expected returns are introduced, then one and low volatility factor. Since the smart
may also envision maximising risk-adjusted factor indices have been engineered to
returns, either from an absolute perspective serve as proxy-replicating portfolios for
(maximising the portfolio Sharpe ratio) or these factors, one would ideally expect that
from a relative perspective (maximising the each smart factor index has (in a long-short
portfolio information ratio). version) a beta of 1 with respect to the
corresponding factor, and a beta of 0 with
The fourth and last dimension related to the respect to other factors. Besides, applying
presence of constraints is the optimisation. this analysis to a CW index would lead to
A simple constraint that is required by many a better measurement of the factor biases
investors is a long-only position in the smart (e.g. large cap and growth tilts) of this index.
factor index constituents. Additional weight In what follows, we perform this analysis by
constraints could be used so as to shrink a regressing the CW index, as well as all eight
scientifically diversified portfolio towards selected smart factor indices (based on the
a naively diversified portfolio (equal-dollar selection of value stocks, but also growth
or equal-risk allocation) in an attempt stocks; mid cap stocks, but also large cap
to alleviate the concern over impact on stocks; high momentum stocks, but also
parameter uncertainty on out-of-sample low momentum stocks; low volatility stocks,
portfolio performance (see Jagannathan but also high volatility stocks) with respect
and Ma (2003) or DeMiguel et al. (2009) for to a set of long/short factor returns. In
competing forms of shrinkage towards the addition to the market factor, we consider
equally-weighted portfolio, and Deguest, for consistency the value factor, the size
Martellini and Meucci (2013) for shrinkage factor, the momentum factor and the low
towards the risk parity portfolio). Other volatility factor.
types of constraints can/should be added to
enhance the investability of the portfolio, It is important to note at this stage that we
including constraints on maximum turnover, are not computing the factor exposure of
minimum liquidity or capacity levels, etc. the CW index and the smart factor indices
In addition to such weight constraints, with respect to a CW version of these factors,
one may also envision the introduction but instead with respect to an EW version of
of constraints on country, sector or factor these factors. Generally speaking, there are
exposures. In what follows, we precisely at least three different contexts in which


Indices
factor returns are being used in portfolio beta of 0 with respect to other factors.
analysis. On the one hand, factors can be So as to identify and measure the factor
used in a performance attribution context. biases of the CW index, the choice of an EW
In this case, the academic consensus is scheme, which can be regarded as the most
in favour of the use of cap-weighted natural and neutral approach, is commonly
factors, which is consistent with the fact used in academic research (see Plyakha,
that the benchmark implicitly or explicitly Uppal and Vilkov (2014) for a recent paper
used by investors is more often than not documenting the impact of the weighting
a cap-weighted index. On the other hand, scheme, EW versus CW, on standard cross-
factors can be used as investable benchmarks sectional as well as time-series asset pricing
themselves. In this case, the focus should tests).25 In other words, it is equally valid
be on maximising efficiency measured in to claim that the EW index has a value
terms of risk-adjusted performance, and the and small cap bias with respect to the CW
25 - Quoting from Plyakha,
Uppal and Vilkov (2014)): market cap weighting scheme is no longer reference, or to claim that the CW index has
"Equal-weighted mean
returns are used in a large taken for granted since cap weighting often a growth and large cap bias with respect
number of papers on leads to high concentration, and therefore to an EW reference. Since our focus is on
empirical asset pricing (see,
for example, the classical an excess of unrewarded risk. Finally, factors the shortcomings of the CW index, and
work of Fama and MacBeth
(1973), Black, Jensen, and can be used in the context of risk budgeting possible remedies to these shortcomings,
Scholes (1972), and Gibbons, strategies, where the focus is on measuring and because we use factors in an absolute
Ross, and Shanken (1989)),
almost all event-studies, and adjusting the relative contribution of risk budgeting exercise, it is only fitting that
and the research that
relates mean returns to firm
various factors to portfolio risk. In this we use a set of EW factors as regressors.26
characteristics (for reviews of context, the regression of the CW index
this literature, see Campbell,
Lo, and MacKinlay (1997) and onto a set of factors that would include The results of this analysis are shown in
Kothari and Warner (2006)).
For example, examining
the CW index factor itself would only give Exhibit 3.3, which leads to a number of
papers published in only two the trivial result that the CW factor has interesting insights.
premier outlets, The Journal
of Finance and The Journal a beta of 1 with respect to itself and a
of Financial Economics, over
a recent 5-year (2005 to
2009) interval, we are able Exhibit 3.3 – Factor Exposure of Smart Factor-Tilted and CW Indices. The graph shows the betas computed with multivariate linear
to identify 24 papers that
regressions using 2-year rolling windows of daily total returns from 31-December-1972 to 31-December-2012. The factors used to
report EW mean returns
and compare them across
perform linear regressions are Max-Deconcentration smart factor US indices used as proxy for equally-weighted factors.
portfolios."
26 - The market factor is
long-only, but all other four
factors are long-short. More
precisely, they are long a
maximum deconcentration
portfolio of the stocks
selected on the basis of
their positive exposure to a
positively rewarded risk factor
(namely Mid Cap, Value,
High Momentum and Low
Volatility stocks) and short a
maximum deconcentration
portfolio of the stocks
selected on the basis of
their negative exposure to a
positively rewarded risk factor
(namely Large Cap, Growth,
Low Momentum and High
Volatility stocks).


Indices


Indices
The first finding is that the CW index shows Another explanation for the presence of
a pronounced negative exposure to the mid residual factor exposures for the smart
cap factor (thus confirming its large cap factor indices is the presence of non-zero
bias) and a negative exposure to the value correlations between the factors, as can
factor (thus confirming its growth bias). be seen from the Exhibit 3.4, which shows
In addition, the CW index shows a positive the factor correlation estimates over the
exposure to the momentum factor and whole sample. These results confirm for
an exposure sometimes positive, but also example that low volatility firms tend
sometimes negative to the low volatility to be more represented within large cap
factor. All in all, these results suggest that firms versus small cap firms (hence the
holding a CW index leads an investor to negative correlation between the low
holding an inefficient and uncontrolled for volatility factor and the mid cap factor),
bundle of factor exposures. that growth firms and past winners tend
to be over represented within large cap
The second finding is that each factor index firms (hence the positive correlation
(e.g. the value smart factor index) has a between the value and mid cap factors or
beta fluctuating around a value of about the negative correlation between mid cap
0.5 with respect to the corresponding factor and momentum factors). In principle, one
(in this case the value factor) while the could try and orthogonalise the factors, but
smart factor index corresponding to the typical orthogonolisation procedures such
opposite factor tilt (in this case the growth as principal component analysis are typically
smart factor index) has a beta fluctuating highly intrusive, and even some of the least
around a value of about -0.5 with respect intrusive orthogonalisation techniques
to the same factor. In this sense, an investor (see for example Meucci, Santangelo
going long the smart value index factor and and Deguest (2013) for minimum linear
short the smart growth index factor would torsion techniques) typically involve a
have a beta of about 1 with respect to the substantial cost in terms of robustness. In
long-short value factor, with respect to Section 3.2.2, we show that the absence
the corresponding factor (in this case the of total purity of the factor replicating
value factor). portfolios (i.e. the fact that smart factor
indices have betas not exactly equal to
The third and last key finding is that while zero with respect to the factors that they
most smart factor indices have a low beta are tracking) or of the factors themselves
with respect to factors others than the one (i.e. the fact that the factors are correlated
for which they serve as a factor replicating even in their long-short version), is not a
portfolio (with a positive or negative strong problem per se as long as such biases
exposure), some of these betas are not can be measured and controlled for using
quite zero. For example, most smart factor suitably-designed factor risk budgeting
indices show a positive, albeit not extremely constraints.
high, exposure to the low volatility factor,
which is consistent with the fact that most
weighting schemes that are used lead to
a bias in favour of the least volatile stocks
within a given stock selection.

Indices
Exhibit 3.4 – Equally-Weighted Factor Correlations. This table shows that correlation of equally-weighted factors using daily total
returns from 31-December-1972 to 31-December-2012. Scientific Beta Max-Deconcentration smart factor US indices have been
used as proxy for equally-weighted factors.
Correlations US Long Term EW Factors

(Dec 1972 - Dec 2012) Momentum Low Vol Value
Mid Cap -0.28 -0.29 0.37
EW Factors Momentum 0.18 -0.17
Low Volatility 0.13
As a last comment, let us note that from in unrewarded risks compared to the
the betas obtained in Exhibit 3.3, we can cap-weighted index. Similarly, investing
compute for each smart factor index, or in the equally-weighted allocation of smart
portfolio of such factor indices, the specific factors also leads to a specific risk that is
risk se defined as the mean-squared error lower than the minimal specific risk values
of the multivariate linear regression with among the smart factor indices themselves.
respect to the long-only market factor and This decrease in specific, a priori unrewarded,
the four standard long-short factors (Size, risk translates into superior risk-adjusted
Value, Momentum and Low Vol): performance, as already noted.
27 - We actually argue
that factor risk budgeting
constraints are more useful 3.1.3. A Roadmap for Risk Allocation
in the absolute return
perspective, where the where with Smart Factor Indices
risk allocation exercise is
performed without any
To summarise the previous discussions,
reference, compared to the we shall sequentially consider in what
relative return perspective,
where the CW index already so as to obtain: follows the absolute return approach and
provides an anchor point (see
Section 3.3 below).
the relative return approach, both with
and without factor risk parity/budgeting
constraints.27 In each case, we consider naive
We notice in Exhibit 3.5 that investing in approaches to diversification (maximum
an equally-weighted portfolio of smart deconcentration in terms of dollar or risk
factor indices leads to a strong decrease contributions) and scientific approaches
Exhibit 3.5 – Specific Risk of Smart Factor-Tilted and CW Indices The graph shows the specific risk of four smart factor tilted
constituents, together with the CW, and the equally-weighted multi-beta multi-strategy using daily total returns from
31-December-1972 to 31-December-2012. R-ERC denotes the relative equal risk contribution.


Indices
(minimum risk from the absolute or the Europe Ex UK, Japan, Dev. Asia Pac. Ex Japan.
relative return perspective). Using the same four smart multi-strategy
indices as proxies for the value, size,
As previously mentioned, all these momentum and volatility rewarded tilts
methodologies will be implemented without in each sub-region as in the US experiment,
any active views (expected return forecasts) we obtain a total of (5x4) 20 constituents.
on constituents or factors; they generate Since the results we obtain in both cases are
portfolios that can be regarded as attractive qualitatively very similar, we focus on the
starting points, with very substantial international data in what follows, which
risk-adjusted outperformance benefits allows us to identify and discuss additional
with respect to cap-weighted indices, to relevant questions such as constraints on
which additional benefits could be added by regional exposures. The results for the US
asset managers possessing skills for actively case can be found in a dedicated Appendix.
timing factor exposures.
We have performed two types of empirical 3.2. Absolute Return Perspective

experiments using two different investment As argued previously, following an equally-
universes: the US on the one hand, and weighted allocation is equivalent to holding
world developed markets on the other an equal dollar allocation, which does not
hand. necessarily lead to an equal risk allocation.
Formally, the risk contribution of a stock
First, we use a US dataset over a long to the total risk of a portfolio is given by
period of time starting on 31-Dec-1972 the weight of the stock in the portfolio
and ending on 31-Dec 2012, and consider times the marginal contribution of the
as smart factor indices the classical stock to the total portfolio volatility.
rewarded tilts over long-term period, Qian (2006) shows that decomposing
namely mid cap, value, momentum and total portfolio volatility in terms of its
low volatility. For the weighting scheme constituents’ risk contributions is also
of the constituents, as argued in Section related to the expected contributions to
2, we adopt an agnostic point of view, the portfolio losses, particularly when
and consider the Multi-Strategy smart considering extreme losses. In what
factors which are equally-weighted follows, we consider two approaches to
combinations of the five weighting schemes the management of portfolio volatility-one
available on the Scientific Beta platform: approach based on minimising portfolio
Max-Deconcentration, Max-Decorrelation, volatility (global minimum variance
Efficient Minimum Volatility, Efficient approach) and another approach based
Maximum Sharpe Ratio, and Diversified on imposing an equal contribution of all
Risk-Weighting. constituents to the portfolio volatility
(heuristic equal risk contribution approach).
We also consider a second dataset over the
10-year period going from 31-Dec-2003
to 31-Dec 2013 using five sub-regions of
the global developed universe: US, UK, Dev.


Indices
3.2.1. Absolute Risk Management forms of implementation of agnostic

without Factor Risk Exposure diversification. When looking at the
Constraints empirical analysis performed in the Global
In our attempt to design an efficient Developed Universe shown in Exhibit 3.6,
allocation to smart factor indices, we first we find that the allocation between the
impose that all constituents to the portfolio equally-weighted and the ERC schemes
have the same contribution to portfolio risk. can exhibit strong discrepancies. For
To do so, we first consider the decomposition example, the largest average weight over
of portfolio variance: the period under study is given to the Japan
Low Volatility smart factor index (7.45%),
whereas the lowest weight is given to
the Dev. Europe Ex UK Value smart factor
where wi is the (positive) portfolio weight (3.78%). We also find that the equal risk
of stock i and σp the portfolio volatility. A contribution can lead to regional allocations
natural solution for imposing an equal risk that strongly deviate from the corresponding
allocation which is an agnostic approach allocation within a cap-weighted index,
in terms of risk rewarding, is to implement where the larger markets (e.g. the US)
an equal-risk contribution portfolio (ERC). strongly dominate smaller markets such
If we define the contribution to risk as: as Japan for example. So as not to introduce
overly strong biases with respect to the CW
index, and even though the focus is not on
with relative risk management in this section,
we introduce, in what follows, a set of
constraints dedicated to ensure that each
The ERC portfolio is defined as the allocation sub-region is not too strongly under- or
w that satisfies the following identity: over-represented with respect to its market
capitalisation in the CW global developed
for all i index.
If one makes the explicit assumption that In order to do so, we impose the following
all pairwise correlation coefficients across constraints (with δ = 2) in each region:
constituents are identical, then the ERC
weights can be obtained analytically and
are proportional to the inverse volatility of
the smart factor indices. In the general case,
i.e. without the assumption of identical
pairwise correlations across stocks, the risk where mcap represents the market
parity methodology does not yield a closed- capitalisations of the different sub-regions,
form solution. However, Maillard, Roncalli and wtilt is the weight in each smart
and Teïletche (2010) propose numerical factor index of the same corresponding
algorithms to compute risk parity portfolios. sub-regions.
Overall, ERC and EW are two competing


Indices
Exhibit 3.6 – EW and ERC Allocations to Smart Factor Diversified Multi-Strategy Indices and Risk Contributions (Developed Universe).
The graph compares the allocation and risk contributions of Diversified Multi-Strategy indices: the equal combination of the 20
Diversified Multi-Strategy indices converted in US Dollars with stock selection based on Mid Cap, Momentum, Low Volatility, and
Value in the five sub-regions: US, UK, Dev. Europe Ex UK, Japan and Dev. Asia Pacific Ex Japan, and the ERC combination of the same
20 constituents. The period goes from 31-December-2003 to 31-December-2013.


Indices


Indices
These constraints at the regional level In the end, this process leads to a dynamically
will be used in the context of the design managed portfolio of the 20 constituents
of the scientifically diversified portfolio, that should achieve a low volatility, but
in this case the GMV portfolio of the 20 that is highly concentrated.
constituents, which is the only efficient
portfolio that does not require expected Exhibit 3.8 shows that the portfolio variance
return inputs: is almost exclusively driven by the low
volatility factor, an observation that stresses
the need for the introduction of risk factor
budgeting constraints in order to better
We notice in Exhibit 3.7 that the GMV balance the factor contributions to the risk
allocation with geographical constraints of the portfolio.28
leads to a portfolio almost exclusively
invested in the lowest volatility smart index 3.2.2. Introducing Risk Budgeting
of each sub-region (on average, 52.47% in Constraints
the low volatility smart factor US index, Having an equal contribution of the
8.60% in the low volatility smart factor UK constituents to the overall portfolio risk is
28 - The contribution of
the low volatility factor is
index, 16.42% in the low volatility smart not identical to having an equal contribution
sometimes even greater than factor Dev. Europe Ex UK index, 12.68% in of the factors.29 In the following, we use the
100% while other factors
have a negative contribution the low volatility smart factor Japan index, factor exposure of the smart factor indices
to portfolio variance due to
the presence of non-zero
and 6.74% in the low volatility smart factor presented in Section 3.1.2 to analyse the
correlations between the Dev. Asia Pacific Ex Japan index). question. Again, we will compute exposure
smart factor indices and
also between the long-short
factors. For example, Exhibit 3.7 – GMV Allocations to Smart Factor Diversified Multi-Strategy Indices Under Geographical Constraints (Developed Universe).
increasing the exposure to The graph shows the allocation and risk contributions of the GMV allocation invested in the 20 Diversified Multi-Strategy indices
a factor that is negatively converted in US Dollars with stock selection based on Mid Cap, Momentum, Low Volatility, and Value in the five sub-regions: US, UK,
correlated with other factors
Dev. Europe Ex UK, Japan and Asia Pacific Ex Japan. The period goes from 31-December-2003 to 31-December-2013.
may contribute to decreasing
the portfolio variance.
29 - It is only if both the
factors and the factor indices
are perfectly "pure", that is
uncorrelated, that these two
approaches coincide.


Indices
Exhibit 3.8 – Factor Contribution to GMV Allocation (Developed Universe). The graph shows the contribution of the four long-short
factors to the risk of the portfolio resulting from a GMV allocation to the 20 Diversified Multi-Strategy indices with stock selection
based on Mid Cap, Momentum, Low Volatility, and Value in the five sub-regions: US, UK, Dev. Europe Ex UK, Japan and Dev. Asia
Pacific Ex Japan. Both risk parity and geographical constraints are imposed onto the resulting portfolios. The period goes from
with respect to the equally-weighted As a neutral target, we may seek to impose

version of the factors, since they are the an equal contribution of the factors to the
most neutral reference portfolios. variance coming from these K factors. This
extension of the ERC approach from the
In order to measure the contribution of constituents to the factors leads to the
the factors to the portfolio variance, we go following K linear constraints in the design
back to the decomposition of the portfolio of the portfolio:
return as the sum of K+1 factors leading to: for all ≤ i, j ≤K
In what follows, we introduce factor risk

where budgeting constraints to the portfolio
allocation process so as to avoid the
domination of any one particular factor
(such as the domination of the low volatility
Then, focusing only on the contribution of factor implicit in Exhibit 3.7). When the
the K long-short factors to the portfolio number of constituents N is greater than
variance leads to the following expression the number of factors constraints K, and
for the contribution of factor i to the long-short solutions are allowed, an infinite
variance coming from the K factors: number of portfolios satisfy a given set of
factor risk budgets (e.g. factor risk parity
exposure). In a long-only context, we may
have zero or multiple solutions. When no
solution exists, then one can start with the


Indices
long-short version and rescale the weights In Exhibit 3.9 we show the result of the
to avoid short positions. Max-Deconcentration and GMV allocations
under risk parity as well as geographical
On the other hand, when multiple solutions constraints.
exist, one can address the diversification
of specific risks, e.g. from a scientific First of all, we notice that factor risk parity
perspective, by minimising portfolio is satisfied, and that the portfolio is no
variance subject to factor risk parity longer simply invested in the low-volatility
constraints: constituents as in Exhibit 3.7.
such that Similarly to the allocation we obtained

for all ≤ i, j ≤K in Exhibit 3.7, we also notice that the
aggregated weights in the different
We may also maximise portfolio sub-regions appear to represent the
deconcentration, measured by the effective sub-region market capitalisations more
number of constituents, again subject to fairly than in the EW or ERC allocations
factor risk parity constraints:30 of Exhibit 3.6, due to the presence of
30 - Remember that in
the absence of constraints,
regional constraints. We also note that
maximising deconcentration the Max-Deconcentration approach
simply leads to giving a
weight of 1/N to each such that shows a more stable allocation over time
constituent in the universe.
for all ≤ i, j ≤K compared to the GMV which is still sensitive
Exhibit 3.9 – Max Deconcentration and GMV Allocations Under Risk Factor and Geographical Constraints (Developed Universe).
The graph shows the allocations and factor contributions of the Max-Deconcentration and GMV Diversified Multi-Strategy indices
invested in the 20 Diversified Multi-Strategy indices with stock selection based on Mid Cap, Momentum, Low Volatility, and Value in
the five sub-regions: US, UK, Dev. Europe Ex UK, Japan and Dev. Asia Pacific Ex Japan. Both risk parity and geographical constraints
are imposed onto the resulting portfolios. The period goes from 31-December-2003 to 31-December-2013.


Indices


Indices
to changes in input parameters. Also we with a tracking error hardly greater than 5%
see that the addition of factor risk parity (see Section 3.3 for explicit management of
constraints forces the allocations to spread relative risk leading to a further decrease
more evenly the country weight among in tracking error).
the different tilts compared to Exhibit 3.7.
Interestingly we note that the introduction
In Exhibit 3.10 we report the risk and return of factor risk parity constraints leads to
characteristics of the various portfolios, 100% outperformance probabilities over
and compare the results that have been a 3-year horizon.
obtained so far.
In Exhibit 3.11, we analyse the performances
We note that the GMV allocation process in bull versus bear market regimes (defined
leads to the lowest volatility. Also, we as positive versus negative returns for the
notice that the EW and ERC allocations CW index). We observe that the addition
have higher returns and higher volatilities of risk parity constraints to the GMV
than the GMV, as is often the case. Also we allocation tends to stabilise the returns
note that the introduction of factor risk across market conditions. For example, in
parity constraints has led to a substantial the absence of factor risk parity constraint,
improvement in information ratios, with the GMV allocation leads to a massive
an information ratio above 1 for the outperformance of 11.94% with respect to
Max-Deconcentration allocation under the CW index in bear markets, which is due
geographical and risk parity constraints. to the almost exclusive domination of the
This shows that the introduction of factor low volatility factor, with a defensive bias
risk parity constraints leads to a stabilisation that proves extremely useful in such market
of the portfolio that has resulted in strong conditions. On the other hand, the relative
outperformance (3.37%) over the CW index return in bull markets is negative at -3.90%
Exhibit 3.10 – Multi Beta Allocations across Smart Factor Indices (Developed Universe). The table shows the allocations of the EW,
ERC, GMV under geographical constraints, and both the Max-Deconcentration and GMV Diversified Multi-Strategy indices under
geographical and risk parity constraints, invested in the 20 Diversified Multi-Strategy indices with stock selection based on Mid
Cap, Momentum, Low Volatility, and Value in the five sub-regions: US, UK, Dev. Europe Ex UK, Japan and Dev. Asia Pacific Ex Japan.
The period goes from 31-December-2003 to 31-December-2013.
(2004-2013)
CW Multi Beta EW Multi Beta ERC Multi Beta Multi Beta Multi Beta
(All Stocks) Allocation Allocation GMV MDecon-Fact GMV-Fact
Allocation Allocation Allocation
Ann. Returns 7.80% 11.37% 11.07% 10.57% 11.17% 10.88%
Ann. Volatility 17.09% 15.32% 14.33% 12.84% 17.23% 17.21%
Sharpe Ratio 0.36 0.64 0.66 0.70 0.56 0.54
Max Drawdown 57.13% 54.40% 51.82% 45.07% 55.22% 55.32%
Excess Returns - 3.56% 3.27% 2.76% 3.37% 3.07%
Tracking Error (TE) - 6.75% 7.51% 6.36% 3.08% 3.34%
95% TE - 13.84% 14.89% 11.85% 5.19% 5.55%
Information Ratio - 0.53 0.44 0.43 1.09 0.92
Outperf. Prob (3Y) - 98.36% 89.34% 89.07% 100.00% 100.00%
Max Relative Drawdown - 6.35% 9.54% 13.10% 4.03% 5.47%


Indices
due to the performance drag associated 3.3. Relative Return Perspective

with exclusively holding defensive equity It is often the case that investors maintain
exposure in bull market conditions. In the cap-weighted index as a benchmark,
this context, one key advantage of the which has the merit of macro-consistency
introduction of factor risk parity constraints and is well-understood by all stakeholders.
is that it leads to a much more balanced In this context, a multi smart beta solution
return profile across market conditions can be regarded as a reliable cost-efficient
with positive outperformance in both substitute to expensive active managers,
bear and bull markets (at 2.66% and 3.18% and the most relevant perspective is not
respectively). an absolute return perspective, but a
relative perspective with respect to the
We have shown that simple allocations cap-weighted index.
that do not balance their expositions to
the factors may be too exposed to the In what follows, we focus on two
low-volatility factor, which may lead to approaches:
lower relative returns with respect to the • Naive diversification: relative equal risk
cap-weighted index, particularly in bull allocation (R-ERC) portfolio, which focuses
market regimes. Another way to design on equalising the contribution of the
allocation strategies that take into account smart factor-tilted indices to the portfolio
the presence of a CW index benchmark is tracking error.
to directly perform risk management using • Scientific diversification: relative global
relative returns. minimum variance portfolio (R-GMV), also
known as minimum tracking error portfolio,
which focuses on minimising the variance
of the portfolio relative returns with respect
to the cap-weighted index.
Exhibit 3.11 – Multi Beta Allocations across Smart Factor Indices in Bull/Bear Regimes (Developed Universe). The table shows the
allocations of the EW, ERC, GMV under geographical constraints, and both the Max-Deconcentration and GMV Diversified Multi-
Strategy indices under geographical and risk parity constraints, invested in the 20 Diversified Multi-Strategy indices with stock
selection based on Mid Cap, Momentum, Low Volatility, and Value in the five sub-regions: US, UK, Dev. Europe Ex UK, Japan and Dev.
Asia Pacific Ex Japan in bull and bear market regimes. The period goes from 31-December-2003 to 31-December-2013.
(2004-2013)
Multi Beta EW Multi Beta ERC Multi Beta Multi Beta Multi Beta
Allocation Allocation GMV Allocation MDecon-Fact GMV-Fact
Allocation Allocation
Ann. Ret. Bull 31.58% 29.55% 25.18% 32.89% 32.26%
Ann. Vol. Bull 11.71% 11.09% 9.42% 12.85% 12.85%
Ann. Rel. Ret. Bull 2.50% 0.48% -3.90% 3.81% 3.18%
Tracking Error Bull 5.03% 5.94% 5.06% 2.53% 2.74%
Ann. Ret. Bear -24.51% -22.42% -17.23% -26.67% -26.50%
Ann. Vol. Bear 21.33% 19.79% 18.40% 24.42% 24.38%
Ann. Rel. Ret. Bear 4.65% 6.74% 11.94% 2.50% 2.66%
Tracking Error Bear 9.64% 10.27% 8.62% 4.08% 4.43%


Indices
It should be noted that controlling for relative GMV allocation to the parameter
factor exposure biases from an absolute estimates, confirming a higher degree of
risk budgeting perspective is no longer a robustness with the ERC approach. Even
key required ingredient since the CW index though both allocation strategies rely
now already provides a proper anchor point, on risk parameter estimates, scientific
which is an implicit (as opposed to explicit) diversification tends to over-use input
reference set of factor exposures. In the same information compared to the more agnostic
manner, we find that regional constraints risk budgeting diversification, which makes
are no longer needed, since a portfolio a more parsimonious use of input estimates
seeking to equalise the contributions of the (see Roncalli (2013) for more details and
20 constituents to the portfolio tracking interpretations for the higher robustness
error, or seeking to minimise the tracking of ERC portfolios with respect to errors in
error, will not lead to a severe overweighting risk parameter estimates).
of smaller regions with respect to larger
regions, in contrast to what has been found Secondly, by construction, we observe
from an absolute risk perspective. that the relative ERC leads to identical
constituents’ contributions to the
In Exhibit 3.12, we show the allocations of tracking error. However, the relative GMV
the relative GMV and relative ERC portfolios. portfolio involves non-equal time-varying
First of all, we find again that the relative contributions from various constituents
ERC allocation is more stable over time, to the tracking error of the portfolio,
which is due to the higher sensitivity of the even though it is less exposed to the
Exhibit 3.12 – Relative GMV and ERC Allocations to Smart Factor Indices and Risk Contributions (Developed Universe). The graph
compares the allocation and risk contributions of Diversified Multi-Strategy indices: the relative GMV and ERC allocations invested
in the four Diversified Multi-Strategy indices with stock selection based on Mid Cap, Momentum, Low Volatility, and Value. The
period goes from 31-December-2003 to 31-December-2013.


Indices


Indices
low-volatility factor even without factor 1.22, which is the highest level among all
risk parity constraints than the GMV portfolio strategies tested so far, with an
portfolio shown in Exhibit 3.7 because of outperformance probability of 100% over
the relative risk focus. any given three-year investment horizon
during the same period. We also find that
Finally, we display in Exhibit 3.13 the risk the focus on relative risk leads to lower
and return characteristics of the relative tracking errors in bull and bear market
ERC and GMV allocation strategies. regimes compared to their absolute risk
counterparts.
We note that the focus on relative return
leads to lower tracking error levels compared
to the portfolios that had an absolute 3.4. Investability Considerations
return focus. For example, the ex-post The multi-factor allocations analysed do not
tracking error is around 2.50% for these only to provide efficient management of
portfolios (2.43% for the minimum tracking risk and return, but also allow for genuine
error portfolio and 2.56% for the relative investability. In fact, each of the smart factor
equal risk contribution portfolio). Such indices has a target of 30% annual one-way
low tracking error levels, associated with turnover which is set through optimal
substantial outperformance (more than control of rebalancing (with the notable
300 basis points per annum for the R-ERC exception of the momentum tilt, which
portfolio), eventually leads to exceedingly has a minimal target of 60% turnover).
high information ratios. In particular, the In addition, the stock selections used to
relative ERC has an information ratio of tilt the indices implement buffer rules in
Exhibit 3.13 – Relative ERC and GMV Allocation across Smart Factor Indices (Developed Universe). The table compares performance
and risk of the Scientific Beta Diversified Multi-Strategy indices converted in US dollars. We look at relative ERC and relative
GMV allocations invested in the 20 Diversified Multi-Strategy indices with stock selection based on Mid Cap, Momentum, Low
Volatility, and Value in the five sub-regions: US, UK, Dev. Europe Ex UK, Japan and Dev. Asia Pacific Ex Japan. The period goes from
Developed CW Diversified Multi-Strategy Developed Diversified Multi-Strategy
(2004-2013) (All Stocks) (2004-2013)
Multi Beta Multi Beta Multi Beta Multi Beta
Relative ERC Relative GMV Relative ERC Relative GMV
Allocation Allocation Allocation Allocation
Ann. Returns 7.80% 10.92% 9.96% Ann. Ret. Bull 31.38% 31.02%
Ann. Volatility 17.09% 16.10% 16.64% Ann. Vol. Bull 11.95% 12.25%
Sharpe Ratio 0.36 0.58 0.50 Ann. Rel. Ret. 2.30% 1.95%
Bull
Max Drawdown 57.13% 54.14% 55.50% Tracking Error 2.09% 2.01%
Bull
Excess Returns - 3.12% 2.15% Ann. Ret. Bear -25.25% -26.93%
Tracking Error (TE) - 2.56% 2.43% Ann. Vol. Bear 22.88% 23.75%
95% TE - 4.70% 4.27% Ann. Rel. Ret. 3.92% 2.23%
Bear
Information Ratio - 1.22 0.88 Tracking Error 3.41% 3.21%
Bear
Outperf. Prob (3Y) - 100.00% 89.34%
Max Rel. Drawdown - 5.10% 4.95%


Indices
order to reduce unproductive turnover due is very reasonable. In fact, managing

to small changes in stock characteristics. a mandate on each smart factor index
The component indices also apply weight separately would yield a turnover which
and trading constraints relative to market- is higher than the average turnover across
cap weights so as to ensure high capacity. the smart factor indices. This is due to the
Finally, these indices offer an optional High fact that rebalancing each component index
Liquidity feature which allows investors to to the allocation target would induce extra
reduce the application of the smart factor turnover. However, implementing the multi-
index methodology to the most liquid beta index in a single mandate exploits
stocks in the reference universe. Amenc et the benefits of natural crossing arising
al. (2014a, 2014b) present a more detailed across the different component indices,
explanation on how including carefully and it actually reduces the turnover below
designed rules at different stages of the the average level observed for component
index design process eases implementation indices. In the table, as opposed to managing
of investments in smart beta indices. the same allocations separately, we provide
the amount of turnover that is internally
In addition to these implementation rules, crossed in multi-beta indices for each
which are applied at the level of each smart multi-beta allocation. We see that about
factor index, the multi-beta allocations 6% turnover is internally crossed by the
provide a reduction in turnover (and hence EW allocation and that the ERC allocation
of transaction costs) compared to a separate which tends to generate more turnover also
investment in each of the smart factor exploits natural crossing effects more than
indices. This reduction in turnover arises the EW allocation (around 7.8% is crossed
from different sources. First, when the internally). These cancelling trades result
renewal of the underlying stock selections in an average one-way annual turnover
takes place, it can happen that a stock being that can be even lower than for the EW
dropped from the universe of one smart allocation, as is the case in the Developed
factor index is being simultaneously added universe.
to the universe of another smart factor
index. Second, for constituents that are In addition to turnover, the exhibit also
common to several smart factor indices, the shows the average capacity of the indices
trades to rebalance the weight of a stock in terms of the weighted average market-cap
in the different indices to the respective of stocks in the portfolio. This index capacity
target weight may partly offset each other. measure indicates very decent levels with
an average market-cap of around US$ 10bn
Exhibit 3.14 displays statistics relative to for the multi-beta index, while the highly
the investability of the multi-beta EW and liquid version further increases capacity to
relative ERC allocations along with the levels exceeding US$ 15bn in the case of
average of the mid cap, momentum, low the US Long Term Track Records. In the case
volatility and value smart factor indices. For of the Developed universe, the weighted
comparison, we also show the same analytics average market caps are higher since the
for their Highly Liquid counterparts. We see period under scrutiny is more recent (last 10
that the turnover of multi-beta indices years) – around US$ 16.3bn for the standard


Indices
Exhibit 3.14 – Implementation of Multi-Beta Allocations across Standard or Highly Liquid Factor-Tilted Indices The analysis is
based on daily total return data from 31-December 1972 to 31 December 2012 (40 years) in Panel A and from 31-December-2003
to 31-December-2013 (10 years) in Panel B. The S&P 500 index and SciBeta Developed CW index are used respectively as the cap-
weighted reference for US Long Term Track Records and SciBeta Developed Investable Indices. Days to Trade is the number of days
necessary to trade the total stock positions, assuming a US$1bn AUM and that 100% of the Average Daily Dollar Traded Volume
can be traded every day. The weighted average market capitalisation of index is in $million and averaged over the 40-year period.
All statistics are average values across 160 quarters (40 years). The net returns are the relative returns over the cap-weighted
benchmark net of transaction costs. Two levels of transaction costs are used - 20 bps per 100% one-way turnover and 100 bps per
100% one-way turnover. The first case corresponds to the worst case observed historically for the large and mid-cap universe of
our indices while the second case assumes 80% reduction in market liquidity and a corresponding increase in transaction costs.
The risk-free rate is the return of the 3-month US Treasury Bill. (*)Due to data availability, the period is restricted to last 10 years of
the sample for Scientific Beta US indices.
Panel A
Diversified Multi-Strategy
All Stocks High Liquidity Stocks
US Long-Term Track Records
Average Average
(Dec 1972 – Dec 2012) Relative
of 4 Smart EW Multi ERC Multi of 4 Smart EW Multi
ERC Multi
Factor Beta Beta Factor Beta
Beta
Indices Indices
One-Way Turnover 34.19% 28.94% 31.53% 38.00% 33.21% 36.78%
Internally Crossed Turnover - 5.65% 7.55% - 5.53% 7.63%
Days to Trade for $1 bn Initial Investment
0.20 0.12 0.12 0.16 0.07 0.07
(Quantile 95%)(*)
Weighted Avg. Market Cap ($m) 9 378 9 378 10 280 13 295 13 295 15 283
Information Ratio 0.67 0.76 0.77 0.59 0.79 0.80
Relative Returns 3.90% 3.98% 3.79% 3.32% 3.43% 3.04%
Relative Returns net of 20 bps transaction
3.84% 3.92% 3.73% 3.24% 3.37% 2.96%
costs (historical worst case)
Relative Returns net of 100 bps
transaction costs (extreme liquidity stress 3.56% 3.69% 3.47% 2.94% 3.10% 2.67%
scenario)
Panel B
All Stocks High Liquidity Stocks
SciBeta Investable Developed Indices
Average Average
(Dec 2003 – Dec 2013) Relative
of 4 Smart EW Multi ERC Multi of 4 Smart EW Multi
ERC Multi
Factor Beta Beta Factor Beta
Beta
Indices Indices
One-Way Turnover 45.69% 39.63% 38.59% 45.85% 39.83% 38.36%
Internally Crossed Turnover - 6.22% 7.76% - 6.27% 8.12%
Days to Trade for $1 bn Initial Investment
0.48 0.27 0.27 0.20 0.09 0.09
(Quantile 95%)
Weighted Avg. Market Cap ($m) 16 047 16 047 16 493 22 391 22 391 23 737
Information Ratio 0.78 0.98 1.05 0.68 1.12 1.22
Relative Returns 2.57% 2.61% 2.55% 2.35% 2.40% 2.38%
2.48% 2.53% 2.47% 2.25% 2.32% 2.31%
costs (historical worst case)
2.11% 2.21% 2.16% 1.89% 2.00% 2.00%
costs (extreme liquidity stress scenario)
Source: www.ScientificBeta.com


Indices
indices and US$ 23bn for the highly liquid

ones. In both regions, we provide an estimate
of the time that would be necessary to set
up an initial investment (i.e. full weights) of
US$1bn in the indices, assuming that the
average daily dollar traded volume can be
traded (100% participation rate) and that the
number of days required grows linearly with
the fund size.31 Overall, this does highlight
the ease of implementation of the multi-
beta indices and the effectiveness of the
high liquidity option. Indeed, the Days to
Trade required for the initial investment on
US indices are very manageable (about 0.12
days for the standard multi-beta indices,
and 0.07 days with the highly liquid feature).
Even in the Developed universe, the highly
31 - The Days to Trade (DTT)
measure is computed for all
liquid multi-beta indices would require
stocks at each rebalancing about 0.09 days of trading. In addition, one
over the last 10 years
(40 quarters). Based on should keep in mind that the number of
the estimated DTT for all
constituents of a given index,
days needed to rebalance the indices (i.e.
we can derive an estimate of trade the weight change rather than the full
the required days to trade for
the index itself, by using, for weight on each stock) would be much lower.
example, extreme quantiles
of the DTT distribution over
Even though the excess return is reduced by
time and constituents, such a few basis points, which can be explained
as the 95th percentile that
we report. by a potential illiquidity premium, it should
be noted that the highly liquid multi-beta
indices do maintain the level of relative
risk-adjusted performance (information
ratio) of the standard multi-beta indices
in the US case and provide even stronger
information ratios in the Developed universe.
Finally, even when assuming unrealistically
high levels of transaction costs, all the smart
factor indices deliver strong outperformance
(from 2% to 3.69%) net of costs in both
regions. Compared to the average stand-
alone investment in a smart factor index,
the multi-beta indices almost always result
in higher average returns net of costs due
to the turnover reduction through natural
crossing effects across their component smart
factor indices.

Conclusion

Conclusion
We find that well-rewarded factor-tilted be achieved even on the absence of views

indices constitute attractive building on factor returns. The portfolio strategies
blocks for the design of an improved we have presented in this paper can be
equity portfolio. First-generation smart regarded as robust attempts at generating
beta investment approaches only provide an efficient strategic factor allocation
a partial answer to the main shortcomings process in the equity space. Active portfolio
of cap-weighted indices. Multi-Strategy managers may generate additional value by
factor indices, which diversify away incorporating forecasts of factor returns
unrewarded risks and seek exposure to at various points of the business cycle in
rewarded risk factors, address the two the context of tactical factor allocation
main problems of cap-weighted indices decisions.
(their undesirable factor exposures and their
heavy concentration) simultaneously. The While our approach has been focused on
adoption of a simple and consistent portfolio an asset-only perspective, it is feasible, and
construction methodology allows for the potentially desirable, to design dedicated
avoidance of data mining risks. The results efficient multi-factor equity portfolios
suggest that such Multi-Strategy factor that are optimised from an asset-liability
indices lead to considerable improvements management perspective. For example, a
in risk-adjusted performance. For long-term mature pension fund facing a stream of
US data, smart factor indices for a range bond-like pension obligations may find it
of different factor tilts roughly double the useful to select stocks that show an above
Sharpe ratio of the broad cap-weighted average degree of "liability-friendliness",
index. Moreover, outperformance of such which can be measured, for example, in
indices persists at levels ranging from terms of their correlation or tracking error
2.92% to 4.46%, even when assuming with respect to a liability proxy and/or their
unrealistically high transaction costs. The ability to pay a high and predictable stream
outperformance of Multi-Strategy factor of dividends. Once these stocks are selected,
indices over CW factor indices is observed a dedicated efficient factor index can be
for other developed stock markets as well. By designed, and used as an additional building
providing explicit tilts to consensual factors, block in allocation exercises dedicated to
such indices improve upon many current achieving the optimal trade-off between
smart beta offerings where, more often liability-hedging benefits and performance
than not, factor tilts result as unintended benefits (see Coqueret, Martellini and
consequences of ad hoc methodologies. Milhau (2014) for a detailed analysis of
these questions).
Moreover, additional value can be added at
the allocation stage, where the investor can
control for the dollar and risk contributions
of various constituents or factors to
the absolute (volatility) or relative risk
(tracking error) of the portfolio. As a result,
extremely substantial levels of risk-adjusted
outperformance (information ratios) can

Appendix

Appendix
The resulting ERC allocation to the four selected US smart factor indices is shown in Exhibit
A.1. We note that the ERC portfolio is not very different from the EW portfolio, which
shows that the covariance structure of the constituents is quite homogeneous. Indeed,
there is a relatively high level of correlation and a relatively low dispersion of volatility
levels. However, we notice a slightly higher weight for the low-volatility constituent
which is due to its lower volatility compared to the other constituents.
Also, as expected, the ERC portfolio generates an identical risk contribution of the four
constituents as illustrated in the bottom graphs of Exhibit A.1.
Exhibit A.1 – EW and ERC Allocations to Smart Factor Diversified Multi-Strategy Indices and Risk Contributions (US Universe). The graph
compares the allocation and risk contributions of Diversified Multi-Strategy indices: the equal combination of the four Diversified
Multi-Strategy indices with stock selection based on mid cap, momentum, low volatility, and value, and the ERC combination of
the same four constituents. The period goes from 31-December-1972 to 31-December-2012.

Appendix
If we also look at the factor risk contributions to the EW and ERC portfolios (see Exhibit
A.2) in the US universe, we notice that the factor contributions are far from being equal,
and can even become negative. We note a strong domination of the low volatility factor
in terms of risk factor contributions, which can be explained in particular by positive
betas of all smart factor indices with respect to this factor.

Appendix
Exhibit A.2 – EW and ERC Factor Risk Contributions (US Universe). The graph compares the factor risk contributions of Diversified
Multi-Strategy indices: the Multi Beta Diversified Multi-Strategy with equal combination of the four Diversified Multi-Strategy
indices with stock selection based on mid cap, momentum, low volatility, and value, and the ERC combination of the same four
constituents. The period goes from 31-December-1972 to 31-December-2012.
The first three panels of Exhibit A.3, show the allocation of a Max-Deconcentration
portfolio under risk parity constraints in a long-short framework. First, it is worth noting
that since the number of constraints is equal to the number of constituents, then the
allocation is fully driven by the constraints (no need for a diversification objective in
this experiment, but it will be proven useful in the following ones). We notice that the
short positions are quite small, and concern mostly the low-volatility constituent. Also,
we confirm that the equality of the factor risk contributions is satisfied, i.e. that the risk
parity constraints are not violated over the 40-year sample period.
If we now move to the last three panels of Exhibit A.3, where we impose no-short sale
constraints on the allocation strategy, then we notice that there are periods where the
Appendix
allocation to the low-volatility constituent is equal to 0 (1998-1999 and 2002-2003),

which is not surprising since it coincides with periods of time where the long-short
allocation was shorting the same constituent. However, if long-only constraints are
imposed, we can no longer satisfy the factor risk parity constraints, and we notice that
the contribution of the low-volatility factor is dominating the other contributions on the
same periods of time where the allocation to the low-volatility factor is reduced to 0.
Given that short positions are often undesirable, one practical way to reduce the exposure
of the low-volatility factor in a long-only context is to introduce a high-volatility
constituent. In this case we would be dealing with a portfolio with N>K constituents,
which would leave room for a diversification objective.
Exhibit A.3 – Max-Deconcentration Allocation under Risk Parity Constraints (US Universe). The graphs show the allocations,
the risk contributions and the factor contributions of a Max-Deconcentration strategy under risk parity constraints in the four
Diversified Multi-Strategy indices with stock selection based on mid cap, momentum, low volatility, and value. The first three panels
show the case of a long-short allocation, whereas the last three panels focus on a long-only allocation. The period goes from

Appendix

Appendix
In Exhibit A.4, we consider a Max-Deconcentration allocation to five smart factors

indices under risk parity constraints (in a long-only environment). We notice that the
introduction of the high-volatility constituent allows us to recover the parity of the
factors’ contributions. Also we see that the need for the high-volatility constituent is
increased over the periods where the factor risk parity was being violated in the case of
four constituents only.
Exhibit A.4 – Max-Deconcentration Allocation under Risk Parity Constraints (Lo with the HiVol Constituent). The graphs show the
allocations, the risk contributions and the factor contributions of a Max-Deconcentration strategy under risk parity constraints in
the five Diversified Multi-Strategy indices with stock selection based on Mid Cap, Momentum, Low Volatility, High Volatility, and
value. The period goes from 31-December-1972 to 31-December-2012.

Appendix
In Exhibit A.5, we present the risk and return characteristics of the various allocation
strategies to the four US smart factor indices. It is interesting to notice that the introduction
of risk parity constraints has led to increases in excess returns, while maintaining a
level of tracking error that is comparable to those of the EW and ERC allocations. This
feature leads to a higher information ratio, which is desirable for investors that have the
cap-weighted index as reference.
If we focus on the bull and bear market regimes (see Exhibit A.6), we observe that the
Max-Deconcentration scheme with factor risk parity constraints leads to much higher
annual returns in bull market regimes (more than 1.3% above the EW allocation), and
displays slightly lower annual returns in bear market regimes (less than 1% below the EW
allocation), which is consistent with the intuition that the introduction of factor risk parity

Appendix
constraints has led to a less defensive portfolio strategy. Besides, the information ratio
of the strategy is much higher than those of the competing allocations (0.81 compared
to 0.76 for the EW or 0.74 for the ERC).
Exhibit A.5 – Multi Beta Allocations across Smart Factor Indices (US Universe). The table compares performance and risk of Scientific
Beta Diversified Multi-Strategy indices. We look at equally-weighted, ERC and Max-Deconcentration under risk parity constraints
allocations in the four Diversified Multi-Strategy indices with stock selection based on mid cap, momentum, low volatility, and
value respectively. All statistics are annualised and daily total returns from 31-December-1972 to 31-December-2012 are used for
the analysis. CRSP S&P500 index is used as the cap-weighted benchmark. Yield on Secondary US Treasury Bills (3M) is used as a
proxy for the risk-free rate.
US Long Term
(Dec 1972 - Dec 2012) Avg of 4 Smart Factor Multi Beta Equal Multi Beta ERC Multi Beta MDec-Fact
Indices Weight Allocation Allocation
Ann. Returns 13.65% 13.72% 13.63% 14.01%
Ann. Volatility 15.99% 15.75% 15.67% 16.41%
Sharpe Ratio 0.51 0.52 0.52 0.52
Max Drawdown 53.91% 53.86% 53.62% 56.56%
Excess Returns 3.90% 3.98% 3.89% 4.27%
Tracking Error (TE) 5.92% 5.23% 5.25% 5.27%
95% TE 10.45% 8.95% 9.10% 8.69%
Outperf. Prob (3Y) 78.61% 80.38% 80.43% 78.83%
Max Rel. Drawdown 33.87% 33.65% 43.46% 33.87%
Exhibit A.6 – Multi Beta Allocations across Smart Factor Indices in Bull/Bear Regimes (US Universe). The table compares performance
and risk of Scientific Beta Diversified Multi-Strategy indices in bull and bear market regimes. We look at equally-weighted, ERC and
Max-Deconcentration under risk parity constraints allocations in the four Diversified Multi-Strategy indices with stock selection
based on mid cap, momentum, low volatility, and value respectively. All statistics are annualised and daily total returns from
31-December-1972 to 31-December-2012 are used for the analysis. CRSP S&P500 index is used as the cap-weighted benchmark.
Yield on Secondary US Treasury Bills (3M) is used as a proxy for the risk-free rate.
US Long Term
(Dec 1972 - Dec 2012) Multi Beta Multi Beta ERC Multi Beta MDec-Fact
Equal Weight Allocation Allocation
Ann. Ret. Bull 34.83% 34.57% 36.14%
Ann. Vol. Bull 12.94% 12.85% 13.42%
Ann. Rel. Ret. Bull 3.03% 2.76% 4.34%
Tracking Error Bull 4.45% 4.46% 4.54%
Ann. Ret. Bear -20.17% -20.04% -21.13%
Ann. Vol. Bear 20.23% 20.14% 21.14%
Ann. Rel. Ret. Bear 4.83% 4.96% 3.87%
Tracking Error Bear 6.57% 6.58% 6.53%
We now turn to the relative return perspective and present in Exhibit A.7 the weights
allocated to the four US smart factor indices by a relative GMV allocation strategy and
by a relative ERC allocation strategy.

Appendix
Exhibit A.7 – Relative GMV and ERC Allocations to Smart Factor Indices and Risk Contributions (US Universe). The graph compares
the allocation and risk contributions of Diversified Multi-Strategy indices: the relative GMV and ERC allocations invested in the
four Diversified Multi-Strategy indices with stock selection based on mid cap, momentum, low volatility, and value, and the ERC
combination of the same four constituents. The Relative GMV strategy has been derived with the following additional weight
constraints:1/δN<w<δ/N, where N=4 constituents and δ=2. The period goes from 31-December-1972 to 31-December-2012.

Appendix
When analysing the risk and performance indicators in Exhibit A.8, we observe lower
levels of tracking error for relative approaches than absolute approaches shown in Exhibit
A.5. Moreover, we see that the relative GMV, which is supposed to minimise the tracking
error, indeed achieves its goal when compared to other strategies.
Furthermore, we see that the levels of information ratios of relative strategies are higher
than their absolute counterparts, but remain lower than allocations implemented with
risk parity constraints. Finally, when we compare the levels of tracking error obtained by
the relative approaches in the bull and bear regimes with those obtained in Exhibit A.8,
we see that relative approaches leads to lower tracking error levels in each regime. We
also find that they lead to positive outperformance in both bull and bear markets, with
stronger outperformance in the latter market conditions, as expected.

Appendix
Exhibit A.8 – Relative ERC and GMV Allocation to the CW Index across Smart Factor Indices (US Universe). The table compares
performance and risk of Scientific Beta Diversified Multi-Strategy indices. We look at relative ERC and relative GMV allocations in
the four Diversified Multi-Strategy indices with stock selection based on mid cap, momentum, low volatility, and value respectively.
All statistics are annualised and daily total returns from 31-December-1972 to 31-December-2012 are used for the analysis. CRSP
S&P500 index is used as the cap-weighted benchmark. Yield on Secondary US Treasury Bills (3M) is used as a proxy for the risk-free rate.
US Long Term Diversified Multi-Strategy US Long Term Diversified Multi-Strategy
(Dec 1972 - Dec 2012) (Dec 1972 - Dec 2012)
Multi Beta Multi Beta Multi Beta Multi Beta
Relative ERC Relative GMV Relative ERC Relative GMV
Allocation Allocation Allocation Allocation
Ann. Returns 13.53% 13.45% Ann. Ret. Bull 34.72% 34.39%
Ann. Volatility 15.69% 15.60% Ann. Vol. Bull 12.89% 12.97%
Sharpe Ratio 0.51 0.51 Ann. Rel. Ret. Bull 2.92% 2.59%
Max Drawdown 53.30% 52.64% Tracking Error Bull 4.20% 3.78%
Excess Returns 3.79% 3.71% Ann. Ret. Bear -20.44% -21.02%
Tracking Error (TE) 4.91% 4.79% Ann. Vol. Bear 20.14% 20.15%
95% TE 8.11% 7.99% Ann. Rel. Ret. Bear 4.56% 3.97%
Information Ratio 0.77 0.77 Tracking Error Bear 6.12% 5.49%
Outperf. Prob (3Y) 80.90% 81.31%
Max Rel. Drawdown 28.74% 27.00%


References

References
• Amenc, N. 2011. A Comment on Better Beta Explained: Demystifying Alternative Equity

Index Strategies. The Journal of Index Investing 2(): 12-18.
• Amenc, N., and F. Goltz. 2013. Smart Beta 2.0. The Journal of Index Investing. 4(3): 15-23.
• Amenc, N., F. Goltz, and A. Lodh. 2012. Choose Your Betas: Benchmarking Alternative
Equity Index Strategies. Journal of Portfolio Management 39(1): 88-111.
• Amenc, N., F. Goltz, A. Lodh, and L. Martellini. 2012. Diversifying the Diversifiers and
Tracking the Tracking Error: Outperforming Cap-Weighted Indices with Limited Risk of
Underperformance. The Journal of Portfolio Management 38(3): 72-88.
• Amenc, N., R. F. Goltz, A. Lodh, and L. Martellini. 2014a. Towards Smart Equity Factor
Indices: Harvesting Risk Premia without taking Unrewarded Risks. EDHEC-Risk Institute
Working Paper.
• Amenc, N., R. F. Goltz, N. Gonzalez, and A. Thabault. 2014b. Investability of Scientific
Beta Indices. ERI Scientific Beta White Paper.
• Amenc, N., F. Goltz and L. Tang. 2011. EDHEC-Risk European Index Survey 2011. EDHEC-
Risk Institute Publication (October).
• Amihud, Y., and H. Mendelson. 1986. Asset Pricing and the Bid-Ask Spread. Journal of
Financial Economics 17(2): 223-249.
• Ang, A., R. Hodrick, Y. Xing, and X. Zhang. 2006. The Cross-Section of Volatility and
Expected Returns. The Journal of Finance 61(1): 259-299.
• Ang, A., Hodrick, Y. Xing, and X. Zhang. 2009. High Idiosyncratic Volatility and Low
Returns: International and Further U.S. Evidence. Journal of Financial Economics 91: 1-23.
• Arnott, R.D., J.C. Hsu, and P. Moore. 2005. Fundamental Indexation. Financial Analysts
Journal 61(2): 83-99.
• Asness, C. 1992. Changing Equity Risk Premia and Changing Betas over the Business
Cycle and January. University of Chicago.
• Asness, C., A. Frazzini, and L. Pedersen. 2013. Quality Minus Junk. Working Paper.
• Asness, C., T. Moskowitz, and L.H. Pedersen. 2013. Value and Momentum Everywhere.
Journal of Finance 68(3): 929-985.
• Baker, M., B. Bradley, and J. Wurgler. 2011. Benchmarks as Limits to Arbitrage: Understanding
the Low-Volatility Anomaly. Financial Analysts Journal 67(1): 40-54.
• Bali, T.G., and N. Cakici. 2008. Idiosyncratic Volatility and the Cross Section of Expected
Returns. Journal of Financial and Quantitative Analysis 43(1): 29-58.
• Bali, T.G., N. Cakici, N, and R.F. Whitelaw. 2011. Maxing Out: Stocks as Lotteries and the
Cross-Section of Expected Returns. Journal of Financial Economics 99: 427-446.
• Bender, J., R. Briand, D. Melas, R. Subramanian and M. Subramanian. 2013. Deploying
Multi-Factor Index Allocations in Institutional Portfolios. MSCI White Paper.
• Bernstein, P. L. 1996. Against the Gods: The Remarkable Story of Risk. Wiley.
• Black, F., M. Jensen, and M. Scholes. 1972. The Capital Asset Pricing Model: Some
Empirical Tests. Studies in the Theory of Capital Markets. Praeger.
References
• Blitz, D., and L. Swinkels. 2008. Fundamental Indexation: An Active Value Strategy in
Disguise. Journal of Asset Management 9(4): 264-269.
• Carhart, M.M. 1997. On Persistence in Mutual Fund Performance. Journal of Finance
52 (1): 57-82.
• Chan, K.C., N. Chen, and D.A. Hsieh. 1985. An Exploratory Investigation of the Firm Size
Effect. Journal of Financial Economics 14(14): 451-471.
• Chan, L.K.C., J. Karceski, and J. Lakonishok. 1999. On Portfolio Optimization: Forecasting
Covariances and Choosing the Risk Model. Review of Financial Studies 12(5): 937-74.
• Choi, J. 2013. What Drives the Value Premium?: The Role of Asset Risk and Leverage.
Review of Financial Studies 26(1): 2845-2875.
• Cochrane, J. 2001. Asset Pricing. Princeton, NJ: Princeton University Press.
• Cohen, R., C. Polk, and T. Vuolteenaho. 2003. The Value Spread. Journal of Finance
58(2): 609-641.
• Coqueret, G., L. Martellini and V. Milhau. 2014. Beyond Liability-Driven Investing -
Designing Equity Benchmarks with Improved Liability-Hedging Properties. EDHEC-Risk
Institute Publication.
• Daniel, K., D. Hirshleifer, and A. Subrahmanyam. 1998. Investor Psychology and Security
Market Under- and Over-reactions. Journal of Finance 53(5): 1839-1886.
• Deguest, R., L. Martellini, and A. Meucci. 2013. Risk Parity and Beyond – From Asset
Allocation to Risk Allocation Decisions. http://ssrn.com/abstract=2355778. Working Paper.
• DeMiguel, V., L. Garlappi, J. Nogales and R. Uppal. 2009. A Generalized Approach
to Portfolio Optimization: Improving Performance by Constraining Portfolio Norms.
Management Science 555: 798-812.
• Ducoulombier, F., F. Goltz, A. Lodh, and V. Le Sourd. 2014. The EDHEC European ETF
Survey 2013. EDHEC-Risk Institute Publication.
• Faber, M.T. 2007. A Quantitative Approach to Tactical Asset Allocation. Journal of Wealth
Management 9(4): 69-79.
• Fama, E., and K. French. 1993. Common Risk Factors in the Returns on Stocks and Bonds.
Journal of Financial Economics 33(1): 3-56.
• Fama, E., and K. French. 1998. Size and Book-to-Market Factors in Earnings and Returns.
Journal of Finance 50: 131-155.
• Ferson, W.E., and M. Qian. 2004. Conditional Performance Evaluation, Revisited. Working
paper. Boston College.
• Frazzini, A., and L.H. Pedersen. 2014. Betting Against Beta. Journal of Financial Economics
111(1): 1-25.
• Goetzmann, W., L. Li, and G. Rouwenhorst. 2005. Long-Term Global Market Correlation.
Journal of Business 78(1): 1-38.
• Goltz, F., and N. Gonzalez. 2013. Risk Managing Smart Beta Strategies. ERI Scientific
Beta White Paper.
References
• Harvey, C., Y. Liu, and H. Zhu. 2013. … and the Cross-Section of Expected Returns.
Working Paper. Duke University.
• Huang, W., Q. Liu, G. Rhee, and L. Zhang. 2010. Return reversals, idiosyncratic risk, and
expected returns. Review of Financial Studies 23: 147-168.
• Jagannathan, R. and T. Ma. 2003. Risk Reduction in Large Portfolios: Why Imposing the
Wrong Constraints Helps. Journal of Finance 58(4): 1651-1684.
• Kan, R., and G. Zhou. 2007. Optimal Portfolio Choice with Parameter Uncertainty. Journal
of Financial and Quantitative Analysis 42(3): 621-656.
• Kogan, L., and M. Tian. 2013. Firm Characteristics and Empirical Factor Models: A
Data-Mining Experiment. Working Paper. MIT.
• Lakonishok, J., A. Shleifer, and R. Vishny. 1994. Contrarian Investment, Extrapolation
and Risk. Journal of Finance 49 (5): 1541–1578.
• Leland, H.E. 1999. Beyond Mean-Variance: Performance Measurement in a Non-symmetrical
World. Financial Analysts Journal 55: 27-36.
• Levy, R.A. 1974. Beta Coefficients as Predictors of Returns. Financial Analysts Journal
30(1): 61-69.
• Liu, L.X., and L. Zhang. 2008. Momentum Profits, Factor Pricing, and Macroeconomic
Risk. The Review of Financial Studies 21(6): 2417-2448.
• Maillard, S., T Roncalli, and J. Teiletche. 2010. The Properties of Equally Weighted Risk
Contribution Portfolios. The Journal of Portfolio Management 36(4): 60-70.
• Markowitz, H. 1952. Portfolio Selection. The Journal of Finance 7(1): 77-91.
• Martellini, L., and P. Priaulet. 2002. Competing methods for option hedging in the
presence of transaction costs. Journal of Derivatives 9(3): 26-3.
• McLean, D., and J. Pontiff. 2013. Does Academic Research Destroy Stock Return
Predictability? Working Paper. MIT.
• Merton, R. 1973. An Intertemporal Capital Asset Pricing Model. Econometrica 41(5):
867-887.
• Merton, R. 1980. On Estimating the Expected Return on the Market: An Exploratory
Investigation. Journal of Financial Economics 8: 323-361.
• Meucci, A., A. Santangelo, and R. Deguest. 2013. Measuring Portfolio Diversification Based
on Optimized Uncorrelated Factors. http://ssrn.com/abstract=2276632. Working Paper.
• Novy-Marx, R. 2013. The Other Side of Value: The Gross Profitability Premium. The
Journal of Financial Economics 108(1): 1-28.
• Plyakha, Y., R. Uppal, and G. Vilkov. 2014. Equal or Value Weighting? Implications for
Asset-Pricing Tests. Working Paper.
• Qian, E. 2006. On the Financial Interpretation of Risk Contribution: Risk Budgets Do
Add Up. Journal of Investment Management 44.
• Roncalli, T. 2013. Introduction to Risk Parity and Budgeting. Chapman & Hall. CRC
Financial Mathematics Series.
References
• Ross, S.A. 1976. The Arbitrage Theory of Capital Asset Pricing. Journal of Economic
Theory 13: 341-360.
• Sharpe, W.F. 1964. Capital Asset Prices: A Theory of Market Equilibrium under Conditions
of Risk. Journal of Finance 19(3): 425-442.
• Tu, J., and G. Zhou. 2011. Markowitz meets Talmud: A Combination of Sophisticated and
Naive Diversification Strategies. Journal of Financial Economics 99(1): 204-215.
• Turner, C., R. Starz, and C. Nelson. 1989. A Markov Model of Heteroskedasticity, Risk,
and Learning in the Stock Market. Journal of Financial Economics 25: 3-22.
• Xiong, J., T. Idzorek, P. Chen, and R. Ibbotson. 2009. Impact of Size and Flows on
Performance for Funds of Hedge Funds. Journal of Portfolio Management 35(2): 118-130.
• Zhang, L. 2005. The Value Premium. The Journal of Finance 60(1): 67-103.

References

About Amundi ETF & Indexing

About Amundi ETF & Indexing
Amundi Amundi ETF & Indexing

Amundi ranks first in Europe1
and ninth With a long-standing experience combined
worldwide in the asset management
1 with a strong pricing power, we offer first-
industry with AUM of close to €800 billion class replication on more than 100 indices
worldwide2. to internationally renowned institutions.
Located at the heart of the main investment The Indexing expertise is built on the search
regions in almost 30 countries, Amundi for value-added sources within strict risk
offers a comprehensive range of products framework. It comprises a wide range
covering all asset classes and major of open-ended funds as well as having
currencies. the capacity to implement customised
mandates, including SRI and smart beta
Amundi has developed savings solutions approaches.
to meet the needs of more than 100
million retail clients worldwide and designs In the ETF segment, Amundi has also
innovative, high-performing products for successfully become a major player thanks
institutional clients which are tailored to its strategy of competitive prices,
specifically to their requirements and risk innovation and high-quality tracking. Our
profile. ETF business has been growing consistently
in recent years to rank1 among the top
The group contributes to funding the five European providers by assets under
economy by orienting savings towards management.
company development.
Amundi ETF & Indexing experienced team
Amundi has become a leading European of dedicated index fund managers is based
player in asset management, recognised for: in Europe and in Japan, with a recognized
• Product performance and transparency; track record, and benefiting from Amundi
• Quality of client relationships based on dealing capabilities and research teams’
a long-term advisory approach; excellence.
• Efficiency in its organisation and teams’
1 - Amundi ETF/ Bloomberg as at 31 December 2013
promise to serving its clients;
• Commitment to sustainable development
and socially responsible investment policies.
1 - No.1 of the Portfolio Manager having their registered office in

Europe - Amundi group’s total assets under management - Source IPE
˝Top 400 asset managers active in the European marketplace˝ published
in June 2013, based on figures as at December 2012. Interviews of
asset management companies on their assets as at end-December
2012 (open-end funds, dedicated funds, mandates).
2 - Amundi Group figures as at 31st March 2014 Issued by Amundi
- Société anonyme with a share capital of €596 262 615 - Portfolio
manager regulated by the AMF under number GP04000036 - Head
office: 90 boulevard Pasteur – 75015 Paris – France – 437 574 452
RCS Paris

About EDHEC-Risk Institute

Founded in 1906, EDHEC is one The Choice of Asset Allocation Six research programmes have been
of the foremost international and Risk Management conducted by the centre to date:
business schools. Accredited by
EDHEC-Risk structures all of its research • Asset allocation and alternative
the three main international
academic organisations, work around asset allocation and risk diversification
EQUIS, AACSB, and Association management. This strategic choice is • Style and performance analysis
of MBAs, EDHEC has for a applied to all of the Institute's research • Indices and benchmarking
number of years been pursuing
programmes, whether they involve • Operational risks and performance
a strategy of international
excellence that led it to set up proposing new methods of strategic • Asset allocation and derivative
EDHEC-Risk Institute in 2001. allocation, which integrate the alternative instruments
This institute now boasts a class; taking extreme risks into account • ALM and asset management
team of over 95 permanent
professors, engineers and
in portfolio construction; studying the
support staff, as well as 48 usefulness of derivatives in implementing These programmes receive the support of
research associates from the asset-liability management approaches; a large number of financial companies.
financial industry and affiliate or orienting the concept of dynamic The results of the research programmes
professors..
“core-satellite” investment management are disseminated through the EDHEC-Risk
in the framework of absolute return or locations in Singapore, which was
target-date funds. established at the invitation of the
Monetary Authority of Singapore (MAS);
the City of London in the United Kingdom;
Academic Excellence Nice and Paris in France; and New York in
and Industry Relevance the United States.
In an attempt to ensure that the research
it carries out is truly applicable, EDHEC has EDHEC-Risk has developed a close
implemented a dual validation system for partnership with a small number of
the work of EDHEC-Risk. All research work sponsors within the framework of
must be part of a research programme, research chairs or major research projects:
the relevance and goals of which have • Core-Satellite and ETF Investment, in
been validated from both an academic partnership with Amundi ETF
and a business viewpoint by the Institute's • Regulation and Institutional
advisory board. This board is made up of Investment, in partnership with AXA
internationally recognised researchers, Investment Managers
the Institute's business partners, and • Asset-Liability Management and
representatives of major international Institutional Investment Management,
institutional investors. Management of the in partnership with BNP Paribas
research programmes respects a rigorous Investment Partners
validation process, which guarantees the • Risk and Regulation in the European
scientific quality and the operational Fund Management Industry, in
usefulness of the programmes. partnership with CACEIS
• Exploring the Commodity Futures
Risk Premium: Implications for
Asset Allocation and Regulation, in
partnership with CME Group

• Asset-Liability Management in Private The philosophy of the Institute is to

Wealth Management, in partnership validate its work by publication in
with Coutts & Co. international academic journals, as well as
• Asset-Liability Management to make it available to the sector through
Techniques for Sovereign Wealth Fund its position papers, published studies, and
Management, in partnership with conferences.
Deutsche Bank
• The Benefits of Volatility Derivatives Each year, EDHEC-Risk organises three
in Equity Portfolio Management, in conferences for professionals in order to
partnership with Eurex present the results of its research, one in
• Structured Products and Derivative London (EDHEC-Risk Days Europe), one
Instruments, sponsored by the French in Singapore (EDHEC-Risk Days Asia), and
Banking Federation (FBF) one in New York (EDHEC-Risk Days North
• Optimising Bond Portfolios, in America) attracting more than 2,500
partnership with the French Central professional delegates.
Bank (BDF Gestion)
• Asset Allocation Solutions, in EDHEC also provides professionals with
partnership with Lyxor Asset access to its website, www.edhec-risk.com,
Management which is entirely devoted to international
• Infrastructure Equity Investment asset management research. The website,
Management and Benchmarking, which has more than 65,000 regular
in partnership with Meridiam and visitors, is aimed at professionals who
Campbell Lutyens wish to benefit from EDHEC’s analysis and
• Investment and Governance expertise in the area of applied portfolio
Characteristics of Infrastructure Debt management research. Its monthly
Investments, in partnership with Natixis newsletter is distributed to more than 1.5
• Advanced Modelling for Alternative million readers.
Investments, in partnership with
Newedge Prime Brokerage EDHEC-Risk Institute:
• Advanced Investment Solutions for Key Figures, 2011-2012
Liability Hedging for Inflation Risk, Nbr of permanent staff 90
in partnership with Ontario Teachers’ Nbr of research associates 20
Pension Plan Nbr of affiliate professors 28
• The Case for Inflation-Linked Overall budget €13,000,000
Corporate Bonds: Issuers’ and Investors’ External financing €5,250,000
Perspectives, in partnership with Nbr of conference delegates 1,860
Rothschild & Cie Nbr of participants
640
• Solvency II, in partnership with Russell at research seminars
Investments Nbr of participants at EDHEC-Risk
182
Institute Executive Education seminars
• Structured Equity Investment
Strategies for Long-Term Asian Investors,
in partnership with Société Générale
Corporate & Investment Banking

The EDHEC-Risk Institute PhD in School of Management to set up joint

Finance certified executive training courses in
The EDHEC-Risk Institute PhD in Finance North America and Europe in the area of
is designed for professionals who aspire investment management.
to higher intellectual levels and aim to
redefine the investment banking and asset As part of its policy of transferring know-
management industries. It is offered in two how to the industry, EDHEC-Risk Institute
tracks: a residential track for high-potential has also set up ERI Scientific Beta. ERI
graduate students, who hold part-time Scientific Beta is an original initiative
positions at EDHEC, and an executive track which aims to favour the adoption of the
for practitioners who keep their full-time latest advances in smart beta design and
jobs. Drawing its faculty from the world’s implementation by the whole investment
best universities, such as Princeton, industry. Its academic origin provides the
Wharton, Oxford, Chicago and CalTech, foundation for its strategy: offer, in the
and enjoying the support of the research best economic conditions possible, the
centre with the greatest impact on the smart beta solutions that are most proven
financial industry, the EDHEC-Risk Institute scientifically with full transparency in
PhD in Finance creates an extraordinary both the methods and the associated
platform for professional development and risks.
industry innovation.
Research for Business

The Institute’s activities have also given
rise to executive education and research
service offshoots. EDHEC-Risk's executive
education programmes help investment
professionals to upgrade their skills with
advanced risk and asset management
training across traditional and alternative
classes. In partnership with CFA Institute,
it has developed advanced seminars based
on its research which are available to CFA
charterholders and have been taking
place since 2008 in New York, Singapore
and London.
In 2012, EDHEC-Risk Institute signed two

strategic partnership agreements with
the Operations Research and Financial
Engineering department of Princeton
University to set up a joint research
programme in the area of risk and
investment management, and with Yale

EDHEC-Risk Institute
Publications and Position Papers
(2011-2014)

EDHEC-Risk Institute Publications

(2011-2014)
2014
• Martellini, L., V. Milhau and A. Tarelli. Towards Conditional Risk Parity — Improving Risk
Budgeting Techniques in Changing Economic Environments (April).
• Amenc, N., and F. Ducoulombier. Index Transparency – A Survey of European Investors
Perceptions, Needs and Expectations (March).
• Ducoulombier, F., F. Goltz, V. Le Sourd, and A. Lodh. The EDHEC European ETF Survey
2013 (March).
• Badaoui, S., Deguest, R., L. Martellini and V. Milhau. Dynamic Liability-Driven Investing
Strategies: The Emergence of a New Investment Paradigm for Pension Funds? (February).
• Deguest, R., and L. Martellini. Improved Risk Reporting with Factor-Based Diversification
Measures (February).
• Loh, L., and S. Stoyanov. Tail Risk of Equity Market Indices: An Extreme Value Theory
Approach (February).
2013
• Lixia, L., and S. Stoyanov. Tail Risk of Asian Markets: An Extreme Value Theory Approach
(August).
• Goltz, F., L. Martellini, and S. Stoyanov. Analysing statistical robustness of cross-
sectional volatility. (August).
• Lixia, L., L. Martellini, and S. Stoyanov. The local volatility factor for asian stock markets.
(August).
• Martellini, L., and V. Milhau. Analysing and decomposing the sources of added-value
of corporate bonds within institutional investors’ portfolios (August).
• Deguest, R., L. Martellini, and A. Meucci. Risk parity and beyond - From asset allocation
to risk allocation decisions (June).
• Blanc-Brude, F., Cocquemas, F., Georgieva, A. Investment Solutions for East Asia's
Pension Savings - Financing lifecycle deficits today and tomorrow (May)
• Blanc-Brude, F. and O.R.H. Ismail. Who is afraid of construction risk? (March)
• Lixia, L., L. Martellini, and S. Stoyanov. The relevance of country- and sector-specific
model-free volatility indicators (March).
• Calamia, A., L. Deville, and F. Riva. Liquidity in european equity ETFs: What really
matters? (March).
• Deguest, R., L. Martellini, and V. Milhau. The benefits of sovereign, municipal and
corporate inflation-linked bonds in long-term investment decisions (February).
• Deguest, R., L. Martellini, and V. Milhau. Hedging versus insurance: Long-horizon
investing with short-term constraints (February).
• Amenc, N., F. Goltz, N. Gonzalez, N. Shah, E. Shirbini and N. Tessaromatis. The EDHEC
european ETF survey 2012 (February).


(2011-2014)
• Padmanaban, N., M. Mukai, L . Tang, and V. Le Sourd. Assessing the quality of asian
stock market indices (February).
• Goltz, F., V. Le Sourd, M. Mukai, and F. Rachidy. Reactions to “A review of corporate
bond indices: Construction principles, return heterogeneity, and fluctuations in risk
exposures” (January).
• Joenväärä, J., and R. Kosowski. An analysis of the convergence between mainstream
and alternative asset management (January).
• Cocquemas, F. Towar¬ds better consideration of pension liabilities in european union
countries (January).
• Blanc-Brude, F. Towards efficient benchmarks for infrastructure equity investments
(January).
2012
• Arias, L., P. Foulquier and A. Le Maistre. Les impacts de Solvabilité II sur la gestion
obligataire (December).
• Arias, L., P. Foulquier and A. Le Maistre. The Impact of Solvency II on Bond Management
(December).
• Amenc, N., and F. Ducoulombier. Proposals for better management of non-financial
risks within the european fund management industry (December).
• Cocquemas, F. Improving Risk Management in DC and Hybrid Pension Plans (November).
• Amenc, N., F. Cocquemas, L. Martellini, and S. Sender. Response to the european
commission white paper "An agenda for adequate, safe and sustainable pensions"
(October).
• La gestion indicielle dans l'immobilier et l'indice EDHEC IEIF Immobilier d'Entreprise
France (September).
• Real estate indexing and the EDHEC IEIF commercial property (France) index (September).
• Goltz, F., S. Stoyanov. The risks of volatility ETNs: A recent incident and underlying
issues (September).
• Almeida, C., and R. Garcia. Robust assessment of hedge fund performance through
nonparametric discounting (June).
• Amenc, N., F. Goltz, V. Milhau, and M. Mukai. Reactions to the EDHEC study “Optimal
design of corporate market debt programmes in the presence of interest-rate and
inflation risks” (May).
• Goltz, F., L. Martellini, and S. Stoyanov. EDHEC-Risk equity volatility index: Methodology
(May).
• Amenc, N., F. Goltz, M. Masayoshi, P. Narasimhan and L. Tang. EDHEC-Risk Asian index
survey 2011 (May).
• Guobuzaite, R., and L. Martellini. The benefits of volatility derivatives in equity portfolio
management (April).


(2011-2014)
• Amenc, N., F. Goltz, L. Tang, and V. Vaidyanathan. EDHEC-Risk North American index
survey 2011 (March).
• Amenc, N., F. Cocquemas, R. Deguest, P. Foulquier, L. Martellini, and S. Sender. Introducing
the EDHEC-Risk Solvency II Benchmarks – maximising the benefits of equity investments
for insurance companies facing Solvency II constraints - Summary - (March).
• Schoeffler, P. Optimal market estimates of French office property performance (March).
• Le Sourd, V. Performance of socially responsible investment funds against an efficient
SRI Index: The impact of benchmark choice when evaluating active managers – an update
(March).
• Martellini, L., V. Milhau, and A.Tarelli. Dynamic investment strategies for corporate
pension funds in the presence of sponsor risk (March).
• Goltz, F., and L. Tang. The EDHEC European ETF survey 2011 (March).
• Sender, S. Shifting towards hybrid pension systems: A European perspective (March).
• Blanc-Brude, F. Pension fund investment in social infrastructure (February).
• Ducoulombier, F., Lixia, L., and S. Stoyanov. What asset-liability management strategy
for sovereign wealth funds? (February).
• Amenc, N., Cocquemas, F., and S. Sender. Shedding light on non-financial risks – a
European survey (January).
• Amenc, N., F. Cocquemas, R. Deguest, P. Foulquier, Martellini, L., and S. Sender. Ground
Rules for the EDHEC-Risk Solvency II Benchmarks. (January).
• Amenc, N., F. Cocquemas, R. Deguest, P. Foulquier, Martellini, L., and S. Sender. Introducing
the EDHEC-Risk Solvency Benchmarks – Maximising the Benefits of Equity Investments
for Insurance Companies facing Solvency II Constraints - Synthesis -. (January).
• Amenc, N., F. Cocquemas, R. Deguest, P. Foulquier, Martellini, L., and S. Sender. Introducing
the EDHEC-Risk Solvency Benchmarks – Maximising the Benefits of Equity Investments
for Insurance Companies facing Solvency II Constraints (January).
• Schoeffler.P. Les estimateurs de marché optimaux de la performance de l’immobilier
de bureaux en France (January).
2011
• Amenc, N., F. Goltz, Martellini, L., and D. Sahoo. A long horizon perspective on the
cross-sectional risk-return relationship in equity markets (December 2011).
• Amenc, N., F. Goltz, and L. Tang. EDHEC-Risk European index survey 2011 (October).
• Deguest,R., Martellini, L., and V. Milhau. Life-cycle investing in private wealth
management (October).
• Amenc, N., F. Goltz, Martellini, L., and L. Tang. Improved beta? A comparison of index-
weighting schemes (September).


(2011-2014)
• Le Sourd, V. Performance of socially responsible investment funds against an

Efficient SRI Index: The Impact of Benchmark Choice when Evaluating Active Managers
(September).
• Charbit, E., Giraud J. R., F. Goltz, and L. Tang Capturing the market, value, or momentum
premium with downside Risk Control: Dynamic Allocation strategies with exchange-traded
funds (July).
• Scherer, B. An integrated approach to sovereign wealth risk management (June).
• Campani, C. H., and F. Goltz. A review of corporate bond indices: Construction principles,
return heterogeneity, and fluctuations in risk exposures (June).
• Martellini, L., and V. Milhau. Capital structure choices, pension fund allocation decisions,
and the rational pricing of liability streams (June).
• Amenc, N., F. Goltz, and S. Stoyanov. A post-crisis perspective on diversification for risk
management (May).
• Amenc, N., F. Goltz, Martellini, L., and L. Tang. Improved beta? A comparison of index-
weighting schemes (April).
• Amenc, N., F. Goltz, Martellini, L., and D. Sahoo. Is there a risk/return tradeoff across
stocks? An answer from a long-horizon perspective (April).
• Sender, S. The elephant in the room: Accounting and sponsor risks in corporate pension
plans (March).
• Martellini, L., and V. Milhau. Optimal design of corporate market debt programmes in
the presence of interest-rate and inflation risks (February).

EDHEC-Risk Institute Position Papers

(2011-2014)
2012
• Till, H. Who sank the boat? (June).
• Uppal, R. Financial Regulation (April).
• Amenc, N., F. Ducoulombier, F. Goltz, and L. Tang. What are the risks of European ETFs?
(January).
2011
• Amenc, N., and S. Sender. Response to ESMA consultation paper to implementing
measures for the AIFMD (September).
• Uppal, R. A Short note on the Tobin Tax: The costs and benefits of a tax on financial
transactions (July).
• Till, H. A review of the G20 meeting on agriculture: Addressing price volatility in the
food markets (July).

For more information, please contact:
Carolyn Essid on +33 493 187 824
or by e-mail to: carolyn.essid@edhec-risk.com
EDHEC-Risk Institute
393 promenade des Anglais
BP 3116 - 06202 Nice Cedex 3
France
Tel: +33 (0)4 93 18 78 24
EDHEC Risk Institute—Europe

10 Fleet Place, Ludgate
London EC4M 7RB
United Kingdom
Tel: +44 207 871 6740
EDHEC Risk Institute—Asia

1 George Street
#07-02
Singapore 049145
Tel: +65 6438 0030
EDHEC Risk Institute—North America

One Boston Place, 201 Washington Street
Suite 2608/2640 — Boston, MA 02108
United States of America
Tel: +1 857 239 8891
EDHEC Risk Institute—France

16-18 rue du 4 septembre
75002 Paris
France
Tel: +33 (0)1 53 32 76 30
www.edhec-risk.com

EDHEC-Risk Publication Risk Allocation Factor Investing Smart Beta PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

EDHEC-Risk Publication Risk Allocation Factor Investing Smart Beta PDF

Uploaded by

Copyright:

Available Formats

An EDHEC-Risk Institute Publication

Risk Allocation, Factor Investing

with the support of

1. Selecting Desired Factor Exposures................................................................19

2. Designing Efficient and Investable Proxies for Risk Premia.................... 25

3. Risk Allocation with Smart Factor Indices.....................................................39

About Amundi ETF & Indexing.............................................................................83

About EDHEC-Risk Institute.................................................................................84

EDHEC-Risk Institute Publications and Position Papers (2011-2014).........89

An EDHEC-Risk Institute Publication 3

About the Authors

Noël Amenc is a professor of finance at EDHEC Business School, director of

Romain Deguest is a senior research engineer at EDHEC-Risk Institute.

Felix Goltz is Head of Applied Research at EDHEC-Risk Institute. He carries

Ashish Lodh is Senior Quantitative Analyst. He does research in empirical

Lionel Martellini is professor of finance at EDHEC Business School and

About the Authors

An EDHEC-Risk Institute Publication 5

About the Authors

6 An EDHEC-Risk Institute Publication

An EDHEC-Risk Institute Publication 7

Introduction: From Cap-Weighted Designing Efficient and Investable

On the one hand, starting with a focus

8 An EDHEC-Risk Institute Publication

An EDHEC-Risk Institute Publication 9

10 An EDHEC-Risk Institute Publication

An EDHEC-Risk Institute Publication 11

12 An EDHEC-Risk Institute Publication

An EDHEC-Risk Institute Publication 13

other developed stock markets as well. By

Moreover, additional value can be added at

14 An EDHEC-Risk Institute Publication

An EDHEC-Risk Institute Publication 15

16 An EDHEC-Risk Institute Publication

This analysis of the dual challenges to (equal-dollar contribution or equal risk

An EDHEC-Risk Institute Publication 17

18 An EDHEC-Risk Institute Publication

An EDHEC-Risk Institute Publication 19

1. Selecting Desired Factor Exposures

20 An EDHEC-Risk Institute Publication

1. Selecting Desired Factor Exposures

An EDHEC-Risk Institute Publication 21

1. Selecting Desired Factor Exposures

Risk-Based Explanation Behavioural Explanation

Value Zhang (2008) provide empirical evidence

22 An EDHEC-Risk Institute Publication

1. Selecting Desired Factor Exposures

to liquidate these assets in bad times when

Behavioural explanations for the low-risk

An EDHEC-Risk Institute Publication 23

1. Selecting Desired Factor Exposures

24 An EDHEC-Risk Institute Publication

An EDHEC-Risk Institute Publication 25

2. Designing Efficient and Investable Proxies

2.1. Conventional Approaches to index, simply because the integration of

26 An EDHEC-Risk Institute Publication

2. Designing Efficient and Investable Proxies

An EDHEC-Risk Institute Publication 27

2. Designing Efficient and Investable Proxies

2.2. Smart Beta 2.0 Approach to to unrewarded strategy specific risks.

28 An EDHEC-Risk Institute Publication

2. Designing Efficient and Investable Proxies

An EDHEC-Risk Institute Publication 29