Equity Investing Insights Into A Better Portfolio

Equity investing:
Insights into a better portfolio
Equity investing:
Insights into a better portfolio
unique qualities of
...the
equity make it a likely

building block of most
asset portfolios for the
foreseeable future.
towerswatson.com/equity-investing
2 towerswatson.com
Contents
06
Section one:
Portfolio construction
30
Section two:
Manager selection
60
Section three:
Manager monitoring
08
Best-in-class equity structure
What is best practice when
considering equity portfolios?
32
Assessing investment skill
in equity managers
A necessary task prior to
making any investment and
as part of an ongoing
monitoring process
62
Benchmarks matter
It is important to ensure
that they continue to
be appropriate
16
Using smart beta in equities
Smart beta can greatly
improve an investors ability
to achieve its objectives
20
Understanding emerging
market equity
How to best access
the expected growth in
these markets
25
Low volatility equity
strategies should
you include them in
your portfolio?
These strategies target
market returns but with
much lower risk
40
Do not hire managers
for past performance
More evidence of why this
approach can lose value
66
What results should
dictate firing a manager?
At what point is the result
so bad that you just have
to move on
43
Quantitative investing
will quants strike back?
Our current views
on quantitative
investment strategies
48
Concentrated equity products
Why we generally prefer them to
diversified products
52
Low volatility equity from
smart idea to smart execution
How to avoid the
potential pitfalls with
this investment idea
55
Sustainability in
investment research
a pragmatic approach
Adding transformational
change to the asset
owners agenda
56
What to look for in
a passive manager
Rigorous qualitative research
can lead to better outcomes
Equity investing: Insights into a better portfolio 3
It
is hard to find managers capable
of sustained outperformance, combine
them in a risk-proportionate portfolio
and manage this through different
economic and market conditions.
4 towerswatson.com
Introdction
Ever since trading began in 17th century coffee houses,

equities have provided an ecient means of raising capital
and transferring ownership. Every day, billions of shares
change hands, each of them representing an ownership
interest in a business.
Critically, the clearing price depends upon the
collective opinion of investors about the future
cash flows of these assets. Of course, future
returns are not certain. It is the dissenting
views over future income streams that create
the opportunity to produce outperformance.
And that creates the challenge, and also the
opportunity, for the manager researcher and
investment committee to find skilled investors
that can produce superior returns on a
sustainable basis.
It is hard to find managers capable of
sustained outperformance, combine them
in a risk-proportionate portfolio, and manage
this through different economic and market
conditions. This task is complicated by many
factors, such as change in the sources of
uncertainty when predicting investment
outcomes, evolution within financial markets
leading to the creation of analogues to
direct equity investment, globalisation and
technological innovation all of which may
reduce inefficiencies in market pricing.
Nevertheless, the unique qualities of equity

make it a likely building block of most asset
portfolios for the foreseeable future. Indeed,
with fixed income yields at historical lows, it is
incumbent on investors to improve the chances
of asset growth from their equity portfolios.
In this compendium of articles, we provide
practical insights about three functional
areas in which asset owners typically
operate: portfolio construction, manager
selection and manager monitoring.
I hope you will find these articles a
thought-provoking and practical aid to
understanding how to construct an equity
portfolio, and select and evaluate its
managers. If you have questions on any of
the articles, please contact either your usual
Towers Watson consultant, or contact me.
James MacLachlan
Global Head of Equity Manager Research,
Towers Watson
Section on
Portfolio construction
6 towerswatson.com
Investors now have more options than

ever to consider when constructing an
equity portfolio. With greater choice
comes the risk of losing sight of the bigger
picture. Investors should therefore begin by
creating a framework for equity portfolio
construction. We advocate a holistic
view of risk and return. We prefer to
blend best-in-class active managers with
smart beta solutions. The latter can help
to eciently manage portfolio biases or
substitute style-oriented active managers at
lower fees. Here, we offer some practical
considerations around topical areas such as
emerging markets and style-based investing.
01 Portfolio construction
Best-in-class equity structure

What next for equities?
...we
believe this is a good time for

institutional investors to revisit their
approach to investing in equities and
to consider what is best practice when
constructing equity portfolios.
8 towerswatson.com
Over the last decade or so, there has been an

explosion in the available options for institutional
investors outside the traditional asset classes of
equities and bonds. Asset owners, including those
under the advice of Towers Watson, have been
diversifying their risk-seeking assets away from
the equity risk premium and across other return
drivers. However, Towers Watson recognises that
the equity risk premium still makes an important
contribution to institutional investors long-term
returns and that there have been significant
developments within equities over the last few
years. As such, we believe this is a good time for
institutional investors to revisit their approach to
investing in equities and to consider what is best
practice when constructing equity portfolios.
Evolution of equity investing

In principle, the task of investing in long-only
equities has evolved very little. However, some
areas have seen considerable development.
For example, the metrics one might consider to
assess a companys fundamentals are largely
unchanged over the past 80 years, but the
availability of this data, the way it is used and the
development of broader explanations for market
behaviour has dramatically changed. In the last
15 years, the market has witnessed at least
two significant market bubbles/cycles with the
dot-com bust and the global financial crisis. As
a result, investors understanding of behavioural
biases has evolved and we have been reminded
of the importance of macroeconomic, social and
political issues (at least in the developed markets
if this was never forgotten in emerging ones).
So whilst the task at hand may look similar, the
solution and its execution may be very different.
The essential building blocks of an equity portfolio
have, for a long time, been bulk beta (market
capitalisation-weighted passive equity) and
alpha (active management). Due to some of the
developments discussed above, we can now add
another pillar to this framework, that of smart
beta. Many of the concepts behind smart beta
have been around for some time, often embedded
within the investment processes of active
managers and rolled up in the reported alpha.
However, recent improvements in data availability,
quantitative management techniques and lower
costs have created the opportunity to isolate
these sources of alpha and construct dedicated
smart beta products. Figure 01 provides an
indication of just how rapid the growth of smart
beta has been.
Figure 01. Number of low volatility equity products over time

90
80
70
60
50
40
30
20
10
0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Source: eVestment Alliance

Methodology: Analysis based on searching for equity products on the eVestment Alliance
database containing the terms low volatility or defensive
The main objective of smart beta is to capture

a particular risk premium for a low cost. These
strategies are expected to improve investment
efficiency relative to bulk beta and aid portfolio
construction (see Using smart beta in equities1
for further information on this topic).
When considering the use of smart beta,
investors must research the validity of the risk
premium or smart beta strategy as there can
be many different strategies and definitions.
Care also needs to be taken in implementation
as there is now many implementation options
for broadly similar strategies and differences in
implementation can lead to very different results.
Towers Watsons smart beta research team has
completed extensive research in this area and
can assist investors through this process.
Importance of portfolio construction

Questioning and examining the ability of equity
managers to construct portfolios is something
that asset owners have been doing, relatively
routinely, for a long time. However, given the
developments we have seen in equity investing,
it is now also important for asset owners
themselves to have a robust portfolio
construction process. It is no longer generally
considered sufficient to have an allocation to
bulk beta and one or two active managers to
construct an equity portfolio. The onus is now
firmly on the asset owner to develop their own
portfolio construction skills, or delegate this
task to third parties.
The need for multiple products

As investment markets broaden and deepen,
the breadth of opportunity around the world has
become vast and complex. To access these
opportunities, asset owners should consider
multiple products across the building blocks
available to them (bulk beta, smart beta
and alpha).
Smart beta and portfolio construction

There are various different types of smart beta
strategies, including fundamentally weighted
indices and risk weighted indices. There are
also thematic approaches that would give the
investor exposure to a particular long-term driver
such as scarcity of natural resources. Different
types of smart betas come with different portfolio
characteristics, structured product investors
should pay greater attention to the associated
implications for risk.
Asset owners need to take a holistic view when

constructing portfolios, building diversified
exposure to multiple opportunities (risks)
through multiple products. To use all of the tools
described, portfolio construction skill becomes
an essential pre-requisite to building robust
long-term risk-proportionate equity portfolios.
Figure 02. Benefits of diversification in a manager structure

Assuming skilled managers expected probability of meaningful
underperformance from the overall structure at some point over
a 10 year period
100%
For example, to reduce the likelihood of 10%
underperformance to less than 10%, at least
eight managers are needed
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Probability of at least:
10 11 12 13 14 15
Number of managers
5% underperformance 7.5% underperformance

10% underperformance 15% underperformance
Probability of underperformance from the overall structure decreases rapidly initially. Assuming you
can find enough high conviction managers there is a strong case for including at least 4 managers in
a manager structure to protect against the possibility of extreme underperformance.
Probabilities are approximated assuming an individual manager tracking error of 5% pa, a net
information ratio of 0.33, an average manager active correlation of 0.3 and a normal distribution of
manager returns. Underperformance is defined as a peak to trough cumulative, not annualised.
Style diversification
Equity market returns do not follow the standard

statistical pattern of a normal distribution as
demonstrated in Figure 03. More unusual price
movements are observed in equity markets with
greater frequency than a normal distribution
would suggest and this further supports the
argument for increased diversification.
10 towerswatson.com
20%
15%
10%
5%
0%
-5%
-10%
-15%
-20%
Observed S&P500 distribution Normal distribution
2012
In reality, far greater diversification is

often required than investors typically expect
(see Figure 02).
Figure 03. Daily price movements
1928
1932
1936
1940
1944
1948
1952
1956
1960
1964
1968
1972
1976
1980
1984
1988
1992
1996
2000
2004
2008
Active managers have a wide variety of investment

styles, philosophies and opportunity sets. Single
word descriptors, such as value and growth,
often used by investors to describe managers
approaches, are inadequate. Investors need to
understand the approach of the active manager
in detail and build sufficient diversification across
the full spectrum of investment styles. In addition,
investors should determine the extent to which the
alpha is driven by the same underlying style risk
premium that can be captured (at lower cost) by
a smart beta alternative.
Portfolio construction considerations

To generate superior long-term returns, we believe
that asset owners require skill in manager selection
and portfolio construction.
Best-in-class manager selection is required to
identify skilled active managers from a universe of
many thousands of competing products. Similarly,
the number of smart beta products has also been
increasing rapidly, meaning expertise is also

required to select the most appropriate smart
beta approach. Simply identifying the beta and
using any product that will give you exposure is not
sufficient and can have negative consequences.
We believe portfolio construction should include
the principles outlined in Figure 04 below.
Figure 04. Portfolio construction principles

Principle
Rationale/benefits
Holistic view of return

driver framework
Ensure returns from smart beta and alpha are expected to come from differentiated
return drivers
Integration of smart beta
Low-cost solution allows better use of fee budget

Facilitates targeted use of (potentially more expensive) alpha strategies
Enables asset owners to change portfolio characteristics to suit prevailing market
conditions more simply, with lower cost
Use of best-in-class
active managers
Appropriate diversification
Combining smart beta and multiple alpha strategies to achieve more

efficient diversification
Risk and scenario analysis
Risk management framework in place

Consideration of scenario analysis to account for non-quantitative risk metrics
(such as sentiment, political, social)
Timely decision making
Capitalise on market volatility and act decisively to take advantage of opportunities

or control risk
Contrarian hiring and firing
Academic studies have shown and behavioural analysis supports the view
that investors often destroy value by hiring managers with good performance,
and firing managers with poor performance. Conversely a contrarian approach
should add value2,3
Incorporating longer-term
views of equity market
risk factors and current
market conditions
Portfolios should reflect longer-term capital market views in the context of current
market valuations. Towers Watsons views are provided by its Global Investment
Committee and Asset Research Team
Consideration given to attractiveness of style risk factors (see Towers Watsons
articles on Value investing4 and Low volatility equity5)
Continuous and dynamic
Ongoing monitoring of the underlying strategies and the combined portfolio

Consideration of portfolio rebalancing
Focus on best-in-class opportunities

Niche alpha strategies as beta exposure is gained elsewhere
Consider long-short alongside long-only
Commission new strategies from skilled managers that are expected to offer
improved risk/reward
Expanding further on some of the key points in

Figure 04:
The impact of using appropriate smart beta
products on an overall fee budget can be
significant. Allocating a material proportion of
assets to low-cost products frees this budget to
allocate to best-in-class, highly active strategies.
It is important to use best-in-class active
managers. The best specialist equity managers
are often found in boutiques that the investors
have established to provide greater focus. In
some cases, they may be managing money
for high-net-worth individuals or running hedge
funds. Often the mind-set or skill-set of these
investors is different. For example, high-net-worth
managers tend to think in terms of absolute
risk rather than relative to market indices.
Hedge fund managers often place a different
emphasis on risk management techniques but
can sometimes be persuaded to apply their
approach to long-only products.
Significant research resource is required to find
these skilled managers. Investing with them may
require tolerance for portfolios which have very
high relative risk (but not necessarily absolute
risk) and are potentially highly concentrated.
In addition, the asset management firm is often

not a recognised brand name. Studies have
shown that portfolio managers typically add
value in their high conviction stock picks
but often destroy value with the unintended
underweight positions in the portfolio. Having
more concentrated portfolios with assets
focused in the managers highest conviction
ideas should offset unintended underweight
positions and lead to better outcomes. (See
Towers Watsons article Concentrated equity
products for further information.) 6
Diversifying across these products and using risk
management frameworks to construct portfolios
is important as these strategies can exhibit strong
style biases due to their focused approaches.
Style or beta exposures can overwhelm alpha if
not managed properly. In the five years since the
financial crisis we have seen the extent to which
styles have diverged (see Figure 05), and this
style risk needs to be controlled.
A risk management framework can help investors
to consider, manage and exploit the various risks
discussed above, such as market, style, macro
and political risks, sentiment and manager
selection risks.
Cumulative relative returns
Figure 05. Divergence of style returns
The impact of using

appropriate smart beta
products on an overall fee
budget can be significant.
150%
125%
Growth
100%
75%
50%
Value
Value
25%
0%
Value
Quality
Value
Growth
-25%
-50%
Quality
-75%
Dot-com
2000
bear market
00s
bull market
Global
financial crisis
Source: Style Research, Towers Watson

All returns as in US$ measured relative to a market capitalisation portfolio of developed market stocks.
Style portfolios are the top quartile of stocks selected from a developed market universe based on a
composite ranking of typical fundamental traits associated with that investment style (for example high
earnings yield, high ROE or high trailing earnings growth). Portfolios are rebalanced quarterly and market
capitalisation weighted. The times periods are defined as: dot-com 01/02/1996 to 31/03/2000, 2000 bear
market 01/04/2000 to 30/09/2002, 00s bull market 01/10/2002 to 31/10/2007, global financial crisis
01/11/2007 to 31/10/2013.
12 towerswatson.com
Complexity and governance

It is reasonable to ask is this additional complexity
worth it? We have advised clients for many years
to either simplify their strategy or raise their game.
At one end of the spectrum is a market capitalisation
passive portfolio that is suitable for many investors.
However, if asset owners are going to compete
effectively for investment returns, they should
consider every tool available and aim for best practice.
Figure 06. Internal skill central to building high

governance portfolios
Alpha
(manager skill)
Whilst there are greater expected rewards from this

approach, it requires more internal governance and
portfolio construction skill from the asset owner.
Therefore, this approach may not be suitable
for everyone.
Bulk beta
Internal
portfolio
management
skill
Smart beta
Asset owners should determine what level of

complexity is appropriate, given their requirements
and their governance levels. If necessary, efforts
can be made to increase or decrease the internal
governance levels, as required.
Figure 07. Impact of governance levels

Low
Governance
High
Low
Complexity
High
Bulk beta
Bulk beta, smart beta, alpha

Niche strategies
Broad global
Lower
Expected returns/
financial efficiency
Higher
Towers Watsons equity

management services
Figure 08. Model portfolio performance
Unconstrained equity
model portfolio
Performance
net of fees
Tracking
error
Net information
ratio
+1.9% pa
3.4% pa
0.6
Performance data shown since inception, as at end June 2013

Benchmark: MSCI AII Country World Index
Inception date: 1 January 2005
Base currency: US$
Towers Watson works

with asset owners in many
different ways to achieve
their objectives.
14 towerswatson.com
Towers Watson works with asset owners in many

ways to achieve their objectives. In some cases,
we assist asset owners as they develop these
in-house skills and raise their governance levels.
In other cases, with more limited governance
structures, we assist with developing less complex,
lower expected risk/return solutions. For asset
owners that want higher expected returns but lack
the internal governance and resources to build
these portfolios, Towers Watsons Delegated
Investment Services (DIS) business may be a
suitable option. By delegating the manager
selection and construction of the equity portfolio
to Towers Watson, the asset owner immediately
harnesses Towers Watsons governance, research
and portfolio construction resources.
DIS is a bespoke service, as we recognise that
different investors face different challenges.
Most notably, investors with smaller asset
portfolios are often prohibited from adopting a
best-in-class approach due to lack of scale or
bargaining power. By negotiating on behalf of our
entire DIS client base, those smaller clients that
delegate to Towers Watson, benefit from greater
resources and collective bargaining power that
they would otherwise be unable to access.
Towers Watson has demonstrated success

in its management of active equity portfolios
through ongoing manager selection and portfolio
construction. In particular, we have highlighted
the long-term track record of the unconstrained
equity model portfolio that is constructed from
best ideas around the world.
Example equity portfolio

With an internal portfolio management capability,
asset owners can employ the full range of portfolio
construction principles and approaches discussed
here. As an example, we have blended the results
of the unconstrained equity model portfolio, the
performance of some of our favoured long-biased
specialist activist managers and three smart beta
portfolios. The results are as follows:
Annualised alpha
3.7%
Tracking error
4.0%
Information ratio
0.94
Approximate fee
0.9%
Construction of the portfolio can be tailored to suit

investor preferences. For example, a desire for
lower cost would most likely employ more smart
beta and less specialist or highly active long-only
managers. A lower absolute risk solution would
probably use certain types of smart beta and
potentially more long-short managers but
fewer high beta long-only managers.
Developments in equity markets

and the industry have added
complexity and breadth, in terms
of available products and portfolio
construction tools.
Conclusion
Whilst investors have typically been diversifying
away from equities in recent years, it is important
to recognise that the equity risk premium remains
an important driver of investors returns. As such,
we believe investors should revisit their approach
to constructing equity portfolios to ensure that
it keeps pace with the opportunity set around
the world and takes full advantage of the new
innovations seen in the industry.
Developments in equity markets and the industry
have added complexity and breadth, in terms
of available products and portfolio construction
tools. Most notable is the relatively recent rise
of smart beta. In a highly competitive world,
we believe asset owners should simplify their
strategy (for example, go passive) or raise their
game, in order to deal with this complexity and
benefit from it. Asset owners therefore need
to consider whether they have the appropriate
internal expertise to adopt a best-in-class portfolio
construction approach. If insufficient governance
is a constraint, then a possible solution is to
outsource some of these increased governance
requirements to Towers Watson. Our DIS approach
provides a bespoke solution to clients, allowing
them to benefit from our scale and expertise to
achieve their investment goals.
References
1
Using smart beta in equities. Towers Watson Limited, 2013.
Replacing managers for performance reasons.

Towers Watson Limited, 2011.
Goyal, A. and Wahal, S. The Selection and Termination

of Investment Management Firms by Plan Sponsors, 2005.
Value investing an old idea, but probably a good one.

Low volatility equity from smart idea to smart execution.

Concentrated equity products why we generally prefer them

to diversied products. Towers Watson Limited, 2013.
Using smart beta in equities
There is strong evidence that viewing markets as simple

systems is wrong. And yet investors continue to use models
based on this view of markets to make investment decisions.
This creates opportunity for those investors that are able to
think beyond these models.
A prime example of such a model is the Capital
Asset Pricing Model (CAPM). This model embraces
beta as the sole measure of risk and says (under
numerous, somewhat unrealistic assumptions),
that the market capitalisation-weighted equity
portfolio is the optimal investment strategy
for equities.
Smart beta is about thinking beyond the CAPM
model. By seeking to exploit the behavioural
biases that affect investors and capture risk
premiums outside the CAPM framework, these
strategies have the scope to be valuable additions
to the tools in an investors toolbox, be it for
portfolio construction, risk management, cost
management, or as alternatives to a passive
market capitalisation index strategy.
Here, we explain some of the main advantages
of using smart beta but also address the risks
and requirements to use it successfully in an
equity portfolio.
16 towerswatson.com
Figure 01. The alpha/beta continuum
Bulk beta
Smart beta
Diversifying
Market
capitalisation
passive
Alpha
Stock selection
Thematic
Systematic
Market timing
What is smart beta in equities?

We divide smart beta strategies into three broad
groups: diversifying, thematic and systematic.
Diversifying smart beta

Assets that provide diversification to an investors
existing portfolio are valuable. However, directly
investing in these assets (for example, infrastructure
or real estate) can be complicated and, in some
cases direct investment is too illiquid to be viable for
an investors portfolio. A properly structured portfolio
of listed equities allows an investor the potential
to gain exposure to these diversifying assets by
owning companies that are exposed to the long-term
underlying economics of these assets.
Thematic smart beta

The objective of thematic smart beta is to benefit
from secular shifts in financial markets. We believe
that financial markets are too focused on the
short term and struggle to consider the potential
range of future outcomes that current uncertainties
create. This results in the mispricing of risk for
long-term themes and potential secular changes.
Thematic smart beta accesses this mispricing
by investing in a portfolio which targets themes
or groups of themes, such as the development of
the emerging markets or future resource scarcity.
For investors with a sufficient belief framework and
governance structure these strategies may be used
to incorporate some of the themes in a portfolio as
described in our recent Secular Outlook1 paper.
Figure 02. Company variations tend to

mean-revert over some time period
Overvalued
Investors extrapolate
good news
Fair value
Over time companies

are fairly priced, but
not all companies
are fairly priced all
the time
Investors overreact
to bad news
Undervalued
Systematic smart beta

Systematic smart beta aims to exploit recurring
market opportunities that are expected to persist
over time. Key drivers of smart beta tend to be the
rebalancing effect and risk premiums that have been
well documented.
1. Rebalancing effect. The observation that market
participants are driven by the emotions of fear and
greed is widely accepted. Both are powerful forces
and to some degree all financial manias (both
bull and bear) can be explained through this lens.
The result is self-reinforcing feedback that moves
market prices to extremes in the short term followed
by a return to more normal conditions over the
medium term. Put another way, momentum moves
prices away from fair value in the short term
and prices mean-revert towards fair value in the
medium term.
Investment strategies that weight securities by their
market capitalisation (both passive and active) are
vulnerable to the risk generated by these effects.
Securities that have performed well become large
weights in the portfolio while the prospects for
future relative underperformance increase.
By using a different weighting methodology
(which is lowly correlated to the change in price)
a rebalance premium can be extracted from this
momentum/mean-reversion effect in the market.
In essence a buy low, sell high discipline is
systematically imposed on the strategy.
We believe that this effect is likely to persist over time
but it will not necessarily be beneficial over all time
periods or rebalance frequencies depending
on the market environment and transaction costs.
2. Capturing risk premiums. There is a lot of
academic research identifying return opportunities
from certain types of securities that are periodically
mispriced, for example, the mispricing of value stocks
allows investors to earn a value premium. These
premiums result from market participants being
less willing to own certain types of companies for
a number of reasons linked to behavioural biases
or principle-agent conflicts inherent in the business
models of some market participants. Therefore, those
investors that are able to own the securities of those
undesirable companies are typically able to earn a
premium for doing so.
As with any premium it is important to consider the
risks for which that premium is compensation and
to recognise that premiums can be both negative
and positive for long-term periods.
Source: Towers Watson
Using smart beta in portfolio construction

In addition to risk and return characteristics of
a specific smart beta, these strategies can also be
useful portfolio construction tools. From a portfolio
construction perspective, its role is often broadly
described by a combination of the following
three reasons.
1. Replacement for active management. Some
active managers use strategies that are largely smart
beta in nature, or are closet indexers to a style index
while charging active management fees. An example
of this would be an active manager that employs
simple value screens to construct a value strategy.
If there is no incremental alpha beyond the screening
then such a strategy is merely capturing the value
premium. Rather than paying active fees, this strategy
could be replicated at a lower cost using a smart
beta approach.
Figure 03. Reasons to use smart beta
Replace active
management
Governance
saving
2. Remove an unwanted bias or completing a

portfolio. Finding active managers that are able
to generate sustained outperformance is not easy.
Once such managers are identified and combined
together in a portfolio there may be residual
undesirable biases or risks. Smart beta provides
an approach to address this issue.
Consider a portfolio of active managers that
create concentrated portfolios based on their
best investment ideas. All things being equal,
such a portfolio might be expected to exhibit
a small company bias (there are many more
small companies than large ones) that might be
undesirable if the overall structure is measured
against a market capitalisation benchmark.
To resolve this, an investor could use a smart
beta strategy that has a bias to larger companies.
This offsets the unwanted bias at the portfolio level
while allowing the investor to retain control, without
constraining the selected managers.
3. Creating additional governance capacity.
Governance is the amount of investment decisions
an investor is able to effectively make. We believe an
investor is most likely to achieve its objectives when
its portfolio reflects its level of governance.
If smart beta reduces the governance requirement
of the equity portfolio, an investor can spend this
governance saving on new strategies that would
previously have exceeded its governance budget.
Portfolio
completion
Governance is consumed by active managers

because they repeatedly require close monitoring.
Therefore, switching from active management
to smart beta reduces a portfolios governance
requirement while maintaining its overall biases
and risk-premium exposures.
Conversely, a market capitalisation passive portfolio
has very low governance costs. Replacing such a
portfolio with smart beta will normally increase the
governance required because smart beta requires
some governance, albeit with a different focus to a
traditional active or passive portfolio.
18 towerswatson.com
While it brings new governance demands, if used effectively,

smart beta can greatly improve an investors ability to achieve
its objectives and so merits its place in the investors toolbox.
Decisions and responsibilities
Risk oversight and monitoring
An investor that uses smart beta assumes greater

responsibility for the investment process used in
its portfolio. The asset manager in a smart beta
approach is often implementing a set process
selected by the investor. This is a departure from
active management, where the active manager
is responsible for the approach followed in order
to deliver a particular objective, such as the
outperformance of a particular index.
Constructing a smart beta portfolio requires a level

of risk management comparable to the construction
an active manager portfolio.
By their nature, smart beta strategies can have

relatively high active risk, and investors must
therefore fully understand and buy into the rationale
of a smart beta strategy before using it. This is
because poor decision making (hiring and firing)
can erode the gains of an otherwise successful
strategy as easily in a smart beta strategy as
in an active strategy.
What to monitor depends on the smart beta

approach being followed. Examples include:
a. Market conditions: These may present a headwind
or tailwind for certain strategies. For example,
compressed valuation spreads may present a more
challenging environment for a value strategy.
b. Crowding: Many investors pursuing a common
strategy may create endogenous risk similar to the
event of 2007 when active quantitative investors
using very similar strategies suffered large losses.
c. Evolving biases: As markets evolve, the diversity
benefits of combining different smart betas in a
portfolio, or their fit with active managers, will vary
over time.
Conclusion
The smart beta concept has significantly expanded
the tools available to an investor, particularly within
equities. While it brings new governance demands,
if used effectively, smart beta can greatly improve
an investors ability to achieve its objectives and so
merits its place in the investors toolbox.
References
1 Global Investment Commitee: secular outloook 2013
assimilating thematic thinking, Towers Watson Limited.
Understanding emerging market equity

Emerging market equities
where are we now?
Structure of the emerging

markets and indices
While the emerging markets (EMs) theme seems

to come in and out of fashion every few years, the
thesis for taking advantage of the long term, yet
volatile growth that these markets are forecast to
deliver remains robust.
One of the first things for investors to understand

when they consider the EMs is the contrast between
the breadth of the markets and the concentration of
the MSCI emerging market (MSCI EM) index.
Nevertheless, many investors maintain a structural

underweight to this area both specifically in their
equity allocation and asset allocations, more
broadly. Although EMs represent almost 50% of
the worlds GDP, their share of world equity market
indices is only about 11%, and they also tend
to represent a substantially smaller portion of
most institutional equity portfolios. For example,
BlackRock indicated recently that the average
US investor only has a 5% allocation to EMs.
Anecdotally, we believe this low allocation may
be fairly representative of the wider institutional
investor base. We believe that investors should
be considering their current allocations to EMs.
20 towerswatson.com
Developed market investors often think of

EMs as a somewhat homogeneous group,
characterised (simplistically) by high growth and
higher volatility, but lower governance standards.
In reality, the EMs represent diverse countries,
spread across five continents with different
cultures, languages, political systems, regulatory
regimes, demographic profiles and stages of
development. When frontier markets (markets that
are not classified as developed or emerging) are
included, this universe becomes more diverse.
While there are many thousands of companies

listed in EMs, for reasons such as liquidity, the
MSCI EM index contains only about 800 stocks.
The China A-share market is also not represented
in the index, as it is not yet a freely traded market.
The MSCI EM index has some notable areas of
concentration in terms of stock, sector and country.
For example, the largest four countries in the MSCI
EM index represent about 56% of the index, despite
the index containing 21 countries. Furthermore,
the index is typically dominated by a relatively
small number of very large companies. Figures 01
and 03 illustrate the concentration in certain parts
of the market.
As the charts show, the energy, financial
and telecommunication sectors comprise a small
number of large companies that are, perhaps,
over-represented in the index. Conversely, there
are many industrial, consumer, and healthcare
companies that seem under-represented.
Figure 01. Sector concentration in emerging markets

0
20
40
Utilities
Telecommunication services
Information technology
Financials
Healthcare
Consumer staples
Consumer discretionary
Industrials
Materials
Energy
% of companies in universe MSCI EM weight
State Owned Enterprises (SOEs)

Many companies in emerging markets have
large controlling shareholders. These may be
families, entrepreneurs or state-controlled entities.
Companies that are partly owned by these state
entities are called State Owned Enterprises
(SOEs). Many of the large energy, financial and
telecommunication companies that dominate
the index tend to be SOEs. Figure 02 shows
the dominance of SOEs in emerging markets
compared to developed markets. Investors
question whether SOEs are always run for the
benefit of the minority shareholder or whether
they are the most efficient allocators of capital.
We can also see from Figure 03 that similar levels
of concentration exist at the country level.
Figure 02. Proportion of SOEs in market indices

30%
25%
20%
15%
10%
5%
0%
Emerging markets
Share of state companies in market capitalisation
Figure 03. Number of companies and index weight by country
Investing in emerging markets

Many institutional investors seek to invest in
emerging markets to access the expected growth
in these countries, particularly through increasing
consumption. However, as we have seen, EM
indices may give investors higher concentration
than they might expect in companies that are
not purely exposed to these drivers of EM growth.
For example, a large portion of the index comprises
energy/commodity companies and exporters, which
may be more reliant on global growth (rather than
EM growth) and SOEs which may not deliver the
desired level of shareholder return.
Accessing EM equity
The usual ways of accessing equity markets apply
similarly in emerging markets. There are investment
options for traditional bulk beta (passive market
capitalisation approaches), smart beta and alpha
(active managers). However, there are important
considerations for implementation which are
specific to emerging markets.
Passive market capitalisation approach

Market capitalisation weighted passive exposure is
now reasonably cheap and offers a high standard
of index tracking (although perhaps not as close
as in developed markets). Institutional passive
product fees have fallen in price, with a standard
fee for a US$50 million mandate now typically
below 0.15% per annum. These products offer
broad exposure to emerging markets and are, by
their nature, not capacity constrained. However, we
remain concerned about passive exchange traded
funds (ETFs) for emerging markets equities, as
these tend to have high fees (a leading EM ETF has
a total expense ratio of 0.67% per annum) and have
displayed fairly high tracking errors over time.
Developed markets
10
20
30
Morocco
Hungary
Czech Rep
Egypt
Peru
Philippines
Colombia
Poland
Turkey
Chile
Thailand
Indonesia
Malaysia
Mexico
Russia
India
South Africa
Taiwan
Brazil
Korea
China
% of companies by listing (excluding A-shares)

% of companies in universe
MSCI EM weight
22 towerswatson.com
40
Passive market capitalisation products represent

suitable implementation options for many clients
for reasons such as simplicity, low cost and broad
exposure. However, for reasons set out in this
article, Towers Watson believes that other
smart beta or active management approaches
may, in theory, be preferable.
Smart beta
The main objective of smart beta is to capture
a particular risk premium for a low cost.
These strategies are expected to provide
superior investment efficiency than bulk beta
(passive market capitalisation). Please see
Towers Watsons article Using smart beta in
equities1 for further information on this topic.
These smart betas include value-weighted
strategies, low volatility strategies and
diversification-based strategies. All require
careful consideration and have implications
for portfolio implementation. Fees for these
strategies are more expensive than their
developed market equivalents but, at around
0.30%, represent reasonable value in
emerging markets.
Active management
Active management is, in theory, very attractive
in emerging markets. It is often argued that EM
equity is a less efficient asset class, in part due
to the issues discussed above, but also because
sell-side analyst coverage has typically been lower
in emerging markets (although this is increasingly
less the case). Active managers are potentially
better able to assess the quality of company
governance and the actions of management,
which can be very important in emerging markets.
Active managers may also have freedom to
invest in emerging market stocks that are listed
on developed market stock exchanges. These
can include multi-national companies with high
emerging market presence/exposure. But more
importantly, genuine emerging market businesses
that have, for various reasons, chosen to list
elsewhere. Active managers are therefore able to
select those companies that are most exposed
to key emerging market drivers, which may not be
some of the large EM index stocks.
However, there are also some structural challenges
facing active management in emerging markets.
1. Large teams: Diverse markets, cultures and
languages often lead managers to hire more
staff to gain broader local knowledge. Bigger
teams can typically present more challenges to
communication and culture and can potentially
become bureaucratic.
2. Large asset bases: Large teams typically

require larger asset bases to support them.
Additionally, they are often found in large
businesses rather than specialist boutiques.
3. Capacity problems and reduced alignment:
Large firms with weaker alignments of interest
may be more motivated to grow their asset
base. Indeed we have observed that many of the
leading EM investment managers are now closed
to new business or have prohibitively large asset
bases. Figure 04 demonstrates the reduction
in opportunity set experienced when managers
grow their assets under management in excess
of US$20 to 30 billion. There are many managers
that are of this scale.
4. High fees: There is a relatively limited universe
of skilled and experienced emerging market
managers compared to US or global equities.
As more assets flow in to emerging markets
this limited supply has greater demand. As
such, standard fees are high. Screening the
eVestment Alliance database 2 global EM equity
managers suggest an average annual fee for a
US$50 million account of 0.92%. This compares
to 0.73% for the global equity universe. It is not
obvious that this higher fee is justified by higher
(expected) alpha.
5. Transaction costs tend to be higher:
This is another hurdle that active managers
need to overcome to deliver added value
net-of-fees to investors.
Figure 04. Percentage of universe stocks that are investable

Based on 1% position at various assets under management,
given the desire to hold <5% of a companys market capitalisation
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
0
10
00
00
00
,0
,0
,0
30
10
20
Assets under management (US$ millions)
0
00
1,
00
2,
00
5,
Frontier markets EM small capitalisation EM
Types of active managers

Active EM managers tend to fall into two broad
categories. The first are those that provide
broad market coverage across most sectors and
regions in the market. These products tend to
be larger and more scalable, and often come
with modest tracking errors. Skilled managers
can be found in this category although few are
expected to deliver attractive returns net of fees,
given the higher fees for the level of risk and
capacity issues. The second category comprises
managers that are less benchmark sensitive and
focus more on maximising long-term absolute
returns or accessing a particular theme (such as
domestic EM growth). These types of managers
will often eschew, for example, global commodity
companies, large index constituents or perhaps
SOEs. Their portfolios can therefore often
look very different to market indices and have
high tracking errors, and perhaps biases to
smaller capitalisation stocks or certain sectors
(for example consumer goods) relative to the
EM index.
As well as global EM managers, there are various
regional managers and frontier market managers.
Towers Watson has research coverage in this
area, but these investors rarely have the size of
allocation or governance to construct portfolios
containing many regional managers.
Frontier markets
In a similar manner to emerging markets, frontier
markets offer a broad and diverse opportunity
set that we believe can benefit from long-term
growth drivers. However, lack of liquidity, relatively
poor governance standards and earlier stages
of economic development make these markets
unattractive for many investors. We believe that
significant investor skill is required when investing
in these markets due to the above constraints and
so avoid passive allocations in these markets.
We believe that allowing skilled emerging market
managers the ability to invest in frontier markets
can give further potential for added value in the
long term.
24 towerswatson.com
How Towers Watson can help investors

Towers Watsons Asia and emerging market research teams
have conducted extensive research into these markets and
the investment managers that operate in them. We believe
that the emerging markets offer exciting long-term investment
opportunities, but implementing investments in this region is
difficult. More detailed understanding and awareness of the key
issues is required. Despite the challenges, we have identified
a select group of active managers that we consider to have
attractive value propositions net of fees. These managers,
together with passive and smart beta offerings, may provide
investors various ways to exploit the attractive structural
opportunity that we believe emerging market equities present.
References
1 Using smart beta in equities. Towers Watson Limited, 2013.
2 As at Q2 2013.
Low volatility equity strategies

Should you include them in your portfolio?
Investors often ask us about low volatility equity

strategies. That is hardly surprising, as these strategies
target market returns but with much lower risk.
While no two low volatility strategies are

identical, most share many traits. For example,
they often target similar anomalies in the equity
market. Here, we ask: Do low volatility equity
strategies improve an investors portfolio?
We show that they can in some cases. Even
then, these strategies still raise concerns that
all prospective investors should resolve before
proceeding further. We discuss these concerns,
before stating the attributes of these strategies
that we tend to seek and avoid.
What is a low volatility

equity strategy?
Many equity managers now run traditional
active products and low volatility products.
In both cases, they have the same investment
philosophy they just apply it differently in
each case.
With traditional mandates, managers aim to
outperform a benchmark, while bearing a similar
absolute level of risk. Yet in most low volatility
mandates, managers aim to generate a return
like the benchmark but with, say, 30% less risk.
This risk/return profile for low volatility products
resonates with many investors who feel unable
to bear large losses. 1
While products with this risk/return profile differ

to some extent, they resemble each other in
the main ways. After all, they tend to exploit
the same perceived systematic mispricings of
stocks. One difference, however, is cosmetic.
Managers use different names, often for the
same thing. For that reason, the terms low
beta, low volatility or even better indices tend
to feature, not, small capitalisation value, which
often reflects the type of stocks being chosen. 2
(For the sake of brevity, we will refer to them
all as low volatility strategies in the rest of
this article.)
As for similarities, most low volatility strategies
target an investment in stocks with low return
volatility. They do so because, counter-intuitively,
low volatility stocks have historically tended to
outperform high volatility stocks on a risk-adjusted
basis, and particularly so in some recent years.
For that reason, many managers run portfolios
that are overweight low volatility stocks and
underweight high volatility stocks. We now
consider why this counterintuitive performance
occurred, and whether it might recur.
What is the investment rationale

for low volatility equity?
As with all systematic active management, the
rationale is that investment markets are inefficient
and include anomalies that investors can exploit for
better risk-adjusted returns. The longer explanation
requires some historical context. Clearly, when
combined together, all investors own the mix
of every conceivable asset. Early investment
academics called this mix the market portfolio.3
As a result, most academics thought that the
optimal investment strategy would be to invest
in some mix of cash and this market portfolio.
This view spawned the use of index funds and
diversification across many different assets.
More recently, though, many academics have
come to see this market portfolio as inefficient.
They reason that the earlier assumptions, which
resulted in the optimality of the market portfolio,
were unrealistic. These earlier assumptions,
for example, held that information is without cost
to obtain and that investors can always borrow
cash without difficulty.
With more realistic assumptions, academics have
shown that patient investors can outperform the
market portfolio, in risk-adjusted terms, by holding
mixes of assets in non-market weights.4 In other
words, investors can improve their outcomes by
exploiting anomalies.5
Excess return of stock
Figure 01. A stocks excess return and its

beta in theory and in practice
As derived from
classical theory
As observed
empirically
Beta of stock
26 towerswatson.com
Although not accepted by every academic, theories

abound on the evidence and reasoning for different
anomalies. These theories often stem from biases
in investor behaviour or from constraints faced
by rational agents.6 We now present one such
anomaly the low volatility or low beta anomaly
framed in both of these ways:
People usually prefer to gamble on long-shots
than on favourites. Studies show the impact
of this bias in different aspects of our lives.
In horse racing, data from many decades and
countries shows that the average return on a
long-shot is dismal, while that on a favourite is
only poor.7 In finance, classical theory calculates
a stocks expected excess return relative to
cash (see Figure 01). This outperformance
relates to the stocks beta, which depends upon
the stocks volatility and the correlation of its
returns with those of the market. Specifically,
classical theory expects high beta stocks to
outperform low beta stocks. Yet, in keeping with
our general preference for long-shots, regardless
of their fundamental value, finance studies
show that high-beta stocks underperform
the return expected by classical theory, while
low-beta stocks outperform those classical
expectations. Assuming that these results recur
because human behaviour doesnt change,
investors choose to overweight low-beta stocks
and underweight high-beta stocks in order to
outperform on a risk-adjusted basis.
The constraints that investors face force them
to act as if they seem irrational. Other parties
take a different view, even though they use the
same evidence and build similar portfolios.
They argue that investors face boundaries on
how they can act, either through rules or
conventions. To circumvent these boundaries,
the theory goes, investors generally alter their
behaviour in a way that would appear irrational
without a prior understanding of the boundaries
that they face. One such example comes from
Baker et al (2010). 8 They begin with the classical
expectations outlined above, based on the
portfolios beta. They then change these
calculations to reflect a real-world scenario.
Specifically, they assume that markets are
rational and that some investors manage their
own money and seek to find an optimal portfolio
in the classical way. Yet they also assume the
presence of another type of investor, which hires
an agent (an investment manager) to run its
money relative to a fixed benchmark. As a
rational agent, this investment manager is
assumed to maximise the information ratio for
its investors portfolio a different goal to the
other type of investor. What becomes clear
from the study is that the interplay between
the different goals for these two types of

investor explains much of the historical low
volatility bias. Furthermore, these aspects of
the real-world scenario are unlikely to change
anytime soon, meaning that the low volatility
bias may well persist, absent a massive flow
of assets to try and capture it.9
We tend to agree with the academic consensus
about anomalies. On paper, it seems that very
patient investors can improve their risk-adjusted
returns by taking advantage of these anomalies.
Experience suggests that doing so in practice is
tougher. For that reason, we now discuss some
of the implementation challenges.
Challenges that investors face when

trying to exploit an anomaly
At best, some of the following challenges can
dilute an investors outperformance. At worst, they
can lead to a major loss of wealth. Prospective
investors should therefore satisfy themselves
that they can overcome these obstacles before
proceeding further.
Popularity tends to worsen a strategys future
prospects. Collectively, investors often prefer
a certain investment approach, given its recent
performance. That can be dangerous, as too
much money may be chasing too narrow a part
of the market. In August 2007, for example,
many quantitative investment strategies
suffered when large hedge fund investors
scrambled for liquidity. This caused a sharp fall
in the prices of some of the large value stocks
that were popular in quantitative strategies.
In turn, this led to huge losses for some
leveraged funds in this area. 10 As it happens,
most low volatility strategies have performed
very well in the last few years and have
attracted assets, which we consider potentially
troublesome. For more detail on the recent
performance and valuation of these low
volatility portfolios, see Low volatility equity:
from smart beta to smart execution. 11
Much of an anomalys excess return
occurs at times that make its investors
most uncomfortable. Consider the small
capitalisation anomaly, for example. The
smaller the stock, the tougher it is to sell.
This illiquidity makes most investors uneasy,
particularly in difficult times, making the
timing of hiring and firing managers in this
area particularly important. Unsurprisingly,
the small capitalisation anomaly seems to
provide its greatest risk-adjusted returns for
the smallest capitalisation stocks in times of
greatest market stress. In order to make their
products seem more palatable or investible to
investors, managers often trim the investment
universe. In doing so, they unduly dilute the

likely payoff from exploiting the anomaly. Worse
still, many investors cannot withstand protracted
underperformance and so may change strategy
in these difficult times, just before an anomalys
return potential is realised.
Does such a strategy merit inclusion

in an investors portfolio?
From now on, we assume that investors
believe that:
Systematic anomalies can be exploited to
enhance risk-adjusted returns12
They do not want to re-engineer their
entire portfolio
They have satisfied themselves that they
can overcome the challenges that we outlined
in the previous section
Whether low volatility strategies should form part
of the investors portfolio then depends on its
existing arrangements:
A low volatility strategy may make sense as
a replacement manager for investors with a
traditional active equity mandate. After all, low
volatility strategies can provide better value for
money than many traditional equity mandates
in accessing the same equity anomalies. Some
strategies go further still and aim to capture
many different anomalies within the same
portfolio. This approach is reasonable, as long
as it remains inexpensive. However, we know
of some investment managers that market
multi-anomaly portfolios as if they added true
alpha, instead of exposing their investors to
factors that could be garnered simply and less
expensively. When products are marketed in this
way, we advise investors to place even more
effort than usual on due diligence.
No matter how different these products purport
to be, many low volatility or low beta strategies
look statistically similar. (For example, low
volatility and fundamental indexation.) As such,
combining two of these strategies in a portfolio
would improve risk-adjusted return less than by
hiring a different type of skilled active manager.
If investors were going to spend their resources
to hire another manager, they would probably be
better to do it elsewhere.
Collectively,
investors often prefer a

certain investment approach, given
its recent performance. That can be
dangerous, as too much money may be
chasing too narrow a part of the market.
Of course, once investors have opted for a low

volatility strategy, they must find the specific type
of strategy and a suitable manager. As we cannot
comment on every manager in one article, we now
consider the attributes of low volatility strategies
that we generally seek and avoid. As with other
strategies, we generally prefer products with
lower costs and fees, wherever possible.
But what should investors avoid?
Investors often measure success too narrowly,
using only the average return and variability
of returns. Most tests of anomalies measure
the average and variability of the returns they
generated. This approach is unfortunate, as
these tests ignore the fat tailed outcomes that
occur in stressed markets, which can cause
investors to exit a strategy with a major loss.
Worse still, empirical studies show that these
strategies generate more very poor outcomes
than their respective market capitalisation
benchmark. 13 This combination of effects could
harm uninformed investors. After all, they
could be attracted to low volatility mandates,
given their association with lower risk, even
though extreme downside events may be more
likely than with a pure market capitalisation
approach. As is the case in other mandates,
investors in low volatility strategies should be
able to endure prolonged periods of painful
performance. Otherwise, they will not be able to
remain invested in the toughest times and so
benefit from the improvement in returns that the
anomaly generates.
Low volatility products expand the

options available to equity investors.
Even better, some of them give
investors cheap access to important
investment anomalies.
What attributes of low volatility

products should investors avoid?
While there are some clear exceptions to the
following preferences, we generally advise
investors to avoid products that:
Have been made too investible. We noted earlier
that most of an anomalys excess return seems
to arise from the aspects that make its investors
uncomfortable. As such, any strategy that only
focuses on comfortable stocks may not provide
enough excess return to justify its fees over
the longer term. We therefore tend to prefer
products that do not unduly exclude the smallest
stocks from their investible universe.
Are highly dependent on an opaque model.
Sadly, not every investment manager acts in
the best interests of its investors. Of course,
there is no link between managers with opaque
models and managers that are nefarious.
However, managers with opaque models
could act improperly and their investors might
never know until it is too late. Lo (2001) 14, for
example, designs a simple but ultimately evil
strategy that could be marketed as a proprietary
opaque model. It is designed to perform well in
the medium term, accruing assets and income
for the manager, before blowing-up in the long
term, to the great detriment of the investor.
Rely heavily on an optimiser. The traditional
concern with optimisers is that they maximise
the impact of errors in the managers forecasts
of expected return. 15 However, they can also
generate a portfolio with an undue, and often
unrecognised focus on a particular factor.
As we noted before, such a narrow focus is
unwelcome in these strategies.
Focus on too narrow a subset of the market.
We base this view on our earlier concern that
low volatility products can become fragile and
prone to asset flows in times of market stress.
Conclusion
Low volatility products expand the options
available to equity investors. Even better,
some of them give investors cheap access
to important investment anomalies. Even
then, these strategies raise concerns that
all prospective investors should resolve
before proceeding further, such as the
focus on too narrow a subset of the market
by strategies that are increasingly popular.
28 towerswatson.com
References
Further reading
The preference for this mandate raises some broader issues, including
some about investment strategy. For example, should loss-averse investors
just allocate less to equity? Would the equity allocation change dramatically
with a low volatility mandate rather than a traditional quantitative equity
mandate? A key consideration is whether the low volatility product
produces a sizeable fall in the volatility of the total portfolio. After all,
reducing the volatility of a certain product in a portfolio may not reduce the
volatility of the portfolio itself by much. However, reducing the expected
return of a product within a portfolio will certainly lower the portfolios
expected return. For that reason, an investor may be able to achieve a
better overall risk-adjusted return by hiring a traditional equity mandate
that targets the same anomalies. The answer to this question will differ by
investor, given their existing allocation. However, investors should consider
these questions before proceeding with such a mandate.
Brav, A. and Heaton, J B. Competing theories of nancial anomalies.

The Review of Financial Studies (2002): 575-606.
While the index of small capitalisation value stocks has a high beta
relative to the broad market index, the same is not true for all of its
constituent stocks. As it happens, it is these lower beta stocks that
often appear in low volatility products.
Schleifer, A. and Vishny, R W. The limits of arbitrage. The Journal of

Finance LII, no. 1 (1997).
Sharpe, W. Capital asset prices: A theory of market equilibrium under

conditions of risk. Journal of Finance 19 (1964): 425-442.
A broader question also arises at this point. That is, why focus only on
cheap equity anomalies? If you wish to be consistent, shouldnt you
be considering as many anomalies as possible in your portfolio? After
all, many of these anomalies, like the illiquidity premium, may present
themselves across multiple assets. For that reason, some expert
investment boards now base their investment strategy not on asset
allocation but on an allocation of factor exposures (that they implement
by investing in certain subsets of an asset class). Ang et al (2009)
describe such an approach in great detail. Other investors use risk parity
funds to achieve a similar aim, although doing so raises more questions,
such as the potential fragility of a risk parity portfolio in the face of
steepening yield curves.
In equities, for example, see Fama and French (1992) for details of the
value-growth and large-small anomalies. (Note that, under specific
assumptions, Fernholz and Karatzas (2006) disagree and view the small
capitalisation and illiquidity premium as identical.) Carhartt (1997) then
considers an additional anomaly, called momentum, while Pastor and
Stambaugh (2003) reflect on an anomaly for illiquidity.
Brav and Heaton (2002) skilfully show the uncanny similarity between
these two types of theory.
Snowberg and Wolfers (2010) reflect this consistent bias in more detail
and conclude that it likely results from behavioural biases.
Baker, M. Bradley, B. and Wurgler, J. Benchmarks as limits to arbitrage:

Understanding the low volatility anomaly. Financial Analysts Journal
(2011): 40-54.
This form of argument is based on the work of Shleifer and Vishny (1997),
concerning the limits of arbitrage.
Carhart, M. On persistence in mutual fund performance. Journal of

Finance 52 (1997): 57-82.
Fama, E. and French, K. The cross-section of expected stock returns.
Journal of Finance 46 (1992): 427-466.
Fernholz, R. and Karatzas, I. The implied liquidity premium for equities.
Annals of Finance, January 2006.
Pastor, L. and Stambaugh, R. Liquidity risk and expected stock returns.
Journal of Political Economy 111 (2003): 642-685.
Ross, S. The arbitrage theory of capital asset pricing. Journal of Economic
Theory 13 (1976): 341-360.
Snowberg, E. and Wolfers, J. Explaining the favorite-long shot bias:

Is it risk-love or misperceptions? Journal of Political Economy (2010): 723-746.
10 See Lo, A W. Interview by House Committee of Oversight and Government

Reform. Hedge funds, systemic risk, and the financial crisis of 2007-08
(13 November 2008), as an excellent account of this period of market turmoil.
11 Low volatility equity: from a smart idea to smart execution.
12 In this article, we make the distinction between systematic and
idiosyncratic anomalies and focus on the former. For example, an investor
may have good reason to expect value stocks to continue to outperform
growth stocks on a risk-adjusted basis. That view refers to a perceived
systematic anomaly. However, the same investor may also have a view on
the diverging prospects of two supermarkets, which she does not believe
is in the two prevailing stock prices. That latter view refers to a perceived
idiosyncratic anomaly.
13 Ang, A. Goetzmann, W N. and Schaefer, S M. Evaluation of active
management of the Norwegian Government Pension Fund Global. 2009.
14 Lo, A W. Risk management for hedge funds: Introduction and overview.
Financial Analysts Journal, November/December 2001.
15 See Finding the right asset allocation: Is optimisation the solution?
Towers Watson Limited, 2012, for further details on why optimisation
is often not sensible.
Section two
Manager selection
30 towerswatson.com
In this section, we discuss our framework

for assessing investment skill, which we
apply consistently across our team of
over 50 equity manager researchers in
15 locations. We base our approach on the
belief that a qualitative, evidence-based
assessment of investment skill outperforms
approaches based on proxies for skill,
such as past investment performance or
the existence of a recognised brand. We
discuss several forms of equity portfolio
management, their characteristics, when
they work, when they do not, and the
importance of long-term thinking.
02 Manager selection
Assessing investment skill

in equity managers
Assessing investment skill in

equity managers
At the heart of any decision to employ an active
equity fund manager is a belief that the portfolio
will produce favourable investment returns above
some benchmark. This is commonly known as
alpha. At best, the judgment of a managers
ability to produce alpha will be based on a direct
assessment of his skill in selecting investments
and his ability to put these investments together
in a portfolio that efficiently balances return
relative to risk.
However, there are significant challenges to
forming judgments of investment skill that lead
many to adopt more easily sourced but spurious
surrogates, such as short-term investment
32 towerswatson.com
performance or a recognised brand, to guide them

when forming a view of investment skill. Such
approaches can be typical of how retail investors
select investment products but should have no
place in the professional market.
In this article we have set out our beliefs on how
investment manager skill should be assessed
both prior to making any investment and, equally
importantly, as part of an ongoing monitoring
process once an investment has been made.
Although a number of the examples in this article
are taken from the equity sphere, our comments
apply across all asset classes and should provide
the reader with a framework for assessing both
their own investment decisions and the robustness
of the advice that they receive.
The perils of past performance

Past performance in particular can sound a
plausible basis upon which to form an opinion.
This is because we are programmed to recognise
patterns in nature and to extrapolate what we
believe we have observed. However, studies
have shown that there is a high degree of
randomness in relative investment returns and
that to be statistically significant, a performance
record should be intact for nearly 15 years. 1, 2
Few investors meet this criterion. Fewer still
meet this requirement and have not experienced
other changes which have a direct impact on
future performance, such as staff turnover or
growth in assets under management which can
affect portfolio construction. Consequently, we
strongly believe that considered in isolation
past performance is a poor basis for assessing
investment skill.
Similarly, it is human nature to want to avoid
standing out in a crowd. Investors take comfort
from the fact that others have invested with the
same manager. As Keynes said it is better for
reputation to fail conventionally than to succeed
unconventionally.3 This behavioural bias often
causes investors to assign value to a recognised
brand that is not based upon the skill of the fund
manager or the likelihood of achieving alpha. In
fact for a variety of reasons the opposite is very
often the case.
Selecting a manager is made more difficult
because it is possible that a skilled manager
will generate poor returns for an uncomfortably
long time before his skill again manifests itself.
This is true even of great investors such as
Warren Buffett who, while amassing his
spectacularly successful 45-year track record of
compound annual outperformance of 10.86%,
still recorded two rolling three-year periods where
his relative performance was negative. It is also
probable that an unskilled manager will produce
positive returns for multi-year periods. Although
survivorship bias means that many of these
investment failures have been dropped from the
databases, there are still many that we could
cite as being value destructive over the long term
whilst demonstrating rolling three-year periods of
positive relative performance. The challenge for
investors and advisors is to be able to differentiate
luck from skill.4, 5
Figure 01. Excess returns are random in aggregate
-15%
-10%
-5%
0%
5%
10%
15%
This chart presents the expected returns (shaded area) assuming returns are random and normally
distributed, and the distribution of actual returns achieved by all large capitalisation US equity portfolio
managers (the red line) over the past three years (net of fees at an estimated 0.5%). As can be seen, the
realised excess returns have a random distribution with a mean return just above 0%.
Source: eVestment Alliance, Towers Watson
Figure 02. Probabilities of skilled and unskilled investors

performance over three years
100%
75%
50%
25%
0%
Skilled
Unskilled
Positive excess returns Negative excess returns

This chart shows the probability of a skilled manager (defined as having a net information ratio
of 0.5) and an unskilled manager (defined as having a net information ratio of -0.1) outperforming the
market over a three-year period. While the skilled manager has a higher probability of producing positive
excess returns it still has a 19% probability of underperforming over a three-year period. Similarly,
although not expected to outperform, the unskilled manager has a 43% probability of outperforming
in a three-year period.
...we
strongly believe that, considered in isolation, past

performance is a poor basis for assessing investment skill.
The process of identifying true skill

Explicit determination of investment skill in an
equity fund manager is complicated. Differences
in investment philosophy, process and execution
mean that the analytical framework applied
to each manager and product must have a
degree of flexibility. Formulaic approaches to
assessing skill do not have the sophistication
needed to recognise and differentiate the
subtle differences in philosophy, process
and execution between managers.
A focus on the long term is particularly important,
as some asset managers appear to rely on the
randomness of returns in markets to produce a
positive short-term track record, which is then
used to market products to advisors and clients
who are attracted by short-term performance.
As John R Minahan stated while most
managers have an investment philosophy
statement, in my experience these statements
are more often marketing slogans or product
positioning statements than they are thoughtful
encapsulations of the investment insights an
investment process is designed to exploit. Leading
to the view that many investment managers are
alpha-pretenders. That is, generating positive
ex-ante alpha is a secondary consideration for
many investment management firms.6
Figure 03. Excess returns decline as assets grow

30%
20%
10%
0%
- 10%
- 20%
- 30%
March
1970
March
1975
March
1980
March
1985
March
1990
March
1995
March
2000
March
2005
March
2010

The scalability of the asset management business model provides financial incentive to grow assets under
management beyond the point at which diminishing marginal returns to investors set in. This is a graph of the
excess returns generated by one of the largest US mutual funds over its 40-year history based on publicly
available information. The size of the portfolio peaked at more than US$100 billion in early 2000. Since 2000
the portfolio has more than halved in size and excess returns have begun to expand.
The role of investment philosophy

To form a view of the investment skill of a
manager it is important to build a comprehensive
understanding of the three key areas mentioned
above. This typically starts with a discussion
of the managers beliefs about what creates
investment opportunities, how a process can be
constructed that enables the manager to capture
such opportunities sustainably, what competitive
advantage the manager may have and how this
process can be expected to evolve.
An investment philosophy is important because
we know nothing with certainty. In particular, no
one knows what a business will earn or what
equity risk premium should be applied in the
future. Consequently, investment managers need
a relatively constant framework of beliefs to guide
them in the formulation and execution of their
process. To confirm the underlying philosophy and
form a view of its robustness, direct questions
may be asked regarding the creation of alpha
opportunities, competitive advantages, and so on.
Perhaps surprisingly, many managers are unable to
address these fundamental issues convincingly.
34 towerswatson.com
A focus on the long term is particularly

important, as some asset managers appear
to rely on the randomness of returns in
markets to produce a positive short-term
track record, which is then used to market
products to advisors and clients who are
attracted by short-term performance.
Translating a philosophy into a portfolio

From as early a stage as possible, it is important
to identify evidence that there is logic and
consistency between how the investment
manager believes alpha opportunities arise
and the process through which these are
identified and the portfolio constructed.
The defined process will also form the foundation
for critical assessment of investment skill.
This assessment involves discussions with the
portfolio manager, and members of the wider
analytical team to determine what criteria are
considered and how they are applied when making
investment decisions. The level of detail must be
sufficient to enable the researcher to understand
the steps the manager/analyst goes through from
idea conception and proof of thesis, to deciding
how to weight the investment in the portfolio, and
the determination of possible sell triggers.
The importance of constructing a comprehensive
understanding of the process cannot be
overemphasised. Vague descriptions are
inadequate for forming high conviction
evidence-based judgments. For example,
one process description we have encountered
stated the process as being investing in
companies with sustainable and high return
on invested capital and high free cash flow.
In this case it would be necessary to know
what is meant by high and sustainable.
An understanding of how the manager calculates

free cash flow is also necessary if the researcher
is to verify that the manager does what he says
he does. Once the terms have been defined it
then becomes possible to screen the portfolio for
what at first sight appear exceptions to the stated
process, challenge the manager more effectively
on his investments and gather the evidence
needed to form high conviction views on depth
of investment analysis and knowledge.
Analysis of the process identifies the primary
sources of expected alpha generation. In some
products this expected alpha generation may
reside with a single individual who has complete
responsibility for investment analysis and portfolio
construction. Alternatively, other products may
involve inputs from many different individuals.
For example, one organisation has a team of
three global equity co-portfolio managers based
in different parts of the world. They source their
ideas from the top two quintiles of investment
recommendations as ranked by their team of
60 equity analysts, who are themselves scattered
across the globe. The majority of products will
fall somewhere between these two extremes.
Processes that rely on broadly-sourced alpha
are often more complicated to assess and
consequently it can be more difficult to build
similar levels of conviction when compared to
less complex processes.
Assessing whether the portfolio

will deliver the expected alpha
After reviewing the investment process and
cross-referencing this with the philosophy, the
primary sources of expected recurring alpha
generation can be confirmed. Once they have
been identified, a decision has to be taken as
to whether it is possible to assess the level
of skill in the alpha critical functions.
In some instances it may be determined that due
to limited disclosure from the manager (or for
some other reason), it will not be possible to form
a high-conviction view of the investment skill and
further research may be abandoned. However,
in the vast majority of cases, expected alpha
generation will be based on the exploitation of a
belief through a process for which a framework can
be established to assess the level of skill.
In the case of equity managers this is usually
relatively straightforward and a range of factors
that are critical to the successful execution of
the process can be identified and measured.
The analysis typically starts with a detailed
examination of the investments within the portfolio
for any obvious inconsistencies with the process.
If there are investments that appear to contradict
the process then these may be prioritised
for discussion. Examining such exceptions
deepens our understanding of the process and
the manager. Alternatively, poorly performing

investments may be selected on the basis that we
are likely to learn more from a managers mistakes
than from his successes. Meetings with the
appropriate individuals (portfolio manager, analyst,
macro specialist, and so on) can then be arranged
to discuss the investments identified.
Other asset classes can entail a somewhat
different approach (indeed not all equity managers
fit neatly into the framework set out above).
However, regardless of asset class or approach,
all require an analysis of the portfolio to confirm
consistency with the stated process together
with an assessment of whether there are skilled
investment professionals who are able to create
and evolve a methodology to gain and maintain
an advantage.
Often, our research meetings will take the form
of a conversation regarding the rationale for a
particular investment. This provides the portfolio
manager or analyst with a blank canvas on which
he can illustrate how he reached a decision. This
then creates an opportunity to explore the key
arguments and to discuss how the rationale has
changed through time. A discussion of the risks to
the thesis may provide additional insights. Asking
some factual questions to which the answer is
already known can help to keep the conversation
honest. Often there are particular aspects to a
process which are worth exploring in more detail.
Figure 04. Complex versus simple process structure

Simple structure
(few alpha sources)
Complex structure
(many alpha sources)
Analysts
Portfolio managers
Products
36 towerswatson.com
...we are likely to learn more from a

managers mistakes than his successes.
The critical point from a research perspective
is to have enough knowledge of the underlying
investment to be able to follow up on open
questions with second, third and fourth
questions to expose more of the basis
upon which the managers judgments
have been made.
It is important to recognise that, at no point
in these interviews, do we form a judgment
on the attractiveness of the underlying
investment. Our sole objective in discussing
investments is to gather evidence to support
an opinion of the managers investment skill
and ability to generate alpha. We are not
attempting to second-guess the managers
decisions, but rather to test the process
and the quality of its execution. That is, we
do not meet the manager to understand the
portfolio, we analyse the portfolio to better
understand the manager.
It is also good practice to ask for the managers
internal research on the investments discussed.
This provides another point of reference to
confirm the veracity of the stated process and
to assess the extent to which an unsatisfactory
investment discussion may be down to poor
communication skills.
We are not attempting to

second-guess the managers
decisions, but rather to test
the process and the quality
of its execution.
Case study 1
We discussed a managers investment in Kingfisher, a UK-based home
improvement retailer. As part of our discussion regarding the investment
rationale, the manager suggested that he expected operating profit
margins to reach 12%. Through research undertaken prior to the meeting
we were aware that in the previous year margins were 6.5% and that
during the past 20 years they had never been above 8.8%. Knowing
this information enabled us to recognise that the forecast for operating
performance was aggressive and led us to drill into the basis for this
forecast to build a clearer picture of the managers depth of analysis
and objectivity. This exposed weaknesses in the application of the
process and behavioural flaws.
Case study 2
We discussed a small UK-based technology company, CSR, with a
manager. During this interview the manager stated that his competitive
advantage in assessing this company was largely based on the markets
preoccupation with short-term earnings momentum when valuing IT
hardware stocks and his discipline in exploiting the mispricing that could
occur as a result. He was clear that he had no better understanding
of the technologies applied by this company than the market. He
comprehensively illustrated both the case for investment in the business
and the case for the shares to underperform thereby demonstrating
balance and emotional detachment from the investment decision,
something we believe helps to avoid behavioural errors.
Implications for manager

research businesses
For a detailed discussion on individual investments
to take place it is necessary to have some current
knowledge of the investment. Annual reports
and other regulatory filings are readily available
and provide excellent research material. There
is also a wealth of accounting and news data
available through sources such as Bloomberg or
Reuters. Undoubtedly the greatest challenge to
the manager researcher is setting aside the time
needed to prepare for these discussions. Thorough
preparation for a discussion with a manager, which
results in questions targeted to produce useful
insights and evidence of skill and behavioural
characteristics, can take many hours. However,
it is this that separates those organisations who
prioritise the quality of their manager research
from those for whom manager research is just
an ancillary service to support a broader
consulting relationship.
Among the many parallels that exist between

investment research and investment manager
research is the fact that both the portfolio manager
and the manager researcher are making judgments
based on incomplete and imperfect information.
Neither knows anything with certainty. In order
to improve the probability of forming a correct
judgment a good manager researcher or asset
manager will look for opportunities to gain additional
insight. In the case of the manager researcher this
often involves observing aspects of the process
in action through sitting in on internal meetings
or even external research meetings between
the manager and the management of existing or
prospective investments. Such opportunities can be
invaluable. But again, to be effective this requires
the manager researcher taking responsibility for
acquiring enough knowledge of the investment
managers process and investments to appreciate
the often subtle distinctions between a strategy that
can credibly sustain alpha generation from one that
is destined to fail.
Figure 05. Cumulative outperformance of the Towers Watson Global Equity Model Portfolio
45%
40%
35%
30%
25%
20%
15%
10%
2013
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
0%
2000
5%
This chart shows the cumulative excess returns achieved by the Towers Watson Global Equity Model Portfolio since its inception net of all fees and expenses.
In
order to improve the probability of forming a
correct judgment a good manager researcher or
asset manager will look for opportunities to gain
additional insight.
38 towerswatson.com
Explicit assessment of the investment skill of

an asset manager is complex and challenging.
It involves confirming the validity of the managers
investment philosophy and defining in detail
the process by which the manager seeks to
capitalise on alpha opportunities. The primary
sources of alpha generation must be understood
and a comprehensive framework for assessing
skill in these areas established. The researcher
must have the ability to determine some of the
investment drivers and key issues for reaching
an investment decision on selected investments
based on publicly available information. Time must
be made available to do the necessary preparation
prior to meetings and evidence of skill, behavioural
weaknesses and other issues must be interpreted
and documented. The challenges in doing this
represent a material cost to the manager research
organisation and are no guarantee of success.
However, they may provide the edge that prevents
forming the wrong judgment about whether an
asset manager is an alpha generator or more
likely, an alpha pretender.7
Explicit
assessment of the investment

skill of an asset manager is complex
and challenging.
References
1 Stewart, S D. Neumann, J J. Knittel, C R. and Heilser, J. Absence of
Value: An Analysis of Investment Allocation Decisions by Institutional Plan
Sponsors. Financial Analysts Journal, Vol. 65, No. 6 (2009): 34-51.
2 Based on a hypothesis test (one tail t distribution analysis at 95%
confidence and IR=0.5) , one would need more than 13 years of yearly
observations. t stat = (X 0)/(X/0.5/sqrt(n)) should be bigger than
t(0.05, n-1) to accept the hypothesis that annual outperformance
X is positive.
3 Keynes, J M. The General Theory of Employment
Interest and Money, 1936.
4
Replacing managers for performance reasons.
5 Goyal, A. and Wahal, S. The Selection and Termination of Investment
Management Firms by Plan Sponsors, 2005.
6 Minahan, J R. The Role of Investment Philosophy in Evaluating
Investment Managers. The Journal of Investing, Summer 2006.
7 The Search for Alpha. Watson Wyatt Limited, January 2007.
Do not hire managers for past performance

More evidence of why this approach can lose value
When it comes to investment managers, the
natural assumption to make is that those who
outperform over the long term are skilled.
Our reasons for saying that

most managers with very strong
performance are unskilled
That may be the natural assumption to make, but

it is wrong. Here, we show that the vast majority
of managers with very strong past performance
lack the skill to outperform reliably in the
future. In fact, only about 20% of the managers
with very strong past performance are skilled
investors. This finding raises the question: why
do so many investors rely on past performance,
when it contains so little information?
To show how we arrived at this view, consider

the example of US equity managers. 2 By most
accounts, there are thousands of different US
equity portfolios being managed today.3 (We have
3,819 US equity products in our database alone.)
To keep the maths simple, we will assume that
there are 4,000 of these products.
We will give more evidence of how value can be

lost when investment managers are hired for
their past performance. 1 Specifically, we observe
that most investment managers with good past
performance lack the skill to outperform reliably
in the future. We also consider the behavioural
biases that tend to make investors assume that
outperformance is a more likely indicator of skill
than it actually is.
40 towerswatson.com
Furthermore, in a recent study, Barras et al

(2010) 4 estimate the proportion of managers
that add value, destroy value and perform like
the benchmark over the long term. We call
these managers skilled, bad and mediocre,
respectively. Using the proportions stated in this
study, we assign relative weightings of 10%, 20%
and 70% to these three manager types.
Knowing the descriptions of these three types

of managers, we then make some assumptions
about how they should perform in the future. First,
we assume that all manager types have excess
returns that follow a standard distribution, with
a tracking error of 5% a year.5 We also expect
that the skilled, mediocre and bad managers
will generate annual excess returns of 2%, 0%
and 2%, respectively, if held indefinitely. With
this information, we can calculate the probability
that these manager types generate good
performance over a three-year period where
good performance is an excess return above 2%
a year.6 We depict these probabilities for skilled,
mediocre and bad managers in Figure 01.
As can be seen, skilled managers have a clear
advantage. They are almost twice as likely
to generate good performance as mediocre
managers, and four times as likely to do so as bad
managers. For that reason, we can understand why
some investors might initially think that most good
performance should come from skilled managers.
Unfortunately, though, these investors are making
a common error: they are not accounting for the
relative prevalence of the three types of manager.
Rather, they implicitly assume that all three types
of manager are as common as each other.
As we showed before, however, this is not the
case. The relative prevalence of skilled, mediocre
and bad managers is 10%, 70% and 20%,
respectively. We can therefore combine these
values with those from Figure 01 to determine
the expected number of managers with good past
performance. We depict these expectations in
Figure 02.
Of those managers with good performance (in
plum), 190 are skilled, 634 are mediocre and 62
are bad. Despite skilled managers being far more
likely to outperform than any other manager, they
represent only 190 (or 21%) of the 886 managers
with good past performance. Put another
way, if you only ever hire managers with good
past performance, only 21% of your candidate
managers will be skilled. Worse still, you would
also overlook the majority of skilled managers.
(After all, 210 of the 400 skilled managers would
not have good past performance.)
At this point, you may be thinking that much of this
discussion is academic. You may think All I need
to do is to ensure that I dont hire managers for
performance reasons.
If only it were so simple. Unfortunately, two
negative effects often hinder investors that want
to behave in this way.
Figure 01. The probability of generating good

performance by manager type
50%
40%
30%
20%
10%
0%
Skilled
Mediocre
Bad
Figure 02. The expected number of performers

by manager type
3,000
2,000
1,000
0
Skilled
Mediocre
Bad
Good performers Not good performers

First, past performance can subtly influence an
investors perception of a manager. After all,
managers might make better presentations after
good performance (as they have more confidence
and can use better examples). Likewise, all
managers have some weaknesses in their
process, but these weaknesses usually come
under more focus when they have underperformed.
Second, as investment firms seek higher profits,
they will market the products that they think
will win the most business (that is, the good
performers). As such, investors (and investment
consultants) will tend to see more marketing
material on good performers than would be ideal.
The confluence of these two effects should lead
investors to be highly mindful of the unwanted
influence that a managers past performance can
have on them.
Why might investors assume that

outperformance is a more likely
indicator of skill than it is in reality?
As we saw in Figure 01, some investors might
initially think that most good performance should
come from skilled managers. Instead, we showed
that this is untrue, as there are more mediocre
and bad managers than good managers.7
To psychologists, this situation is an example
of representative or base rate bias. But why
does this bias occur?
According to the experts in this field, people
rely on a limited number of heuristic principles,
which reduce the complex tasks of assessing
probabilities and predicting values to simpler
judgemental operations. In general, these
heuristics are quite useful, but sometimes
they lead to severe and systematic errors. 8
Consider another popular example of base
rate bias, this time from Michael Pompien.
His example involves a shy man, called Simon.
Pompien asks for opinions on whether Simon is
more likely to drive a BMW or be a stamp collector.
People generally respond that Simons shyness
is more representative of being a stamp collector
than a BMW driver. What they neglect to consider,
though, is that there are far more BMW drivers
than stamp collectors in the world, and so Simon
is much more likely to drive a BMW.
...investors should downplay the

importance of past performance
and focus on reliable drivers of
future excess returns.
42 towerswatson.com
Conclusion
Using our standard assumptions, we showed that
nearly 80% of US equity managers with good past
performance lack the skill to outperform reliably
in the future. In other words, only 20% of the
managers that boast good past performance did
so because of skill. That begs the question: why
do so many investors rely on past performance,
when it contains so little information?
In our view, investors should downplay the
importance of past performance and focus
on reliable drivers of future excess returns.
Comparing these factors across many firms
is more time consuming than just measuring
performance, but can provide genuine insight into
the quality of the manager. Investors can then
use this insight to improve their forecast of their
managers likely excess return.
References
1
We have made this point many times before, most recently in

How much value will you likely gain or lose by replacing your manager
for performance reasons? Towers Watson Limited, 2011.
That said, our approach works with all other mandates.
Of course, some managers run more than one portfolio, but well
ignore the distinction between manager and product, for the sake of
simple storytelling.
Barras et al studied the performance patterns of all managers to

show how many were skilled, mediocre and bad, after making some
sensible adjustments for luck. Their results show what happened
(for example, 10% of all managers were skilled) but not really whether
this happened because of their earlier performance. That said, the
authors findings provide a reasonable guide to the proportions of
skilled, bad and mediocre managers across the entire manager
population. We therefore use the historical average of these
proportions as the general basis for our calculations in this
article. Our overall conclusion, however, is robust to material
changes in these proportions.
Specifically, we assume that these excess returns are drawn

independently from a lognormal distribution. Our results are not
especially sensitive to this distributional assumption or to sensible
changes in the tracking error of the managers.
Of course, one could also see these differently-weighted distributions

of three manager types as one distribution of the entire manager
population. If we do so using the parameters that we assign above,
this population distribution resembles a lognormal distribution,
centred on slight underperformance.
Whilst our view on this matter may seem to come from only one paper,
Barras, L. Scaillet, O. and Wermers, R. False discoveries in mutual
fund performance: measuring luck in estimated alphas. Journal of
Finance, February 2010. It also reflects our experience of dealing
with institutional investment managers.
See Tversky, A. and Kahneman, D. Judgment under uncertainty:

Heuristics and biases. Science, September 1974.
Quantitative investing
Will quants strike back?
Quantitative equity managers (quants) have

experienced a roller-coaster ride in the past 10 years.
Throughout the middle of the last decade, life was
good for quants: performance was strong, asset inflows
were significant, and even fundamental managers
were embracing the merits of quantitative tools in
screening and portfolio construction. Today, however,
investors often exclude quants from their agenda.
What went wrong? How has the industry

reacted? What is Towers Watsons current
view on this group of asset managers?
Here, we answer these questions and
suggest what opportunities may lie ahead
for quantitative approaches.
Quants on a roller-coaster
Quants, who use systematic factor-based
models to analyse and invest using widely
available data, have a long history in equity
management. Financial market historians
ascribe the beginning of quantitative
investment strategies to Harry Markowitzs
seminal work on portfolio theory in 1952. 1
From a few early adopters, quants slowly found
their place as providers of niche investment
approaches alongside traditional fundamental
managers. Much of that popularity stems from
the growing support of finance academics. 2
The early 2000s changed the scene. Having

gone through the dot-com bubble unscathed,
quant strategies gained widespread legitimacy,
becoming popular with investors and attracting
dramatic growth in assets on the back of strong
returns.3 Strategies employing leverage were
particularly successful. Many fundamental
managers adapted their processes to make
greater use of quantitative insights. The end for
traditional fundamental approaches seemed nigh.
But the history of financial markets is
replete with humbled investors and the
extraordinary success enjoyed by quantitative
strategies proved to be short-lived. Many
quantitative managers have performed poorly
(see Figure 01), and just as strong performance
led to strong inflows, the opposite has come
to pass. Quantitative investing is, for many
today, out of fashion.
Return (US$ per annum)
Figure 01. Relative rolling five-year performance of a representative group of large

active quants 4
6%
4%
2%
0%
-2%
-4%
-6%
2007
44 towerswatson.com
2008
2009
2010
2011
2012
What happened?
Extreme market events, such as the quant crunch

during the summer of 2007, were caused by
crowding in other words, too many assets chasing
the same return factors. Quantitative approaches,
relying largely on historical information to forecast
risk and return, were not well-suited to capture the
systemic risks created by these crowded trades.
For example, before the financial crisis, few
quantitative managers used valuation spreads
to assess the attractiveness of value. Even fewer
had developed indicators to monitor crowdedness.
The trend is not always a quants friend

The risk-on/risk-off macro-led market environment
seen over the past few years has been challenging
for active equity strategies, whether quantitative or
fundamental. Two investment styles in particular
faced headwinds in this environment: value and
momentum. We discussed our views on value
in a recent article.6 Momentum is a trend-based
investment style that struggles at major market
inflection points (as, by definition, it lags behind
changes in market leadership). Many quants use
momentum to add diversification to their often
value-biased approaches. As the two styles are
expected to be complementary, the theory is that
a combined approach should generate smoother
returns. This had been the case for many years.
However, when both styles underperform there
are few places for a quant to hide, particularly
if the strategy is expected to add value over the
short to medium term. As Figure 03 suggests,
it is no coincidence that recent style headwinds
in value and momentum coincided with negative
performance for quantitative managers.
Assets (US$ billion)
In our January 2008 article, Quant management

at an inflection point, we painted a cautious view
on quants. Our concern was primarily driven by the
significant increase in assets, often leveraged and
managed using quantitative strategies. We felt
that structurally low barriers to entry and prolonged
favourable market conditions had encouraged
excessive asset gathering (see Figure 02) in
broadly similar strategies, thereby reducing the
attractiveness of the opportunity set.
Figure 02. Increase in quantitatively-managed assets 5

70
60
50
40
30
20
10
0
2004
2005
2006
2007
Source: GMO, Towers Watson
Figure 03. Relative rolling five-year performance of value

and momentum 7
Return (US$ per annum)
Too many assets
20%
15%
10%
5%
0%
-5%
-10%
1998
2000
2002
2004
2006
2008
2010
2012
Momentum Value
Source: Style Research Limited, Towers Watson
...it
is no coincidence that recent style
headwinds in value and momentum
coincided with negative performance
for quantitative managers.
All change please

Over the past five years, quants have had to face
a challenging environment. Many quantitative
managers decided they had to change their
approaches significantly, in order to maintain a
competitive edge. By and large, these efforts have
focused on three areas: uniqueness of insights,
style-timing and judgemental input.
Unique insights. To avoid the issues resulting
from the crowding of factors, quantitative
managers have worked hard to identify unique
insights. Some managers sought newer,
differentiated data sources or worked to build
proprietary signals, leveraging less exploited
relationships between data and expected returns.
Style-timing. This has been at the top of the
research agenda. Quants have increasingly
attempted to make use of dynamic risk/return
frameworks that adapt to changes in equity
market leadership.
Judgemental input. Some managers have
decided to broaden their risk/return framework
by adding macro-based signals or to rely
more on fundamental or thematic judgement
to compensate for the limitations of their
quantitative models.
What we look for in quants

Not all innovation has resulted in improvements.
Whilst still relying on traditional factors,
quantitative equity strategies have become
increasingly heterogeneous and complicated.
This has raised the bar for asset owners and
consultants when seeking to understand and
assess these strategies.
However, we believe there are a few quantitative
managers that are ahead of their competitors and
can demonstrate credible differentiation in their
approach. Some of the key things we look for are:
Introspection
The best quant managers are highly reflective
and well aware of the natural shortcomings
in quantitative approaches. They know that
quantitative tools may introduce discipline but are
not inherently better than traditional, subjective
methods. These managers are more likely to
foresee problems rather than react to events.
Pragmatic market awareness

It is important that managers complement robust
historical studies with a pragmatic and intuitive
understanding of markets. For example, we think
46 towerswatson.com
...quantitative equity strategies have

become increasingly heterogeneous
and complicated.
that some style-timing indicators may have
long-term signalling power. However, many
managers are required to deliver alpha over
the shorter term, and style inflection points are
very difficult to predict with accuracy. Back-tests
of style rotation indicators are, by definition,
period-specific and we have observed a number
of these processes struggle when used in
real-life scenarios.
Factor differentiation
We are cautious of managers claiming an edge
through the exploitation of unique factors. Such
factors are, in fact, rarely unique. We see similar
insights spreading rapidly across the quantitative
investment community, inevitably impairing their
effectiveness. In order to prolong its competitive
advantage, a manager may become more
secretive. But this reduces transparency and
can make it more difficult to confirm competitive
advantage. Data mining can also be an issue
as ever-growing swathes of data series are
scrutinised. Finally, even when we see more
differentiated factors, they frequently do not
account for a significant part of the quantitative
models overall risk budget.
Low assets
Everything else being equal, a modest level of
assets under management is an advantage. Our
view remains that it is easier to deliver alpha with
total assets of US$1 billion than it is with, say,
US$20 billion. This is particularly true for higher
portfolio turnover approaches which use factors
with a short-time horizon for potential added value.
Suitable fees
Fees for quants are often too high for the likely
level of value added. Managers often have
over-optimistic assumptions about the future
information ratio (the ratio of relative return per
unit of relative risk taken) of their strategies.
Some managers also develop products with
very low active risk in order to optimise gross
information ratios of the strategy, effectively
ignoring the real-world drag from fees paid
by clients.
...we believe that quantitative investing

We do not subscribe to the belief that traditional
can still play a useful, and expanded,
quantitative factors have been permanently arbitraged
away. We expect the well-documented behavioural
role in portfolios via greater use of
phenomena behind these factors, such as the
smart beta strategies.
premium associated with basic valuation ratios,
The rise of smart beta
to recur over the long term and provide opportunities

for patient investors.
Market innovation is making quantitative investing
more commoditised. This may present challenges for
some types of active manager. However, for the asset
owners, this trend is good news. Systematic equity
exposures can increasingly be accessed through
cost-effective, transparent and easily-implementable
investment strategies, leading to improvements in
overall investment efficiency. We have long been
advocates of such solutions which we call smart
beta that bridge the gap between passive and
traditional active approaches.
Index providers and passive managers have so
far been at the forefront of the smart beta trend.
There is now a widening range of indices and
products that offer alternatives to the default
market capitalisation-weighted index. Of course,
some of the caveats that apply to quants also
apply to systematic smart beta approaches.
Indeed we believe that traditional quantitative
managers have a role to play within smart beta,
given their extensive experience of the practicalities
of quantitative investing.
In summary
Following disappointing performance from some
products, traditional active quantitative equity
investing is now far less popular. We have a positive
view of some strategies, but remain cautious on this
group as a whole. Despite efforts by managers to
differentiate themselves, innovation rarely remains
unique for long and process enhancements often
lead to greater complexity.
Nonetheless, we believe that quantitative investing
can still play a useful, and expanded role in
portfolios via greater use of smart beta strategies.
Asset owners can use systematic strategies to
target style exposures inexpensively and in a way
that is consistent with their beliefs or portfolio
construction needs. To achieve this, it may not
be necessary for quantitative approaches to use
unique inputs or to be very complicated if they are
well-grounded and available at reasonable fees.
A smarter quant could be a simpler quant.
References
1 Markowitz, H. Portfolio selection. The Journal of Finance, March 1952.
2 For example: Fama, E. and French, K. The cross-section of expected stock returns. The Journal of
Finance, June 1992; or Jegadeesh, N. and Titman, S. Returns to buying winners and selling losers:
Implications for stock market efciency. The Journal of Finance, March 1993.
3 For example, Casey Quirk & Associates LLC. The geeks shall inherit the Earth?, 2005.
4 Average relative returns of representative active global or international equity strategies from 10
large quantitative managers. Manager selection based on assets managed in active quantitative-only
strategies. Performance is relative to stated strategy benchmark, gross of fees.
5 US$ value of coincident holdings, as taken from 13F filings in US, of the largest eight
quantitative-only investment managers.
6 Value investing an old idea, but probably a good one. Towers Watson Limited, January 2013.
7 Simulated performance of selected strategies; momentum: highest quintile by 12 months
price momentum; value: highest quintile by earnings yield. Universe: Largest 2500 global
companies in global universe, market capitalisation weighted, quarterly rebalance.
Concentrated equity products

Why we prefer them to diversified products
For a typical investors portfolio, adding skilled active

management increases expected return without much
change in total risk. We therefore encourage equity
investors to seek products with higher expected returns.
That means concentrated rather than diversified products.
Investor preferences drive the choice of
concentrated or diversified equity products.
Most investors want to improve the return
efficiency of their entire portfolio. We call this a
portfolio-level preference. That requires them
to consider the contribution of asset allocation
and active managers. We generally find that
using skilled active managers helps to improve
overall investment efficiency. That is because
expected active returns add to expected asset
returns, whilst overall risk barely changes with the
introduction of active risk. We therefore expect
active management to have positive marginal
impact on overall risk-adjusted return. 1
48 towerswatson.com
All other things being equal, we expect investors

with these portfolio-level preferences to favour
products with higher levels of expected excess
return. They should, therefore, favour concentrated
products to diversified products. This is our main
point in this article.
Some investors have a narrower focus, though.
Their focus is at the product level rather than at
the total portfolio level. They just want each active
manager to generate its excess returns efficiently,
with the best information ratio. This goal also
minimises the chance that any one manager
will underperform. 2 By doing so, these investors
underplay the contribution of asset risk to
total risk.
The assumptions of Grinold and Kahn (2000) 3

show that a manager can increase its
information ratio by making more decisions
that are independent. That would imply a higher
information ratio for diversified products, as
investment managers make more decisions in
them than in concentrated products. In other
words, diversified products are more suitable
to investors with product-level preferences.
Whilst we applaud the ingenuity of Grinold
and Kahns thinking, we doubt some of their
simplifying assumptions. As we show later,
real-world evidence suggests that these
simplifying assumptions are unrealistic
and lead to an overly favourable view of
diversified portfolios relative to their
concentrated counterparts.
We now expand upon these points, by:
Using a simple example to show how investment
preferences can drive the choice of concentrated
or diversified equity products
Explaining our preference for concentrated
portfolios in other ways
Asking whether diversified products have better
information ratios than concentrated products
Outlining why some investors with
product-level preferences may fare better
with passive management
An example of how investor

preferences can drive the choice
of concentrated or diversified
equity products
Imagine two investors, Don and Carl, who only
invest in equities. Their portfolios are almost the
same. They split their assets equally between
the same two skilled managers, ABC Investors
and DEF Partners. The only difference is that Don
uses the diversified products of these managers,
whilst Carl uses their concentrated products. Each
product has the same benchmark, which reflects
the broad equity market. The correlation between
the excess returns of ABC and DEF is 0.2, be that
for diversified or concentrated products.
In keeping with our previous statements, we
expect the concentrated products to have double
the excess returns and triple the tracking error
compared to the diversified products.4
Given these values, an investor with
product-level preferences would favour Dons
portfolio of diversified products, with its better
information ratio (of 0.43 versus 0.29).
We now consider the total picture. We assume
that equities return 5% a year on average, with a
variability of 15% a year and with no correlation to
the excess returns of the managers.
Figure 01. The excess returns of diversified and concentrated products

Expected
excess return
Tracking error
Information
ratio at the
product level
Information
ratio at the
portfolio level
Dons diversified products
1% pa
3% pa
0.33
0.43
Carls concentrated products
2% pa
9% pa
0.22
0.29
Figure 02. The overall returns of diversified and concentrated products

Expected
return
Risk
Ratio of expected return to

risk at the portfolio level
Dons diversified products
6% pa
15.2% pa
0.40
Carls concentrated products
7% pa
16.5% pa
0.42
Figure 02 shows the impact on the overall

ratio between expected returns and risk, a
measure favoured by those with portfolio-level
investment preferences.
Using this measure, we expect Carls concentrated
products to beat Dons diversified products. The
higher excess return and tracking error of the
active products increases expected total return
far more than it increases total risk. That is
particularly true for the concentrated products.
This is the main reason why we generally advise
most investors to use concentrated rather than
diversified equity products.
Other reasons for preferring

concentrated portfolios
Our other main reason relies less on theory and
more on common sense. If you think that an equity
manager is skilled, then you should get its best
ideas into your portfolio. In a world of uncertainty
and imperfect managers, though, you do not
just want your best manager to hold a one-stock
portfolio. Instead, you should add some diversity.
You should let your best manager put only its best
ideas into your portfolio. Once the managers
conviction in its next-best stock idea falls, then
you should move on to the second-best manager
and get its top ideas. You should repeat this
process until any more diversity unduly dilutes
your expected outperformance.
We can also make the same point in a similar way.
Consider two equally weighted portfolios that both
seek to beat the S&P 500 Index. The first portfolio
is concentrated and holds 20 stocks, each at a
5% weighting. The other portfolio is diversified and
holds 1% in each of 100 stocks. To hold these long
positions, each portfolio has explicit or implicit
underweight positions in the remaining stocks in
50 towerswatson.com
the index. However, the size of the average active

bet is about five times larger with the concentrated
portfolio than with the diversified portfolio. Any
stock selection edge from the concentrated
manager will therefore be magnified relative to
the outcome of the diversified manager.
Contrary to the simplifying assumptions of
Grinold and Kahn, we also find that many
managers have particular expertise in certain
pockets of the market. In these cases, the active
bets of a concentrated approach would extract
more value for the investor than those of a
diversified approach.5
For what it is worth, empirical work also
suggests that managers with more concentrated
stock decisions tend to outperform other types
of managers.6
Do diversified products have better

standalone information ratios than
concentrated products?
In the summary, we said that Grinold and Kahn
showed how to influence the best information
ratio that a product can attain. We discuss these
influencing factors below, comparing their values
in concentrated and diversified products from the
same manager.
The skill of the forecaster. We expect no
difference in skill between these concentrated
and diversified products, as they share the same
manager and benchmark.
The number of independent forecasts made
in the product. By design, the manager
makes more independent forecasts in the
diversified product. Theory also assumes
that successive forecasts are subject
to diminishing marginal returns.
The level of implementation efficiency in

converting these forecasts into allocations
within the product. Portfolio constraints play
a large role here. Guidelines for managers, for
example, often compel them to hold a given
amount in certain parts of the investment
universe. Traditional equity mandates also
impose a long-only constraint to prevent
negative allocations to stocks.7 Theory
suggests that this long-only constraint hinders
implementation efficiency far more in products
with higher tracking error. 8 For that reason,
concentrated products tend to fare worse
than diversified products in implementation
efficiency. Perhaps for that reason, concentrated
products often have fewer constraints than
diversified products.
Taking these points together and accepting the
simplifying assumptions, the diversified product
of a skilled manager would be expected to produce
a higher information ratio than its concentrated
counterpart, even though it should not outperform
by as much.9
Given our views in the previous section of this article,
we do not accept all these simplifying assumptions.
We therefore do not have strong views on whether
diversified or concentrated products generate better
standalone information ratios.
Some investors with product-level

preferences may perform better
with passive management
In The cost of trigger-happy investing, 10 we followed
two investors in a typical equity product. We showed
how the product only benefited the more disciplined
investor of the two. The less disciplined investor,
who could not tolerate prolonged underperformance,
would have fared better with an index fund.
We repeat this conclusion here, as it relates to a
flaw in the reasoning of some investors when they
argue for diversified products. According to these
investors, they prefer diversified products because
they lead to fewer instances of underperformance,
and so to fewer occasions when they will fire the
product. They then argue that they will not lose
as much value from poor decisions but will still
benefit from outperformance from most of their
active products.
Further reading
Clarke, R. de Silva, H. and Thorley, S. Portfolio Constraints and the
Fundamental Law of Active Management. Financial Analysts Journal,
September/October 2002: 48-66.
Cremers, K J M. and Petajisto, A. How Active Is Your Fund
Manager? A New Measure that Predicts Performance. The Review
of Financial Studies, 2009.
Track records Luck or judgement? Introducing hit rates
and win loss ratios. Inalytics Limited.
Scherer, B. and Xu, X. The Impact of Constraints on Value-Added. Journal of
Portfolio Management, Summer 2007.
Concentrated Portfolios: An Examination of their Characteristics and
Effectiveness. The Brandes Institute, 2004.
Remapping our investment world. Watson Wyatt, 2003.
References
1 This article concerns equity portfolios but many of the principles
also apply to other asset classes, even though the definitions of
concentrated and diversified may differ.
2 When combined with a managers commercial goals, these conclusions
may help to explain why managers often target the best information
ratio for their product. After all, an investment manager usually gains
most commercial benefit if it retains its mandate with the client. As it
can improve its odds of retention by trying to minimise its odds of
underperformance, it does so by targeting the maximum information
ratio for its product. Of course, this action makes commercial sense
for the manager, but does not necessarily benefit the client.
3 Grinold, R C. and Khan, R N. Active Portfolio Management.
McGraw-Hill Professional, 2000.
4 We base these assumptions on performance after fees. We recognise,
however, that the fees for concentrated products typically exceed those
of diversified products in absolute terms, but not in risk-adjusted terms.
5 The article from Inalytics also considers this dynamic. Put briefly,
it shows that managers collectively tend to win more than they lose
in overweight positions. However, they lose more than they win in
unintended short positions.
6 See Cremers and Petajisto (2009) for further details. Seemingly running
counter to these views, other studies have shown that the average
concentrated manager does not outperform the average diversified
manager, both before and after a risk adjustment. One such example
is the paper from The Brandes Institute. To us, these findings are
reasonable but potentially misleading, as we condition our preference
for concentrated managers on them being skilled, not average.
7 Investors can improve the information ratios of many skilled
managers, however, by relaxing the constraint against holding
a negative weight in a stock. That said, fees are often higher and
these long/short portfolios tend to suit the investment approach
of only a subset of fundamental managers.
8 For further details, see Clarke et al (2002), and Scherer and Xu (2007).
9 See Kroll, B. Reach for more excess return but dont hurt yourself.
Part II: The paradox of portfolio concentration are your managers
best ideas good enough? JP Morgan Asset Management, 2004,
for further details.
10 The cost of trigger-happy investing. Towers Watson Limited, 2012.
Given the conclusions of our other article, this

double negative argument will harm the investor.
Investors that prefer diversified products for this
reason should improve their likely outcomes and
invest in an index fund.
Low volatility equity

From smart idea to smart execution
Low volatility equity strategies have gained a lot of attention

in the last two years, with proponents claiming they can
deliver market level returns for below market level risk.
We have seen significant interest from our clients worldwide
and the volume of new product launches from asset managers
has been remarkable.
The low volatility idea has merits
Towers Watsons Thinking Ahead Group was one
of the first to advocate inexpensive, transparent,
systematic strategies (smart beta) as an attractive
alternative to some forms of traditional active
management. For example, our institutional
investment clients have allocated over US$18 billion
to smart beta solutions. In addition, our active equity
manager research team has long been a proponent
of skilled managers with defensive, quality-oriented
approaches particularly during periods of high
market uncertainty. Combining the broad smart beta
initiative and defensive equity investing brings
us to smart beta low volatility equity strategies.
52 towerswatson.com
The potential merits of the low volatility

thesis are quite compelling: it is supported by
long-term empirical evidence with plausible intuitive
rationales; there is the prospect of reduced
volatility, reduced absolute drawdowns and
improved reward per unit of risk; it has the
potential for use in de-risking programmes; and,
finally low volatility strategies can be inexpensive,
transparent and widely available. That said,
plausible arguments also exist that a large part
of the low volatility observation can be explained
by existing investment return drivers which have
long been recognised such as market beta, value,
country or sector effects.
Potential pitfalls
While the low volatility concept is worth investigating,
there are potential hazards along the path from the
initial idea to final implementation. There is a very
wide range of strategies to consider and smart beta
investments are a new avenue for many investors.
Furthermore, the effect of impressive results from
the strategy tends to raise expectations, rather
than raise caution which may be a more appropriate
reaction. We believe it is essential to navigate the
potential pitfalls, which we highlight here.
Realistic expectations
Empirical evidence suggests that there is a trade-off
between return and risk, but that this relationship
may be flatter than traditional market theory
suggests. If so, a low volatility equity portfolio may
have a higher risk-adjusted return than the market as
the reduction in risk is achieved with only a moderate
sacrifice of return. Note that we still expect a
somewhat lower than market return from (unlevered,
long-only) low volatility equity strategies over the
long term. Having an expectation of significant risk
reduction without any sacrifice of long-term return
is, in our view, counterintuitive and based on an
overemphasis on period-specific historical backtests.
Hence, we are cautious of the statement used in
marketing presentations: market level returns for
below market level risk.
Empirical
evidence suggests that there

is a trade-off between return and risk,
but that this relationship may be flatter
than traditional market theory suggests.
Measurement should not be an afterthought
It is important for investors to determine, in
advance, their objectives from an investment in a
low volatility equity strategy. Different approaches
to measurement may determine whether such an
investment is considered a success or a failure,
as demonstrated in Figure 01.
Current market conditions

The case we most typically hear supporting
low volatility equity is a strategic one based on
long-term historical evidence. However, we believe
long-term past factor performance should only
be a tentative guide to the future. As demand
for low volatility equity strategies increases
then a portfolio of low volatility stocks becomes
more expensive thus reducing future returns,
all else equal. Our analysis of current market
conditions considers the valuation of a global low
volatility strategy from several angles. Recently
we have observed that, while the strategy
looks reasonable from an absolute valuation
perspective, the relative valuation appears
stretched compared to historical levels.
Figure 01. Success or failure?
The market returns

14% per annum
with 20% volatility
Investor A
Low volatility
strategy
(beta 0.8)
Investor B
5 years
Investor A notes the lower

achieved volatility but is
more focused on the
(non risk-adjusted) relative
return of -2% per annum.
Investor A
is unhappy
Low volatility
strategy returns
12% per annum with
15% volatility
Cash returns
2% per annum
Investor B looks at the

risk-adjusted return of the
low volatility strategy versus
the market. Against a
composite 75% market,
25% cash benchmark
he sees outperformance
of 1% per annum.1
Investor B
is happy
Figure 02. Number of low volatility products on eVestment database over time
90
80
70
60
50
40
30
20
10
0
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Source: eVestment Alliance

Methodology: Analysis based on searching for equity products on the eVestment Alliance database containing the terms low volatility or defensive.
Growth is understated as we are aware of products which are not yet on the database.
Selecting the right product

Perhaps the greatest challenge to an investor that
wants to invest in a low volatility strategy is the
overwhelming choice available. We have observed
huge growth in low volatility equity products over
time as shown in Figure 02.
We believe there is a devil in the detail challenge
when evaluating low volatility equity strategies.
One key differentiating factor between approaches
is the breadth of methodology. We are cautious
of processes that focus on low historical price
volatility to select securities. In some global
equity simulations of this approach, we observed
a large allocation to the financial sector prior to
the global financial crisis. This would probably
surprise investors thinking they were buying
a safer portfolio of equities. This example
demonstrates that low historical stock price
volatility is not a watertight guide to the future
resilience of a business. The criterion for a stock
to have low historical price volatility is simply that
there were few sudden changes in the balance
between buyers and sellers over the particular
period in question. We are also concerned that
undifferentiated low volatility strategies may
suffer from poor liquidity owing to the crowding
of similar approaches buying the same universe
of stocks (as experienced by quantitatively driven
approaches in 2007).
Products in this space differ considerably in terms

of their complexity. Examples include embedded
quantitative alpha models, fundamental risk
models, statistically-based principal component
models, implied volatility models and various
optimisation techniques. The more complex the
process, the less transparent, and the more
effort is required to evaluate the relative expertise
of the manager. The existence of more moving
parts within a process often leads to higher
turnover and potentially higher transaction costs.
Furthermore, greater complexity usually results in
greater management fees. We would argue that
fees in this area should be low given the moderate
running costs, high capacity, and the intense level
of market competition.
Summary
We advocate smart beta investments as a new
avenue for many investors. In our view, low volatility
equity can be smart, but only by avoiding the
potential pitfalls.
References
1 In the example given, Investor B could also have used alternative
risk-adjusted performance measures such as Jensens alpha.
This article is based on a Towers Watson article first published in Investment Pensions Europe (IPE) in November 2012.
54 towerswatson.com
2011
2012
Sustainability in investment research

A pragmatic approach
Capital markets, economic linkages and political realities
are likely to be starkly different in 10 to 20 years from now.
Several interconnected converging mega-trends will lead
to transformational change in coming decades and impact
directly on long-term economic growth prospects.
Adverse demography, economic imbalances and
degradation of natural capital will all shape the
investment landscape as will changing roles
and the influence of business, government and
innovation. Whilst some of these trends are
reasonably well understood, others, such as
environmental and social trends, are more complex
and difficult to predict. Given these prospects
we believe that prudent long-term investors
need to consider how their investments might
be affected over time. This goes hand-in-hand
with the growing focus on long-term investment
and responsible stewardship of capital.
In this context we are incorporating sustainability
considerations into our investment research, both
from an asset class and investment manager
perspective. Within our manager research
process we consider the relevance of longer-term
sustainability trends on differing investment styles,
philosophies and portfolios. Active equity strategies
with long time horizons, of five to ten years for
example, will undoubtedly be more sensitive to
sustainability factors than trading style strategies
which have a much higher portfolio turnover. By
their very nature, passive strategies will feel the
full force of market wide impacts such as changing
demographics and degradation of natural capital.
Strategies with high exposure to resource intense
sectors, such as mining, cement, oil and gas will
be affected by carbon pricing, greater legislation
and liabilities from environmental damage. Similarly
portfolios investing in companies with strong
consumer brands could be negatively affected by
poor labour standards in supply chains or fast-food
companies by efforts to curb the rise of obesity.
Using a propriety analytical tool we assess the
sustainability risk profile of equity portfolios drawing
on stock specific performance data supplied by
a third party. We can identify where the most

significant sustainability risks lie within a portfolio
from a regional, sector and stock perspective.
Alongside more traditional risk attribution this
analysis provides another lens through which to
assess portfolios.
In our discussions with investment managers we
explore how they identify, assess and act on the
sustainability risks inherent in their portfolio. As we
noted earlier, the degree to which these risks are
relevant to any given strategy is a function of its
time horizon, style and particular exposures. Where
sustainability trends could realistically impact asset
prices over the possible holding period we expect
managers to try and reflect this in their investment
thesis, financial models and/or ownership activities.
Clear thinking on sustainability risks may also
influence a managers portfolio construction and risk
assessment processes.
Predicting the future is not a precise science and the
impact of sustainability trends is highly uncertain.
With this in mind the use of scenarios and stress
testing can be useful tools to explore how portfolios
will respond to changing parameters. Identifying
and collating key data points will be critical in
understanding how and when dynamics are shifting.
Given future uncertainty and pace of change the
need for diversity within and/or between portfolios
becomes even more important.
Were in the early stages of understanding the
impact of sustainability trends on investment risk
and return and we encourage and welcome an open
debate with asset owners and managers around
how we can advance best practice. One thing that
we do know however is that it is important to start
the research and analysis now to reduce the risk of
being caught out in the future.
What to look for in a passive manager

Passive management, or indexation, is often seen as a
commodity product, where choices should be made on price
alone. We do not believe this to be true. Here, we explore
points of differentiation between passive equity managers
and aims to illustrate why we believe rigorous qualitative
research can lead to better outcomes.
In passive equity management, the primary

objective is to track an index of stocks, whether
it is a bulk beta or a smart beta index. It is
often used in a portfolio to deliver broad, low-cost
equity exposure and, as such, it can be easy to
focus on headline costs. However, doing passive
management well is more complicated than many
people realise and we believe it is well worth
spending time researching the nuances of various
products and the methodologies being applied.
Poor passive management can, for example, lead
to higher transactions costs and/or unexpected
performance deviations from the chosen
benchmark. Good passive management involves
focusing on the detail. In Figure 01 we have set
out what we believe to be the key criteria for
passive managers and expand on each of these
criteria below.
A common theme that runs across a number
of these criteria, and contrasts with active
management, is the advantage that accrues from
size. Size is no guarantee of quality in an index
manager, but it often helps. Indexation, more
than any other form of fund management, is a
low margin, high capacity product. Competition
between managers and the perception of
indexation as a commodity means that managers
rarely see the kind of profit margins that are
seen in other areas of fund management.
Therefore, in many of the areas that we highlight,
there are advantages of scale, both at the
manager and the product level.
56 towerswatson.com
Figure 01. Key criteria for passive managers

Investment professionals
Broad, deep, experienced team
Philosophy and insight generation

Methods to add value within acceptable
risk parameters
Portfolio management
High quality systems
Efficiency of trading
Securities lending
Firm and team stability

Size of the indexation business
Opportunity set
Size of individual index tracking funds
Appropriate fund structures
Alignment
Low fees and costs
A range of ancillary services
Strong client service
Investment professionals
Indexation, unlike other areas of fund
management, does not tend to attract star
portfolio managers. However, index tracking
requires a specific skill set, so a dedicated and
high-quality team can make a lot of difference
to the strength of the offering. Experience
is an important characteristic, as well as a
good understanding of both quantitative risk
systems (their strengths and limitations) and
more lateral risk metrics. Depth of team is
also important, to ensure a comfortable level
of coverage for each portfolio manager. In
our opinion, portfolio managers should not
have too many different accounts to manage
and should also have an understanding
of different types of client accounts.
makes it more cost-effective to hold a greater

proportion of the index constituents a manager
may be able to offer full replication, where every
stock in the index is held. A pool that is smaller,
or is tracking an index of stocks that are more
expensive to buy and sell, may be run on an
optimised basis, where the manager aims to
generate the returns of the index by holding a
representative sample of stocks. Any kind of
sampling is likely to result in higher tracking error
or tracking difference than full replication.
In addition, managers may seek to add
small amounts of value to offset cost within
acceptable risk parameters, for example
by timing trading around index changes.
The best managers will balance total
returns after fees against minimisation of
tracking error (or tracking difference).
Philosophy and insight generation

While the primary function of an index manager
should always be to provide returns that are
close to those of an index, the art and science
of indexation is to hold a range of securities that
will track the performance of the underlying index
closely, without incurring a disproportionate
cost in accessing and holding these securities.
Techniques that a manager will employ will vary
depending on the index being tracked, the size
of the product and the managers investment
philosophy. A large pool of assets generally
...managers
may seek to add small

amounts of value to offset cost
within acceptable risk parameters,
for example by timing trading around
index changes.
Portfolio management
High quality systems
The nature of indexation management means
there is a strong focus on cost and risk. The
best managers have systems that allow them
to see their portfolios easily, understand what
the key risks are and transact efficiently within
their defined parameters. Pre- and post-trade
compliance checks that are built into the portfolio
and order management systems are another
important risk mitigation feature. A high level of
commitment to buying, building and maintaining
these systems is a differentiating factor for the
leading firms. While a high specification and easy
to use system may seem nice-to-have rather than
essential, we believe that these tools materially
reduce risk of human error.
Efficiency of trading
Almost no index that is tracked, be it in equities,
bonds or alternatives, takes account of transaction
costs. Every penny that is spent on transactions,
administrative costs, fees and taxes represents a
negative tracking difference. All managers should
be aware of trading costs, but for indexation
managers that often hold large numbers of
securities in relatively small proportions and
have no alpha in which to hide costs, reducing
these is a crucial efficiency. Again, scale can help
here, as large managers are likely to have more
money moving around in order to match buyers
and sellers of stock, and are able to negotiate
favourable terms with brokers. Some managers
have periodic days where they try to encourage
all clients to trade, so as to maximise crossing
opportunities. As technology and trading venues
continue to broaden and expand with the growth of
direct market access, algorithmic trading and dark
pools, we would expect to see the best indexation
managers leading efforts to drive trading cost
efficiencies for their clients.
Securities lending
Securities lending within the portfolio is an
additional method that a manager may use to add
return to the portfolio, but it is not without risks.
These risks were graphically illustrated in 2008
when many funds had problems with their stock
lending programs, mostly where cash collateral
had been reinvested in assets that became
illiquid. This caused underperformance and, in a
number of cases, withdrawal restrictions.
Securities lending programmes in particular should
be reviewed to ensure that investors are happy with
the risks they bring, looking for details such as:
What limits there are on lending

What collateral is taken and how it is held
Whether indemnification is offered
How revenues are split between the manager
and the investors and whether that brings
conflicts of interest
The degree of oversight a manager retains
of a program, if it is outsourced
A good securities lending programme cannot make
up for a poor manager and any such programme
must be monitored. Done well, we believe it can be
a beneficial addition to a good product. However,
if investors are not comfortable with a managers
approach to securities lending, non-lending funds
should be sought.
Firm and team stability

As we have noted, indexation, more than any other
form of fund management, is a low margin, high
capacity product. Where profit margins are low,
gaining confidence that the team and systems will
continue to be supported by the business for the
long term is key. Firms with a large, established
product base and a business structure where
indexation products are important to the firm
have an advantage here.
As technology and trading venues continue to broaden

and expand with the growth of direct market access,
algorithmic trading and dark pools, we would expect to
see the best indexation managers leading efforts to drive
trading cost eciencies for their clients.
58 towerswatson.com
Opportunity set
Size of individual index tracking funds
Within each index tracking fund, size is almost
always a benefit. A large pool of assets can mean
a manager can hold more stocks in the index, but
it also gives a broader base over which to spread
the (often largely fixed) administrative costs.
When funds are very large they can also use this
scale to create further efficiencies. A bigger fund
is likely to have more buyers and sellers than a
smaller one and so may offer opportunities for
investors to trade at mid-price (through matching
these buyers and sellers together). A large
fund may also be able to use its scale to earn
fees for sub-underwriting new stock issuance
(which the fund has to buy as the stock enters
the index), and to trade more opportunistically
than a smaller fund might be able to.
Appropriate fund structures

to meet specific client needs
There are clearly benefits in all types of fund
management in offering the appropriate vehicle,
but within passive management this can be
one of the key differences between products.
Different types of institutional investors in various
markets have certain tax statuses. To invest in
a vehicle that mimics their tax status or offers
tax transparency can bring material benefits in
areas such as withholding tax on equity dividends,
for example. This is particularly important as
a differentiator for certain categories of client,
where the tax savings in an appropriate vehicle
can swamp the manager fees a more expensive
manager with appropriate vehicles may be a better
choice than a cheaper manager with inappropriate
fund structures.
Alignment
Low fees and costs
As we have illustrated, headline fees do not tell
the full story when assessing a passive product.
Nonetheless, it is clear that one of the key
benefits that passive management offers when
compared to active management is its low cost.
Understanding the total cost involved (fees and
other costs) will factor into any assessment.
It is important to note the details, such as cut
off points on sliding fee scales and discounts
offered for holding more than one index product
with a manager: typically it is cheaper to find one
manager than can offer multiple indexed products,
as fees can be negotiated on total assets.
A range of ancillary services

and good client service
Ancillary services are also an important part
of the value proposition of indexation managers.
The ability to monitor and rebalance a multi-asset
class mandate against a benchmark, offer help
in transitioning assets in and out of funds
and offer currency hedging, are often key
reasons for employing a passive manager in
the first place. As a low-cost alternative to active
management, it is important that the manager
provides an efficient and helpful service and is
available to support the client where necessary,
providing representatives for meetings and
general assistance.
More broadly, alignment captures a managers
overall commitment to its client base. A manager
may demonstrate positive alignment through its
ownership structure or through fair distribution
of securities lending revenue. Investors who
prioritise environmental, social and governance
factors may also evaluate managers on their
alignment with these values, for example through
established programmes of engagement with
company management.
Summary
Passive management is a fast-growing sector in
fund management, as many institutional investors
focus on costs and efficiencies within their funds.
It is often regarded as easy and argued that the
selection decision should be based solely on fees.
Here, we have demonstrated that while price is
an important factor, there is a lot more to good
passive management than low fees. We believe
that broader qualitative due diligence, using
the detailed criteria we have set out should be
undertaken to ensure that appropriate managers
are hired to meet an investors needs, now and
in the future.
Section t hree
Manager monitoring
60 towerswatson.com
When constructing a portfolio, it is

important to consider the level of
governance resource that an investor can
apply. The same is true when establishing
the parameters for monitoring the
implementation of decisions and
evaluating their success (or otherwise).
In this section, we ask if the decisions
taken by investment committees alter
their performance. Establishing an
appropriate benchmark and time-frame
to measure success is a necessary, but
often misunderstood, first step. Evidence
suggests that asset owners struggle to
make timely hiring and firing decisions,
so we address these here.
03 Manager monitoring
Benchmarks matter
Investors use benchmarks to:
1. Define an investment universe for their managers
2. Measure and evaluate performance
It is therefore important for investors to review the
benchmarks that are being used and ensure that they
continue to be appropriate.
Benchmarks are not static and often it is unclear
what the differences are between similar-sounding
benchmarks. The increase in absolute return
mandates has made benchmarking decisions
even harder.
As an example, consider Asian equity indices.
There are well over 100 different Asian equity
indices available to investors covering a range
of styles, sizes and regional variations. From a
sample of 50 of these indices, the narrowest of
these indices contains two countries, the broadest
thirteen, and there are large differences in
performance over prolonged periods of time.
62 towerswatson.com
Given the considerable time and cost spent on

monitoring, hiring and firing managers, it is
important that the inputs to this process are as
accurate as possible.
Among other things, the choice of benchmark
used can affect:
Your opinion of whether or not a manager has
demonstrated skill by outperforming the market
Whether and how much you pay in performance
fees to your managers
The type and level of market exposures in
your portfolio
What makes a good benchmark?

An ideal benchmark index should represent the
risk and return characteristics of the asset class
and meet the criteria shown in Figure 01.
Figure 01. What makes a good benchmark?

1. Clarity
The index construction rules and constituents

should be known and clearly defined.
Monitoring managers: Does

outperforming a benchmark mean
your active manager is skilful?
2. Completeness
The index should have broad, diversified

coverage of the securities available for
investment in the asset class.
Measuring performance against a market

benchmark is a common way to assess whether a
manager is able to add value by outperforming the
market. The benchmark used for comparison needs
to reflect the universe of investments from which
the manager is choosing its holdings. Arguably,
benchmarks should also reflect factors such as
leverage, currency trades and style biases that
affect performance but do not necessarily represent
skill and should not be rewarded as such.
3. Investability
The constituents of the index should be

tradable that is, have a relatively liquid market
with regular pricing. The constituents and
characteristics of the index should be relatively
stable over time.
4. Independent and
regular data
Ideally, indices should also be:

Calculated regularly by independent providers
Available historically to enable managers to
understand the characteristics of the index.
As an example, consider the manager in Figure 02

who was appointed in January 1999 to manage a
100 million global equities portfolio. Performance
was patchy for the first few years, but strong
relative to the MSCI World Index in 2008. Then
in 2009 they gave that all back and more, with a
peak to trough fall of 6.8% in a year.
Figure 02 does not actually show the performance
of an active manager but shows the performance
of the MSCI All Countries World Index against the
MSCI World Index (the first includes emerging
markets and the second does not). If this was
the performance of a manager who was closely
tracking the MSCI All Countries World Index but
was measured against the MSCI World Index,
that manager may have been terminated in 2006
for poor performance, or may have received a
performance fee in 2009, when in fact neither
of these actions is justified.
Figure 02. Does this performance tell us anything about skill?

3.0%
2.0%
1.0%
0.0%
-1.0%
-2.0%
-3.0%
-4.0%
-5.0%
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Source: eVestment, MSCI
Monitoring managers: have your

managers styles drifted over time?
Usually, an appropriate manager benchmark is
considered as part of the initial manager selection
process. However, a managers approach can
change gradually over time. The make up of a
particular benchmark can also change over time,
as securities and markets drop out of an index
and new ones are added. Therefore, even if a
managers benchmark was the most appropriate
choice when the manager was appointed, manager
benchmarks should be reviewed to make sure they
are still the right fit several years on.
...
even if a managers benchmark was the
most appropriate choice when the manager
was appointed, manager benchmarks
should be reviewed to make sure they
are still the right fit several years on.
Market-weighted benchmarks
One example of how benchmarks are important for passively
managed assets is the recent development of alternatives to
the traditional market-weighted benchmarks. The majority
of commonly-used benchmarks are weighted according to the
market value of the securities that a company has in issuance.
The benefit of this approach is that it gives exposure to the
whole of the available universe at low cost.
However, there are also several drawbacks with this approach:
Market-weighted benchmarks hold more of a security as it
increases in value. If a security becomes overpriced and then
falls in value again, the benchmark holds the largest amount
of that security right before the price begins falling. Similarly,
a market-weighted benchmark holds less of a security as it
falls in value, and its allocation to a security is at its lowest
right before the price begins to increase in value again.
This means that market-weighted benchmarks tend to have
high allocations to over-priced assets and low allocations
to under-priced assets.
For bond markets, a market-weighted benchmark has the
highest allocations to the most indebted borrowers (the
companies or countries with the most bonds in issuance),
which seems counterintuitive.
In recent years, investors have become increasingly aware
of these issues, and index providers have developed
alternative ways to construct a passive benchmark. For
equity benchmarks, these tend to involve gaining exposure
to different factors or approaches, such as value-weighting
or targeting a certain level of volatility. For bonds, the focus
has been on improving diversification and considering metrics
which measure a companys ability to repay its debt.
The
purpose of a review is to better

understand what exposures you
are trying to achieve, and to better
equip yourself to identify and reward
genuine manager skill.
64 towerswatson.com
Is the combination of market exposures

in your portfolio appropriate?
At an individual manager level, benchmarks
provide a means for monitoring performance and
for identifying manager skill. However, benchmarks
can also be used to help to define and monitor
your long-term investment strategy. Because of
this, benchmarks are as important for passive
managers as for active managers.
A review of manager benchmarks can also
provide an opportunity for investors to consider
whether the combination of their active and
passive managers styles and opportunity sets is
appropriate when viewed at a total portfolio level.
What about alternative asset classes?

There are a number of issues specific to individual
asset classes that require consideration. Here
we consider private markets and hedge funds
as examples where choosing an appropriate
benchmark is particularly challenging. There are
the three main types of benchmarks that are used
for these investments; we highlight some key
advantages and drawbacks of each.
1 Public market indices: these are among the
most common benchmarks used for private
markets in particular.
Market indices can be useful to investors as
they provide a comparison of how a manager is
performing versus a portfolio of similar securities
that are listed on regularly traded markets.
However, the drawbacks to using public
indices include:
Public market performance is not directly
linked to the main drivers of return for private
market and hedge fund investments.
Public markets tend to exhibit higher volatility
than private markets and hedge funds (or at
least appear to do so).
For private market investments, portfolio
returns are typically measured using a
different calculation method to the method
used for calculating market returns (private
market returns are measured on an internal
rate of return basis whereas public market
returns are calculated as time weighted
rates of return).
2 Peer group indices: these are compiled by

aggregating the performance of a group of
investment managers in the same industry.
This type of index is particularly popular for
measuring hedge fund performance.
Measuring a manager against their peers provides
an indication of how other investments of the same
nature have performed and avoids many of the
issues related to the use of public market indices.
However, there are many weaknesses of peer
group indices which include:
Peer group indices are not investable, as
investors do not have access to all of the
managers that make up the index.
The valuation and performance methodologies
can vary between managers, and often the
managers reported performance is not verified
by the index provider.
Peer group indices typically include only
those managers who choose to report their
performance. This results in two key problems.
Firstly, only the strongest performing funds
report their performance generally. Secondly,
funds that cease to exist possibly following
a period of poor performance are removed
from the index, and those managers past
performance may also be removed. Both of
these problems tend to result in peer group
index performance being artificially high.
3 Absolute return benchmarks: instead of
comparing a manager to its market or peer group,
some investors choose to measure manager
performance against a long-term performance
target, such as 5% per annum. Other return
targets include a margin above cash returns
or a margin above inflation.
The drawbacks to this type of benchmark include:

They are not investable.
Absolute return indices do not allow for
market returns. Given all manager performance
is linked to some extent to what happens
in markets, an absolute return benchmark
is unlikely to give a good indication of how
that manager has performed given
market events. This is particularly true for
short-term performance.
Given the pros and cons of each benchmark type
considered above, it is important for investors to
review their manager benchmarks on a case-by-case
basis, considering both the characteristics of the
particular asset class being invested in, and each
managers investment style and time horizon.
As no single metric can provide a complete analysis
of a managers competence, we also think it is
useful for investors to take a balanced scorecard
approach: this approach supplements a quantitative
benchmark with a qualitative monitoring process
including a number of short- and long-term factors
such as the quality of a managers investment
processes and various risk measures.
Using the right benchmarks is important

Reviewing whether your managers are being
measured against the right benchmarks isnt about
penalising managers, or trying to create benchmarks
that are more difficult to beat. The purpose of a
review is to better understand what exposures you
are trying to achieve, and to better equip yourself to
identify and reward genuine manager skill.
Put simply, better design means better outcomes.
These benchmarks can be useful for assessing

the long-term success of a manager or portfolio
relative to broader goals.

What results should dictate

firing a manager?
Since the financial crisis, we have seen a number of
managers with strong long-term track records produce
performance way off the scale in other words
underperformance of more than two times the expected
tracking error. The key question clients should be asking
in extreme circumstances is at what point is the result so
bad they just have to move on. Here, we offer a general
framework for answering that question.
66 towerswatson.com
Some context
Some commentary
This question has arisen in large part on account

of the credit crunch. The much higher market
volatility and the fact that markets have generally
been driven by highly technical factors and
sentiment have combined to produce a large
number of extremes in manager performance
when viewed against their benchmarks (both
on the positive and negative side). Many clients
understandably have been surprised and
concerned by the combination of exceptionally
poor market performance and relative poor
manager performance in some cases. The
question itself is representative of a perfectly
understandable attitude that trustees and other
fund owners should be naturally biased to action.
The philosophy is that if someone in business
produces a poor outcome, they are accountable,
so we should terminate their appointment
and move on. For pension fund trustees to be
action-oriented on behalf of their beneficiaries is
particularly compelling, as there is a legitimate
fear that the beneficiaries would look at any
inaction as inadequate. But, is a bias to action,
the right response in this context?
The principle is widely accepted that interpretation

of past performance is complex. It is essential
that we do some analysis, and the first thing to do
with any performance figure is try to understand
the reasons behind it. Was it a case of very poor
work by the manager?
We think this question sits alongside some other

questions to which we provide answers below:
Is poor performance indicative of poor work
and what investigation is needed to decide?
To what extent is failure to meet targets
an outcome that suggests the termination
of a manager?
What action does extremely bad performance
suggest if it is not termination?
What buy/sell discipline in changing managers
produces the best financial outcomes in the
long run?
Let us make the assumption that you started the

period thinking your manager was skilful. In this
situation it could be argued that there are four
possible explanations for the bad performance.
1 Poor skill. This manager has not been skilful,
either because something has happened to
affect his skill or the original assessment of
that skill was incorrect.
2 Poor decisions. This manager has made some
misjudgements. The manager remains skilful,
but has simply made mistakes. A few bad
judgements should be expected and do not
make a manager bad overall.
3 Poor luck. This manager has made
some sensible decisions which proved
unsuccessful due to events that could
not have been anticipated.
4 Timing. This manager has a number of
positions that have so far been unsuccessful,
but there are reasonable hopes that they
will turn around (this is not possible in some
situations, such as companies that have
been nationalised).
Looked at this way, the appropriate action falls
naturally into place. If 1 applies, the manager
should be fired. In the three other cases, the jury
is still out and changes do not appear to make
much sense unless there are reasons other than
simply a view of the managers skill (for example,
their fit in the manager line-up). It clearly calls
for keeping things under more significant review,
however. In the case of 2, the manager is definitely
under a cloud and needs to regain your confidence.
There is a caveat. You never quite know if the
cause was poor skill, poor decisions, poor luck or
simply timing. The best you can do is perhaps to
put these in order of their likelihood, or where you
see a mixture of more than one factor at work,
then decide on their respective weights. Given that
typically only poor skill should lead to termination,
the decision is still reasonably straightforward.
Why have so many managers

underperformed since the
financial crisis?
Manager performances in 2008 were poor relative
to their benchmarks in many cases. Within bonds,
an unusually high proportion of active managers
underperformed (often through being overweight
non-benchmark assets); within equities a number
of fundamental managers suffered.
The credit crunch has produced highly abnormal
conditions for all managers. Most investment
processes analyse securities through the lens
of fundamental values. They assess future
income. But since the credit crunch the
valuations of securities have reflected a
number of non-fundamental factors like:
Lack of liquidity
Securities caught up in forced deleveraging
Securities affected by government interventions
A flight to (perceived) quality
Other aspects of sentiment, in which fear has
played a big part
Could managers have adjusted their preferences

to these conditions? We do not think that this was
ever a realistic or, possibly, even desirable change.
Many of these special considerations have been
impossible to anticipate, have not been part of the
managers natural philosophy and have involved
taking a dangerously short-term view. Furthermore,
the poor liquidity and increased transaction costs
have meant any change of stance would have been
very expensive to implement, for some managers
prohibitively so.
This account above does not absolve all
managers of responsibility for their poor
performance far from it and it will be the
managers whose underperformance is extreme
that have most to account for. Indeed one of the
key questions managers need to be able to answer
is whether their philosophy and process have been
designed simply to work in an environment like
the one that has been in place over the last 20
years, or whether they believe the approach is
sufficiently adaptable regardless of the long-term
market environment.
We should also be careful about exceptionally poor

past performance triggering manager changes which
are destabilising.
68 towerswatson.com
What about when the performance

is off the scale?
The right response is that you have to investigate
more deeply. When we refer to off the scale, we
have in investment the concept of tracking error
to guide our thoughts. It leads to the common rule
of thumb that underperformance worse than twice
the tracking error is off the scale. However, that
rule of thumb needs lots of qualifications.
First, it is important to note that through chance
alone one should expect underperformance
worse than twice the tracking error around
one year in 40. Second, statistics in investment
are often badly behaved. Tracking error is a
backward-looking measure, which will be ill-suited
to unusual investment conditions. That does not
lead us to discard the measure, as it does a rough
and ready job in bringing some semblance of
clarity to bear on a subject that would otherwise
be unworkable. Nevertheless, if we have a twice
tracking error result, there is work to be done to
decide why it has occurred. Most likely it will be
caused by one or more of the following:
A Chance (although as mentioned above, the
likelihood of this is relatively low).
B A so-called fat tail result, where circumstances
have made a number of managers look like
this because of their style or convictions of
the moment.
C A result attributable to one or more significant
positions going against the manager (which may
suggest a failure of risk control).
D A result from leverage in the individual strategy,
or more likely, the unwinding of leverage in
stressed conditions.
We should also be careful about exceptionally poor

past performance triggering manager changes
which are destabilising. So the assessment of a
managers underperformance will need to cover
those factors which could produce a down-rating
of the view about skill:
Client losses producing damaging business
changes, reduced resources, inappropriate
leadership changes.
Style responses that compound the problems,
losing nerve at a critical moment.
That said, we would expect well-managed
investment organisations on the whole not to react
in the wrong sorts of ways. All managers should be
conditioned by the idea that they will underperform
at times. All managers have had some experience
of this condition. No well-managed business
makes itself too vulnerable to performance
down turns. All of these factors should provide
some sort of protection so that a skilful manager
with poor performance can rebound.
All this guides us to the rational view that
performance should not be a firing issue on its
own. But this thinking risks missing an important
emotional aspect of this extreme situation.
When past performance is extreme, a clients
exasperation will be real and large given the
large loss (noting that it may be permanent or
temporary, but it is far from clear which). Our
instinct is to have a bias to action. We are all
pressured by the thought that if this does not turn
around we will be in a worse place. Yet we should
take care with this reaction as it turns out that a
bias to action generally loses the fund more than
it gains.
Can these be sorted by our 1 to 4 division earlier?

The interpretations of C and D might well suggest
a failure of skill which should call for termination.
The others appear more likely to fall in the
jurys out response.
Is there a period of underperformance

that should always lead to termination?
Most people would argue that it seems reasonable
that managers should be judged over five-year
periods. That is certainly a period in which skill
should generally emerge, but there is no law
that it will. Five years is arbitrary and if used too
slavishly it will produce premature judgements with
resulting penalties. Really there is no substitute
for assessing managers on both shorter- and
longer-term measures.
What buy/sell discipline in changing

managers produces the best financial
outcomes in the long run?
The Goyal and Wahal research study in 2005 1
covered a very large number of US pension
plans and their hiring and firing decisions.
It demonstrated quite clearly the difficulties of
changing managers. This picked out very strong
evidence that investment committees were on
average firing underperforming managers who
subsequently saw their performance rebound.
The record of hirings on average added a
little value, but not sufficient to overcome the
conclusion that the funds would on average have
done better to stay put.
The research is suggestive of the best process we
can use to hire and fire managers. It incorporates
these rules:
Hire in numbers to get diversification
and limit the impact of any single
managers underperformance.
Hire based on a prospective view of skill,
track record comes well behind.
Fire when the evidence mounts that the
manager has lost his skill advantage,
again track record comes well behind.
What
is past is past. The manager

line-up should always be managed
with respect to the future.
What beliefs should trustees have?

While there is a need for trustees to review a
list of beliefs for themselves, we think this is a
concise set to build upon:
Past performance does not tell you much directly
about skill, but it does help inform the debate
about skill.
Manager line-up decisions need to be based on a
prospective assessment of future performance.
The psychology displayed in investment
committees is often too emotionally engaged,
which suggests the need to diversify across a
number of managers to reduce the distress from
any single managers results.
For a more advanced set of beliefs that
Towers Watson employs in extreme conditions,
you might consider these too:
Tracking error, while broadly indicative of
return dispersion, is highly state dependent,
and so will tend to understate the probability
of extreme events.
At times of change, alpha (manager
outperformance) may be extremely
difficult to anticipate.
In stressed conditions (when asset prices are
materially impacted by shorter-term credit and
solvency perceptions rather than by longer-term
fundamentals), manager alpha may suffer in the
short term with pricing uncertainties, but alpha
is likely in subsequent periods to undergo some
catch-up if those concerns dissipate, as has
often been the case in the past.
In stressed conditions, there are likely to be
short-term correlations between alpha and
beta (market returns).
And the answer to our question at what point
is the result so bad we just have to move on?
Performance is not a necessary or sufficient
condition to fire a manager, however bad it is.
Only when the past performance is clearly
happening alongside a loss of skill or a failure
of risk control do we reach a point of no return
or when a managers results produces negative
changes to business or style.
What is past is past. The manager line-up should
always be managed with respect to the future.
References
1
G oyal, A. and Wahal, S. The Selection and Termination of Investment
Management Firms by Plan Sponsors, 2005.
70 towerswatson.com
Further information
For further information, please contact your usual
Towers Watson consultant or:
James MacLachlan
Global Head of Equity Manager Research
+1 212 309 3876
james.maclachlan@towerswatson.com
Fabio Cecutto
Senior Investment Consultant
+1 212 309 3867
fabio.cecutto@towerswatson.com
About Towers Watson

Towers Watson is a leading global professional services
company that helps organisations improve performance
through effective people, risk and financial management.
With more than 14,000 associates around the world, we
offer consulting, technology and solutions in the areas
of benefits, talent management, rewards, and risk and
capital management.
Towers Watson
21 Tothill Street
Westminster
London
SW1H 9LL
This document was prepared for general information purposes only and should
not be considered a substitute for specific professional advice. In particular,
its contents are not intended by Towers Watson to be construed as the
provision of investment, legal, accounting, tax or other professional advice
or recommendations of any kind, or to form the basis of any decision to do
or to refrain from doing anything. As such, this document should not be relied
upon for investment or other financial decisions and no such decisions should
be taken on the basis of its contents without seeking specific advice.
This document is based on information available to Towers Watson at the date
of issue, and takes no account of subsequent developments after that date.
In addition, past performance is not indicative of future results. In producing
this document Towers Watson has relied upon the accuracy and completeness
of certain data and information obtained from third parties. This document
may not be reproduced or distributed to any other party, whether in whole
or in part, without Towers Watsons prior written permission, except as may
be required by law. In the absence of its express written permission to the
contrary, Towers Watson and its affiliates and their respective directors,
officers and employees accept no responsibility and will not be liable for any
consequences howsoever arising from any use of or reliance on the contents
of this document including any opinions expressed herein.
To unsubscribe, email eu.unsubscribe@towerswatson.com
with the publication name as the subject and include your
name, title and company address.
Copyright 2014 Towers Watson. All rights reserved.
TW-EU-2013-34601. January 2014.
towerswatson.com

Equity Investing Insights Into A Better Portfolio

Uploaded by

Copyright:

Available Formats

Equity Investing Insights Into A Better Portfolio

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

Equity Investing Insights Into A Better Portfolio

Uploaded by

Copyright:

Available Formats

What are the main topics covered in the document?

What are the main topics covered in the document?

What are some strategies discussed for constructing equity portfolios?

What are some strategies discussed for constructing equity portfolios?

Equity investing:

Insights into a better portfolio

equity make it a likely

Equity investing: Insights into a better portfolio 3

Ever since trading began in 17th century coffee houses,

Nevertheless, the unique qualities of equity

Equity investing: Insights into a better portfolio 5

Investors now have more options than

Best-in-class equity structure

believe this is a good time for

Over the last decade or so, there has been an

Evolution of equity investing

Figure 01. Number of low volatility equity products over time

Source: eVestment Alliance

The main objective of smart beta is to capture

Equity investing: Insights into a better portfolio 9

Importance of portfolio construction

The need for multiple products

Smart beta and portfolio construction

Asset owners need to take a holistic view when

Figure 02. Benefits of diversification in a manager structure

 5% underperformance  7.5% underperformance

Equity market returns do not follow the standard

 Observed S&P500 distribution  Normal distribution

In reality, far greater diversification is

Figure 03. Daily price movements

Active managers have a wide variety of investment

Portfolio construction considerations

increasing rapidly, meaning expertise is also

Figure 04. Portfolio construction principles

Holistic view of return

Integration of smart beta

Low-cost solution allows better use of fee budget

Combining smart beta and multiple alpha strategies to achieve more

Risk and scenario analysis

Risk management framework in place

Timely decision making

Capitalise on market volatility and act decisively to take advantage of opportunities

Contrarian hiring and firing

Continuous and dynamic

Ongoing monitoring of the underlying strategies and the combined portfolio

Focus on best-in-class opportunities

Equity investing: Insights into a better portfolio 11

Expanding further on some of the key points in

In addition, the asset management firm is often

Cumulative relative returns

Figure 05. Divergence of style returns

The impact of using

Source: Style Research, Towers Watson

Complexity and governance

Figure 06. Internal skill central to building high

Whilst there are greater expected rewards from this

Asset owners should determine what level of

Figure 07. Impact of governance levels

Bulk beta, smart beta, alpha

5% underperformance 7.5% underperformance

Observed S&P500 distribution Normal distribution

% of companies in universe MSCI EM weight

Share of state companies in market capitalisation

% of companies by listing (excluding A-shares)

Frontier markets EM small capitalisation EM