White Paper - HSBC Global Asset Management

Transcription

White Paper - HSBC Global Asset Management
White Paper
June 2015
The next stage in smart beta investing
Harnessing factor purity
Authored by:
Alexander Davey, Director, Senior Equity Product Specialist
The next stage in smart beta investing
Harnessing factor purity
Background: The search for diversification
3
Traditional beta, factor beta, and alpha
5
Factors and persistent risk premia
6
Styles as a driver of returns
7
Factor investing
8
Factor risks and factor efficiency
8
Harnessing factor purity
11
Conclusion
14
Executive summary
Relying on the premise that markets are inefficient,
factor investing aims to take advantage of certain
anomalies to potentially achieve excess returns
relative to a cap-weighted benchmark. While factor
exposures have been available through active
management for a long time, they can now be
accessed in a systematic, rules-based manner with
lower cost smart beta strategies.
Further, we think that combining pure factors enables
investors to capture those risk premia that are
persistently rewarded, and to mitigate the effects of
unwanted residual exposures, or ultimately, to create
complementary factor combinations.
Therefore, if style factors have a natural place within
the investment toolkit, the increased control and
capacity to engage with the market cycle should look
attractive to investors, especially when added to the
advantages of periodic rebalancing and the lower
implementation costs of smart beta strategies.
The challenge for the investor is to identify which
factors can offer persistent risk premia, and which
can be combined to take advantage of different
market cycles. In looking at the use of factors, it
becomes clear that investors should have a welldefined framework for determining which strategies
they wish to deploy, before delving into the
methodologies and construction behind the products
that may appear to match their objectives.
In particular, an increased focus on the efficiency of
factor exposures and smart beta construction
methodologies should lead investors to favour
portfolios that have been built to exclude unrewarded
risks. We believe that strategies offering such ‘pure’
factor exposures represent potentially better
solutions for the investor.
2
Background
The search for diversification
As of the end of March 2015 there were USD554
billion of assets invested in smart beta strategies
globally.1 This is set against the backdrop of an
expanding passive market share and an increasing
level of granularity across the smart beta products
now available to investors.
This research concluded that systematic risk factors
were the issue and that, more broadly, exposure to
these factors was a positive thing for long-term
investors as it enabled them to earn a risk premium.
Today, what has become increasingly important to
investors is the question of how to take advantage of
factors and factor combinations across various
market cycles.
Investors may be concerned that “smart beta” will
require them to disregard what has gone before and
to take on a new set of precepts or values. In fact,
“smart beta” is simply about accessing the factors
identified by Ang et al., and others (factors that have
been present in the investment industry for decades)
and to benefit from the specific risk premia they offer.
So how is this relevant to smart beta?
A search for diversification
One may argue that the continual search for
diversification lies at the heart of many of the
developments in the investment industry over the last
25 years. The reaction to the market drawdown of
2000-2001 was a shift towards alternative assets –
believed to be additive to performance and less
correlated to equities and bonds. These included
hedge funds, commodities, infrastructure and private
equity, for example.
The smart beta index or strategy construction
methodology thus enables investors to potentially
achieve excess returns over the long term relative to
a capitalisation-weighted benchmark, within the
framework of passive, cost-effective, rules-based
products.
Significantly, in recent quarters investors have been
focusing on which of these styles or factors are
driving these excess returns and whether returns are
due primarily to factor bias or stock selection, or
some combination of the two.
In many instances the financial crisis of 2008 showed
these investments to be more correlated with
traditional equities and bonds than had been
expected. This was perhaps attributable in part to the
macroeconomic picture – i.e., a function of low
interest rates and economic growth.
To analyse this, a majority of products in the smart
beta universe can be classified under three core
groupings, making it easier to compare various
product designs or investment processes within each
group:
Additionally, the smaller size of areas such as private
equity or some of the hedge fund strategies
generated benefits through the illiquidity premium
they earned. The influx of investors eroded these
benefits and then exacerbated illiquidity as many
headed for exits at the same time. In the wake of
this, asset allocators often began classifying
investments based on the underlying exposure:
private, public, long/short or market neutral strategies
were ultimately all holding equity, and the search for
diversification took a new direction.
 Fundamentally weighted
 Portfolio construction/reweighting (equally
weighted/GDP weighted)
 Risk or factor based – low volatility, quality, value,
growth, yield, size, momentum
The central tenet of all three groups is the use of
rebalancing, a theme that lies at the heart of smart
beta strategies. Where index funds offer an investor
an elegant means of capturing “the market” as
defined by price, with the exception of index changes
at quarter end (or company insolvency, mergers and
acquisition or privatisation) they have no fundamental
weight or base to return to. Price is that base and,
importantly, systematic (or market) risk is diversified
as a result of a wider number of stocks being
included in the index.
Investors began looking in far greater detail into a
principle researched as early as in the 1970s and
revisited by various academics thereafter: that of
separating systematic factors or risk premia. At the
vanguard of this was a study by Ang, Goetzmann
and Schafer for the Norwegian Government fund to
analyse its poor performance.2
1 Source:
Morningstar. This figure only includes mutual funds and
ETFs. It is thus reasonable to suggest that AUM including
segregated mandates could be considerably higher, in the region of
USD700 billion.
2 Ang, Goetzmann, Schafer, 2009, Evaluation of Active
Management of the Norwegian Government Pension Fund - Global
3
Exhibit 1: Weight of largest companies (by market cap) in local indices
Weight of largest company in local indices
Weight of top 5 largest companies in local indices
Spain
57.42%
Germany
IBEX 35
43.10%
France
DAX
35.45%
UK
20%
10.75%
FTSE 100
8.56%
0%
9.80%
CAC 40
23.69%
USA
17.65%
6.36%
DJ US Total SMI
40%
60%
80%
3.17%
0%
5%
10%
15%
20%
Sources: Factsheets – IBEX 35 (at 20.03.2015), DAX, CAC 40, FTSE 100, Dow Jones US Total Stock Market Index (all at 31.03.2015).
For illustrative purposes only.
However, as illustrated in Exhibit 1, the dynamics of
price can do strange things to an index – individual
stocks can dominate as momentum and exuberance
decouple from standard measures of value. Investors
have increasingly worried that their systematic risk is
anything but diversified as single stocks or sectors –
and with it the relative risk of ownership – dominate
an index.
96% of the random portfolios outperformed the capweighted index, unequivocally demonstrating that the
simple act of rebalancing has the capacity to add
substantial value.
However, at the heart of the debate on primate stockselection skills is how much of the excess return is
actually attributable to factor exposures of the market
and how much to innate primate skill. The work of
Fama and French in 19934, and Carhart in 19975,
provided a factor framework for decomposing excess
returns and indicating those derived from exposure to
market, size, value and momentum versus those that
were a result of stock selection skills. In analysing
the return of many active strategies it became clear
that factor returns did explain the bulk of
performance.
In contrast, rebalancing requires the investor to
define a desired asset allocation and to periodically
realign the portfolio to continue matching this
allocation. An active manager aims to use skill as his
medium to buy and sell at an optimal time and by
extension rebalance the portfolio to achieve this. The
more traditional stock-picker has historically relied on
either a measure of value or target price to initiate
this process. A quantitative fund may have a more
mechanistic approach to construction by trading on a
monthly or quarterly basis around a set of underlying
‘signals’. If an investor is not simply buying and
holding a stock in perpetuity then the system of sale
and re-entry must be a medium for adding value:
rebalancing must help drive excess return and
reduce risk.
Regardless of implementation, while the concept of
rebalancing a portfolio is not new, its systematic use
as a driver of return is more nuanced. In their paper
of 2012, Arnott, Hsu, Kalesnik and Tindall3 analysed
the excess return generated by the proverbial dartthrowing monkey using 30 randomly selected stocks
from the Wall Street Journal and they regularly
rebalanced these stocks over a 47-year period.
Arnott, Hsu, Kalesnik, Tindall, 2012, The Surprising “Alpha” from
Malkiel’s Monkey and the Upside-Down Strategies
4 Fama, French, 1993, Common Risk Factors in the Returns on
Stocks and Bonds, Journal of Financial Economics, 33
5 Carhart, 1997, On Persistence in Mutual Fund Performance,
The Journal of Finance, 52
3
4
Traditional beta, factor beta, and alpha
Robert Schiller popularised the theory of market
inefficiency, based on the idea that market prices are
noisy and excessively volatile relative to
fundamentals. In particular, Schiller suggests that
these noisy prices fluctuate or “mean revert” around
fair value.
In contrast to beta, the search for factor tilts and for
alpha rely on the same premise, namely that markets
are inefficient.
Exhibit 2: Types of equity exposure
If investors believe markets are “inefficient,” the
discussion moves to selecting a strategy to take
advantage of this. Traditionally, there are three key
contributions to the return equation which can be
identified as passive (traditional beta), “factor-based”
beta and stock selection, or alpha. (See Exhibit 2.)
Traditional
Beta
Smart
Beta
Active
Fundamental
Stock
Selection
ALPHA
Investment decisions are often factor-led. Smart beta
strategies enable investors to access factor tilts in a
systematic, rules-based manner. This has resulted in
much greater transparency as to the sources of
returns.
FACTORS
(Explicit)
FACTORS
(Explicit)
BETA
BETA
RETURNS
FACTORS
(implicit)
Undoubtedly, many talented managers focus on stock
selection; and the rise of unconstrained mandates and
hedge funds has allowed investors to focus on the
“alpha” part of the equation and even to disconnect it
from “beta” altogether. Indeed, at an extreme, we see
the concept of “portable alpha” where uncorrelated
excess returns generated by manager skill can be
placed over an unconnected beta index to achieve
outperformance in a reliable manner.
BETA
Source: HSBC Global Asset Management, May 2015.
For illustrative purposes only.
5
Factors and persistent risk premia
A common explanation of value, size and momentum
is based on the idea that if a group of stocks has
higher returns than the broad market, then these
stocks must have some systematic risk characteristic
that makes them less desirable than the “average”
stock. To induce investors to hold them, they must
have higher returns than the average stock (i.e. they
carry a risk premium).
Exhibit 3 indicates several points:
 Holding the market capitalisation-weighted index
is not a bad option, offering a reasonable
risk/return ratio, albeit one that is lower than other
factor exposures and fundamentally weighted
smart beta indices
 An equally weighted index may provide a bias to
the smaller size factor which shows good longterm performance, but may prove more volatile in
shorter timeframes
Schiller would argue that if prices are noisy, then a
consequence of these fluctuations is that value stocks
(i.e. the shares of a company whose market value is
low relative to a fundamental measure of the
company’s intrinsic value) are more likely to be
trading below fair value than the shares of a company
which is expensive relative to such a measure. The
correction of this under-pricing over time is what can
boost the returns of value stocks.
 The low volatility anomaly is illustrated using a
minimum volatility index
 The fundamentally weighted HSBC Global
Economic Scale Index provides limited style bias
and shows volatility that is comparable to market
cap indices, while compounding healthy excess
returns
Similarly, if one divides the universe of stocks into
small caps and large caps based on the market value
of the issuing company, then the small cap universe
will statistically contain more undervalued than
overvalued companies. Conversely, the large cap
universe will contain more overvalued companies
than undervalued companies. The correction of
pricing errors will boost the returns of the small cap
stocks and drag down the returns of the large cap
stocks.
Exhibit 3. Risk/ Return of MSCI ACWI relative to the
MSCI Value Weighted, Equal Weighted (small cap
bias), and Minimum Volatility Indices and the HSBC
Economic Scale Index Worldwide*,
from December 2001 to March 2015
Return %
12
It is this kind of insight that investors can leverage to
drive returns using style factors. Identifying the factors
that have persistence can enable them to take the
process further.
MSCI
Equally
Weighted
10
A caveat exists, which is that the historical starting
point of the data sample can have a meaningful
impact on the relative attractiveness of one style
versus another. As such, using MSCI indices, we will
look at the long-term performance of value, low
volatility, equally weighted (small size) and
fundamentally weighted factors.
HSBC
Economic
Scale Index
Worldwide
MSCI
Minimum
Volatility
MSCI Value
MSCI ACWI Weighted
Cap
Weighted
8
Volatility %
6
10
12
14
16
18
Source: HSBC Global Asset Management., Bloomberg,
Datastream, MSCI Barra, All returns are monthly total returns in
GBP (with gross dividends re-invested) from 31 December 2001
to 31 March 2015. HSBC Economic Scale Index data prior to 15
June 2012 is back tested (simulated) data calculated by the
independent calculation agent, the Euromoney Indices team. Data
subsequent to the Index launch date has been calculated daily by
Euromoney
*HSBC Economic Scale Index Worldwide is a fundamentally
weighted smart beta index (using GDP contribution).
6
Styles as a driver of returns
The Reinhart and Rogoff book, “This Time Is
Different,” has much to tell us regarding the negative
impact of investor behaviour across all manner of
market manias, panics and crashes:6 by failing to
properly define their current market environment,
investors may make their allocations at precisely the
wrong point in a market cycle.
By implication, a rebalancing strategy would have
recovered as well. In the right conditions, not only
can rebalancing help drive positive returns, it can
also create a compounding effect to further enhance
returns.
In looking at the use of styles, it becomes clear that
investors must have a robust framework for
assessing which strategy to deploy. As a corollary,
they must then look closely at the construction of
available indices or products that match their
strategy.
As we start to break down returns over more
meaningful time periods, or perhaps market regimes,
we may be able to define a picture of when particular
styles will be most in favour. What is evident is that
post-2007, the divergence of factors has made this
decision-making far more relevant, enabling
investors to capture added returns. We will examine
this further in the following section.
If an investor considers various stock market metrics
such as price-to-book or dividend ratios, it is intuitive
to think that a resulting portfolio will have a built-in
bias to value, and potentially to small cap stocks as
well. This may be preferable at certain times.
The fact that the behaviour of factors is not
synchronised across regions can be very
interesting if we take a global investment
approach. Factors can then be a powerful tool to
realise excess returns from each region
depending on where it is in the market cycle. In
fact, the larger the playing field, the more options
can be explored.
6 Reinhart,
Rogoff, 2009, This Time Is Different: Eight Centuries of
Financial Folly, Princeton University Press
But what about the other major driver in smart beta,
i.e. rebalancing?
We know it is critical to driving returns, but should we
exercise caution around its use? The answer actually
lies less in its use than in understanding when it may
expose investors to periods of underperformance
against a cap-weighted benchmark. In the tech boom
of 1998-2000, when prices were the key determinant
of market performance, a fundamental strategy
based on stock value and rebalancing away from
“winners” would have underperformed significantly.
This is because the rebalancing framework
essentially works as a contra trade, by redeeming
from holdings that have increased in value and
buying into those that have fallen. If all prices are
driving upward, then this methodology will fail to keep
up.
However, it should be remembered that such periods
are transient and often ‘bubble-like’ in nature. Indeed,
in the traditional active investing community at the
time, “value” investors took a lot of criticism from the
proponents of a “new paradigm shift.” When reality
hit in 2000, the technology bubble hurt. The stock
falls were painful for all market participants, but as
markets recovered so did the value strategy.
7
Factor investing
Factor risks and factor efficiency
Much has also been written on the concept of
potentially crowded trades in the lower
volatility/minimum variance sector.10 Certainly, it is
very important that investors understand how these
portfolios are constructed, what inputs are used to
create them, and what the larger stock holdings are.
The behaviour of some of these lower volatility
indices in response to the taper tantrum of April-May
2013 suggests that the rationale for ownership has
changed investor behaviour towards these stocks.
If academia has provided a robust framework for
analysing returns, it has also provided further insights
into styles or premia that appear to have persistence
in generating excess returns. We have observed that
the bulk of investor interest in the smart beta market
focuses on certain factors which expand on the initial
model devised by Fama and French:7
 Value
 Market
 Size
 Quality
 Momentum
 Income
 Low Volatility
For example, many income stocks belong to
relatively dull and potentially undervalued industries
such as utilities or tobacco companies, for example.
These stocks have traditionally delivered several
benefits: lower volatility relative to the market, higher
(and stable) dividends and lower valuations. They
have as a result become widely owned by a broad
range of investors – theoretically for differing reasons
and in differing factor portfolios including lower
volatility.
The market capitalisation-weighted index is seen as
representative of the ‘market’ premium and, as such,
should enable an investor to capture the equity risk
premium. However, using market capitalisation as a
weighting methodology can be stylised using
momentum and large size as factors.
The investor seeking lower volatility stocks can
therefore find these stocks held by other investors—
those chasing yield or those riding the stock’s
valuation. The latter two investor categories have
more capacity to sell quickly and/or more frequently,
potentially driving up volatility. Are investors exposed
to the factor they expected – lower volatility – or are
they inherently accessing value and income as well?
On the other hand, the first generation of smart beta
indices focused on the benefits of rebalancing and
further posited that small size and value are the more
rewarding risk premia over time. These style effects
were embedded – typically through the weighting
method – into the construction of the index and
delivered with rebalancing as a package.
This raises the question as to how “pure” investors
expect their factor exposures to be, especially when
they may ultimately be blended with other factors in
the end portfolio. One could argue that early smart
beta adopters may have been more focused on
capturing the positive effects of rebalancing and
considered factor exposures as a by-product of this.
Today, with the increased granularity of the market it
is apparent that understanding the true make-up of a
strategy is of critical importance.
A healthy debate can be had as to whether both
small size and low volatility are risk premia in their
own right or anomalies that investors may exploit –
the essence being that they perhaps do not hold the
long-term premium of other factors. Indeed, research
exists to support the assertion that, after the original
paper by Rolf Banz8 in 1981 on the ‘small size effect,’
this long-term premium has decreased or has been
removed.
Haugen and Heins9 characterised low volatility
stocks as an anomaly, initially using data from 1926
to 1971 to support their theory, and going on to
conduct various subsequent studies to cover the
following 40+ years. The rationale for this continues
to fill many academic column abstracts, but is not the
focus of this paper. The plethora of minimum
variance, low volatility, and managed volatility
products that have been launched suggests that
asset managers and investors have embraced the
concept.
7 Fama,
French, 1993, Common Risk Factors in the Returns of
Stocks and Bonds, Journal of Economics, 33
8Banz, 1981, The Relationship Between Return and market Value
of Common Stocks, Journal of Financial Economics, 9
9Haugen , Heins, 1972, On the Evidence Supporting the
Existence of Risk Premiums in the Capital Market
10See for example: Davey, Tong, 2015, Factor investing: smarter
than beta - A Low Volatility case study, HSBC Global Asset
Management
8
Exhibit 6: Decomposition of Total Active Risk in a Crosssectional Factor Risk Model
In their work for the Journal of Portfolio Management,
Amenc, Goltz et al., explored the concept of
rewarded and unrewarded risk in the context of
capitalisation-weighted portfolios and various
generations of smart beta products.11
Total Active Risk
Market cap indices are sub-optimal on the basis
that they are often over-concentrated and have
embedded style exposures (large cap,
momentum/growth) which are not controlled.
This reduces the overall risk/return ratio
compared to a portfolio constructed so as to
exclude unrewarded risks and/or unwanted
factors.
Systematic Active Risk
Style
Industry
Group
It is our belief that the increased focus on both the
efficacy of factor exposures and what really lies
under the hood of many smart beta strategies will
lead investors to gain much purer exposures.
Investors would be best allocated to portfolios where
unrewarded risks are reduced or removed as far as
possible and where rewarded risks are diversified.
This assertion provides an interesting litmus test for
many of the smart beta products in the marketplace.
The inability to address both sides of the statement is
not a failing for a product but it should indicate to
investors that it may come associated with some
additional risk.
Country
Currency
Source: HSBC Global Asset Management, May 2015.
For illustrative purposes only.
Using this rationale, the authors highlight the
disparity between active exposure and factor
efficiency; they show that an index with high
exposure to a particular factor does not necessarily
have high factor efficiency.
In a 2014 paper, Hunstad and Dekhayser12
addressed this issue by proposing a new measure
called the ‘Factor Efficiency Ratio.’ Their contention
is that the drive for methodological transparency and
simplicity is not always compatible with factor
efficiency. As has been noted within academic
literature,13 a common smart beta strategy such as
minimum variance can often come with unintended
time-varying exposures. The basic Factor Efficiency
Ratio is shown below:
𝐹𝑎𝑐𝑡𝑜𝑟 𝐸𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 𝑅𝑎𝑡𝑖𝑜 =
Specific Active Risk
AR 𝐷
AR − AR 𝐷
The numerator in the equation relates to the sum of
Active Risk Contributions while AR is the total Active
Risk of the portfolio. For a single factor this would be
calculated as follows:
𝐹𝑎𝑐𝑡𝑜𝑟 𝐸𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 𝑅𝑎𝑡𝑖𝑜𝑉𝑎𝑙𝑢𝑒 =
AR 𝑉𝑎𝑙𝑢𝑒
AR − AR 𝑉𝑎𝑙𝑢𝑒
11
Amenc, Goltz, Lodh, Martellini, 2014, Towards Smart Equity
Factor Indices: Harvesting Risk Premia without Taking
Unrewarded Risks, Journal of Portfolio Management
12 Hunstad, Deckhayser, 2014, Evaluating the Efficiency of ‘Smart
Beta’ Indexes, Working Paper
13 Goldberg, Leshem, Geddes, 2013, Restoring Value to Minmum
Variance, forthcoming in the Journal of Investment Management
The factor efficiency ratio is by construction an exante ratio which attributes the risk a portfolio hopes
to deliver into desirable and undesirable active risk.
9
Exhibit 8: Decomposition of Total Active Risk for
an ETF based on the MSCI World Enhanced
Value Index
Exhibit 7 below makes this comparison more
explicitly by showing the decomposition of active risk
for a pure/smart factor index. This decomposition is
the common risk attribution output from portfolio risk
analytics software. The index14 has been specifically
designed to maximise exposure to the factor (value,
in this instance) and reduce/minimise the effect of
other residual factors. It should be noted that this and
other illustrations focus on long-only indices.
100%
90%
80%
70%
60%
50%
The highest contribution to risk comes, as one may
hope, from style and more specifically, as the bottom
graph demonstrates, from the ‘value’ component.
Country is also a notable risk component and a
further analysis indicates that exposure to Japan
(typically a value market) would explain this.
40%
30%
20%
10%
0%
Style
Countries
Industries Currencies Non-Factor
Decomposition of the Style Components
Exhibit 7: Decomposition of Total Active Risk for
Efficient Value Index
20%
100%
15%
90%
80%
70%
10%
60%
50%
5%
40%
30%
Growth
Size
30%
25%
Positive Contribution
20%
Dividend Yield
Decomposition of the Style Components
Earnings Variability
Industries Currencies Non-Factor
Leverage
Countries
Momentum
Style
Profitability
0%
Trading Activity
Volatility
10%
Value
0%
20%
Negative Contribution
Source: HSBC Global Asset Management, May 2015.
For illustrative purposes only.
15%
10%
This naturally has an important impact on the
behaviour of the factor portfolio. Portfolios that pick
up exposures to multiple factors are more likely to
show higher volatility and variability of return, as
each factor exposure adds to the strategy’s tracking
error.
5%
Dividend Yield
Growth
Leverage
Momentum
Negative Contribution
Earnings Variability
Positive Contribution
Profitability
Size
Trading Activity
Value
Volatility
0%
For example, if an investor has selected value to take
advantage of a specific point in the market cycle,
contaminating this with additional and significant
exposure to size or quality negates both the theory of
investing and, potentially, the likely portfolio makeup.
Source: HSBC Global Asset Management, May 2015.
For illustrative purposes only.
We then compare this to an ETF using the MSCI
World Enhanced Value Index as its underlying15
(Exhibit 8). Value once again explains the style
exposure of the active risk component, but we see
far higher residual influence from factors such as
volatility, size, momentum and growth.
“The index” here refers to an HSBC proprietary value factor
strategy
15 The ETF is the iShares MSCI World Value Factor UCITS ETF
14
10
Factor investing
Harnessing factor purity
Exhibit 9: Annualised excess returns of factors
Annualised Excess Returns
20%
15%
10%
5%
0%
-5%
Value
Small Cap
Momentum
Low Volatility
2014
2013
2012
2011
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
-10%
Quality
Source: HSBC Global Asset Management., May 2015 For illustrative purposes only.
Past performance is not indicative of future returns.
 Analysis of the period post-2007 supports the
theory that individual factors add value versus the
market cap index, but that it is critical to isolate a
factor (i.e. to have factor purity) and to keep other
factors from eroding its excess returns.
This idea that exposure to a pure factor is perhaps a
safer allocation is further driven by reference to the
change in factor behaviours over the last 15 years.
We noted above that by timing an allocation to an
inflexion point it would be possible to add, or lose,
considerable value.
As explored earlier in the paper, there is a third issue
to consider:
Additionally, investors often speak of the cyclicality of
markets. However, when we look at the performance
of individual factors over the period 2001-2007, it is
striking to see the almost complete absence of any
negative style effect. Tilting to pretty much any style
or blend would potentially have achieved excess
returns.
What if the selected factors not only had a clear
exposure to the desired risk premia (e.g. value,
quality, small size, etc.), but also had low correlation
of their excess returns to the other available factors?
In this way investors would have the opportunity to
make meaningful combinations, but on their own
terms and without suffering from the presence of
other embedded factors which have been inherent in
some smart beta products.
This has two important ramifications that investors
may wish to consider:
 Some earlier articles on smart beta observed that
strong excess returns can be generated through
exposure to value and small cap factors. The data
supporting this naturally covered the period before
2007, when a structural tilt to these factors would
have offered very strong performance. However, it
would be trickier to support the same thesis if an
investor concentrated more heavily on the last
seven to ten years of data.
11
Using more widely available MSCI style factor data –
and making no assertion around their factor
efficiency – its possible to see how factors have
performed across longer time horizons.
These exhibits also show the importance of time
period. When we align our graph to include the
minimum variance index (Exhibit 10) we are forced to
start from 1999 (the back-tested start date of this
index), which has the effect of truncating the tech
bubble and adversely affecting the impact of
momentum. In contrast, Exhibit 11 shows the
performance of value, momentum and quality factors
since January 1996: momentum now becomes far
stronger and quality gains at the expense of value.
Exhibit 10: MSCI Factor Indices: performance,
1999-April 2015
1.5
1.4
1.3
Overall, this picture appears to suggest that quality
would be the most effective style or factor to select at
the expense of others. However, when we start to
select specific market regimes – tech bubble versus
the financial crisis, for example – we gain a far
clearer picture of the potential duration of individual
factor underperformance and of the ability of factors
to offset each other.
1.2
1.1
1
0.9
0.8
Value
Momentum
Quality
12/2014
02/2014
04/2013
06/2012
08/2011
10/2010
12/2009
02/2009
04/2008
06/2007
08/2006
10/2005
12/2004
02/2004
04/2003
06/2002
08/2001
10/2000
12/1999
0.7
Minimum Volatility
Exhibit 12: April 1997-March 2000 excess returns
relative to MSCI World Index
80%
60%
Sources: HSBC Global Asset Management, MSCI, Bloomberg,
May 2015. Past performance is not indicative of future returns.
40%
Exhibits 10 and 11 touch on an issue in factor
investing that is likely at the forefront of an investor’s
mind, i.e. the risk of underperformance. Long
timeframes provide an illusion of stability which may
mask prolonged months or quarters of
underperformance.
20%
0%
-20%
A lack of uniformity across the factors shown in
Exhibit 10 suggests that each factor has a different
performance pattern and that one factor may have
the ability to potentially offset the underperformance
of another.
-40%
Value
Momentum
Quality
Sources: HSBC Global Asset Management, MSCI, Bloomberg,
May 2015. Past performance is not indicative of future returns
Exhibit 13: May 2008-June 2011 excess returns
Exhibit 11: MSCI Factor Indices: excess returns
relative to the MSCI World Index 1996-April 2015
relative to MSCI World Index
300%
50%
250%
40%
200%
30%
150%
20%
100%
10%
50%
0%
0%
-10%
Value
Momentum
Quality
Value
Sources: HSBC Global Asset Management, MSCI, Bloomberg,
May 2015. Past performance is not indicative of future returns.
Momentum
Quality
Sources: HSBC Global Asset Management, MSCI, Bloomberg,
May 2015. Past performance is not indicative of future returns.
12
“It is precisely because factors episodically lose
money in bad times that there is a long-run
reward for being exposed to factor risk. Factor
premiums are rewards for investors enduring
losses during bad times.”16
The optimisation of factors could lead to approaches
such as risk parity, dynamic timing or a even a
change to redefine a portfolio benchmark or segment
to accommodate the factor investment activity.
Factors can be combined together overall, or by
region, or different factor combinations can be
selected for different regions in an effort to
outperform a global index. For example, an investor
can design a portfolio combining a momentum tilt on
the US and a combination of value and low volatility
factors in Europe.
For many investors there is a natural appeal for
tactical timing, especially for those who have looked
to use several active managers based on their style
biases. This is precisely where the growing suite of
smart beta products can offer an alternative.
However, as observed, to time factor exposures
dynamically, investors or asset allocation committees
must be confident that they understand the current
market regime. While looking back at historical data
on returns, volatility, and economic data may provide
some comfort, this is ultimately a predictive exercise.
Many institutions and private clients run their
portfolios to long time horizons and liability streams
and, despite the disquiet, these investors can suffer
periods of underperformance to more effectively gain
exposure the risk premia that various factors may
offer.
If factors have a place in an investor’s asset
allocation palette, then the increased control and
capacity to engage with the market cycle should
appeal, especially when considered in tandem with
rebalancing and the lower implementation costs of
the smart beta universe of products.
“Prediction is very difficult, especially if it’s
about the future,” said Nobel laureate physicist
Nils Bohr.
Investors tend to adopt factor-based approaches to
their portfolio in two ways:
As previously discussed, the aim is to maximise the
risk/return ratio of the chosen strategy while
discarding unwanted or unrewarded factors. To
optimise the implementation, a tactical overlay may
be the best option.
 by identifying the factor exposures that are
inherent in their portfolio (a more risk-based
approach) and monitoring these; or
 by integrating the use of factors into the strategic
or the tactical asset allocation process
Of course, not every investor will be aiming for this
level of granularity or control, nor will they wish to
review and interpret a full range of factor
combinations and overlaps. However, for those that
have conducted a detailed review of factor-based
investment products, used style analytics and sought
to create an optimal blend of factor exposures, we
would contend that the simplicity of a pure style – or
a combination of pure styles – in a transparent format
at a lower cost will prove attractive.
The former tends to focus on checking for
unwarranted concentrations rather than maximising
the benefits of a particular factor.
By integrating factor exposures into the asset
allocation process, investors can take more control of
their outcomes. However, this approach does require
a long-term view as to which factors can add a
consistent premium. Additionally, it requires some
thought in terms of the rebalancing of factors. It may
be more accurate to describe this approach in the
context of tactical allocations, or tilts.
16
Factor Investing, Ang 2009
13
Conclusion
?
What becomes increasingly obvious as investors
investigate the smart beta market further is that a
single marketing term covers a myriad of investment
approaches. Acquiring perspective on their origin,
their academic grounding, their similarities and
ultimately the more efficient implementations will help
investors see through the tangle of strategies and
gain clarity to optimise their investment decisions.
Pure factors can then be combined so as to
create complementary exposures, giving
investors increased control with the
transparency and cost-efficiency of a smart beta
approach.
The potential of factor-based investing can thus be
harnessed by discretionary wealth managers, private
banks and advisors seeking specific exposures for
their clients, as well as by pension plans
(government or corporate) and insurance companies
seeking long-term excess returns.
Undoubtedly, as investors ask ever more probing
questions on factor combinations and how various
smart beta strategies aim to add excess returns, it
becomes easier to begin drawing clear distinctions
between strategies and providers.
In particular, favouring portfolios that avoid exposure
to unrewarded risks will increase the efficiency of
factor tilts.
14
Authors
?
Alexander Davey
Ioannis Kampouris
Director, Senior Equity Product
Specialist
Quantitative Research Analyst
HSBC Global Asset Management
(UK)
HSBC Global Asset Management
(UK)
Alexander Davey is the Director of Alternative Beta
Strategies with HSBC Global Asset Management
(UK) Ltd., joining the firm in 2014. He has 17 years
industry experience having held both sales and
investment focused roles with Barclay Global
Investors, Morgan Stanley Investment Management
and most recently Barclays Wealth and Investment
Management.
Ioannis Kampouris is a Quantitative Research
Analyst at HSBC Global Asset Management,
specialising in quantitative strategies research and
development, statistical modelling and application of
advanced optimisation methods. Prior to joining
HSBC in February 2012, Ioannis worked as a
Quantitative Analyst at Gulf International Bank (GIB).
Ioannis holds an MSc in Computational Statistics and
Machine Learning from University College London
(UCL) and MEng in Computer Engineering and
Informatics from University of Patras (Greece).
Alexander has a BA Honours in History from the
University of York and is a full Member of the
Chartered Securities Institute.
Inez Khoo
Quantitative Research Analyst
HSBC Global Asset Management
(UK)
Inez is a Quantitative Research Analyst at HSBC
working on global quantitative strategies and
research. Prior to joining HSBC, she worked in
Quantitative Equity Research at Macquarie
Securities focusing on alpha and portfolio
construction research.
Inez has a Masters in Finance from London Business
School and a BSc in Econometrics and Mathematical
Economics from the London School of Economics.
15
Important information
?
This document is intended for Professional Clients only and should not be distributed to or relied upon
by Retail Clients.
The contents of this document are confidential and may not be reproduced or further distributed to any person
or entity, whether in whole or in part, for any purpose. The material contained herein is for information only and
does not constitute investment advice or a recommendation to any reader of this material to buy or sell
investments. This document is not intended for distribution to or use by any person or entity in any jurisdiction
or country where such distribution or use would be contrary to law or regulation. This document is not and
should not be construed as an offer to sell or the solicitation of an offer to purchase or subscribe to any
investment.
HSBC Global Asset Management (UK) Limited has based this document on information obtained from sources
it believes to be reliable but which it has not independently verified. HSBC Global Asset Management (UK)
Limited and HSBC Group accept no responsibility as to its accuracy or completeness
The views expressed above were held at the time of preparation and are subject to change without notice. Any
forecast, projection or target where provided is indicative only and is not guaranteed in any way. HSBC Global
Asset Management (UK) Limited accepts no liability for any failure to meet such forecast, projection or target.
The value of any investments and any income from them can go down as well as up and your client may not get
back the amount originally invested. Where overseas investments are held the arte of currency exchange may
also cause the value of such investments to fluctuate. Investments in emerging markets are by their nature
higher risk and potentially more volatile than those inherent in some established markets.
Stock market investments should be viewed as a medium to long term investment and should be held for at
least five years.
Any performance information shown refers to the past should not be seen as an indication of future returns. It is
important to remember that these alternative indices do not outperform all the time. In particular in a momentum
driven bubble (such as with technology stocks in the late 90s) share prices can diverge from fair value for an
extended period. In such cases alternative index strategies will underperform capitalisation weighted indices as
rebalancing does not improve returns. However when the bubble bursts and share prices drop back towards fair
value then alternative index strategies are more likely to outperform
HSBC Global Asset Management (UK) Limited provides information to Institutions, Professional Advisors and
their clients on the investment products and services of the HSBC Group. This document is approved for issue
in the UK by HSBC Global Asset Management (UK) Limited who are authorised and regulated by the Financial
Conduct Authority. Copyright © HSBC Global Asset Management (UK) Limited 2014. All rights reserved.
26950/0416ex310716/FP15-0867
16