Asia Risk QIS Special Report 2019 PDF
Asia Risk QIS Special Report 2019 PDF
Asia Risk QIS Special Report 2019 PDF
Investment Strategies
Special report 2019 Risk.net
Sponsored by
Opinion
Outperformance
in an evolving and
growing landscape
Q
Marcel Chambers uantitative investment strategies (QIS) open up a new frontier for
Head of APAC Sales
marcel.chambers@infopro-digital.com
asset managers, allowing investors to acquire diversified returns
thanks to a combination of rigorous financial research and
Stuart Willes
Commercial Editorial Manager
advanced technical capabilities.
stuart.willes@infopro-digital.com This report surveys the QIS landscape, defining the range of investment
Alex Hurrell
strategies under the QIS label and delving into how and why they can
Commercial Subeditor provide outperformance. It also analyses the means by which investors can
alex.hurrell@infopro-digital.com
access these strategies and the challenges associated with gauging their
Philip Harding integrity and benchmarking their performance.
Head of Content, Marketing Solutions
philip.harding@infopro-digital.com
The report also peeks behind the scenes at a leading QIS vendor to break
down the various stages of development that go into making an effective QIS
David Pagliaro
Group Managing Director
product, as well as a leading analytics platform used by the sell side and buy
david.pagliaro@infopro-digital.com side alike to compare and contrast different strategies.
Ben Wood
Quantitative strategies are gaining prevalence among asset managers
Group Publishing Director under pressure to generate absolute returns at lower costs for investors.
ben.wood@infopro-digital.com Traditional hedge fund allocations are out of favour as a result of prolonged
Ben Cornish underperformance and high fees, yet investors still crave the uncorrelated
Senior Production Executive
ben.cornish@infopro-digital.com
returns they can provide. QIS offers these exposures through rules-based
asset allocation mechanisms rooted in well-evidenced financial research and
Infopro Digital (UK) readily explained to investment management boards – unlike often opaque
Haymarket House, 28–29 Haymarket
London SW1Y 4RX hedge fund mandates – and at a fraction of the cost.
Tel: +44 (0)20 7316 9000 This report eliminates any misconceptions about these strategies by clearly
Infopro Digital (US) laying out what they can provide – diversified returns over the long term –
55 Broad Street, 22nd Floor and what they cannot: alpha without risk.
New York, NY 10004-2501
Tel: +1 646 736 1888 Readers will also be better equipped to navigate the growing QIS
inventory. Investors have an abundance of choice when it comes to QIS
Infopro Digital (Asia-Pacific)
Unit 1704-05, Berkshire House products, from vanilla single-factor, long-only strategies to multi-factor,
Taikoo Place, 25 Westlands Road multi-asset and long/short variants.
Hong Kong, SAR China
Tel: +852 3411 488 Defining an investment objective is the first step of any selection exercise,
but determining which QIS is best suited to fulfil this requires a granular
Cover image
agsandrew/Getty understanding of how risk premia derive their returns, the degree to which
these correlate with the market, the nature of their idiosyncratic cycles and
the ways they slot into portfolio construction.
Depending on the objective, an asset manager could choose one of a
number of QIS to replace its hedge fund allocation, or build up a portfolio of
Published by Infopro Digital
strategies with negative correlation to its existing equity and/or credit
© Infopro Digital Risk (IP) Limited, 2019 exposures. Virtually all goals can be pursued through QIS, but asset
managers need to understand the capabilities and limitations of the available
products before putting money down.
The evolution of QIS is accelerating. This report offers a helpful guide to this
rapidly changing asset class and can assist managers in making informed
All rights reserved. No part of this publication may be choices when adding these products to their portfolios.
reproduced, stored in or introduced into any retrieval system,
or transmitted, in any form or by any means, electronic,
mechanical, photocopying, recording or otherwise, without
the prior written permission of the copyright owners. Risk is
registered as a trade mark at the US Patent Office
risk.net 1
Contents
8 Q&A
A sound understanding of key drivers of investment returns As QIS becomes increasingly sophisticated and firms constantly
and defining logical rules are fundamental to ensuring QIS is drive for cost-efficiency, the relationship between buy-side
effective. Edhec-Risk Institute has established itself as a go-to firms and investment banks is being altered. PremiaLab offers
resource for institutions creating or investing in QISs, offering a fresh outlook on asset allocation, providing productivity and
diversified portfolios across and between factors transparency for buy- and sell-side market participants
risk.net 3
Strategy selection
I
nvestors’ desire for diversified sources of return at low costs has powered Yet interest in QIS is growing in the region. In the second quarter of 2018,
demand for QIS. But investors should ensure they understand the nuances one of the largest insurers in China said it was in the process of rolling out a QIS
of these products and ask challenging questions of providers before platform for clients. Others are likely to follow as sources of alpha in emerging
selecting a strategy. markets threaten to diminish.
This report identifies the characteristics of the most popular QIS iterations, These QIS are constructed to perform specific functions. Indeed, the initialism
their utility to asset managers and ways in which they can be implemented is a catch-all for a series of asset allocation tools, including smart beta,
within a portfolio. It also elaborates on certain selection criteria that should be alternative risk premia (ARP) and hedging.
considered by institutions planning to implement a QIS and highlights the risks
they could face on their way. Smart beta
Smart beta represents an elementary form of QIS. These strategies are designed
Defining QIS to systematically select, weight and rebalance a portfolio in line with certain
Buy-side firms are increasingly turning to QIS to complement their portfolios. metrics other than an individual security’s market capitalisation.
However, actual investments are concentrated among US and European money Their objective is to combine the benchmark-beating performance of a well-
managers, with Asian-based investors only recently joining the club. crafted, actively managed portfolio with the low expense and simplicity of an
index-linked allocation. The strategies are typically from traditional market risks such as bonds and The efficacy of a hedging QIS should be graded
long-only and have been embedded in a wide range equity. This is not to say an ARP allocation does not against two criteria: its ability to neutralise an
of exchange-traded funds in recent years. have its ups and downs – risk premia are subject to investor’s exposure to the unwanted specific
A common smart beta strategy is equal cycles of their own – but these may not synchronise risk, and its cost of carry – the expense incurred
weighting. Governed by such a methodology, a with the highs and lows of the market at large. assuming the investment positions dictated by
strategy would invest in each constituent of a Their capacity as diversifiers make ARP strategies the strategy.
chosen index as equally as possible, instead of attractive investments, allowing asset managers An exorbitant cost of carry could outweigh the
allocating more to the securities with the highest to achieve higher expected returns for a given risk hedging benefits of a strategy by depriving an
capitalisation and less to those with the smallest. budget or similar returns for a lower risk budget. investor of the very returns they wanted to protect
This decouples the strategy’s performance from An asset manager may select a desired risk premia in the first place. Efficient hedging QIS therefore
that of the very largest securities, as in a market exposure and have that govern the securities it attempt to achieve the investor’s desired level of
cap-based strategy, instead responding evenly to the buys and sells within a segment of its portfolio. ARP coverage while optimising cost of carry.
price movements of each asset in the portfolio. strategies can incorporate multiple asset classes and,
However, doing so increases the exposure to small unlike smart beta, seek to maximise factor-based Focus on ARP
stocks, which are typically more volatile and vulnerable returns through a mix of long and short exposures. For those looking to generate diversified and
to sudden drawdowns than their larger peers. The higher expected returns, ARP strategies could
superior returns possible through an equally weighted Hedging prove especially attractive considering current
strategy might come hand in hand with higher risks. The third broad category of QIS is systematic trends – particularly when correlations that have
This is by design, not by accident. All smart hedging strategies. These are programmatic held true for a decade or more, and form the basis
beta strategies strive to outperform a traditional allocation methodologies designed to efficiently of investors’ diversification strategies, may be
allocation methodology rather than subvert it. This protect investors from specific risks such as market breaking down.
being the case, their performance remains correlated risk, interest rate risk and tail risk. Take, for example, the equities market, which
to the market at large. Hedging QIS are not intended as yield-enhancing erupted in October 2018 and caught fire in
Therefore, while they make attractive alternatives plays, or a means of diversifying an asset manager’s December, when leading bourses entered correction
to the long-only portion of a portfolio, smart beta sources of return, as with an ARP strategy. Instead, territory (see figure 1): as equities fell, so did
strategies cannot offer a high degree of diversification. they can be used to safeguard returns by dynamically bonds – an occurrence that undermined the
rebalancing a portfolio between risky and risk-free negative correlation of the two asset classes that
Alternative risk premia assets in response to predefined market signals. dates back to around 1998.
ARP strategies, on the other hand, can provide
diversification, yield enhancement and lower
risk exposure. These investments offer access to Figure 1 Rebased performance of equity versus bonds (October–November 2018)
systematic sources of return that are well evidenced
through empirical research and academic study. 105 100.5
Traditional risk premia are well known as the
sources of return over the risk-free rate that can be
achieved through a selected asset class. They include
equity risk premia – returns accessible to an investor 100 100
through buying into the stock market – and credit
risk premia, which reward those who invest in debt.
In contrast, ARP returns are linked to other, discrete
risk premia inherent across asset classes and markets. 95 99.5
Well-documented ARP, identified through
Bonds
Equity
Investing in an ARP strategy offers meaningful Equities and bonds initially fell in unison during this period, with negative correlation returning towards the
diversification as returns are statistically decoupled end of 2018
risk.net 5
Strategy selection
Such findings suggest established correlation assumptions are fraying. For This premium is academically sound. Taking the excess performance of a
asset managers wedded to a portfolio composed of 60% equity and 40% long position in 10-year US Treasuries over the past 20 years, two-thirds of the
bonds, this spells trouble as the diversification that such a mix is intended to performance came from coupon payments to investors – the specific, predefined
provide may not prove robust going forward. reward for carrying US government exposure. The other third came from the
Changes to macroeconomic policy could be responsible for recent market decline in interest rates, a reward for market risk, which cannot be perfectly
disruptions, and signal further turbulence ahead. known in advance of investing.
The US Federal Reserve Board began a process of balance sheet A rates-focused ARP strategy seeks to monetise this carry by buying and
normalisation in October 2017, under which its asset portfolio will shrink by selling rate exposures across currencies and tenors while controlling for
as much as $1.9 trillion over five years. This will draw liquidity out of a market idiosyncratic market risk. Essentially, it makes money by borrowing at a low
that has become used to easy money, and may increase short-term funding interest rate and lending at a higher one.
costs for banks. A practicable implementation of such a strategy would dynamically allocate
Expectations have also changed around the Fed’s interest rate rising cycle. In between a basket of interest rate swaps linked to different currencies at a set
September 2018, the central bank hiked rates for the eighth time since 2015, tenor – say, the 10-year rate. Swaps the trading strategy identifies as having the
with more rises anticipated. A seeming U‑turn and more dovish tone from Fed highest carry after controlling for realised volatility are bought and those with
chairman Jerome Powell in November suggests the Fed’s timing and frequency the lowest are sold, with the difference between payments sent and received
will be determined as the macroeconomic and geopolitical outlook becomes generating the return.
clear. Other central banks followed the Fed moves, with the Bank of England Greater diversification can be achieved by dialling up the number of currencies
and the People’s Bank of China adjusting rates upwards in 2018, but an air of included in the basket, although this comes at the expense of performance, as
uncertainty now exists. The European Central Bank, though it has kept rates the carry premium is eroded by spreading the rate exposure more widely.
static, ended net asset purchases on December 13, 2018. This example is founded on a well-known, academically sound risk premium;
Tighter monetary policy implies lower equity valuations, meaning the engine it can be implemented with simple, liquid instruments – in this case, interest
driving stocks higher worldwide could start to sputter. rate swaps; and its source of return is easily explained to an asset manager’s
It is against this tumultuous backdrop that ARP strategies show their quality. investment board.
They can supplement any portfolio to provide the diversified returns sorely
needed at a time when traditional asset prices appear fragile and correlations Steps to implementation
appear unsound. Identifying a reliable ARP should be a prospective investor’s first priority. In
Indeed, research by global investment firm Cambridge Associates shows recent years, a host of factors have been unearthed by quantitative analysts and
that a balanced portfolio with a 30% exposure to ARP funds would have academics. However, just a handful can be empirically proven to have produced
yielded 12% between January 2007 and March 2009 – the peak of the positive returns. In fact, a 2014 study of 600 factors found that 49% produced
financial crisis – compared with only 3% for a traditional portfolio without zero or negative risk premia.1
this allocation. The stable of risk factors with a sufficiently rich history to recommend them to
As noted, this kind of performance is possible because ARP strategies diversify investors is smaller still.
away unrewarded specific risks – such as broad market risks – and focus on The burden of proving a factor’s efficacy lies with the ARP provider, but it is
delivering returns linked to fundamental risk premia. the responsibility of the asset allocator to decide whether a given amount and
quality of evidence is sufficient to greenlight an investment.
It is also the investor’s duty to understand how a given premium is defined
by a provider, and the process by which related exposures are selected by their
QIS unlock a wealth of possibilities for asset managers. provider. ARP with the same names are designed differently by each provider
and can exhibit wildly different behaviours. For example, ‘quality’ strategies rely
ARP implementations in particular offer the attractive on a variety of metrics to select appropriate securities, including, but not limited
opportunity to generate additional returns and inject to, the issuing firm’s leverage, earnings stability, earnings growth, dividend
diversification into a given portfolio strength and governance. Each provider will use a subtly different selection
and weighting taxonomy to extract the ‘quality’ premium, resulting in variable
return profiles.
Investors therefore need to be aware of the signals a provider considers when
Selecting an ARP strategy selecting assets to populate an ARP strategy, as well as those used to prompt
The utility of ARP strategies is clear. But investment should be predicated portfolio rebalancing. Otherwise, they could accrue exposure to unwanted risks
on a clear understanding of how they are constructed and what drives their or miss out on a premium they want to access. They could also impose excessive
returns. A good strategy should be intuitive and readily explainable, and a trading costs if the chosen signals demand frequent asset turnovers.
worthwhile provider should be transparent on how the chosen premia are
defined and accessed. Backtesting bias
Take the following example of a ‘carry’ risk premium strategy. This is designed A popular tool used by providers to guide investors’ ARP selection is the
to maximise the systematic returns investors receive in exchange for holding an backtest. But investors should approach these with a healthy scepticism and
asset and assuming its associated risks. scrutinise them for hidden biases.
This ARP can be found across asset classes. In rates, for instance, the carry Testing an ARP using historical data from a specific time period reflecting
premium is an investor’s reward for taking the duration, interest rate, liquidity
and credit risks linked with a defined bond position. 1. Levi Y and I Welch, 2014, Long-term capital budgeting,” SSRN
particular market conditions may evidence outperformance, but does not will incur costs in the form of a swap maintenance fee to the counterparty, as
guarantee similar results in the here and now. well as an index fee to cover the expense of the third-party calculation agents.
Take, for example, ‘momentum’ strategies: the quantitative methods used
may accurately select high-momentum securities over the backtesting period, The ARP difference
but could deteriorate in live performance as the signals may have been Prospective investors should be wise to the differences between ARP strategies
optimised for that specific timeframe, instead of being designed for robustness and traditional hedge fund allocations.
in all market conditions. Active managers and ARP strategies both seek to extract returns from selected
To avoid buying into ARP strategies that have been overfitted to certain data, risk premia. However, the latter do so systematically in accordance with preset
investors should check backtests for these kind of biases, and demand they cover rules hard-coded into their governing algorithms. Unlike actively managed
extended timeframes embracing a variety of market regimes. strategies, an ARP is free from human discretion – and interference – which
means an investor always has perfect knowledge of how their money is being
Forms of implementation deployed, and why.
Managers have a range of means to gain exposure to ARP strategies, although Best practice among ARP providers is to outsource the administration of
not every wrapper is appropriate for all users. a strategy wholesale to an independent third party, further strengthening
Cash-rich managers can opt to invest in a dedicated ARP fund, thereby its integrity.
gaining real exposure to the assets captured by the strategy. These are subject Naturally, conflicts of interest arise where a bank owns and operates a
to management fees and potentially liquidity fees and redemption gates, strategy that it sells to clients. If a bank’s traders are privy to when an ARP
which may obstruct an investor’s ability to withdraw cash at will. In addition, algorithm buys and sells, they have the ability to front-run the transactions and
investors do not have the option of altering the strategy once committed, as depress the strategy’s returns accordingly, to the detriment of the end‑investor.
implementation is fully outsourced to the fund manager. The presence of an independent calculation agent, alongside the provision of
Strategies can also be embedded within structured notes or certificates. These a comprehensive strategy disclosure document, can minimise these conflicts.
wrappers provide investors with factor exposure under a single Committee on One other consideration remains for prospective investors: the choice of a
Uniform Security Identification Procedures number, and may offer additional benchmark against which to judge an ARP strategy’s performance. The problem
benefits such as a principal guarantee or downside protection ‘buffers’. is that few appropriate benchmarks exist. The purpose of these strategies is to
A structured note wrapper may prove ideal for real money managers offer returns uncorrelated to market risk, meaning that traditional benchmarks,
prevented from investing in synthetics. Although the return of a note is linked including prominent stock and bond indexes, are not fit for purpose.
to the performance of the underlying ARP strategy, the investor is never the Furthermore, as each provider’s definition and means of accessing ARP varies,
asset owner. Instead, the provider runs the strategy in-house and transfers its comparing the performance of one strategy with another can prove fruitless. One
performance to the note by means of a swap. solution is to use an index consisting of all or a large sample of ARP strategies
Money managers permitted to use and comfortably transact derivatives may anchored to a particular factor. This way, an investor can determine whether a
consider accessing ARP through a total return swap or similar instrument. Here, specific momentum strategy, for example, has outperformed or underperformed
the investor gets pure exposure to the return of their chosen strategy minus a the universe of momentum strategies as a whole.
set fee.
This structure requires no upfront cash investment, freeing managers to put
their cash to work elsewhere. Swaps are generally of short duration, but can be
rolled over with changed terms – such as different notionals – offering investors
Identifying a reliable ARP should be a prospective
a flexibility they do not enjoy with other implementations. For certain ARP,
investors can also buy a call option, either in a securitised format or an over-the- investor’s first priority. In recent years, a host of factors
counter derivative, to achieve more leverage. have been unearthed by quantitative analysts and
academics. However, just a handful can be empirically
Costs of implementation
proven to have produced positive returns
An essential consideration when choosing between ARP strategies is cost.
A central appeal of these strategies is that they can deliver returns above
those of the broader market at a lower cost than an active manager or hedge
fund allocation. However, these can be crimped by excessive strategy and Conclusion
wrapper fees. QIS unlock a wealth of possibilities for asset managers. ARP implementations
An ARP provider will typically extract payment for an investor’s right to access in particular offer the attractive opportunity to generate additional returns and
its intellectual property and to cover the costs of implementing and managing inject diversification into a given portfolio.
the investment via a deduction from the strategy’s returns. Potential investors should be wise to the processes that underpin the
Competition between providers has pushed these costs down in recent years, ARP lifecycle, from inception to construction, and favour strategies that are
although multifaceted strategies reliant on a high degree of asset turnover and transparent, simple to explain and underpinned by in-depth, unbiased research
incorporating short positions are typically more expensive. However, an ARP documenting their ability to provide outperformance over the long term.
strategy is still some way cheaper than a typical hedge fund allocation subject to They must also be wary of the implicit and explicit costs of different
a ‘two and 20’ fee structure. strategies, and sensitive to the type of implementation most suited to their
The choice of wrapper will also influence an ARP strategy’s all-in cost. A fund investment mandates.
allocation will attract management fees and transaction costs if and when the An asset manager that keeps these considerations in mind, however, should
chosen strategy signals a portfolio rebalancing, whereas a swap implementation find much to celebrate after taking the decision to develop an ARP strategy. ■
risk.net 7
Q&A
How is your organisation currently using risk premia strategies, How do you tackle issues around liquidity and capacity?
and how has its approach evolved over time? Chen Ni: Capacity is a very important issue, especially for alpha strategies,
Clément Leturgie, Lombard Odier Investment Managers: Lombard because if there is too much flow into the strategy the arbitrage opportunity
Odier Investment Managers has been researching, developing and combining will disappear. China Minsheng Bank constantly observes the market and reacts
in-house alternative risk premia (ARP) for more than four years. Our approach to this diminishing opportunity. For example, in the commodity carry strategy,
has remained consistent in that we research and develop premia and work with we calculate the implied return and see how much room is left. If there isn’t
clients to design solutions that fit their needs. Today, our clients mostly use ARP much room for the arbitrage, we will stop the strategy. Liquidity is another
as a source of uncorrelated returns. issue. We only choose underlyings with very good liquidity in our risk premia
strategies – for example S&P, West Texas Intermediate, US Treasury bonds, and
Chen Ni, China Minsheng Bank: China Minsheng Bank uses beta and alpha so on.
strategies to construct portfolios for clients. Beta strategies are long-only strategies
across assets. Alpha strategies mean long-short strategies. We previously only used How does your organisation perceive crowding, and what tools
beta strategies, but we gradually included alpha strategies because the portfolio is it using to be better informed of the dynamics involved?
would otherwise suffer greatly under disadvantageous conditions. For example, this Qibao Pan: One of the best ways to tackle crowding is to analyse it according
year almost all of the long-only strategies were doing badly, but some long-short to their correlations. For example, I would put the overcrowding strategies into
strategies were performing much better. From 2017 we started to combine the two different designated ‘baskets’ or relevant categories depending on the nature
strategies into our portfolio to achieve a relatively stable performance. or asset class of these strategies. Therefore, these baskets may be crowded
Now about half of our investment is in beta strategies and the other half in internally, but still exhibit reasonably low correlation with each other.
alpha strategies: equity, commodity, foreign exchange, interest rate and volatility.
We allocate our risk evenly among these five alpha strategies. Chen Ni: Crowding is very difficult to perceive. There are two ways to overcome
this: through close observation of market indicators such as implied return
Qibao Pan, China Guangfa Bank: China Guangfa Bank builds its risk premia positions and exchange-traded fund flows, and through discussion with peers
strategies on a mix of beta and alpha, which in turn makes our asset allocation and people on the street.
portfolio broader. Currently, 40% of our risk is in beta and the rest in alpha.
The change in risk premia strategies over the past year – from focusing mainly Clément Leturgie: Crowding is one of the main issues for risk premia – and
on beta to adding factors such as momentum – occurred because we found that for markets in general. In risk premia, crowding often results in lower returns
beta strategies in many scenarios have a relatively large drawdown, especially and potentially more volatility or drawdown in periods of unwind, affecting the
when the stock market performs badly. To mitigate such fluctuation, we bring in risk-return profile of premia over time. Lombard Odier offers four main solutions
alpha strategies, which have a lower correlation, increasing the overall Sharpe ratio. to tackle issues around liquidity and capacity:
1. We focus on the most liquid risk premia, excluding, for example, emerging
What is the main attraction of these strategies? markets and premia investing in less liquid securities such as credit
Qibao Pan: The beta strategies we use are global, which means our asset single securities.
portfolio can be accessed globally – for the most part, in countries in the 2. We aim to be different and innovative in the way we capture and implement
developing world and emerging markets. That also spreads the bank’s alpha our premia to limit crowding exposure. One example of this is our equity
strategies across a variety of asset classes, including fixed income, forex and factor market-neutral premia, where integrating environmental, social and
commodities. Our strategies are diverse and dispersed within clear asset governance considerations helps to differentiate our portfolio. We also avoid
classes, which we consider helpful in increasing the Sharpe ratio. We find those shorting single stocks to prevent the negative impact of a potential crowding
strategies that have lower correlations but are still robust, and tie them to others. unwind short-squeeze phenomenon.
This gives us a decent Sharpe ratio for the portfolio as a whole, while each 3. We have limited our capacity: under current conditions, we would cap our
individual strategy may not have a Sharpe ratio as high. offering at $3 billion in ARP.
4. We continuously research new premia to diversify our offering and
Clément Leturgie: The objective is to offer an uncorrelated source of returns in increase capacity.
a liquid format, with transparency and a low-fee structure.
Does factor timing play a part in your organisation’s strategy?
Chen Ni: The main attraction is transparency – especially compared with Does the risk of crowding outweigh any potential benefits?
actively managed strategies – which makes the strategy easier to explain to Clément Leturgie: Lombard Odier does not try to time premia based
clients. Furthermore, the execution of those strategies lowers trading costs on discretionary views. However, one may argue that using a risk-based
and management fees compared with active strategies. approach to construct a portfolio incorporates, in some way, a timing
risk.net 9
Q&A
component. In our portfolio construction process, we use a proprietary risk when these objectives are delivered. Our objectives are mostly focused on
measure that takes into account short- and long-term premia information. returns targets, volatility targets, Sharpe ratio, maximum drawdown limits and
While the short-term component incorporates a momentum factor, the diversification benefits.
long-term component is more static. The resulting portfolio is probably less We understand the limits of backtesting and follow a simple set of rules to
subject to short-term premia momentum than a portfolio using a simplistic ensure the premia we develop and implement add value to clients’ portfolios:
risk-based approach. 1. Premia should always have a clear economic or financial rationale, or
be derived from market participants’ behavioural biases. Premia offer
Chen Ni: China Minsheng Bank doesn’t use factor timing. We do backtesting compensation for a certain risk, which should be well identified and
and try to construct an all-weather strategy, which is expected to withstand any understood. This prevents the team developing premia that are simply the
market conditions with limited drawdown. fruit of intensive data mining.
2. We avoid overoptimising premia. We like them to be pure, which enables us
Is your strategy managed internally? What are the pros and cons? to better understand and be more efficient in combining them.
Chen Ni: Most of China Minsheng Bank’s strategies are managed internally, 3. We do not develop premia simply to add them to our offering. When
while some counterparties offer us a lot of help. Our advantage is that we know implemented, a new premium should add diversification benefits.
the philosophy behind the strategy, which makes it easy for us to explain it to 4. We set conservative objectives and expect our premia to deliver about half
our clients. Our disadvantage is that, compared with the international investment of the backtest returns.
banks, we lack experience.
Do there need to be new benchmarks in ARP?
Clément Leturgie: Lombard Odier’s strategy is managed internally. Clément Leturgie: Yes, although this is a difficult task considering the
We believe developing our own premia enables us to be innovative and heterogeneity of the space. The simplest way to start with ARP benchmarks
differentiated to limit crowding risk, have a deep understanding of, full control would be to offer indexes that reflect the aggregated performance of the
of, and transparency in our investments, be fully flexible and able to customise industry’s major players. It would also be great to have benchmarks for
solutions for partners, and limit the costs of our solutions by netting premia individual premia, showing the performance of the most simplistic way of
with offsetting positions. capturing a certain premium. It will then be easy to assess the value added
While diversified and relatively broad, our individual premia and trading by a manager, whether in the premia development process or the
capacities for complex premia offerings will remain smaller than investment combination process.
banks can offer. We work with investment banks to leverage their capabilities,
notably on complex premia such as those in the volatility space. What is the biggest opportunity in ARP going forward?
Chen Ni: China Minsheng Bank is planning to implement machine learning
algorithms in the near future. Recently, we have been developing a new multi-
asset strategy – Multi-Asset Cycle Rotation Overseas Index, or Macro – based
on signal processing theory. We found that by applying machine learning
Qibao Pan algorithms, the strategy will be more efficient.
Director of Financial Market Department
China Guangfa Bank What is likely to be the greatest challenge to ARP?
www.cgbchina.com.cn/en Qibao Pan: An obvious pitfall is to judge a strategy based on its high Sharpe
ratio, which is gathered from its backtesting. However, well-performing strategies
on paper backtest may not persist in the future due to various factors from
limitations in historical data or overoptimisation. We shouldn’t pay too much
attention or put too many expectations on strategies that do particularly well in
How does your organisation measure success? What is done to backtesting. Instead, we should select strategies that are robust and consistent
ensure a backtested strategy is right for its portfolio? and combine them.
Qibao Pan: China Guangfa Bank prefers strategies that are more robust and
consistent based on the historical data, but not necessarily with a high Sharpe Clément Leturgie: The biggest challenge today for investors is assessing
ratio. If a strategy’s Sharpe ratio is tested 0.5, I would say it’s a good strategy. managers. The development of benchmarks will help in this endeavour.
Sometimes, a strategy performs well because filters were added and parameters Managers should also work closely with investors to set realistic expectations
twisted, which we should resist. What we need is a simple strategy with just a and determine achievable objectives.
few logical parameters and filters. For managers, the greatest challenge will come during periods of strong
risk and market unwind. A lot of ARP, though market-neutral, can recorrelate
Chen Ni: China Minsheng Bank has absolute return targets for its portfolio. We to equities in periods of stress. Adding crowding to the equation can result in
usually invest in ARP strategies in option format. We consider it a success if the large losses at the same time as the equity market. It will become clear in these
performance of our chosen strategies can beat the option fee we paid. periods which managers have developed the most differentiated premia and
built the most robust and uncorrelated portfolios.
Clément Leturgie: Lombard Odier has been active in developing customised
solutions for clients, offering uncorrelated risk premia and a robust Chen Ni: For beta strategies, decreased economic growth is the
combination process in the context of their global portfolio. When we build main challenge. For alpha strategies, it is crowding and changes in market
a solution, we define clear objectives with partners, and success is achieved structure. n
Analysing consensus
An academic approach to QIS
A sound understanding of key drivers of investment returns and defining logical rules are fundamental to ensuring QIS is effective.
Edhec-Risk Institute has established itself as a go-to resource for institutions creating or investing in QISs, offering diversified portfolios
across and between factors
Quantitative investing is an approach to asset The selected equities are chosen within
allocation grounded in financial research and economically integrated regions rather than on
mathematical modelling. A successful strategy is one a worldwide basis, however. This is to prevent
that is rooted in academic research of the fundamental exposing an index to geographical risks. Disparate
drivers of investment returns and governed by interest rate regimes, currencies and economic cycles
immutable rules dictated by reason, not whimsy. exist between regions, meaning if stock selection
It’s little surprise, therefore, that Edhec-Risk were based globally, certain countries would be
Institute (ERI) has become a premier resource for over- or under-represented.
institutions looking to create or invest in QIS. As a “If you want to create a value factor index, you
leading academic think-tank in investment solutions, shouldn’t look at value stocks across the entire
ERI provides research, outreach, education and developed or developing world, because if you
practitioner partnerships oriented towards the do that you may find your index overweighted
broader application of innovative asset management Japan, as that is where a lot of cheap stocks can be
models – and especially factor-based strategies. found, whereas there are not many cheap stocks in
ERI only promotes factors for which broad the US,” says Shirbini.
academic consensus exists: value, momentum, size, “Our methodology prevents us taking big
low volatility, high profitability and low investment. macroeconomic bets, which would be wrong as
The institute champions simple definitions of these factors are microeconomic exposures,” he adds.
factors and is wary of those that rely on data mining Once factor-rich stocks have been identified,
as opposed to the well-documented empirical ERI’s methodology applies a stock diversification “Our academics spend a long time
research in academic studies on factor investing. weighting to prevent the index tilting too much analysing these factors, looking at
Eric Shirbini, global research and investment towards a small handful of securities.
solutions director at ERI, explains: “Our academics “Instead of relying on individual stock
data typically spanning some 40 to 80
spend a long time analysing these factors, looking characteristics, we capture a wide group of stocks years to see if they are priced or not”
at data typically spanning some 40 to 80 years to exhibiting a factor bias and diversify between them. Eric Shirbini, ERI
see if they are priced or not. This analysis spans the The objective is to capture the factor performance of
period after a given risk factor has been discovered, the group,” Shirbini explains. Benchmarking
so that we can see if it persists once practitioners It is this two-layer approach that generates the Not knowing how to gauge the performance of a
start to invest.” superior risk-adjusted returns investors seek from factor strategy can frustrate a naive investor. Factor
ERI has constructed indexes that access all factor strategies. Shirbini says an ERI index could models promise much, but a manager that uses
six factors as well as a multi-factor product for deliver a Sharpe ratio between 60% and 80% a deficient means of benchmarking their chosen
investors seeking diversification between sources of higher than a corresponding cap-weighted index strategy will receive a flawed assessment of their
alternative risk premia. This allows asset managers over a period of 10 years or more. true performance.
to smooth out the peaks and troughs associated Although the lion’s share of this Shirbini says the appropriate method of
with individual factor exposures. outperformance is provided by the factor comparison will depend on the role a factor strategy
exposure, around 25% to 40% can be traced plays within a manager’s broader portfolio.
Index construction to the high degree of diversification achieved “If you are replacing a cap-weighted index with
ERI has been painstaking in its approach to index between the chosen stocks, and between a factor index, it doesn’t make sense to measure
design. First of all, simple and transparent criteria the factor index as a whole and its cap- the performance of the latter relative to the former,
for identifying those securities rich in a given weighted competitors. because the purpose of this allocation shift is to
factor premium were established and then applied “Cap-weighted strategies are extremely create a portfolio with lower risk than a market
consistently within and between equity markets – concentrated. Our factor strategies stay away from exposure but one that obtains higher returns. The
meaning the same analysis is used on an equity taking too much risk on certain individual names. more appropriate measure is absolute risk-adjusted
regardless of whether it’s listed in Lima or London. That is why we diversify,” says Shirbini. return,” he says.
risk.net 11
Case study: Edhec-Risk Institute
“But if you are looking to use a factor portfolio ERI offers an index overlay that controls for One argument peddled by factor sceptics is that
on a relative return basis, versus the market, it’s market beta, allowing investors to maintain a the returns associated with value, momentum,
important to adjust the benchmark to reflect the constant exposure to the market risk factor within low-volatility premia and the like will evaporate as
different market beta of that factor exposure. their chosen factor strategy. investors pile into the equities that exhibit them.
There’s no point benchmarking to the market Sector bias refers to a factor strategy’s propensity Take, for example, value stocks: these are
without an adjustment, as with these strategies to certain over- or underweight industry segments securities considered cheap relative to their price-
your exposure to the market is changing all the in its allocation. This represents an unrewarded to-book ratios. However, as managers increasingly
time,” he adds. risk, as a concentrated sector investment can lead buy up such stocks, as directed by value strategies,
performance to be influenced by that sector, rather their prices will rise, reducing or even eliminating the
Isolating biases than the chosen factor. initial value premia.
Market beta bias is one of three ERI identifies as A value index, for example, may contain a large Shirbini has little time for this supposed flaw
having ‘implicit risks’ hidden in factor strategies. number of insurance or bank company stocks – far in factor investing. “Crowding applies to alpha,
Another is geographic bias, as previously described, more than in the corresponding cap-weighted index. as this is a transient phenomena, which, once
and the third is sector bias. This would make the strategy overweight financials. uncovered, will disappear. Factors are different.
Market beta bias concerns the extent to which An investor may want to adjust the portfolio to They are specific, rewarded risks. You can have a
a factor strategy’s performance is contingent on pare back this sector risk. However, applying such a factor that performs well for long periods of time
the equity market risk factor. A zero-correlated filter could significantly reduce the very exposures but then the risk they hold materialises, resulting
alternative risk premia strategy would exhibit no that generate the factor-based returns, undermining in drawdowns. Then all the people who jumped
market beta bias, for example. But all long-only the efficacy of the strategy. on the bandwagon when it was doing well will
factor indexes will see their returns governed in part “We wouldn’t recommend adjusting for sectors, jump off again, but the factor premia will still be
by broader market movements. because there is a price to pay in terms of long-term there for the long-term investor who rides out the
Shirbini says it is important investors performance, unless the investor is concerned about cycle,” he says.
understand the role of market beta bias tracking error,” says Shirbini. Another misapprehension is that factors can be
and adjust their allocations, or benchmark, However, this decision – as with all those regarding ‘timed’ – that is, an investor can efficiently rotate
as appropriate. implicit risks – is in the hands of the investor alone, between different factor exposures over a given
“Research suggests that, when you invest in a and depends on their desired outcomes, says Shirbini. time horizon to maximise their returns and minimise
factor, your market beta changes. If you have a value their risks.
exposure, your market beta would be completely The myth of crowding Shirbini says this is wishful thinking. “You cannot
different than if you had a momentum exposure. ERI does not need to put faith in factor investing. It has time factors. They have their own cycles and they
It’s important to recognise this, as that market the empirical data and economic rationale to prove cannot be forecast. This is why the best solution is
beta will still be an important driver of a strategy’s these strategies’ utility, and the institute is keen to to hold a diversified portfolio across and between
performance,” he explains. debunk certain fallacies that have arisen around them. factors. Such a portfolio is very difficult to beat.” ■
A new perspective
Overcoming increasing complexity
As QIS becomes increasingly sophisticated and firms constantly drive for cost-efficiency, the relationship between buy-side firms and
investment banks is being altered. PremiaLab offers a fresh outlook on asset allocation, providing productivity and transparency for
buy- and sell-side market participants
The universe of QIS is vast, complex and expanding. built upon millions of data points and deliver factor strategies using different risk parity allocation
PremiaLab acts as a guiding light for buy-side firms these insights through our lens. This simplifies methodologies, and conduct risk and performance
navigating this space through its series of powerful and streamlines the asset allocation process for simulations in seconds.
analytical tools, designed to assist asset managers institutional investors,” he explains. “We understand that there is no single strategy
and pension funds, among others, to construct and Not only can an institution effectively compare or risk factor that can perform in all market
monitor factor portfolios. strategies, but it can also uncover how different environments. What’s needed is diversification, and
“We provide easy access to data,” says portfolio allocation styles would perform relative our platform gives clients the ability to put together a
Pierre Trecourt, co-founder of Hong Kong-based to one another. For example, an asset manager diversified performance engine for their portfolio that
PremiaLab. “It can be a challenging process for buy- could assess the performance of a group of five produces steady growth,” says Trecourt (see figure 1).
side clients to understand the risk premia offering
and position it within their portfolio. We offer fast-
track understanding and transparency.”
Trecourt and Adrien Geliot, his co-founder,
“There is no single strategy or risk factor that can perform in all market
recognised the frantic growth of the QIS marketplace environments. What’s needed is diversification, and our platform gives
necessitated a fresh analytics infrastructure to clients the ability to put together a diversified performance engine for their
provide order and clarity to institutions active in portfolio that produces steady growth”
this fast-growing market segment. Their solution: a
Pierre Trecourt, PremiaLab
risk premia database and factor analytics suite that
put all the resources a portfolio manager needs to
build a suitable quantitative allocation on a single 1 Multi-factor portfolio construction using the PremiaLab platform
convenient platform.
190 Portfolio
Access to data Strategy 1
180 Strategy 2
PremiaLab has partnered with 14 leading
Strategy 3
investment banks, feeding their QIS strategies into
Strategy 4–5
its database, representing an estimated $480 billion 170
of assets under management. The user interface
allows clients to streamline the analysis of strategies 160
under a series of performance and risk metrics –
including value-at-risk, volatility and Sharpe ratio – 150
and against a set of well-known benchmarks.
Clients can also simulate multi-asset portfolio 140
performance over various time periods, granting
them control over the allocation process. 130
PremiaLab’s mission is to provide independent
access to data to clarify investors’ decision-making 120
when it comes to factor strategies. Trecourt
believes putting all the necessary data through a 110
standardised, fast and intuitive interface is essential
if firms are to design a portfolio that addresses their 100
specific needs.
“When trading with different providers, it’s
90
complicated getting all the relevant data on 2014 2015 2016 2017 2018 2019
strategies and constituents in a comparable form.
As of 31 December 2018
We provide a comprehensive analytic solution
risk.net 13
Case study: PremiaLab