Smart Beta for improving risk
adjusted returns
James Bevan, CIO, CCLA IM
James Bevn, CIO, CCLA IM
16-17
Page 1September 2014 | Millennium Hotel London Mayfair
Attractions of Smart Beta as an alternative to active
and passive investments
•Equity investors like the idea of outperforming but
recognise that being consistently ahead is very challenging
– and active fees often seem to reward the manager and
not the risk taker
•Passive management has grown in popularity because
investors recognise that a properly specified index is hard
to beat, and minimising cost-leakage is really important –
but investors do reckon that they can do better
Page 2
Attractions of Smart Beta as an alternative to active
and passive investments
•Very large funds typically hire a number of active
managers, but this practice can quickly become selfdefeating, as the effects of the managers’ pure stock
selection skills will diversify away very rapidly, and the fund
will effectively be left with a very expensive index fund
overlaid with a small number of Style tilts
•A 2009 performance evaluation study done for the
Norwegian sovereign wealth fund came to the conclusion
that the fund would be better off simply building a set of
Style Factor portfolios themselves
•In recent years, this conclusion has resulted in the
enormous growth of ‘Smart Beta’ funds and ETFs
Page 3
Attractions of Smart Beta as an alternative to active
and passive investments
•Finance academics and practitioners have long since
identified a number of Style Factors that out-perform the
broad equity market, on average, over time
Examples include Value, Growth, Momentum, Quality
•The underlying idea is that each of these Style Factors has a
corresponding factor risk premium, or return, that can be
harvested for investors
•Smart Beta funds purport to deliver these returns to
investors: they can be thought of as index funds with Style
tilts; to a quant, they are Factor portfolios
•A recent internet search in the US market found over 40
Value indices, and 28 Value ETFs
Page 4
Attractions of Smart Beta as an alternative to active
and passive investments
•The idea of Smart Beta funds makes a lot of sense
•Their basic purpose is to deliver Style Factor returns to
investors as cheaply and easily as possible
Return = Risk free
+ beta
+ alpha
Type of risk:
Opportunity set:
Range of return/risk:
Correlations:
Difficulty:
Cost:
Certainty of positive return:
Systematic
Limited
Low
High
Low
Cheap
High over time
Unsystematic
Extensive
High
Low
High
Expensive
None
Time horizon is important – and skill is rare
Page 5
Attractions of Smart Beta as an alternative to active
and passive investments
•One of things that many investors have recognised is that if
you can explain an ‘active bet’, then it almost ceases to be an
alpha
•Portfolios can therefore be seen to contain exposure to many
factors which masquerade as alpha, but which should be
viewed, treated and invoiced as beta
•The disturbing reality is that disciplined investors can on
paper, with enough time, beat indices quite easily
Page 6
Attractions of Smart Beta as an alternative to active
and passive investments
Virtually all ‘price indifferent’ strategies outperform
Investment Strategy
Risk Weighted Strategies
Fundamental Strategies
Smart beta
Random monkeys
Cap weighting
Breaking the
Rebalance
link between
price and weight










Yet many smart beta products and strategies under-deliver
Page 7
Comparing smart-beta strategies
1. All non-cap-weighted strategies have value and small
size tilts
2. Many smart beta strategies suffer from high
implementation costs
3. Issues for investors:
-Time horizon
-Turnover and trading costs
-Economic representation
-Taking only bets that make sense
-Avoiding unintended, unwarranted, unwanted risks
-Capacity/liquidity
Page 8
Comparing smart beta strategies
Scalability really matters
Page 9
Comparing smart-beta strategies
•Apart from cost leakage, many ‘Smart Beta’ funds don’t
deliver
- They have a Style tilt, but don’t give the Style return,
- They usually have more risk than necessary, and
significant exposures to other factors
•We can trace the problem to poor portfolio construction
- Weightings to stocks are often arbitrary and are not
designed to capture the Style Factor premia
- With no effort to optimise factor exposures, performance
ends up being driven by exposure to market, industry and
other factors
Page 10
Comparing smart-beta strategies
•It may be that the marketing imperative to have a simple
story is more important than delivering an efficient Style tilt
•All is not lost – it’s possible to build Smart Portfolios that can
do a much better job of delivering the Style Factor return, with
lower overall risk, and much less exposure to other factors
•Managers can trade off each stock’s (Style-related) expected
return against its risk
Page 11
Comparing smart-beta strategies
• Markowitz proposed that the most efficient way to manage
portfolios was to have holdings whose contribution to portfolio
expected return matched their contribution to risk
• This idea was first published in 1952, and no-one has yet
come up with a better idea
• However most Smart Beta funds don’t do this
• In fact, their construction method often disregards risk
completely, except for having lots of holdings, which is
presumed to give greater diversification
• This makes them inefficient – therefore it should be easy to
improve their performance
Page 12
Comparing smart-beta strategies
Risks that people think that they can avoid –
‘Low vol’ and ‘min vol’
•The CAPM identifies ‘risk’ as the principal driver of premium
returns – yet for extended periods, there seems to have been a
return premium from seeking least risk
•This in part an example of how easy it is to beat a cap
weighted index
•If we look at the data, we can discern a clear link between the
returns to low vol. and the movement in rates and bond yields
– which have been in a downtrend since the early 1980s.
•If the trend in bond yields turns up, low vol. may
underperform.
Page 13
Comparing smart-beta strategies
Risks that people often don’t recognise –
Environmental, social and governance (ESG) factors
•There are systematic risk factors that are not captured in
conventional P&L and balance sheet analyses
•Governance risks relate to audit, board structure and function
and the relationship between boards and shareholders and
dominant and minority shareholders. Exclusions can reduce
risk
•Social and environmental risk factors relate primarily to miscosted activities
•‘Sustainability’ will be a core equity valuation factor for all long
term focused strategies.
Page 14
How should institutional investors set metrics for
manager performance with these strategies, and asset
returns against costs?
•Institutional investors should be focused on
–Whether they get what they need relative to their liabilities
–Whether they get what they’re expecting and deserve from
exposure to the selected ‘smart beta’
–Net returns
–What risks they’re running, whether these are acceptable,
and how they can be managed
–Long only vs. long/short with leverage is a big issue
•Combining smart betas can significantly reduce ‘risk’, but
there needs to be clear focus on diversification opportunities
and realities
Page 15
How should institutional investors set metrics for
manager performance with these strategies, and asset
returns against costs?
•It’s really important to determine how performance will be
considered and who owns what accountability
•Understanding how returns have been achieved is as
important as understanding what the returns are
–Investors need to do a lot of homework before starting a
mandate and combining mandates
–Agreeing what’s to be measured can be (surprisingly)
tricky
–Portfolio construction needs to be robust
–Risk needs to be clearly in focus
Page 16
How should institutional investors set metrics for
manager performance with these strategies, and asset
returns against costs?
•It is not possible to measure Style Factor returns directly. In
practice, they are estimated, either by creating FactorMimicking Portfolios (FMPs), or by running cross-sectional
regressions on stock returns
•The difficulty with using FMPs lies in trying to make the
portfolios independent of other factor effects
•On the other hand, using cross-sectional regressions means
the Style Factor returns will be conditioned on all the other
factors included in the regression
•Style Factor returns are usually conditioned on each other;
however, they are often conditioned on market, industry or
currency factors as well
Page 17
How should institutional investors set metrics for
manager performance with these strategies, and asset
returns against costs?
•Investors should look for an optimal portfolio construction
process to create an efficient Style Factor portfolio
•Optimisers are notoriously prone to error maximisation – to be
useful, we need to have confidence in return and risk estimates
•With Style Factor portfolios, expected returns can be a
sensitivity to the Style factor – we hope that the Style Factor
premium is positive, but at least we are sure about the stock
beta to the factor.
•Using Style betas as the expected return proxy ensures that the
portfolios have a significant Style Factor tilt
Page 18
How should institutional investors set metrics for
manager performance with these strategies, and asset
returns against costs?
•We also need to be confident about the risk numbers
•In practice, this means that we need to be sure that the risk
model has done a good job of identifying the sensitivities of
each stock to the various factors, and of capturing its systematic
common factor risks
•We can filter the candidate universe to screen out stocks with
low R-Squareds
•We can filter out very high risk stocks as their risk
characteristics may be biased estimates of true risk, and may
appear to offer (spurious) diversification
Page 19
How should institutional investors set metrics for
manager performance with these strategies, and asset
returns against costs?
•Liquidity control: it’s wise to set a maximum that can be
bought of any stock, expressed as a limit on the shares
held as a percentage of trade volumes
•There needs to be shared view on scalability and investors
need to keep a watchful eye on ‘shelf life’
•Rebalancing: there need to be triggers for rebalancing
•Cost control: in deciding whether to trade or not there
need to be explicit assumptions on implementation costs
and fees. Total cost leakage needs to be in focus
Page 20
Smart beta: conclusions
•Smart Beta offers plenty of attractions as an alternative to
active and passive investments
•When comparing smart beta offers, we need to focus on
what, why, and how
•We need particular focus on expected net returns,
scalability, cost and risk
Page 21