Thinking About Smart Beta Ralph Goldsticker, CFA July 17, 2014
May 06, 2015
Thinking About Smart Beta
Ralph Goldsticker, CFA
July 17, 2014
Thinking about Smart Beta
Agenda
I. What is Smart Beta?
II. Why might you want to invest in Smart Beta strategies?
Why not?
III. Smart Beta versus Alpha
Smart Beta and Asset Allocation
IV. Digging into some Smart Beta strategies
V. Summary and discussion
2Smart Beta 101 - Goldsticker
Thinking about Smart Beta
Agenda
I. What is Smart Beta?
II. Why might you want to invest in Smart Beta strategies?
Why not?
III. Smart Beta versus Alpha
Smart Beta and Asset Allocation
IV. Digging into some Smart Beta strategies
V. Summary and discussion
3Smart Beta 101 - Goldsticker
What is Smart Beta?
Smart Beta is a term for rule-based investment strategies that provide
systematic exposures that are promised to result in superior risk
adjusted returns.
Examples:
• Fundamental Indexing: Objective: Higher expected return
Approach: Stocks’ weights based on fundamental value per share (e.g. earnings,
dividends, etc.) rather than price per share.
• Minimum Volatility: Objective: Lower portfolio risk, without sacrificing expected return
Approach: Construct portfolio that is expected to have the minimum volatility.
4Smart Beta 101 - Goldsticker
Origins of the term “Smart Beta”
1. The term “Beta” comes from the Capital Asset Pricing Model (CAPM).
Beta describes sensitivity to the market.• A stock with a beta of 1.0 moves in line with the market
• A stock with a beta of 1.5 moves 1½ times the market.
2. The CAPM argues that Beta is the only compensated risk factor.
3. Research discovered that the CAPM doesn’t describe market behavior.• Value stocks and small capitalization stocks outperformed.
• Low volatility stocks had higher returns than high volatility stocks.
4. Researchers created factor portfolios to describe this behavior.• The value factor portfolio tracks the performance of value stocks.
• The size factor portfolio tracks the performance of small cap stocks.
5. The term “Factor Beta” was used to describe sensitivity to a factor.
Marketing turned it into “Smart Beta”
6. Over time “Smart Beta” has come to include any rules-based strategy.
Smart Beta is sometimes referred to as “Alternative Beta”.
5Smart Beta 101 - Goldsticker
This presentation uses only a few Smart Beta strategies
to illustrate the concepts.
Fine Print: • This presentation focuses on long-only equity Smart Beta strategies.
• Designed using similar concepts, there are Smart Beta approaches in fixed income and commodities, as well
as multi-asset class and long-short strategies.
• The Smart Beta strategies in this presentation were selected to illustrate the different objectives and
approaches to portfolio construction.
• They are stylized, and do not represent the specific approach, characteristics and/or results of any fund.
• Inclusion or exclusion does not imply a view on the relative efficacy of a strategy.
6
The Cube
Asset Class, Strategy Style, and Risk Factor
Source:
Expected Returns on Major Asset Classes, Ilmanin, 2011
Smart Beta 101 - Goldsticker
Smart Beta has seen tremendous asset growth
7
Source: Morningstar, “A Sensible Approach to 'Smart Beta”
US exchange traded products.
Net Assets in Smart Beta ETFs Assets by Approach
Strategy % of Category % of Flows
Dividend 30% 24%
Growth 23% 15%
Value 23% 25%
Multi-Factor 6% 21%
Equal Weighted 4% 6%
Volatility 3% -4%
Non-Trad Commod 3% -2%
Fundamental 2% 4%
Momentum 1% 3%
Buyback/Shareholder Yld 1% 5%
Earnings Weighted 1% 1%
Quality 0% 2%
Multi-Asset 0% 0%
Revenue Weighted 0% 0%
Expected Returns 0% 0%
Non-Trad Fixed Income 0% 0%
Low/High Beta 0% 0%
Risk Weighted 0% 0%
Smart Beta 101 - Goldsticker
Thinking about Smart Beta
Agenda
I. What is Smart Beta?
II. Why might you want to invest in Smart Beta strategies?
Why not?
III. Smart Beta versus Alpha
Smart Beta and Asset Allocation
IV. Digging into some Smart Beta strategies
V. Summary and discussion
8Smart Beta 101 - Goldsticker
Why all the fuss about Smart Beta?
Simulated Smart Beta returns appear to be compelling.
Note: While the smart beta strategies have relatively short live track records,
researchers “discovered” the valuation, small cap and low volatility anomalies
decades ago.
9
Source: Goldsticker calculations, Research Affiliates data
Smart Beta 101 - Goldsticker
Frequently heard rationale for allocating to Smart Beta
1. Historical superior performance in simulations is likely to persist.
The “anomalies” have robust explanations and/or long histories.
a) Risk premium: Compensation for risks other than CAPM beta
• Market segmentation
• Imperfect markets
• Different preferences
b) Investors make persistent behavioral errors (and won’t “learn”).
- or -
2. Capitalization weighted portfolios are too concentrated and/or too
exposed to growth and momentum stocks.
10Smart Beta 101 - Goldsticker
Why might you rely on the backtests?
Increased adoption creates risk that overweighted assets in strategy
become expensive (crowded trade).
If that happens, the risk premium could disappear or even reverse.
11
Risk Premium StoriesHigher expected return due to higher risk
Behavioral StoriesHigher returns due to persistent behavior
ValueValue stocks are risker.
I.e. Will do worse in bad timesInvestors overpay for growth.
Small CapSmall cap stocks are riskier and
less liquid
Investors prefer the familiarity of
stocks of large companies.
Low VolOverweighting high vol stocks is
only way to create portfolio with
higher vol than the market.
“Lottery ticket” effect bids up
prices (lowers expected returns)
of high vol stocks.
• The superior performance of Smart Beta strategies primarily came from
value, small cap and/or low volatility tilts.
• There are reasons to believe the tilts’ historical performance will continue.
Smart Beta 101 - Goldsticker
Why might you want to avoid Smart Beta strategies?
1. You don’t “believe” the backtests.
a. Inconsistent with market efficiency
• Historically observed anomalies will not persist.
• Not “real”. Product of data mining.
• Now “discovered”, risk premium will be competed away
• Theory argues only non-diversifiable risks should be compensated.
b. Not macro-consistent
• Who owns the “dumb” portfolios?
• Why?
• Why won’t they “learn”?
2. Cap weighting is lower cost and has higher capacity.
3. Organizational
• Too much tracking error versus market and peer group
• Added complexity
• Aversion to higher fees
• Tendency to sell at bottom of performance cycles
12Smart Beta 101 - Goldsticker
Strategies may become crowded trades
• Most strategies are valuation agnostic
• Crowding of most concern with concentrated approaches
13
Source: “Avoiding Pricey Low Volatility Investing”, Feifei Li, Research Affiliates
Smart Beta 101 - Goldsticker
While the full period simulations look compelling,
all had significant episodes of underperformance.
14
-31%
-50%
-38%-24%
-41%
Source: Goldsticker calculations, Research Affiliates data
Smart Beta 101 - Goldsticker
Thinking about Smart Beta
Agenda
I. What is Smart Beta?
II. Why might you want to invest in Smart Beta strategies?
Why not?
III. Smart Beta versus Alpha
Smart Beta and Asset Allocation
IV. Digging into some Smart Beta strategies
V. Summary and discussion
15Smart Beta 101 - Goldsticker
Smart Beta versus Alpha
Excess return from Smart Beta is a “risk” premium from a “tilt”.
Active managers’ alpha results from their skill, not from persistent tilts.
• Successful active strategies deliver returns in excess of the market plus
any persistent style “tilts”.
Hire active managers because you believe they can outperform the
appropriate index, not because you want their underlying tilts.
• E.g. A skillful active manager with a value style should outperform a value
benchmark such as a fundamental index.
• Use a value-tilted index fund or smart beta strategy if you don’t have
confidence in the manager’s ability to add value.
16Smart Beta 101 - Goldsticker
Disaggregating expected returns from:
Cap Weighting, Smart Beta and Active Management
17
Market Return
Factor/Style
Returns
Alpha
Market Return
Factor/Style
Returns
Market Return
Capitalization
WeightingSmart Beta
Active
Strategies
Smart Beta 101 - Goldsticker
Framework for allocating to:
Smart Beta, Indexing and Active Management
Three potential sources of return:
• Market
• Average factor exposures
• Alpha from manager skill
Optimal mix of cap weighting, smart betas and
active management depends on:
• Expected risks and return for each component
• Risk tolerance
• Tracking error tolerance
• Asset size, liquidity needs, capacity of strategies
• Appetite for complexity
• Fees
Note: By definition, the sources of returns are uncorrelated.
• If smart beta return is correlated with the market, it’s market beta.
• If alpha from active management is correlated with smart beta, it’s smart
beta, not alpha.
18Smart Beta 101 - Goldsticker
Which strategies? How much?
• What are your risk and return expectations for each strategy?
• What are your objectives?
a) Proactive: Tilt the portfolio towards target anomalies?
b) Reactive: Move away from capitalization weights?
• How do the tilts fit into the rest of the portfolio?
• What is your tolerance for underperformance relative to standard indices
and peer group?
• Do you need to be concerned about liquidity and capacity?
Strategies with larger tilts tend to have higher expected returns, but also:
• Higher tracking error
• Less capacity
• More concentration
19Smart Beta 101 - Goldsticker
Thinking about Smart Beta
Agenda
I. What is Smart Beta?
II. Why might you want to invest in Smart Beta strategies?
Why not?
III. Smart Beta versus Alpha
Smart Beta and Asset Allocation
IV. Digging into some Smart Beta strategies
V. Summary and discussion
20Smart Beta 101 - Goldsticker
Evaluating the designs of Smart Beta portfolios *
Portfolio Construction Approach
Rule-Based Optimized
Inve
stm
en
t
Ra
tio
na
le *
* Higher Return • Fundamental Indexing • Maximum Sharpe Ratio
Low Vol Tilt • Inverse Volatility • Minimum Volatility
Improved
Diversification
• Equal Weight
• Risk Parity
• Equal Risk Contribution
• Max Diversification
21
* We use only a few strategies to illustrate the concepts
There are strategies that incorporate objectives besides value, small cap and low vol. ** Many Smart Beta strategies incorporate more than one of the objectives.
For example, equal weighting also tends to have a value tilt.
Smart Beta 101 - Goldsticker
Template for evaluating Smart Beta strategies
Stocks’ weights • How are the stocks’ weights determined?
Tilts relative to cap weights• Which tilts does the strategy produce?
Value Small Cap Low Volatility
Implicit assumptions • Assumptions that would make the strategy optimal
Objections • Implicit assumptions that make strategy questionable
Economic rationale • Arguments used to justify risk premium
Behavioral rationale • Behavioral arguments used to explain performance
Required inputs • Data required to construct portfolio
Capacity / Liquidity• Aggregate capacity in the strategy for all investors
• How liquid is it to enter or exit the strategy
Concentration• Equivalent to number of stocks in an equal weighted
portfolio (Smaller number is more concentrated.)
Turnover / Trading Costs• How much turnover does the strategy entail?
• How expensive will it be to rebalance?
Tracking Error• What is the tracking error of the strategy versus the
capitalization weighted benchmark?
22Smart Beta 101 - Goldsticker
Capitalization Weighted
23
Stocks’ weights Stock′s Market CapTotal Market Cap
= Shares Outstanding × Price
∑ (Shares Outstanding × Price)
Tilts relative to cap weights • None
Implicit assumptions• CAPM: Expected returns are proportional to beta
• Maximum Sharpe Ratio portfolio is appropriate
Economic rationale • CAPM assumptions
Behavioral rationale • Invest like average investor
Required inputs • Shares outstanding, prices
Objections• Heavy concentration in large cap stocks
• Heavy concentration in momentum stocks
Capacity / Liquidity • Maximum
Concentration * • 138 stocks
Turnover / Trading Cost ** • 6.7% / .03%
Tracking error • None
* S&P 500, Average 2000-2009, Calculations: Goldsticker, Data source: Research Affiliates
** Source: “A Survey of Alternative Equity Index Strategies”, Journal of Portfolio Management, Sept/Oct 2011
Equal Weighted
24
Stocks’ weights1
NumberofStocks
Tilts relative to cap weights• Primary: Small cap
• Secondary: Value
Implicit assumptions • Stocks have same returns, risks and correlations
Objections• Heavy concentration in small cap stocks
• High turnover
Required inputs • None
Capacity / Liquidity • Limited
Concentration * • 922 stocks (Less than 1000 due to drift.)
Turnover / Trading Cost ** • 22.6% / .22%
Tracking error vs Cap Wt ** • 6.37%
* US Top 1000, Average 2000-2009, Calculations: Goldsticker, Data source: Research Affiliates
** Source: “A Survey of Alternative Equity Index Strategies”, Journal of Portfolio Management, Sept/Oct 2011
Smart Beta 101 - Goldsticker
Fundamental Indexing
25
Stocks’ weightsFundamental Value per Share × Shares Outstanding
∑ (Fundamental Value per Share × Shares Outstanding)
Tilts relative to cap weights • Primary: Value
Implicit assumptions• All stocks should trade at same multiple
• Optimization is not necessary
Objections• Growth stocks deserve higher multiples
• Portfolio construction ignores risk.
Behavioral rationale• Cap weighting systematically over-prices expensive
stocks.
Required inputs
• Fundamental value per share: earnings, dividends,
sales, etc.
• Shares outstanding
Concentration * • 163 stocks (18% higher than capitalization weighted.)
Turnover / Trading Cost ** • 13.6% / .13%
Tracking Error ** • 4.5%
* US Top 1000, Average 2000-2009, Calculations: Goldsticker, Data source: Research Affiliates
** Source: “A Survey of Alternative Equity Index Strategies”, Journal of Portfolio Management, Sept/Oct 2011
Smart Beta 101 - Goldsticker
Equal Risk Contribution (ignore correlations)
26
Stocks’ weights(1/Volatility)
∑ (1/Volatility)
Tilts relative to cap weights• Primary: Small cap
• Secondary: Low Volatility, Value
Implicit assumptions
• Stocks have same expected Sharpe Ratios.
• All correlations are the same.
• Maximum Sharpe Ratio portfolio is optimal.
Objections
• Constant Sharpe Ratio assumption does not reflect
different correlations
• Risk model dependent
Behavioral rationale • Improved diversification
Required inputs • Volatilities
Concentration • 830 stocks * (Similar to equal weighted)
Turnover / Trading Cost ** • 8.9% / .06%
Tracking Error ** • 2.6%
* US Top 1000, Average 2000-2009, Calculations: Goldsticker, Data source: Research Affiliates data
** Source: “A Survey of Alternative Equity Index Strategies”, Journal of Portfolio Management, Sept/Oct 2011
Smart Beta 101 - Goldsticker
Maximum Diversification
27
Stocks’ weights Maximize ∑(���������������×��������� �)
"�#�$��������������
Tilts relative to cap weights• Primary: Low Volatility
• Secondary: Value
Implicit assumptions• Stocks have same expected Sharpe Ratios
• Maximum Sharpe Ratio portfolio is optimal
Objections
• Ad hoc objective
• Constant Sharpe Ratio assumption ignores correlations
• Prefers high volatility stocks that are uncorrelated
• Maximum Sharpe Ratio portfolio is optimal
Required inputs• Volatilities
• Correlations
Concentration • 58 stocks * (very dependent on the covariance matrix)
Turnover / Trading Cost ** • 56.0% / .53%
Tracking Error ** • 7.1%
* US Top 1000, Average 2000-2009, Calculations: Goldsticker, Data source: Research Affiliates data
** Source: “A Survey of Alternative Equity Index Strategies”, Journal of Portfolio Management, Sept/Oct 2011
Smart Beta 101 - Goldsticker
Don’t be impressed by high Sharpe Ratios or high risk-
adjusted returns.
Focus on the effect on the aggregate portfolio’s total risk and return.• You can’t spend risk-adjusted returns.
• Sharpe Ratios are not returns. (They are the ratio of return to risk.)
• Strategies that appear attractive on a risk-adjusted basis may be low risk
and low return.
• Risk-adjusted comparisons make sense only if both strategies have same
risks and correlations. (In which case you may as well compare returns.)
28Smart Beta 101 - Goldsticker
Thinking about Smart Beta
Agenda
I. What is Smart Beta?
II. Why might you want to invest in Smart Beta strategies?
Why not?
III. Smart Beta versus Alpha
Smart Beta and Asset Allocation
IV. Digging into some Smart Beta strategies
V. Summary and discussion
29Smart Beta 101 - Goldsticker
Additional considerations:
• Be proactive.
• Decide which tilts you want. Assemble portfolio that provides them.
• Consider: Smart Beta, indexing, active strategies and custom solutions.
• Focus on how strategies fit into the overall portfolio, not their stand-alone
characteristics.
• How do Smart Beta’s factor tilts align with your active managers’ styles?
• Are the constraints in portfolio construction rules really necessary?
• Constraints limit strategies’ ability to deliver the desired tilts.
• Understand the “cost” of constraining Smart Beta portfolio weights
and characteristics to be similar to the cap weighted alternative.
• If adopting Smart Beta, do enough to “move the needle”.
• Don’t get caught up in backtests and marketing stories.
• Focus on the underlying characteristics.
• Why do you expect them to add value in the future?
30Smart Beta 101 - Goldsticker
Summary:
What do you believe? What are your objectives?
• Do you “believe” the backtests?
• Was the historical performance due to fundamental relationships?
• Or was it data mining?
• If the historical performance was “real”, is it likely to persist in the future?
• Has investor behavior changed?
• Has the approach become a crowded trade?
• Can you live with the cycles of relative performance?
• Is the incremental expected return large enough to justify the
organizational resources required to select and monitor the strategy?
31Smart Beta 101 - Goldsticker
Discussion
32Smart Beta 101 - Goldsticker
33Smart Beta 101 - Goldsticker