Kepos Capital LP

advertisement
Exotic Beta Revisited
November 2014
This presentation is for informational purposes only and addresses certain points made in the academic paper referenced herein, the full text of which is available in Carhart, M., Cheah, U., De Santis, G., Farrell, H.
and Litterman, B., (2014). Exotic Beta Revisited. Financial Analysts Journal, Volume 70 (Issue 5). doi: 10.2469/faj.v70.n5.4. The opinions, data and analyses contained herein are solely with respect to such academic
paper. Such opinions, data and analyses are as of the date written, and are subject to change without notice. Nothing contained herein should be considered investment advice or an offer or solicitation to buy or sell
any investment.
Outline
1. A Taxonomy of Exotic Betas
2. Empirical Evidence on Exotic Betas
3. Comparing Alternative Risk Premium Approaches
4. A Dynamic Portfolio of Exotic Betas
5. Closing Remarks
Challenge of Policy Allocation
Conventional Approach
Capital-Based Allocations
Challenge:
Allocation
decisions
reflect
unintended
factor
exposures
Proposed
Solution:
Asset
Classes
Purposeful
and explicit
allocation to
risk premia
Better
knowledge
of return
sources
Risk
Premia
Themes
Risk Factor Approach
Multiple Risk Premia within Asset Classes
The Return Spectrum
“…returns exist along a continuum – from beta, to exotic beta, and ultimately to alpha.”
– Litterman (2005)
Alpha
Beta
•
•
Macro statistical arbitrage
Exotic Beta
•
Volatility arbitrage
•
Long term value
•
Systematic events
•
Income
•
Long term momentum
•
Insurance
Global equities
Passive
Active
Lower cost
Higher cost
Well-known
Proprietary
A Taxonomy of Risk Premia
We focus on four risk premia themes across all four major asset classes.
ASSET CLASSES
Equities
Value
Income
RISK
PREMIUM
THEMES
Insurance
Momentum
Bonds
Currencies
Commodities
Cheap assets tend to outperform expensive assets
Investors demand additional yield for lower volatility assets
Risk averse investors pay a premium to insure against extreme events
Asset performance can be persistent
Building an Exotic Beta Portfolio
1
Choose a
carry measure
2 Construct a
3 Hedge away
dollar-neutral FX
portfolio on
carry rank
ACWI beta using
liquid equity index
futures
FX Carry Portfolio
3-Month LIBOR
-100 -50
0.0 1.0 2.0 3.0 4.0
0
FX Carry Portfolio
50 100
-100
Australia
New Zealand
New Zealand
Canada
Australia
Australia
Euro
Norway
Norway
Japan
Canada
Canada
New Zealand
Sweden
Sweden
Norway
United Kingdom
United Kingdom
Sweden
United States
United States
Switzerland
Euro
Euro
United Kingdom
Japan
Japan
United States
Switzerland
Switzerland
ACWI - Hong Kong
ACWI - Japan
ACWI - Eurostoxx 50
ACWI - United Kingdom
ACWI - United States
-50
MSCI ACWI Hedge
0
50
100
Comparing Theoretical and Tradable Exotic Betas
Tradable Exotic Beta Portfolios are obtained as follows:
•
Use only tradable assets (e.g. liquid futures contracts)
•
Penalize tracking error from theoretical Exotic Beta portfolio
•
Charge for transaction costs
Distribution of Equity Beta Estimates for Exotic Beta Portfolios
1.50
1.50
1.00
1.00
0.50
0.50
0.00
0.00
-0.50
-0.50
-1.00
-1.00
Historical Evidence: ACWI and Individual Exotic Betas
Key Takeaways
All portfolios are normalized to 10% target volatility
1.
Attractive potential
returns
2.
Low cross-correlation
3.
Poor performance
does not correspond
with poor
performance in
equities
Summary Statistics for Exotic Beta Portfolios
Sampling period: January 1990 – December 2012
Return Correlations for Exotic Beta Portfolios
•
•
Average correlation is 0.04
Using three-year window estimates:
- Average correlation is between -0.04 and 0.10
- Less than 7% of all the correlations are larger than 0.35 (in absolute value)
Cumulative Excess Retruns: ACWI and Equal Risk Exotic Beta Portfolio
LTCM blow up
and tech bubble
1994 run-up in
sovereign bond
yields
Risk Parity
Main idea:
• Scale investment in different asset classes to the same risk
• Combine scaled asset class investments into an equally weighted portfolio
Risk Parity (Across Asset Classes): Return Correlations
•
•
Average correlation is 0.23
Using three-year window estimates:
- Average correlation is between 0.12 and 0.36
- More than 25% of all the correlations are larger than 0.5 (in absolute value)
Understanding Implied Views
The common idea that the risk parity portfolio has no
views on expected returns is a misconception
Comparing Alternative Investment Strategies
The exotic beta portfolio offers
• Better return/risk profile
• Less extreme DD/vol
• More attractive diversification
Can we explain exotic beta returns?
Conditional alpha is almost identical to
unconditional alpha for the exotic beta portfolio
Risk parity and hedge funds explain very little
of the return variation in the exotic beta portfolio
Estimated Premiums and Correlations
 Most realized returns are biased, due to overfitting, unusual circumstances, survivorship
and backfilling bias, etc.
 We use the Black-Litterman model to shrink historical returns towards a CAPM
equilibrium
Some correlations are high, due to a
common ACWI exposure
Efficient Frontier and Risk-Return Trade-Offs
6%
Exotic Beta
28% Global Bonds
39% Hedge Funds
33% Exotic Beta
Black-Litterman Expected Excess Return
5%
Risk Parity
Optimal at Bond Vol
4%
Hedge Funds
Maximum Sharpe Ratio
60% Global Equities
40% Global Bonds
37% Global Bonds
37% Hedge Funds
26% Exotic Beta
3%
Global Equities
60/30/10
60/40
60% Global Equities
30% Global Bonds
10% Hedge Funds
Global Bonds
2%
1%
0%
2%
4%
6%
8%
10%
Volatility
12%
14%
16%
18%
Risk Budgeting: Out of Sample Evidence
2005 Max SR Portfolio
2008 Max SR Portfolio
41% Hedge Funds
38% Exotic Beta
21% Global Bonds
42% Hedge Funds
35% Exotic Beta
23% Global Bonds
Outline
1. A Taxonomy of Exotic Betas
2. Empirical Evidence on Exotic Betas
3. Comparing Alternative Risk Premium Approaches
4. A Dynamic Portfolio of Exotic Betas
5. Closing Remarks
Exotic Beta Portfolio: Equal-Risk vs. Dynamic-Risk Allocations
Equal-Risk Exotic Beta Portfolio
• Uses only backward-looking estimates of volatilities and correlations
• Targets 10% volatility
• Rebalances on a monthly frequency
• Incorporates t-cost penalties
Dynamic-Risk Exotic Beta Portfolio
• Same as Equal-Risk Exotic Beta Portfolio
• Incorporates conditional views on the risk premiums based on:
− Prior
− Historical means
− Value
− Momentum
− Spillover (Market Disruption)
• Imposes penalties on position and trade size
Conditional Forecasts of an Exotic Beta: Currency Carry
2.0%
1.0%
0.0%
Prior
Historical Mean
Value
Momentum
Spillover
Dec-12
Jun-12
Dec-11
Jun-11
Dec-10
Jun-10
Dec-09
Jun-09
Dec-08
Jun-08
Dec-07
Jun-07
Dec-06
Jun-06
Dec-05
Jun-05
Dec-04
Jun-04
Dec-03
-1.0%
Exotic Beta Portfolio: Equal vs. Dynamic Risk Allocation
STATISTIC
Average Excess Return
Annualized Volatility
Sharpe Ratio
Maximum Drawdown
MaxDD/Vol
Correlation with ACWI
Equal Risk
Dynamic
14.6%
19.2%
10.4%
9.5%
1.41
2.03
-21%
-14%
(2.0)
(1.4)
0.21
0.13
Average Marginal Contributions to Risk:
Equity Value
13.1%
Bond Yields
9.6%
Bond Slope
7.2%
Commodities
9.3%
Real Assets
13.0%
Currency Value
10.9%
Volatility
11.6%
Credit
15.3%
Catastrophe Bonds
10.1%
13.5%
10.3%
5.5%
13.6%
9.7%
24.1%
18.4%
3.1%
1.8%
 Marginal contribution to risk from timing model is 27%
 Average turnover of the Dynamic portfolio is 9 months
Outline
1. A Taxonomy of Exotic Betas
2. Empirical Evidence on Exotic Betas
3. Comparing Alternative Risk Premium Approaches
4. A Dynamic Portfolio of Exotic Betas
5. Closing Remarks
Download