Asset Allocation Overview- Callan Research.pdf

download Asset Allocation Overview- Callan Research.pdf

of 24

Transcript of Asset Allocation Overview- Callan Research.pdf

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    1/24

    0Independent Adviser Group

    Mark AndersenVice President415-274-3023 [email protected]

    Callan Associates

    101 California StreetSuite 3500San Francisco, CA 94111

    The Chemistry of Asset Allocation

    Are Traditional Asset Classes the Right Building

    Blocks for Portfolio Diversification?

    2011 Callan IAG National Conference

    September 2011

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    2/24

    1Independent Adviser Group

    Agenda

    Overview of traditional approaches

    Benefits and shortfalls

    Implementation

    Introduction to factor-based allocation

    Classifying factors

    Combining factors

    Implementation

    A hybrid approach to asset allocation

    Economic Scenario Buckets

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    3/24

    2Independent Adviser Group

    Why theInterest in Factors?

    Traditionally portfolios are built with asset classes such as equity, fixed income,

    real estate, etc.

    Ideally portfolios would be made up of many components, each independently

    risky, and each compensated with return for taking risk.

    It is possible to break down asset classes into building blocks, or factors, which

    explain the majority of their return and risk characteristics.

    Asset classes are an indirect way to invest in factors, but it is possible to invest

    directly in certain factors.

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    4/24

    3Independent Adviser Group

    Asset classes are bundles of exposures divided into categories such as

    equities, bonds, real assets, etc.

    Asset classes should be as independent as possible (minimal overlap) and in

    aggregate should cover the investment universe (minimal gaps).

    Correlations among asset classes are driven by many common factor

    exposures.

    The Basics Why the Interest in Factors?

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    5/24

    4Independent Adviser Group

    The finer the distinctions between various asset classes, the higher the resultingcorrelations.

    Typical asset allocation relies on sub-asset classes (such as large cap or small capU.S. equity or non-U.S. developed equity).

    Equity

    U.S.

    Large

    Small

    Non-

    U.S.Developed

    Emerging

    Debt

    U.S.

    Investment

    Grade

    HighYield

    Non-

    U.S.Developed

    Emerging

    Asset Class

    Sub-Asset Class

    The BasicsDistinction Between Asset Classes

    and Sub-Asset Classes

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    6/24

    5Independent Adviser Group

    The output of a mean-variance optimizer is a frontier composed of efficientportfolios.

    Portfolios are classified as efficient if they provide the greatest expectedreturn for a given level of expected risk.

    Optimization, and efficient portfolios, rely heavily on the quality of the inputs

    Capital market forecasts are the basis of this type of model.

    The Basics Mean Variance Optimization

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    7/24

    6Independent Adviser Group

    The frontier is composed of efficient portfolios across the risk spectrum.

    Less than 100% correlation among asset classes lead to diversification benefits whichefficient portfolios maximize.

    0% 5% 10% 15% 20% 25% 30% 35%

    0%

    2%

    4%

    6%

    8%

    10%

    U.S.

    Equity

    Non-U.S. Equity

    U.S. Fixed Income

    Real Estate

    PrivateEquity

    Cash

    Emerging Markets

    Equity

    Expected Risk (Standard Deviation)

    Expec

    tedReturn

    24% U.S. Equity

    13% Non-U.S. Equity5% Emerging Markets

    48% Fixed Income

    7% Real Estate

    3% Private Equity

    E(r) = 6.6%, E() = 10%

    Traditional Approach Classic Efficient Frontier

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    8/24

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    9/24

    8Independent Adviser Group

    Correlations across asset classes can be high because many asset classes are

    exposed to similar or common risk factors.

    For example, U.S. equity and U.S. corporate bonds share some common exposures, as doprivate equity and public equity.

    U.S. Equity U.S. Corporate Bonds

    GDPGrowth

    Value Liquidity

    Size Inflation

    Volatility

    Currency

    Momentum

    RealRates

    Liquidity

    DurationDefault

    Risk

    Inflation Volatility

    Currency

    Capital

    Structure

    Traditional Approach Some Limitations

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    10/24

    9Independent Adviser Group

    If asset classes are molecules, then factors are atoms.

    Those factors aggregate into the risk return characteristics of asset classes.

    We are interested in identifying and classifying various factors and exploring

    how they can be used to allocate assets.

    What Are Factors?

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    11/24

    10Independent Adviser Group

    Factors are the basic exposures which make up investments.

    They can be thought of as building blocks or risk premiums.

    Periodic Table of Factors

    Macroeconomic Regional Equity Fixed Income Other

    GDP GrowthSovereign

    ExposureSize Duration Liquidity

    Product ivit y Currency Value Convexity Leverage

    Real Interest

    Rates

    Emerging

    Markets

    (Institutions +Transparency)

    Momentum Credit Spread Real Estate

    Inflation Default Risk Commodities

    VolatilityCapital

    StructurePrivate Markets

    Classifying Factors A Sampling

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    12/24

    11Independent Adviser Group

    Ironically, even though risk factors are the basic building blocks of investments,

    there is no natural way of investing in many of them directly.

    Most risk factors can be accessed through derivatives or long/shortpositions.

    Simple examples:

    Inflation: Nominal Treasuries less TIPS

    Real Interest Rates: TIPS

    Volatility: VIX futures

    More complex examples:

    Value: Long Developed Equity Value, Short Developed Equity Growth

    Size: Long Developed Equity Small Cap, Short Developed Equity LargeCap

    Credit Spread: Long U.S. High Quality Credit, Short U.S.Treasury/Government

    Duration: Long U.S. Treasury 10+ Year, Short US Treasury 1-3 Year

    Emerging Markets: Long Emerging Equity, Short Developed Equity

    Not exactly possible:

    GDP Growth, Productivity, Momentum, Leverage

    Risk Factors How can one get exposure?

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    13/24

    12Independent Adviser Group

    Factor Risk and Return

    Factor CorrelationsEquity Value Size EM HY Spread Default Duration Real Rates Inflation Volatility

    Equity 1.00

    Value 0.07 1.00

    Size 0.39 0.15 1.00

    EM 0.43 -0.14 0.42 1.00HY Spread 0.69 0.06 0.40 0.41 1.00

    Default 0.54 0.03 0.38 0.34 0.74 1.00

    Duration -0.18 0.12 0.07 -0.04 -0.38 -0.28 1.00

    Real Rates 0.09 -0.04 0.32 0.19 -0.02 0.21 0.57 1.00

    Inflation -0.43 0.12 -0.41 -0.35 -0.49 -0.60 0.12 -0.67 1.00

    Volatility -0.69 0.04 -0.22 -0.32 -0.49 -0.40 0.15 -0.03 0.32 1.00

    Correlation < -0.3 Correlation between 0.4 and 0.6 Correlation > 0.6

    Factor Exposure Long Position Short Position Return Risk

    Equity MSCI World 3.20% 17.58%

    Value MSCI World Value MSCI World Growth 1.07% 6.31%

    Size MSCI World Small Cap MSCI World Large Cap 8.80% 8.22%

    EM MSCI Emerging Markets MSCI World 13.69% 11.47%

    HY Spread BC High Yield BC Intermediate Credit (IG) 2.57% 9.97%

    Default BC Aaa BC BBB 1.22% 5.08%

    Duration BC 20+ Year Treasuries BC 1-3 Year Treasuries 2.51% 11.95%Real Rates BC TIPS 7.03% 6.86%

    Inflation BC Treasuries BC TIPS -1.74% 5.05%

    Volatility CBOE VIX -4.05% 66.54%

    Historical

    Based on 10 Years of Monthly Data ending 12/31/2010

    Risk FactorsReturn, Volatility, and

    Correlation Historical Experience

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    14/24

    13Independent Adviser Group

    Factor returns (or premiums) are fairly low, most have returned less than 5% overthe past decade.

    Factor characteristics are extremely time sensitive, so varying the data horizoncan materially impact relationships.

    Factor standard deviations range widely between 5% to 66% (for the VIX).

    Correlations among factors are low, typically ranging from -0.5 to +0.6.

    Relatively highly correlated factors include Equity vs. High Yield and HighYield vs. Default.

    The average correlation for the 10 factors on the previous page is 0.03.

    This is significantly less than many asset class correlations which range from -

    0.15 to more than +0.90.

    Sub-asset classes like U.S. Small Cap vs. U.S. Large Cap are the mostcorrelated while relatively unrelated pairings such as U.S. 1-3 YearTreasuries vs. Private Equity are also the least correlated.

    Factor Assumptions Notes on the Historical Experience

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    15/24

    14Independent Adviser Group

    Equal weighting 10 sample factors into a portfolio with monthly rebalancing results in historicalequity-like returns with considerably less risk.

    Even with this simple weighting scheme, the factor portfolio exhibits relatively low volatility vs. atypical 60/40 portfolio.

    Both portfolios have a very low correlation with each other (-0.18).

    40%Russell 3000

    40%BC Aggregate

    20%MSCI ACWI ex-US

    vs.

    Historical Return: 5.16% 5.66%

    Historical Risk: 11.16% 5.75%

    Traditional 60/40 Portfolio

    Value

    Size

    EM

    HY

    SpreadDefault

    Duration

    Real Rates

    Inflation

    Volatility

    Factor Portfolio

    Based on 10 Years of Monthly Data ending 12/31/2010

    Equity

    Traditional vs. Factor Portfolio comparison

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    16/24

    15Independent Adviser Group

    Optimizing a factor portfolio using historical risk, return, and correlations results in a best-fit portfolio tuned for the 10 year data sample.

    This portfolio was selected from the efficient frontier based on its 5.75% expected risk.

    This example is meant to illustrate that optimization with factors is possible, but high qualityforward-looking inputs are necessary.

    Historical Return: 5.66% 10.25%

    Historical Risk: 5.75% 5.75%

    Equity

    Value

    Size

    EM

    HY

    SpreadDefault

    Duration

    RealRates

    Inflation

    Volatility

    Simple Factor Portfolio

    Real Rates

    VolatilityValue

    Size

    EM

    OptimizedFactorPortfolio

    vs.

    Based on 10 Years of Monthly Data ending 12/31/2010

    Simple vs. Optimized Portfolio Comparison

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    17/24

    16Independent Adviser Group

    Which factors to include?

    How to weight factors?

    Factor portfolios must be implemented using long and short exposures, oftenvia derivatives allocations.

    Active high frequency rebalancing among a large number of risk factors isnecessary to maintain the desired exposure.

    Factor portfolios resemble certain styles of hedge funds and risk parityapproaches (sans leverage).

    Portfol io Construction Factor Portfolios Present Notable Challenges

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    18/24

    17Independent Adviser Group

    Examine asset classes through a factor lens and group like asset classes

    together under macroeconomic scenarios.

    Inflation

    Eco

    nomicGrowth

    Low or Falling Growth

    High or Rising Inflation

    Inflation Linked Bonds (TIPS)

    Commodities

    Infrastructure

    HighGrowth

    HighInflation

    Real Assets: Real Estate,Timberland, Farmland,Energy, MLPs

    LowGrowthLowInflationorDeflation

    Cash

    Government Bonds

    HighGrowthLowInflation

    Equity

    Corporate Debt

    A Hybrid Approach Can We Still Apply Factors to Traditional Methods?

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    19/24

    18Independent Adviser Group

    Bucketing asset classes based on their response to macroeconomic scenarioscombines the transparency of investing with asset classes with the granularity offactor-based approaches.

    Sample Groupings:

    Capital Growth: Grow assets through relatively high long-term returns

    Global equity

    Private equity

    Income / Flight to Quality: Provide current income and protect capital in times

    of market uncertainty

    Global fixed income

    Cash equivalents

    Volatility Hedge: Earn returns between stocks and bond while attempting to

    protect capital and temper market volatility

    Diversified hedge funds

    Real Assets: Support the purchasing power of assets

    Real estate

    TIPS

    Commodities

    Natural Resource Equities

    Economic Scenarios What Roles do Asset Classes Play?

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    20/24

    19Independent Adviser Group

    0% 5% 10% 15% 20% 25% 30% 35%

    2%

    4%

    6%

    8%

    10%

    Optimization Set: 2011 ESB PvtAsset Class Risk-Reward

    Expected Volatility

    ExpectedR

    eturn

    Capital Growth

    Private Equity

    Income

    Real Assets

    Real Estate

    Volatility Hedge

    Cash

    Economic Scenarios Working Within the Mean Variance Optimizer

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    21/24

    20Independent Adviser Group

    Economic Scenarios The Building Blocks

    0% 5% 10% 15% 20% 25% 30% 35%

    2%

    4%

    6%

    8%

    10%

    Optimization Set: 2011 ESB PvtAsset Class Risk-Reward

    Expected Volatility

    ExpectedR

    eturn

    Broad Domestic EquityGlobal Ex-US Equity

    Private Equity

    Domestic Fixed

    High Yield

    Non US FixedTIPS

    Commodities

    Natural Resources

    Real Estate

    REITsVolatility Hedge

    Cash

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    22/24

    21Independent Adviser Group

    Factor Asset Class

    Target

    Al lo cation CapitalA

    ppreciati

    on

    InflationProtection

    Deflation/Crisi

    s

    Protection

    CapitalP

    reservati

    on

    Purcha

    sing

    Power

    Preservation

    IncomeGen

    eratio

    n

    Diversific

    ation

    Alpha

    Cash 2% Interest Rates 6%

    U.S. Government Bonds 4% International Government Bonds 2%

    Company Exposure 53%

    Global Credit 11% Global Equity 36% Private Equity 6%

    Real Assets 18%

    Real Estate 12% Infrastructure 3% TIPS 3%

    Special Opportunities 21%

    Absolute Return 6% Real Return 7% Distressed Debt 1% Mezzanine Debt 1% Structured Credit 1% Other 5%

    This is how one large plan combines factors and asset classes:

    Source: Adapted from the Alaska Permanent Fund Corporations Investment Policy Statement dated 12/1/2010.

    Economic Scenarios A Hybrid Example

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    23/24

    22Independent Adviser Group

    Practitioners will place increased emphasis on understanding how differentportfolios react to specific economic and capital market outcomes.

    Asset classes will be increasingly defined by their expected reactions to thesedifferent economic environments.

    Liquidity will be an explicit consideration both in strategic policy developmentand implementation.

    Next Steps

    in Asset AllocationWhat Does the Future Hold?

  • 7/29/2019 Asset Allocation Overview- Callan Research.pdf

    24/24

    23Independent Adviser Group

    Building portfolios of factors is challenging, but using factors to understandtraditionally constructed portfolios is very useful.

    Limitations, including basic investability, preclude most market participantsfrom solely investing along factor lines.

    However, factors provide a useful way to group traditional asset classesin macroeconomic buckets.

    Several multi asset class managers can engineer strategies with specificfactor exposures; these can be used to augment existing frameworks.

    Some factor exposures are explicitly incorporated within manager structureanalysis (such as liquidity, leverage, duration, currency, size, momentum,

    etc.).

    Conclusion