FINM 3008/8016 Applied Portfolio Construction
Applied Portfolio Construction
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
Asset Allocation: More Methods
Alpha & Beta
FINM 3008/8016
Applied Portfolio Construction
1
Today’s lecture
What you can expect to learn:
• Some more methods to assist in making asset
allocation decisions
• About the taxonomy of alpha and beta
2
Some asset allocation methods
• Mean-variance analysis
• Benchmark-relative
• Two-stage approach (Russell Investment)
• Dynamic strategies (Ibbotson Associates)
• Scenario analysis (Gosling, JPM, Fall 2010)
• Liability-driven investing
• Fundamental risk approach (Russell Investment)
• Factor investing (JP Morgan)
• Hierarchical clustering multi-asset multi-factor asset allocation
(Invesco)
Message: They are just models! Apply multiple methods to get
more robust results and exercise your judgment!
3
Theoretical issues with traditional MPT
Issue Problem
1. Investment horizon is undefined Yet it matters, especially if returns are
not iid (identically and independently distributed)
2. Single period model Real world is multi-period, with
stochastically changing investment
opportunities
3. Assumes risk aversion is only
investor difference that matters
(‘separation’ => investors hold
combination of M and Rf)
Other investor differences matter,
e.g. objectives, liabilities, investment
horizon, opportunities, costs, taxes
(=> separation unlikely to hold in practice,
i.e. different portfolios may be optimal)
4. Portfolio optimization across all
available assets
Not necessarily feasible (curse of
dimensionality); multi-factor models
may be more effective
4
Benchmark-Relative
• Estimate tracking error: TE = Std Dev(rPortfolio – rBenchmark)
– Data-based estimates: create and analyze two time series
– Parametric-based: analyze portfolio defined by wi,Portfolio – w
i,Benchmark
• Simplest approach is to impose a TE constraint
• Chow (FAJ, 1995) suggests this objective function:
• TP – values of 0.50 to 1.00 typically used in industry
• TTE – values of 0.10 to 0.50 have been suggested
TE
TE
P
P
P TT
rEUtility
22
E[rP] expected return for portfolio
σP
2 portfolio variance
σTE
2 portfolio tracking error variance
TP portfolio risk tolerance
TTE tracking error risk tolerance
5
Two stage approach
• Setting the stage: collect inputs (goals,
preferences, circumstances, capital markets)
• Stage 1: decide broad asset class exposures
– Optimization is safe
– Use equal expected returns for major equity markets
– Resulting AA will meet plan risk-return objectives
– Evaluate against client-specific objectives
• Stage 2: decide performance enhancing
exposures
– Optimization is unsafe
– Rely on good judgment: supportable investment beliefs, logic,
experience, simulations and sensitivity analysis
– Evaluate against client-specific liabilities
6
Dynamic strategies: Lifetime asset
allocation
7
Dynamic strategies: typical lifecycle of
human capital and financial capital
8
Scenario analysis
• Defining the individual scenarios: economic
growth, inflation, investor sentiment
• Generating return assumptions
• Assigning scenario probabilities
• Generating scenario output: provide rich
information of return distribution, risk and
diversification
9
10
Liability-Driven Investing (LDI)
• Liability can be viewed as a negative asset
• Difference: it is not usually a choice variable
• The trade-off: Surplus risk versus expected return (or cost)
Surplus (Deficit) = Assets – Liabilities
Funding Ratio = Assets / Liability
• One approach: (more detail on next slide)
a) Identify minimum risk portfolio, i.e. best liability hedge
b) Find preferred position – the final asset allocation decision is
always linked to stakeholder’s objectives, i.e. the sponsoring entity
may push to increase asset risk for higher returns, hence less
contribution is needed
• Another approach is based around cash flow matching 11
Source: Submission to the Financial System Inquiry”, RBA, March 2014
12
What drives DB liabilities
1. Discount rate applied to projected benefits
– approaches vary across jurisdictions
– often tied to bond yields, e.g. Aa corporate, treasury bonds
– expected return on assets used in US public sector
2. Drivers of benefit projections
– salary growth
– turnover of beneficiaries
– longevity (where full-life pension)
– options offered
3. Inflation – where it has differential effects on discount rate
and benefit projections
4. Other – accounting and regulation (eg AASB119)
13
VLiability = Benefits
(1+r)n
14
Wage, years of service, etc.
Discount rate = real
interest rate + inflation
Implementing LDI
1. Identify the measure of liability value (and hence surplus)
– For a DB fund, this might be the actuarial valuation (NPV) of future
benefit payments, or Projected Benefit Obligation (PBO)
2. Characterize how liability value relates to asset values:
– Mean-variance framework: covariance with assets
– Duration-matching (see Siegel & Waring, FAJ, 2004)
– Economic or factor-based
3. Locate the minimum risk portfolio
4. Characterize the trade-off (return vs surplus volatility / shortfall)
5. Choose preferred portfolio, given objectives & preferences
(all stakeholders have a say in this decision!)
15
Fundamental Risk Approach
• Economic diversification
• A way of thinking about portfolios, and how they might
be improved at the margin:
• Step 1: Identify the common, fundamental risks to
which the overall portfolio is most exposed
• Step 2: Consider how the portfolio could be modified to
reduce risk exposure without sacrificing too much E[r]