Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 55

Black-Litterman

Asset Allocation Model

QSS Final Project

Midas Group Members


Bo Jiang, Tapas Panda, Jing Lin, Yuxin Zhang

Under the Guidance of Professor Campbell Harvey

April 27, 2005


Agenda

 Part 1: Motivation and Intuition


 Part 2: Analytics
 Part 3: Numerical Example
 Part 4: BL in Practice
 Part 5: Test the Model
 Epilogue: 3 Recommendations
Part 5: Test the Model

The best way to test the model is…


Introspection…
Part 1: Motivation & Intuition
The Problems of Markowitz Optimization

 Highly-concentrated portfolios
 Extreme portfolios

 Input-sensitivity
 unstable
 Estimation error maximization

 Unintuitive
 No way to incorporate investor’s view
 No way to incorporate confidence level
 No intuitive starting point for expected return.
 Complete set of expected return is required.
Black-Litterman Model

 B-L model uses a Bayesian approach to


combine the subjective views of an investor
regarding the expected returns of one or
more assets with the market equilibrium
vector of expected returns (the prior
distribution) to form a new mixed estimate of
expected returns (the posterior distribution).
How does BLM work?
 Start with the market returns using reverse
optimization and CAPM.

 Apply your own unique views of how certain


markets are going to behave.

 The end result includes both a set of


expected returns of assets as well as the
optimal portfolio weights.
Intuition of BLM
 If you do not have views, you hold the
market portfolio (the benchmark).

 Your views will tilt the final weights away


from the market portfolio, the degree to
which depending on how confident you
are about your views.
Road Map
Part 2: Analytics
Equilibrium Returns (1)
 Equilibrium Return
 =current Market collective forecasts of next
period returns; i.e., the market’s collective
view on future returns
 =reverse optimized returns
 this Market View is to be combined with Our
View; and the combination (using GLS) will
take the estimation error of either views into
consideration.
Equilibrium Returns (2)

 Assume Market has the following


attributes
 N assets
 Expected Return vector μ[Nx1]
 Expected covariance Matrix ∑[NxN]
Equilibrium Returns (3)

 Today when the trades took place,


market collectively reached the
equilibrium (supply = demand).
 To do this it had ran the Markowitz
mean-variance optimization and
reached the optimized weights w[Nx1]
– which are the current market
capitalization weights
Equilibrium Returns (4)
 Max [w’μ – (λ/2)w’∑w]
 Note: This is derived from the utility theory and
multivariate normal distribution – Financial
Economics 101
λ = risk aversion coefficient (E(M) –rf)/σ(mkt)^2)
E(M) = Expected market or benchmark total return
λ is found from historical data (approx = 3.07)
 Solve δw’μ /δw - δ((λ/2)w’∑w)/ δw = 0
 They got μ = λ ∑w
 Note: two most important matrix derivation formula
δw’μ /δw = μ and δ(w’∑w)/ δw = 2∑w
Equilibrium Returns using
Implied Beta
 Equilibrium Returns can be calculated by
using the “implied Beta” of assets.
 μ = β(implied)*(risk premium of market )
 Implied β = ∑*w(mkt)/(w(mkt)T*∑w(mkt))
The denominator is basically the variance of
market portfolio. The numerator is the
covariance of the assets in the market portfolio.
Asset weights are the equilibrium weights.
Covariance matrix ∑ is historical covariance.
What is the estimation error of
the Equilibrium Returns?
 A controversial issue in BL model.
 Since the equilibrium returns are not
actually estimated, the estimation error
cannot be directly derived.
 But we do know that the estimation error of
the means of returns σE[r(i,t+1)] should be less
than the covariance of the returns.
 A scalar τ less than 1 is used to scale down
the covariance matrix (Σ) of the returns.
 Some say that “τ =0.3 is plausible”.
Forming Our View (1)

 Our view is:


 Q=Pu+η, μ~Φ(0,Ω)
 Note: same as Pu=Q+η, because η~Φ(0,Ω)  -
η~Φ(0,Ω)
 u is the expected future returns (a NX1
vector of random variables).
 Ω is assumed to be diagonal (but is it
necessary?)
Forming Our View (2)

 What does this Q=P*u+η, Or


equivalently P*u=Q+ η mean?
 Look at P*u:
each row of P represents a set of weights on the N
assets, in other words, each row is a portfolio of the
N assets. (aka “view portfolio”)
u is the expected return vector of the N assets
 P*u means we are expressing our
views through k view portfolios.
Forming Our View (3)

 Our Part 3 Numerical Example will


show some examples of the process of
expressing views.
 The Goldman Sachs Enigma is how
they express views quantitatively.
Forming Our View (4)

 Why is expressing views so important?

 Because the practical value of BL model lies


in the View Expressing Scheme; the model
itself is just a publicly available view
combining engine.
 Our view is the source of alpha.
 Expressing views quantitatively means efficiently and
effectively translate fundamental analyses into Views
Forming Our View (5)

 We will try to decode Goldman Sachs


Enigma in Part 4 “Applications”.
Combining Views (1)
{

Generalized Least
μComb
Square Estimator of μ

μComb
Combining Views (1)
{

Generalized Least
μComb
Square Estimator of μ

μComb
Combining Views (2)

Var(μComb)
Now we have a combined forecast of the expected
returns.
The next step is to do Markowitz Mean-Variance
Optimization.

 By using the combined forecasted means

 and the forecasted covariance matrix ∑.

 So we start with Markowitz (reverse optimization) and


CAPM (implied beta).

 Go though Black-Litterman View Combining engine.

 And end up with Markowitz again with predictive means,


(and forward looking return covariance matrix.)
Part 3: Numerical Example
An Eight Assets Example…
Asset Class µHist µP wmkt
US Bonds 3.15% 0.08% 19.34%
Int’l Bonds 1.75% 0.67% 26.13%
US Large Growth -6.39% 6.41% 12.09%
US Large Value -2.86% 4.08% 12.09%
US Small Growth -6.75% 7.43% 1.34%
US Small Value -0.54% 3.70% 1.34%
Int’l Dev. Equity -6.75% 4.80% 24.18%
Int’l Emerg. Equity -5.26% 6.60% 3.49%

Weighted Average -1.97% 3.00%


Standard Deviation 3.73% 2.53%

High 3.15% 7.43% 26.13%


Low -6.75% 0.08% 1.34%

 μHist is historical mean asset returns


 μp is calculated relative to the market cap. weighted portfolio
using implied betas and CAPM model.
 Market portfolio weights wmkt is based on market
capitalization for each of the assets
Market Returns П(nx1)
 Market returns are derived from known
information using Reverse Optimization:
П = ‫∑ ג‬wmkt
 П (nx1) is the excess return over the risk free
rate
 ‫ ג‬is the risk aversion coefficient
 ∑(nxn) is the covariance matrix of excess returns
 Wmkt (nx1) is the market capitalization weight of
the assets
Risk Aversion Coefficient ‫ג‬

 More return is required for more risk

‫(=ג‬E (r) – rf )/σ2=Risk Premium/Variance

 Using historical risk premium and


variance, we got a ‫ ג‬of aprrpoximately
3.07
Coviriance Matrix ∑

 Coviriance Matrix ∑(nxn)

Asset Class 1 2 3 4 5 6 7 8
1. US Bonds 0.001005 0.001328 -0.000579 -0.000675 0.000121 0.000128 -0.000445 -0.000437
2. Intl Bonds 0.001328 0.007277 -0.001307 -0.00061 -0.002237 -0.000989 0.001442 -0.001535
3. US Large Growth -0.000579 -0.001307 0.059852 0.027588 0.063497 0.023036 0.032967 0.048039
4. US Large Value -0.000675 -0.000610 0.027588 0.029609 0.026572 0.021465 0.020697 0.029854
5. US Small Growth 0.000121 -0.002237 0.063497 0.026572 0.102488 0.042744 0.039943 0.065994
6. US Small Value 0.000128 -0.000989 0.023036 0.021465 0.042744 0.032056 0.019881 0.032235
7. Int'l Dev. Equity -0.000445 0.001442 0.032967 0.020697 0.039943 0.019881 0.028355 0.035064
8. Int'l Emerg.Equity -0.000437 -0.001535 0.048039 0.029854 0.065994 0.032235 0.035064 0.079958
Market Returns П(nx1)

Π = λΣwmkt
Asset Class µHist µP Π
US Bonds 3.15% 0.08% 0.08%
Int’l Bonds 1.75% 0.67% 0.67%
US Large Growth -6.39% 6.41% 6.41%
US Large Value -2.86% 4.08% 4.08%
US Small Growth -6.75% 7.43% 7.43%
US Small Value -0.54% 3.70% 3.70%
Int’l Dev. Equity -6.75% 4.80% 4.80%
Int’l Emerg. Equity -5.26% 6.60% 6.60%

Weighted Average -1.97% 3.00% 3.00%


Standard Deviation 3.73% 2.53% 2.53%

High 3.15% 7.43% 7.43%


Low -6.75% 0.08% 0.08%
The Black – Litterman Model
 The Black – Litterman Formula

• E[R] (nx1) is the new Combined Return Vector


• τ is a scalar
• ∑ (nxn) is the covariance matrix of excess returns
• P (kxn) is the view matrix with k views and n assets
• Ω (kxk) is a diagonal covariance matrix of error terms
from the expressed views
• Π (nx1) is the implied market return vector
• Q (kx1) is the view vector
What is a view?
 Opinion: International Developed Equity will be
doing well
 Absolute view:
 View 1: International Developed Equity will have an
absolute excess return of 5.25% (Confidence of view =
25%)
 Relative view:
 View 2: International Bonds will outperform US bonds by
25 bp (Confidence of view = 50%)
 View 3: US Large Growth and US Small Growth will
outperform US Large Value and US Small Value by 2%
(Confidence of View = 65%)
What Is The View Vector Q Like?

5.25% ε1
Q+ε= 0.25% + ε2
2.00% ε3

 Unless a clairvoyant investor is 100% confident in the


views, the error term ε is a positive or negative value
other than 0
 The error term vector does not enter the Black –
Litterman formula; instead, the variance of each error
term (ω) does.
What Is The View Matrix P Like?
US Bonds Intl Bonds US Lg Growth US Lg Value US Sml Growth US Sml Value Int'l Dev. Eqt Int'l Emerg.Eqt
0 0 0 0 0 0 1 0
P= -1 1 0 0 0 0 0 0
0 0 0.9 -0.9 0.1 -0.1 0 0

 View 1 is represented by row 1. The absolute view results in


the sum of row equal to 1
 View 2 & 3 are represented by row 2 & 3. Relative views
results in the sum of rows equal to 0
 The weights in view 3 are based on relative market cap.
weights, with outperforming assets receiving positive weights
and underperforming assets receiving negative weights
Finally, The Covariance Matrix Of The
Error Term Ω
0.0007089 0 0
Ω= 0 0.000141 0
0 0 0.000866

 Ω is a diagonal covariance matrix with 0’s in all of


the off-diagonal positions, because the model
assumes that the views are independent of each
other
 This essentially makes ω the variance (uncertainty)
of views
Go Back to B-L Formula…

 First bracket “[ ]” (role of “Denominator“) : Normalisation


 Second bracket “[ ]” (role of “Numerator“) : Balance between
returns Π (equilibrium returns) and Q (Views). Covariance (τ
Σ)-1 and confidence P’ Ω-1P serve as weighting factors, and
P’ Ω-1Q = P’ Ω-1P P-1 Q
 Extreme case 1: no estimates ⇔ P=0: E(R) = Π i.e. BL-returns
= equilibrium returns.
 Extreme case 2: no estimation errors ⇔ Ω -1→ ∞: E(R) = P -1Q
i.e. BL-returns = View returns.
Return Vector & Resulting Portfolio
Weights
Π
E [ R ] - Π

Asset Class E[R] norm wmkt


US Bonds 0.07% 0.08% -0.02% 29.88% 28.83% 19.34% 10.54%
Int’l Bonds 0.50% 0.67% w
-0.17% 15.59% 15.04% 26.13% -10.54%
=
US Large Growth 6.50% 6.41% 0.08% 9.35% 9.02% 12.09% -2.73%
(
US Large Value 4.32% 4.08% 0.24%
λ 14.82% 14.30% 12.09% 2.73%
US Small Growth 7.59% 7.43% 0.16%
Σ 1.04% 1.00% 1.34% -0.30%
US Small Value 3.94% 3.70% 0.23%
) 1.65% 1.59% 1.34% 0.30%
-
Int’l Dev. Equity 4.93% 4.80% 0.13%
1 27.81% 26.84% 24.18% 3.63%
Int’l Emerg. Equity 6.84% 6.60% 0.24%
E 3.49% 3.37% 3.49% 0.00%
Sum[ 103.63% 100% 100% 3.63%
R
]

Π = λΣwmkt w =(λΣ) -1Π

w =(λΣ) -1E[R]
Combined Return E[R] vs. Equil. Return Π

8%
Π E[R]

6%

4%

2%

0%
US Bonds Int ’l Bonds US La rge US La rge Va lue US S ma ll US S ma ll Va lue Int ’l De v. Equit y Int ’l Eme rg.
Growt h Growt h Equit y
Resulting Asset Allocations Changed A
Lot…
View 1 – Bullish view on Int’l Dev. Equity
View 2: Int’l bonds will outperform increases allocation.
US bonds less than market implied.

35%
Market Cap. Weight New Weight
30%
View 3 – Growth tilt towards value
25%

20%

15%

10%

5%

0%
US Bonds Int ’l Bonds US La rge US La rge Va lue US S ma ll US S ma ll Va lue Int ’l De v. Int ’l Eme rg.
Growt h Growt h Equit y Equit y
Part 4: BL in Practice
Applications
 Just now we presented unconstrained optimization.

 Of course constraints can be added to the optimizer.

 Also, the market portfolio can be replaced with any benchmark


portfolio, and the Mean-Variance objective function can be replaced
by any other risk models (maximize risk adjusted returns.)

 Littleman, “The real power of the BL model arises when there is a


benchmark, a risk or beta target, or other constraints, or when
transaction costs are taken into consideration. In these more complex
contexts, the optimal weights are no long obvious or intuitive”.

 Wai Lee, “The model can be used to combine different models or


signals, ”such as valuation model and technical analysis.
BL Limitation

 What we presented is still in the mean-


variance optimization framework, which
cannot deal with higher moments.

 For ideas of handling both estimation error


and higher moments, see “Portfolio
Selection With Higher Moments: A
Bayesian Decision Theoretic Approach”,
by our professor Campbell Harvey.
Attempt to decode GSQE (1)

 Return generating model is the source


of alpha.
 Ideally, views and their estimation error
should be generated quantitatively.
 That’s what Goldman Sachs
Quantitative Equity does.
 How the heck do they actually do it?
Attempt to decode GSQE (2)

 Credit Swisse’ sort of confirmed our


decoding of GSQE. Previously we
thought there was 30% chance that we
have decoded GSQE; now we are 80%
sure.
 The two companies are doing virtually
the same thing in terms of generating
views quantitatively.
Attempt to decode GSQE (3)

 Ri,t+1 =f(z1,z2,z3,z4,z5,z6), z is firm attributes.


 The factor loading is just partial derivative.
 Credit Swisse uses long-short to get this
partial derivative (5 long-short portfolios)
 Goldman Sachs has another scheme to do it:
a special kind of Characteristic Portfolio (6
view portfolios).
 Whatever, the essence is still to get the
partial derivative for each factor.
Epilogue:
3 Recommendations
1st:
To Our Professor and Fuqua
Fuqua Course 999: “Quantitative Beauty Selection”

Mahalanobis Distance
By the way, this is your long list.
2nd:
To Corporate America
Quantitative Employee Selection

Better Get Rid Of Cover Letters and


Interviews !!
unless …
3rd:
To Ourselves
You can quantify pretty much everything in
the pragmatic world, but

Do not ‘calculate’:
 Compassion

 Friendship and Love

 Aesthetic Value

 Intellectual Curiosity

 Respect for Individuality

You might also like