Amath 546/Econ 589 Risk Budgeting and Risk...
Transcript of Amath 546/Econ 589 Risk Budgeting and Risk...
Amath 546/Econ 589
Risk Budgeting and Risk Reporting
Eric Zivot
April 10, 2013
Risk decomposition using Euler’s theorem
Let RM(w) denote a portfolio risk measure that is a homogenous function ofdegree one in the portfolio weight vector w Euler’s theorem gives the additiverisk decomposition
RM(w) = 1RM(w)
1+ 2
RM(w)
2+ · · ·+
RM(w)
=X=1
RM(w)
= w0RM(w)
w
Terminology:
MCR =RM(w)
= marginal contribution to risk of asset i,
CR = ·MCR = contribution to risk of asset i,
PCR =CRRM(w)
= percent contribution of asset i
Then
RM(w) = 1 ·MCR1 +2 ·MCR2 + · · ·+ ·MCR= CR1 + CR2 + · · ·+ CR
1 =CR1RM(w)
+ · · ·+ CRRM(w)
= PCR1 + · · ·+ PCR
Risk Decomposition for Portfolio SD
RM(w) = (w) = (w0Σw)12 = w0
(w)
wNow
(w)
w=
(w0Σw)12
w=1
2(w0Σw)−122Σw
=Σw
(w0Σw)12=
Σw
(w)
⇒ (w)
= ith row of
Σx
(w)
Remark: In R, the PerformanceAnalytics function StdDev() performs this de-composition based on the sample covariance matrix of returns.
Example: 2 asset portfolio
(w) = (w0Σw)12 =³21
21 + 22
22 + 21212
´12Σw =
Ã21 1212 22
!Ã12
!=
Ã1
21 +212
222 +112
!Σw
(w)=
⎛⎝ ³1
21 +212
´(w)³
222 +112
´(w)
⎞⎠
Then
MCR1 =³1
21 +212
´(w)
MCR2 =³2
22 +112
´(w)
CR1 = 1 ׳1
21 +212
´(w) =
³21
21 +1212
´(w)
CR2 = 2 ׳2
22 +22
´(w) =
³22
22 + 1212
´(w)
and
PCR1 = CR1(w) =³21
21 +1212
´2(w)
PCR2 = CR2(w) =³22
22 +1212
´2(w)
How to Interpret and Use MCR
MCR =(w)
≈ ∆
∆⇒ ∆ ≈ MCR ·∆
However, in a portfolio of assets
1 + 2 + · · ·+ = 1
so that increasing or decreasing means that we have to decrease or increaseour allocation to one or more other assets. Hence, the formula
∆ ≈ MCR ·∆
ignores this re-allocation effect.
If the increase in allocation to asset is offset by a decrease in allocation toasset then
∆ = −∆
and the change in portfolio volatility is approximately
∆ ≈ MCR ·∆ +MCR ·∆
= MCR ·∆ −MCR ·∆
=³MCR −MCR
´·∆
1 2 21 22 1 2 12 120.175 0.055 0.067 0.013 0.258 0.115 -0.004875 -0.164
Table 1: Example data for two asset portfolio.
Consider two portfolios:
• equal weighted portfolio 1 = 2 = 05
• long-short portfolio 1 = 15 and 2 = −05
MCR CR PCR = 01323
Asset 1 0.258 0.5 0.23310 0.11655 0.8807Asset 2 0.115 0.5 0.03158 0.01579 0.1193
= 04005Asset 1 0.258 1.5 0.25540 0.38310 0.95663Asset 2 0.115 -0.5 -0.03474 0.01737 0.04337
Table 2: Risk decomposition using portfolio standard deviation.
Interpretation: For equally weighted portfolio, increasing 1 from 05 to 06decreases 2 from 0.5 to 0.4. Then
∆ ≈ (MCR1 −MCR2) ·∆
= (023310− 003158)(01)= 002015
For the long-short portfolio, increasing 1 from 15 to 16 decreases 2 from-0.5 to -0.6. Then
∆ ≈ (MCR1 −MCR2) ·∆
= [025540− (−003474)] (01)= 002901
Remark
The reallocation does not have to be all in one asset. Suppose, the reallocationis spread across all of the other assets 6= so that
∆ = −∆ s.t.X 6=
= 1
Then X 6=
∆ = −X 6=
∆ = −∆
X 6=
= −∆
and
∆ ≈ MCR ·∆ +X 6=
MCR ·∆
=
⎛⎝MCR ·−X 6=
·MCR
⎞⎠∆
In matrix notation the result can be written as
∆ ≈³MCR0α
´∆
MCR = (MCR1 MCR)0
α = (−1 −−1 1−+1 − )0
Interpreting MCR when (w) = (w) or (w)
• ∆ = −∆
∆ (w) =³MCR
−MCR
´·∆
∆(w) =³MCR −MCR
´·∆
• ∆ = −∆ s.t.P 6= = 1
∆ (w) ≈³MCR 0α
´∆
∆(w) ≈³MCR 0α
´∆
Beta as a Measure of Asset Contribution to Portfolio Volatility
For a portfolio of assets with return
(w) = 11 + · · ·+ = w0R
we derived the portfolio volatility decomposition
(w) = 1(w)
1+ 2
(w)
2+ · · ·+
(w)
= w0
(w)
w(w)
w=
Σw
(w)(w)
= ith row of
Σw
(w)
With a little bit of algebra we can derive an alternative expression for
MCR =(w)
= ith row of
Σw
(w)
Definition: The beta of asset with respect to the portfolio is defined as
=((w))
((w))=
((w))
2(w)
Result: measures asset contribution to (w) :
MCR =(w)
= (w)
CR = (w)
PCR =
Remarks
• By construction, the beta of the portfolio is 1
=((w) (w))
((w))=
((w))
((w))= 1
• When = 1
MCR = (w)
CR = (w)
PCR =
• When 1
MCR (w)
CR (w)
PCR
• When 1
MCR (w)
CR (w)
PCR
MCR CR PCR = 01323
Asset 1 0.258 0.5 1.76 0.23310 0.11655 0.8807Asset 2 0.115 0.5 0.24 0.03158 0.01579 0.1193
= 04005Asset 1 0.258 1.5 0.64 0.25540 0.38310 0.95663Asset 2 0.115 -0.5 -0.09 -0.03474 0.01737 0.04337
Table 3: Risk decomposition using portfolio standard deviation.
Remarks:
• For the equally weighted portfolio, Asset 1 is a risk enhancer and Asset 2is a risk reducer
• For the long-short portfolio, 2 0!
− − Decomposition of Portfolio Volatility
Recall,
MCR =(w)
= ith row of
Σw
(w)=
((w))
(w)
Using
= ((w)) =((w))
(w)⇒ ((w)) = (w)
gives
MCR =(w)
(w)=
Then
CR = ×MCR = × ×
= allocation× standalone risk× correlation with portfolio
Remarks:
• × = standalone contribution to risk (ignores correlation effects withother assets)
• CR = × only when = 1
• If 6= 1 then CR ×
MCR CR PCR = 01323
Asset 1 0.258 0.5 0.90 0.23310 0.11655 0.8807Asset 2 0.115 0.5 0.27 0.03158 0.01579 0.1193
= 04005Asset 1 0.258 1.5 0.99 0.25540 0.38310 0.95663Asset 2 0.115 -0.5 -0.30 -0.03474 0.01737 0.04337
Table 4: Risk decomposition using portfolio standard deviation.
Remarks:
• For the equally weighted portfolio, both assets are positively correlatedwith the portfolio
• For the long-short portfolio, Asset 2 is negatively correlated with the port-folio
Remark:
• Goldberg, Hayes, Menchero and Mitra (2009) “Extreme Risk Manage-ment”, Barra Technical Report, extends the and − − decompo-sitions to ES
Asset $ Asset VaR MCVaR CVaR PCVaRAsset 1 10 .10 .01 .10 -.03 .003 .01 .10Asset 2 20 .20 .02 .12 -.04 .002 .02 .11... ... ... ... ... ... ... ... ...Asset N 5 .05 .01 .07 -.07 .010 .04 .13Portfolio 100 1 .03 .08 .08 1
Table 5: Portfolio VaR Report
Portfolio VaR and ES Reports
A common portfolio risk report summarizes asset and portfolio risk measuresas well as risk budgets