1. Page 1. Daniel HERLEMONT. Financial Risk Management. Portfolio Risk, Analytical Methods. Following P. Jorion,. Value at Risk, McGraw-Hill, Chapter 7.
Portfolio Risk, Analytical Methods Following P. Jorion, Value at Risk, McGraw-Hill, Chapter 7
Daniel HERLEMONT
Portfolio of Random Variables N
Y = ∑ wi X i = wT X i =1
N
E (Y ) = µ p = w E ( X ) = w µ X = ∑ wi µ i T
T
N
N
i =1
σ 2 (Y ) = wT Ωw = ∑∑ wiσ ij w j i =1 j =1
VAR p = αW wT Ωw Daniel HERLEMONT
Page 1 1
Diversified VAR
Daniel HERLEMONT
VAR and correlations
Daniel HERLEMONT
Page 2 2
Exercise
Daniel HERLEMONT
VAR Tools
Marginal VAR
Incremental VAR
Component VAR
Daniel HERLEMONT
Page 3 3
Marginal VaR
How risk sensitive is my portfolio to increase in size of each position? - calculate VaR for the entire portfolio VaRP=X - increase position A by one unit (say 1% of the portfolio) - calculate VaR of the new portfolio: VaRPa= Y - incremental risk contribution to the portfolio by A: Z = X-Y i.e. Marginal VaR of A is Z = X-Y
Marginal VaR can be Negative; what does this mean...?
Daniel HERLEMONT
Daniel HERLEMONT
Page 4 4
Relation to beta and CAPM
Daniel HERLEMONT
Incremental VaR
Risk contribution of each position in my portfolio. - calculate VaR for the entire portfolio VaRP= X - remove A from the portfolio - calculate VaR of the portfolio without A: VaRP-A= Y - Risk contribution to the portfolio by A: Z = X-Y i.e. Incremental VaR of A is Z = X-Y Incremental VaR can be Negative; what does this mean...?
Daniel HERLEMONT
Page 5 5
Daniel HERLEMONT
Daniel HERLEMONT
Page 6 6
Example
Daniel HERLEMONT
Component VAR Objective: to get a risk decomposition of the portfolio Taking individual VAR is not useful since it ignore diversification Rather, the component VAR defined as
indicates how the portfolio would change (approximately) if the component is deleted from the portfolio
Daniel HERLEMONT
Page 7 7
Daniel HERLEMONT
VaR decomposition VaR
Incremental VaR Marginal VaR Portfolio VaR Component VaR 100
VaR95%=1.65⋅σP = $835M VaR99%=2.33 ⋅σP=$1.18B Actual loss was $1.3B Daniel HERLEMONT
Baring's Risk
Daniel HERLEMONT
Page 10 10
Simplifying the correlation Matrix
With 10 assets, the covariance matrix need to estimate 10*11/2=55 elements. With 100 assets, we need to estimate 5050 elements ... lead to estimation errors need for simplification and robustness
One Factor (Market) Model - Sharpe / CAPM Multi Factors Model - APT, Ross & Roll, BARRA, ... Implicit Factors (Principal Component Analysis) Daniel HERLEMONT