You are on page 1of 5

# VaR

## Monte Carlo Simulation

 CMRA
Monte Carlo Simulation
Monte Carlo is most helpful when some or all assets in a
portfolio are not amenable to analytical treatment

## 1 Scenario Generation -produce a large number of

future price scenarios
2 Portfolio valuation - for each scenario, compute a
portfolio value
3 Summary - report the results of the simulation, either
as a portfolio distribution or as a particular risk
measure

 CMRA 2
Monte Carlo Simulation
Scenario generation
Monte Carlo begins with the generation of n normal
variables with unit variance and correlation matrix Σ.

## ♦ Decompose the correlation matrix Σ using the Cholesky

factorization, yielding Σ=ATA

variables

## ♦ Let Y = AZ. The elements of Y will each have unit variance

with the correlation matrix

 CMRA 3
Monte Carlo Simulation
Cholesky Decomposition

we find A?

## s11 s12 s13  a11 0 0  a11 a21 a31 

    0 a a 
s s s = a a
 21 22 23  21 22  0  22 32 
s31 s32 s33  a31 a32 a33 0 0 a33
   

## s11 s12 s13  a11

2
a11a21 a11a31 
   
s =
 21 22 23  11 21 21 22
s s a a a2
+ a 2
a a
21 31 + a32 22 
a
s31 s32 s33  a a a a + a a a2 + a2 + a2 
   11 31 21 31 32 22 31 32 33 

 CMRA 4
General Result
1/ 2
 i−1 
 2
aii = sii −∑aik 
 k=1 
 

i−1
1
aij = (sij −∑aikajk )1/ 2 j = i +1, i + 2,, N
aii k=1

 CMRA 5