Scribd
Upload
68 views

Quantitative Risk Management WS1920 Assignment 6

This document contains an exercise sheet for a quantitative risk management course. It provides 4 exercises involving risk measurement using historical simulation, peak-over-threshold modeling, and copula modeling. The exercises will be held on January 28, 2020 in two classrooms for two groups of students. Students are instructed to attempt the exercises at home before the scheduled class.

Uploaded by

Ka Wing Ho
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
68 views

Quantitative Risk Management WS1920 Assignment 6

This document contains an exercise sheet for a quantitative risk management course. It provides 4 exercises involving risk measurement using historical simulation, peak-over-threshold modeling, and copula modeling. The exercises will be held on January 28, 2020 in two classrooms for two groups of students. Students are instructed to attempt the exercises at home before the scheduled class.

Uploaded by

Ka Wing Ho
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Chair of Mathematical Finance

Department of Mathematics
Technical University of Munich
Quantitative Risk Management
Prof. Dr. Matthias Scherer and Henrik Sloot

Exercise sheet 6
The exercise is held at January 28, 2020 in room BC2 0.01.04 (group 1) and BC2 0.01.05 (group 2).
You should try to solve the exercises at home before the exercise.

Exercise 6.1
In the following we consider the Danish fire insurance data. The dataset consists of 2167 fire insurance
losses over 1000000 Danish kroner from 1980 to 1990 and is contained in the R package evir (to load the
data type data ( danish ) in R). We assume that the observations are realisations of an iid sequence
of random variables.
a) Use the historical simulation method to estimate VaRλ for λ ∈ {0.01, 0.001}.
b) Construct a mean excess plot to select a suitable threshold t. Use the function gpd () to obtain
maximum likelihood estimates for γ and σ. Check graphically the fit of the GPD model to the data
for your chosen threshold (see ? gpd).
c) Using the Peak-over-Threshold method, estimate VaRλ for λ ∈ {0.01, 0.001}.
bHn,k to estimate VaRλ for λ ∈ {0.01, 0.001} based on the approach
d) Now compute the Hill estimator γ
given on Page 141 of the lecture slides (R function hill). Select a suitable number k of upper
order statistics.

Exercise 6.2
Let X1 , . . . , Xn be iid random variables with distribution function F . Further, let Mk := max {X1 , . . . , Xk }
for k = 1, . . . , n. Define a sequence of random variables N1 , . . . , Nn as follows:
n
X
N1 = 1, Nn = 1 + I{Xk >Mk−1 } , n ≥ 2.
k=2

Give an interpretation to the random variable Nn . Compute the expectation E [Nn ] for n ≥ 2.

Exercise 6.3
Let C0 and C1 be copulas and let us consider the parameter θ ∈ [0, 1]. Show that the weighted arithmetic
mean (1 − θ)C0 + θC1 is also a copula.

Exercise 6.4
Sklar’s Theorem (cf. slide 168 ) allows to treat multivariate distribution functions in terms of a corre-
sponding copula and the marginal distribution functions: A function F : Rd → [0, 1] is the distribution
d
function of some random vector (X1 , . . . , Xd ) if and only if there exist a copula C : [0, 1] → [0, 1] and
univariate distribution functions F1 , . . . , Fd : R → [0, 1], such that

C(F1 (x1 ), . . . , Fd (xd )) = F (x1 , . . . , xd ), x1 , . . . , xd ∈ R.

The distribution function of component Xj equals Fj , j = 1, . . . , d, and the link between F and C is
one-to-one if all functions F1 , . . . , Fd are continuous.
Let X1 and X2 be random variables with joint distribution function given by

F (x1 , x2 ) := (1 + e−x1 + e−x2 )−1 , x1 , x2 ∈ R.


c Technical University of Munich, Chair of Mathematical Finance
2

a) Show that X1 and X2 have the following univariate marginal functions

F1 (x1 ) := (1 + e−x1 )−1 ,


F2 (x2 ) := (1 + e−x2 )−1 .

b) Prove that the copula of X1 and X2 is given by


u1 u2
C(u1 , u2 ) := .
u1 + u2 − u1 u2

Quantitative Risk Management — Exercise sheet 6


The exercise is held at January 28, 2020 in room BC2 0.01.04 (group 1) and BC2 0.01.05 (group 2). You
should try to solve the exercises at home before the exercise.

You might also like