Risk theory: Loss models and risk measures, MATH 4280 3.

00 F
Instructor: Edward Furman
Homework 5

1. A portfolio contains 16 independent risks, each with a gamma distribution

with parameters α = 1 and θ = 250. The pdf is

e−x/θ (x/θ)α−1
f (x) = , x > 0.
Γ(α)

m
Give an expression using the incomplete gamma function

er as
∫ v

co
Γ(u, v) = e−x xu−1 dx

eH w
0

o.
for the probability that the sum of the losses exceeds 6,000. Then approxi-
rs e
ou urc
mate this probability using the central limit theorem.

2. The severities of individual claims have the Pareto distribution with pa-
o
aC s

rameters α = 8/3 and θ = 8, 000 (see Homework 1 for the pdf). Use the
vi y re

central limit theorem to approximate the probability that the sum of 100
independent claims will exceed 600,000.
ed d

3. The severities of individual claims have the gamma distribution with param-
ar stu

eters α = 5 and θ = 1, 000. Use the central limit theorem to approximate

the probability that the sum of 100 independent claims exceeds 525,000.
is
Th

4. A sample of 1,000 health insurance contracts on adults produced a sample

mean of 1,300 for the annual benefits paid with a standard deviation of 400.
It is expected that 2,500 contracts will be issued next year. Use the central
sh

limit theorem to estimate the probability that benefit payments will be more
that 101% of the expected amount.

5. Using the methods for comparing the tail heaviness discuss in class (except
for the mean excess loss function), compare the tail weight of the Weibull and

https://www.coursehero.com/file/11434984/Homework-5-solve-for-the-midterm-test/
 inverse Weibull distributions. The pdf of the former is given in Homework
2, while the pdf of the latter is

τ (θ/x)τ e−(θ/x)
τ

f (x) = , x > 0.
x

6. Place the lognormal distribution between the gamma and Pareto distribu-
tions with regard to heaviness of the tail. To reinforce your conclusion,
consider a gamma distribution with parameters α = 0.2, θ = 500; a lognor-
mal distribution with parameters µ = 3.709290, σ = 1.338566; and a Pareto

m
er as
distribution with parameters α = 2.5, θ = 150. First demonstrate that all

co
three distributions have the same mean and variance. Then numerically

eH w
demonstrate that there is a value such that the gamma pdf is smaller than

o.
rs e
the lognormal and Pareto pdfs for all arguments above that value and that
ou urc
there is another value such that the lognormal pdf is smaller than the Pareto
pdf for all arguments above that value. The pdf of X v LN (µ, σ 2 ) is
o

{ }
aC s

1 (ln x − µ)2
√ exp − , x ≥ 0.
vi y re

f (x) =
xσ 2π (2σ)2

7. For a Pareto distribution with α > 2, compare e(x) (the mean excess loss
ed d

function) to e(0) and also determine the coefficient of variation. Confirm that
ar stu

these results are consistent with the Pareto distribution being heavy-tailed.

8. Let Y be a random variable that has the equilibrium density. That is,
is

fY (y) = fe (y) = SX (y)/E[X] for some random variable X. Use integration

Th

by parts to show that

MX (t) − 1
MY (t) =
sh

tE[X]
whenever MX (t) exists.

9. You are given that the random variable X has probability density function
f (x) = (1 + 2x2 )e−2x , x ≥ 0.
(a) Determine the survival function S(x).
(b) Determine the hazard rate h(x).

https://www.coursehero.com/file/11434984/Homework-5-solve-for-the-midterm-test/
 (c) Determine the survival function Se (x) of the equilibrium distribution.
(d) Determine the mean residual life function e(x).
(e) Determine limx→∞ h(x) and limx→∞ e(x).
(f) Prove that e(x) is strictly decreasing but h(x) is not strictly increasing.

10. Assume that X has probability density function f (x), x ≥ 0.

(a) Prove that ∫∞
x
(y − x)f (y)dy
Se (x) =
E[X]

m
(b) Use (a) to show that

er as
∫ ∞

co
yf (u)dy = xS(x) + E[X]Se (x).

eH w
x

o.
(c) Prove that (b) may be rewritten as
rs e ∫∞
ou urc
x
yf (y)dy
S(x) =
x + e(x)
and that this, in turn, implies that:
o
aC s

E[X]
S(x) ≤
vi y re

x + e(x)
(d) Use (c) to prove that, if e(x) ≥ e(0), then
E[X]
ed d

S(x) ≤
x + E(X)
ar stu

and thus
1
S[kE[X] ≤ ,
is

k+1
Th

which for k = 1 implies that the mean is at least as large as the (smallest)
median.
(e) Prove that (b) may be rewritten as
sh

∫∞
e(x) x
yf (y)dy
Se (x) =
x + e(x) E[X]
and thus that
e(x)
Se (x) ≤
x + e(x)
GOOD LUCK!

https://www.coursehero.com/file/11434984/Homework-5-solve-for-the-midterm-test/

Powered by TCPDF (www.tcpdf.org)