Professional Documents
Culture Documents
Risk Models 1 Upload
Risk Models 1 Upload
Risk Models 1 Upload
Risk models
Insurable risks:
The risk that is of interest to policyholder and is quantifiable
Insurable risk events should:
Be independent
Have low probability of occurrence
Be pooled with similar risks
Have an ultimate liability
Avoid moral hazards
Basic ideas:
Definition:
Policy lasts for a fixed, relatively short, period (e.g. often 1 year)
Insurance company receives premium from policyholder(s)
Insurer pays claims arising from policy during term of policy
Important feature:
Premium is set at a level to cover claims during short term of policy only.
Different from life assurance policy (related to mortality rate)
Example
( ) = ( = , )
=0
= ( = , )
=0 =1
= ( = ) ( )
=0 =1
How to model N?
How to calculate (=1 )?
. . . ~(): ( ) = ()
=1
fold convolution of
p.d.f =
() = () ( ) (discrete r.v.)
() = () ( ) (continuous r.v.)
c.d.f =
() = () ( ) (discrete r.v.)
() = () ( ) (continuous r.v.)
Example
1 1 1
If (0) = 2 , (1) = 4 , (2) = 4 and 1 , 2 , 3 are independent and identically
distributed like , find the density of = 1 + 2 + 3 .
Solution
Moments: required in risk premium
[] = [[|]]
() = [(|)] + ([|])
= 1 + 2 + +
= random number of claims
1 , 2 , , independent and distributed like
1 , 2 , , and are independent
[| = ] =
[| = ] =
(|) =
(|) =
Mean:
() =
=
= pure risk premium
Variance:
() = [(|)] + [(|)]
() = ()() + ()[()]2
Example
Suppose (1) = .6 and (2) = .4 gives the distribution of claim amount variable ,
and ( = 0) = .7, ( = 1) = .2 and ( = 2) = .1 gives the distribution of the
claim number variable .
mgf:
Example
Assume the number of claims has a (100,0.01) distribution, and individual claim
sizes are (10,0.2).
(i) Find the mean and variance of the aggregate claim amount.
(ii) Find an expression for the mgf of the aggregate claim amount.
Solution
Cumulant Generating Function
Defined as:
() = log ()
() = (0)
() = (0)
() = (0)
Mean:() =
() =
[( 1 )3 ] =
mgf: () =
There are n portfolios of insurance policies, believed independent of each other, with
the portfolios aggregate claim amounts equal to modelled by a compound
Poisson distribution with parameter and (), = 1,2, , where () denotes
the distribution function of the individual claim amounts.
Show this.
Solution
Compound Binomial Distribution:
Example
A group life insurance policy covering n lives, with each insured life subject to the
same mortality rate q and lives independent with respect to mortality. Thus, the
distribution of the number of deaths in a year is binomial: ~(, )
~(, )
pdf : ( = ) = (1 ) , = 0,1, . . ,
Moments: () = , () = (1 )
mgf: () = ( ) = (1 + )
. . . ~():
= ( ), () = [exp{ }]
Mean: () = ()( ) = 1 .
Variance:
() =
[( [])3 ] = 3 3 2 2 1 + 2 3 13
mgf:
() =
Compound negative binomial distribution:
Example
To compute the risk premium, we need to calculate the mean and variance of the
aggregate claim distribution.
~(, )
pdf : ( = ) = +1 (1 ) , = 0,1, ..
(1)
Moments: () = , () = (1 )/2
mgf: () = ( ) = [1 (1 ) ]
= 1: reduces to ()
. . . ~():
= ( ), () = [exp{ }]
(1)
Mean: () = ()( ) = 1 .
Variance:
() =
3 2 1 2 2 3 13 3
[( [])3 ] = + +
2 3
mgf: () =
Solution