01.06.2015 Views

Actuarial Modelling of Claim Counts Risk Classification, Credibility ...

Actuarial Modelling of Claim Counts Risk Classification, Credibility ...

Actuarial Modelling of Claim Counts Risk Classification, Credibility ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

278 <strong>Actuarial</strong> <strong>Modelling</strong> <strong>of</strong> <strong>Claim</strong> <strong>Counts</strong><br />

(2) even if the policyholder leaves the company after a claim, he has to pay for the deductible<br />

(but in motor third party liability insurance, there may be difficulties linked to the<br />

collection <strong>of</strong> the deductible);<br />

(3) in the mixed system, the r l s and the severity <strong>of</strong> the deductibles may be tuned in an<br />

optimal way in order to attract the policyholders.<br />

The numerical study conducted in this chapter shows, provided appropriate values for the<br />

parameters are selected, that the amounts <strong>of</strong> deductible are moderate in the mixed case<br />

(reduced relativities combined with deductibles per claim).<br />

Although deductibles can be difficult to implement in motor third party liability insurance<br />

(companies indemnify the third parties directly and so cannot simply reduce the amount they<br />

pay), the system is nevertheless easy to implement for first party coverages (for which the<br />

payments are made directly to the policyholders), like material damage for instance. In the<br />

European Union, the premium for the material damage cover is either subject to the bonusmalus<br />

system applying to motor third party liability or to a specific bonus-malus system.<br />

In the latter case, using the techniques suggested in this chapter, one could replace the<br />

bonus-malus scale for material damage with deductibles determined by past claim history.<br />

This chapter will focus on optional coverages.<br />

7.2 Distribution <strong>of</strong> the Annual Aggregate <strong>Claim</strong>s<br />

7.2.1 <strong>Modelling</strong> <strong>Claim</strong> Costs<br />

As explained in the introduction, the strategy designed in this chapter is difficult to apply in<br />

motor third party liability insurance. To fix the ideas, we consider here first party coverages<br />

for material damages, so that the actuary is not faced with possible large losses. The total<br />

claim cost can thus be modelled using a compound mixed Poisson model. Specifically, let<br />

us denote as C 1 C 2 C N the amounts <strong>of</strong> the N claims reported by the policyholder. The<br />

total claim amount for this policy is<br />

S =<br />

N∑<br />

C k (7.1)<br />

k=1<br />

with the convention that the empty sum equals 0. The severities C 1 C 2 , are assumed<br />

to be independent and identically distributed, and independent <strong>of</strong> the claim frequency N .<br />

As in the preceding chapters, N is taken to be oi distributed. This construction<br />

essentially states that the cost <strong>of</strong> an accident is for the most part beyond the control <strong>of</strong><br />

a policyholder. The degree <strong>of</strong> care exercised by a driver mostly influences the number <strong>of</strong><br />

accidents, but in a much lesser way the cost <strong>of</strong> these accidents. Nevertheless, this assumption<br />

seems acceptable in practice, at least as an approximation. Note that the severities C 1 C 2 <br />

are also independent <strong>of</strong> .<br />

Considering Expression (7.1) for the total cost <strong>of</strong> claims S, the pure premium for a<br />

policyholder without claim history is EC 1 . This amount will be corrected by a specific<br />

relativity according to the level occupied in the bonus-malus scale, that is, according to the<br />

claims reported to the company in the past.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!