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

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

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

De Pril (1979) defined L n l k t as the amount that the actual accident must exceed in<br />

order to justify the filing <strong>of</strong> a claim, if the policyholder is at time t <strong>of</strong> period n in level l and<br />

has already filed k claims. The optimal value <strong>of</strong> L n l k t is determined to minimize the<br />

discounted expectation <strong>of</strong> the total future costs (premiums and self-defrayed accidents) for<br />

the policyholder. After De Pril (1979), De Pril & Goovaerts (1983) determined bounds<br />

for the optimal critical claim size when only incomplete information about the claim amount<br />

distribution is available. They considered −1/top bonus-malus scales.<br />

Dellaert ET AL. (1990) proved that under mild conditions the optimal decision rule is<br />

to claim for damages with amount above a certain limit. In some instances, policyholders<br />

are allowed to decide at the end <strong>of</strong> an insurance year which damages occurred during the<br />

year should be claimed; see, e.g., Martin-L<strong>of</strong> (1973). This means that the policyholder<br />

has perfect information about the number <strong>of</strong> accidents and the corresponding damages at<br />

the moment he/she decides which damages to claim. This situation has been investigated<br />

by Dellaert ET AL. (1991). Let us also mention that Dellaert ET AL. (1993) considered<br />

damage insurance (where, in addition to the bonus hunger phenomenon, the optimal stopping<br />

rule to terminate the insurance has to be determined).<br />

Holtan (2001) envisaged the loss <strong>of</strong> bonus after a claim as a rate <strong>of</strong> interest paid from<br />

the customer to the insurer, and studied the hunger for bonus from this viewpoint.<br />

Optimal claiming rules have also been considered in Operational Research, using Markov<br />

decision processes. When a driver is involved in a motor accident decisions have to be made<br />

as to whether or not a claim should be made. Hastings (1976) considered this problem as<br />

a Markov decision process with the expected cost over a finite horizon (where the relevant<br />

costs are repair costs and premium costs) as objective function. See also Haehling von<br />

Lanzenauer (1974) and Hey (1985). Norman & Shearn (1980) proposed the following<br />

simple rule <strong>of</strong> thumb that is shown to work well in their case: irrespective <strong>of</strong> when the<br />

accident occurs, claim only if the amount <strong>of</strong> the claim exceeds the difference over the next<br />

4 years between the total premiums payable if a claim is made and those payable if it is not,<br />

assuming that no further claims will be made. Chappell & Norman (1989) demonstrated<br />

that this simple rule was less efficient in the case <strong>of</strong> protected bonus. Kolderman &<br />

Volgenant (1985) examined the same problem under the assumption that only one claim<br />

is needed to change the insurance premium category, and any extra claims in the year have<br />

no effect on the current repair estimate for an accident.<br />

Walhin & Paris (2000) derived the actual claim amount and frequency distributions<br />

within a bonus-malus system. As explained above, policyholders should defray the small<br />

claims to avoid the penalties induced by the bonus-malus system. Consequently, there are<br />

more accidents than the number <strong>of</strong> claims filed by the insurer: the insurance data are censored.<br />

The kind <strong>of</strong> censorship is nevertheless very particular, and much more complicated than the<br />

phenomena encountered in classical nonlife problems (where losses are censored because<br />

they exceed some policy limit, or fall below a given deductible). The procedure described<br />

in Section 5.4.1 is taken from Denuit, Maréchal, Pitrebois & Walhin (2007a), where<br />

alternative approaches to obtain uncensored accident distributions can be found.<br />

Even if the bonus-hunger phenomenon has been extensively studied in connection with<br />

bonus-malus scales, the same idea applies to credibility systems. See, e.g., Norberg (1975)<br />

and Sundt (1988) for an illustration.

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

Saved successfully!

Ooh no, something went wrong!