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

Bonus-Malus Scales 209
4.8.3 Distances between the Random Effects
A first measure may be to compare the expected a posteriori random effect, which actually
is the relative premium at level l in the bonus-malus scale :
r L=l = L = l
So the closest level l j in scale 2 to the level l i occupied in scale 1 is given by
argmin lj
r L1 =l i
− r L2 =l j
2
This rule simply amounts to placing the policyholder in the new scale at the level with
the closest relativity to the one applicable in the previous scale. Because most commercial
scales are normalized (to associate a unit relativity with the entry level), this means that the
insurer has to compute the relativities for both scales. The implicit assumption is that the
new entrant has the same characteristics as the policyholders in the portfolio (no adverse
selection being allowed).
Another measure of discrepancy consists of comparing the distribution functions of the a
posteriori random effects. This can be done by using the Kolmogorov distance d K or the
total variation distance d TV :
d K L 1 = l i L 2 = l j = max Pr ≤ L 1 = l i − Pr ≤ L 2 = l j
d TV L 1 = l i L 2 = l j =
∫
0
f L1 =l i
− f L2 =l j
d
Summarizing, a policyholder being at level l i in scale 1 and moving to scale 2 will be put
in the level l j of scale 2 that minimizes one of the following distances:
(i) d E L 1 = l i L 2 = l j = r L1 =l i
− r L2 =l j
2
(ii) d K L 1 = l i L 2 = l j
(iii) d TV L 1 = l i L 2 = l j .
4.8.4 Numerical Illustration
In this section we use Portfolio A and its risk classification described in Table 2.7. Let us
assume that we want to move policyholders between the following two bonus-malus scales:
• The −1/top scale with 6 levels, numbered 0 to 5.
• The −1/ + 2 scale of Taylor (1997) with 9 levels, having transition rules given in
Table 4.24.