Credibility theory shows how to combine an individual’s claims record with the premium of their class. The credibility premium calculator takes the years of experience, the Bühlmann constant K, the average claims, the manual premium and a target credibility, and gives the credibility factor Z, the blended premium, the change from the manual premium and the experience needed to reach the target.
How the calculation works
- Experience rating blends two estimates: the policyholder’s own experience and the manual premium set for the whole class of similar risks.
- The Bühlmann credibility factor is Z = n ÷ (n + K), where n is the amount of own experience (years or exposures) and K is a constant that measures how noisy individual experience is compared with the differences between risks.
- The credibility premium is Z × own experience + (1 − Z) × manual premium. The more experience there is, the more weight it gets.
- To reach a target credibility Z, the experience needed is n = K × Z ÷ (1 − Z). When n equals K, Z is exactly one half.
Example
A policyholder has 8 years of experience with average claims of 540 a year, while the manual premium is 480. The credibility constant K is 12. Then Z = 8 ÷ 20 = 0.4, and the credibility premium is 0.4 × 540 + 0.6 × 480 = 504.00, which is 5.00 % above the manual premium. To reach a credibility of 0.8, the policyholder would need K × 0.8 ÷ 0.2 = 48 years of experience.
Why it matters
Experience rating is used for commercial fleets, workers’ compensation and group health plans. The calculation shows how a bad or good run of claims should and should not move a premium, and it is a core topic in actuarial exams.
Good to know
The Bühlmann model assumes that risks are drawn from a population with a constant variance structure and that experience is stable over time. Real data may show trends, so premiums are often capped or limited in how fast they change. The constant K must be estimated carefully.
Related tools
The margin of error calculator shows how data amount affects precision, and the confidence interval calculator covers uncertainty of averages.