Credibility theory
An actuarial framework that weights a risk's own experience against a market benchmark, depending on how statistically reliable the available data are.
Bühlmann, 1967, ASTIN Bulletin (“Experience rating and credibility”): the weight given to a risk's own experience is Z = n / (n + K), where n is the number of observed periods and K the ratio of the risk's own variance to the variance between risks. Credibility theory answers a pricing question: how far to trust an insured's own history against a broader benchmark. With plenty of stable data, the risk's own experience nearly stands on its own. With little, it is too volatile to take at face value, so it gets blended with a market benchmark. Z runs from zero to one and rises with the volume and stability of the data; the complement 1 − Z goes to the benchmark. The rate is a weighted average of individual experience and the collective mean. That keeps the actuary from overreacting to a few isolated losses without ignoring something genuinely specific to the risk. In cyber, a risk's own data carry little credibility because track records are too short and unstable, so market benchmarks and expert judgment weigh heavily and segmentation is less granular.
Textbook case: a company has had a single cyber loss in five years. That is too thin a basis to price on, so the actuary blends this low-credibility experience with the average loss experience of its sector, which is far more stable.
Follow the research
One letter a week. First, the week's underwriting note, when one is published: a dated, argued decision, signed by Paul Bernardski. Then what the observatory published on cyber risk, political risk and insurance. No account, no tracking.
No spam, one-click unsubscribe, no data ever sold.