The County Fair Cyber Loss Distribution

Insurance premiums reflect expectations about the future losses of each insured. Given dearth cyber security loss data, market could shed light on true magnitude despite noise from factors unrelated to losses. To that end, we extract insurance pricing information regulatory filings 26 insurers. We provide empirical observations how vary by coverage type, amount, and policyholder type over time. A method using particle swarm optimisation expected value premium principle is introduced iterate through candidate parameterised distributions with goal reducing error in predicting observed prices. then aggregate inferred models across 6,828 prices all insurers derive County Fair Cyber Loss Distribution . demonstrate its decision support applying it a theoretical retail firm annual revenue $50M. The results suggest liability $428K faces 2.3% chance experiencing between $100K $10M year. resulting estimates help organisations better manage risk, regardless whether they purchase insurance.