Bridging the gap between pricing and reserving with an occurrence and development model for non-life insurance claims

Abstract Due to the presence of reporting and settlement delay, claim data sets collected by non-life insurance companies are typically incomplete, facing right censored count severity observations. Current practice in pricing tackles these via a two-step procedure. First, best estimates computed for number claims that occurred past exposure periods ultimate severities, using historical data. Second, actuaries build predictive models estimate technical, pure premiums new contracts treating as actual observed outcomes, hereby neglecting their inherent uncertainty. We propose an alternative approach brings valuable insights both reserving. As such, we effectively bridge two key actuarial tasks have traditionally been discussed silos. Hereto, develop granular occurrence development model reserving at same time resolves inconsistency traditional techniques between observations imputed estimates. illustrate our proposed on well reinsurance portfolio. The advantages strategy most compelling illustration where large uncertainties originate from long delays, low frequencies heavy (even extreme) sizes.