Skip to content
←Back to glossaryReinsurance

Working layer vs catastrophe layer

The distinction between low layers frequently hit by attritional losses and high layers reserved for catastrophes.

Definition

The distinction between the working layer and the catastrophe layer shapes the construction of an excess-of-loss reinsurance program. The working layer is the low layer, sitting just above the cedant's retention, that ordinary attritional frequency losses hit regularly. Historical data are plentiful there, so it is priced largely on burning cost, and its rate on line is high because the reinsurer responds often. The catastrophe layer is a high layer, rarely reached, built to absorb exceptional events of great severity. At those levels history is too thin to price on, so the work moves to catastrophe modeling and scenario analysis. Its rate on line is low. Its prudential role is decisive, though, since it is what protects the cedant's solvency against major shocks. Knowing where a layer sits on this scale tells you how to price it, what volatility to expect and what prudential job it does, and all of that changes radically from the bottom of the program to the top. In cyber the boundary blurs, because accumulation risk can turn frequency losses into a peak event.

Its 5 cross-references, woven across the glossary's 18 categories
Working layer vs catastrophe layer: Reinsurance glossary illustration, AlgoPolis
Example

Textbook case: a program combines a working layer of 5 million excess of 5, hit frequently, with a catastrophe layer of 50 million excess of 50 that has never been triggered to date but is essential if a systemic event hits.

Read next
Put it to work
Related terms
Related articles
Go further at the Academy
Stay informed

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.