Attachment probability
The annual probability that an event reaches a securitized layer's attachment point and erodes its principal.
Attachment probability is the chance, over one year, that an event reaches the threshold at which a layer begins to pay. Together with exhaustion probability and expected loss, it forms the triptych that every insurance-linked securities prospectus publishes and investors compare across issues. It makes risk legible to an investor who is not an insurer. That investor cannot read a treaty, but can compare a probability of default with that of a rated bond, and this figure allows exactly that. It is always read alongside expected loss, because the two say different things. A thin layer with a high attachment probability will be hit often but lose little each time. A wide high layer will rarely be hit but lose everything when it is. Its limit is the model producing it: two modelers can publish probabilities differing by half on the same risk.
A 2026 cat bond shows an attachment probability of 1.85 percent, an exhaustion probability of 0.74 percent and an expected loss of 1.21 percent, for a spread of 5.8 percent. A second note from the same sponsor on a higher layer shows 0.62 percent attachment and 0.41 percent expected loss, for a spread of 3.1 percent: the multiple rises from 4.8 to 7.6 moving up the tower.
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.