Non-proportional reinsurance is built around loss size rather than a fixed percentage share of every policy. The reinsurer pays only after losses exceed an agreed retention or attachment point, up to the treaty limit.
Munich Re defines excess-of-loss reinsurance as a broad category that includes per-risk, per-occurrence and aggregate forms. Swiss Re similarly distinguishes per-risk, per-event catastrophe and stop-loss structures.
Per-risk excess of loss
Per-risk cover is designed around one insured risk. The treaty retention and limit apply to each risk affected by a loss rather than to the total portfolio for the year.
Simplified example: an insurer retains the first $1 million on each risk and buys $4 million excess of $1 million. If one covered risk produces a $3 million loss, the insurer retains $1 million and the reinsurer can respond to the next $2 million, subject to treaty terms.
Per-event or catastrophe excess of loss
Per-event cover aggregates losses that arise from one defined event. In property catastrophe reinsurance, a hurricane, earthquake or other event may damage many insured locations. The treaty is designed to protect the insurer from the accumulation.
The event definition is critical. Reinsurance contracts often specify how related losses are grouped and may use hours clauses or other event wording for particular perils.
| Structure | Primary aggregation unit | Main purpose |
|---|---|---|
| Per-risk XOL | One insured risk | Protect against severe individual losses |
| Per-event / Cat XOL | One defined event across multiple risks | Protect against catastrophe accumulation |
| Aggregate / Stop-loss | Total covered losses during a stated period | Limit adverse annual or period loss experience |
Stop-loss or aggregate excess of loss
Stop-loss reinsurance looks at aggregate losses over a period, often a year. Munich Re defines aggregate excess of loss as a form that indemnifies the cedent when losses over the period exceed an agreed amount or an agreed percentage of another measure such as premium.
Swiss Re notes that stop-loss can protect the net retained portfolio when the cumulative burden of losses becomes unusually high. Unlike per-risk or per-event protection, the trigger is the aggregate result rather than one large loss or one catastrophe.
Why insurers combine the structures
The covers solve different problems. An insurer may want per-risk protection for large individual commercial accounts, catastrophe protection for a regional hurricane accumulation and aggregate protection against an unusually bad year across the portfolio.
They can therefore sit in the same reinsurance program at different layers.
Attachment point, retention and limit
- Retention: the amount of covered loss the cedent keeps before reinsurance responds.
- Attachment point: the loss level at which the layer begins to pay.
- Limit: the maximum amount the reinsurer will pay for the covered layer, subject to the contract.
- Reinstatement: a provision that can restore some exhausted excess-of-loss capacity after a covered event, often for an additional premium.
- Aggregate deductible: an amount the cedent may have to retain across multiple losses before aggregate protection responds.
How the same loss can look different under each treaty
Suppose a storm causes 500 property losses. A per-risk treaty evaluates each insured risk separately. A catastrophe treaty may aggregate the losses into one event if the event wording is satisfied. An annual stop-loss treaty may count the retained portion of those losses together with other covered losses during the year.
This is why reinsurance analysis cannot stop at the phrase “excess of loss.” The aggregation rule is central to understanding what protection has actually been purchased.
What reinsurance buyers should test
- How does the contract define one risk and one event?
- Which losses count toward the retention?
- Are loss-adjustment expenses inside or outside the limit?
- Are there occurrence hours clauses?
- Does the layer reinstate after a loss, and at what cost?
- For stop-loss, which premiums and losses are included in the ratio or aggregate calculation?
Frequently asked questions
Is stop-loss reinsurance proportional?
No. Stop-loss is generally a non-proportional or aggregate excess-of-loss form.
Is catastrophe reinsurance the same as per-risk reinsurance?
No. Per-risk cover responds to individual risks, while catastrophe or per-event cover is designed around an accumulation from one defined event.
Can one insurer buy all three structures?
Yes. Reinsurance programs commonly combine different treaty types because they address different dimensions of severity and accumulation.
Does excess-of-loss mean the reinsurer pays every loss above the attachment point?
Only covered losses that satisfy the treaty wording count. Exclusions, event definitions, limits, reinstatements and other provisions still apply.
Sources & further reading
Reviewed against Munich Re and Swiss Re reference material in September 2026. Examples are simplified; treaty wording, aggregation definitions and accounting provisions control actual recoveries.
