Reinsurance that responds only above the insurer's retained loss layer.
Excess of loss reinsurance is a form of reinsurance in which the reinsurer pays losses only after the ceding insurer’s retained amount or attachment point has been exceeded, up to the reinsurance limit.
The structure is about severity protection. The reinsurer does not share every ordinary claim dollar from the start. It steps in only when losses move above the insurer’s own retention.
Excess-of-loss reinsurance is central to how insurers protect themselves against severe individual losses, catastrophe events, and accumulation risk without giving up a share of every ordinary premium and claim from first dollar.
In Canada, this matters especially for catastrophe-exposed property portfolios, large liability losses, and event-driven concentrations such as wildfire, hail, or flood. The insurer keeps the layer it is prepared to absorb net and buys protection above that point.
Excess-of-loss reinsurance is a non-proportional structure. The main moving pieces are:
This is different from proportional reinsurance, where premium and losses are shared from the beginning according to percentages. In excess-of-loss arrangements, the reinsurer is mainly there for severity once the agreed threshold has been crossed.
| Layer question | Why it matters |
|---|---|
| Where is the insurer’s retention? | That determines how much loss the ceding company keeps net before reinsurance begins. |
| What is the attachment point and layer size? | These define when the reinsurer pays and how far the protection extends. |
| Is the layer per risk, per event, or aggregate? | The same label can behave very differently depending on the contract design. |
| What concentration or catastrophe exposure is the insurer protecting against? | The layer is usually designed for severity, not everyday claims sharing. |
An insurer retains the first CAD 5 million of catastrophe losses from an event and buys an excess-of-loss layer that responds above that amount. A severe wildfire event pushes losses beyond the retention, and the reinsurer pays within the layer it agreed to cover.
The same concept can apply to large individual losses. A reinsurer may not participate in ordinary attritional claims at all, yet still be critical once a severe single risk or catastrophe event breaches the kept layer.
Excess of loss reinsurance is not the same as a policyholder’s deductible. The deductible applies at the customer-policy level. Excess-of-loss reinsurance applies at the insurer-reinsurer level.
It is also different from proportional reinsurance, where premium and losses are shared from the start according to fixed percentages.
Another common mistake is assuming all excess-of-loss layers work the same way. They do not. Per-risk, catastrophe, aggregate, and clash-style structures can behave very differently, even though the broad concept of attachment above a retention remains the same.
Readers also sometimes assume excess-of-loss support means the insurer is protected against every bad result above one simple number. In practice, the exact trigger, event definition, exclusions, and limit structure still matter materially.
The exact operation depends on whether the layer is per risk, per event, catastrophe-based, or aggregate-based. The phrase gives the broad structure, not the full attachment logic.