Forms of reinsurance

Read the PDF edition (4 pages)

The practice of reinsurance distinguishes contracts along two axes. The first concerns the way in which risks are brought under the contract. In facultative reinsurance, each risk is offered to the reinsurer individually, leaving the reinsurer free to accept or decline it. In treaty or obligatory reinsurance, the parties agree in advance that all risks of a defined portfolio, for instance all fire policies written by the cedant in a given country during a given year, are ceded automatically. The cedant is obliged to cede them and the reinsurer is obliged to accept them.¹

Facultative reinsurance, the older form, remains important for large individual risks that fall outside the scope of the cedant’s treaties or exceed their limits: a refinery, a satellite launch, a major bridge. It requires individual underwriting and is therefore expensive to administer. Treaty reinsurance, by contrast, allows the cedant to write its daily business without consulting the reinsurer and gives the reinsurer a share of a whole portfolio, which is statistically easier to assess than a single risk. The economic relationship under a treaty is closer to a partnership. While the reinsurer depends on the cedant’s underwriting discipline, the cedant depends on the reinsurer’s willingness to stand by the treaty in bad years. This interdependence explains the importance of long-term relationships in reinsurance and the role of good faith as a governing principle.

The second axis concerns the way in which premiums and losses are shared. In proportional reinsurance, the reinsurer takes a fixed share of each risk and receives the same share of the original premium, less a commission that compensates the cedant for its acquisition and administration costs. The simplest form is the quota share, in which the reinsurer takes a fixed percentage of every risk in the portfolio. The Berkshire Hathaway agreement of 2008, under which Berkshire assumed twenty per cent of Swiss Re’s property and casualty business, was a quota share of this kind on a very large scale. In a surplus treaty, the cedant retains each risk up to a fixed amount, its retention, while the reinsurer takes the part of the sum insured that exceeds the retention, up to a multiple of it. Surplus treaties allow the cedant to keep its small risks entirely while balancing its portfolio by ceding the upper parts of its large ones.

In non-proportional reinsurance, the link to the individual risk is broken. The reinsurer pays only when the cedant’s loss exceeds an agreed amount, the deductible or priority, and then up to an agreed limit. The premium is calculated separately, on the basis of the expected loss to the layer and a loading for the reinsurer’s capital costs. The main forms are the excess of loss per risk, which covers the part of a loss on a single risk that exceeds the priority, the catastrophe excess of loss, which covers the aggregate of losses from a single event such as a storm or earthquake, and the stop loss, which protects the cedant against an annual loss ratio above a threshold.² Catastrophe excess of loss programmes are usually arranged in several layers, each placed with a panel of reinsurers, so that a primary insurer in Switzerland may have dozens of reinsurers participating in its protection.

The distinction has practical consequences for the drafting and interpretation of contracts. In proportional treaties the reinsurer shares the cedant’s fortunes from the first franc, so that questions of allocation and aggregation arise mainly in accounting. In non-proportional contracts, by contrast, everything depends on the definition of the loss event: whether a series of claims arising from a storm, a pandemic or a defective product constitutes one loss, subject to one priority and one limit, or several. The World Trade Center litigation concerned a primary policy, but the same questions arise under catastrophe excess of loss treaties, where the “hours clause” defines how long a period of storm or flood damage may be aggregated into one event. The COVID-19 pandemic produced a wave of disputes of exactly this kind, above all in the London market, concerning the aggregation of business interruption losses under catastrophe covers.³

Reinsurers themselves buy reinsurance. This is called retrocession, the reinsurer that accepts it being the retrocessionaire. A reinsurer that has accumulated too much exposure to, say, Californian earthquake risk can retrocede part of it to other reinsurers or to the capital markets. Swiss Re’s net loss of 4.3 million francs from San Francisco, compared with a gross loss of 8.4 million, shows that retrocession was already an established technique in 1906. Today it is also carried out through catastrophe bonds: Swiss Re’s own Matterhorn Re programme, for instance, transfers peak risks from the group’s portfolio to investors.⁴ Retrocession spreads risk across the system but also creates chains of dependency. The collapse of the London retrocession market in the late 1980s, the so-called LMX spiral, showed how risks can circulate between a small number of participants until nobody knows where they lie.⁵

Life and health reinsurance follows partly different rules. Here the reinsurer typically assumes mortality, longevity or morbidity risk, often through proportional treaties on the basis of risk premiums. It may also finance the acquisition costs of a growing life insurer or take over entire closed blocks of policies. The long duration of life business makes the reinsurer’s investment result and its assumptions about future mortality decisive for profitability. The review of underperforming life portfolios in Australia, Israel and South Korea mentioned above illustrates the point: an error in the assumptions for a long-term portfolio can require an adjustment years after the business was written.

Notes

1. Swiss Re, The essential guide to reinsurance, https://www.swissre.com/dam/jcr:d06472ab-2625-48cf-8b4e-7c7ac8aa63f0/The-essential-guide-to-reinsurance.pdf. ↩
2. Swiss Re, The essential guide to reinsurance, https://www.swissre.com/dam/jcr:d06472ab-2625-48cf-8b4e-7c7ac8aa63f0/The-essential-guide-to-reinsurance.pdf. ↩
3. Court of Appeal of England and Wales, Unipolsai Assicurazioni SpA v Covéa Insurance plc [2024] EWCA Civ 1110, judgment of 30 September 2024, https://caselaw.nationalarchives.gov.uk/ewca/civ/2024/1110. The appeal from an arbitration award concerned whether COVID-19 business interruption losses arose out of one “catastrophe” under a catastrophe excess of loss reinsurance with a 168-hour clause. ↩
4. “Swiss Re secures targeted $65m Matterhorn Re 2025-2 retro cat bond at low-end pricing”, Artemis, https://www.artemis.bm/news/swiss-re-secures-targeted-65m-matterhorn-re-2025-2-retro-cat-bond-at-low-end-pricing/. ↩
5. English Wikipedia, “Lloyd's of London”, https://en.wikipedia.org/wiki/Lloyd%27s_of_London, which describes how after the Piper Alpha explosion of 1988 the underwriters “initially had no idea how extensive their exposure was: the loss was passed around in what became known as the London market excess of loss (LMX) ‘spiral’”. ↩

For general information only. This article does not constitute legal or tax advice, does not replace advice on an individual case and creates no lawyer-client relationship. It reflects the legal position as of September 2025.