Reinsurance interview questions

Reinsurance is insurance for insurers: a contract in which one insurer (the cedent) transfers part of a risk it has already underwritten to another carrier (the reinsurer) in exchange for a share of the premium. The original policy is untouched — the cedent still owes the claim whether or not the reinsurer pays.

Example. An insurer covers a $10M factory. It keeps $2M for itself and cedes the other $8M to a reinsurer, who is paid a share of the premium to cover that $8M if the factory burns down.

Four reasons come up in almost every interview: capacity to write bigger risks than its own capital allows, smoother year-to-year results, protection against a single catastrophe, and capital relief from regulators who reward an insurer for laying off risk. It is a transfer of risk, not of responsibility.

Example. An insurer with $50M of capital could be wiped out by one $80M hurricane. Reinsurance lets it keep writing that much coastal business anyway, because the loss above what it can absorb is passed on.

The cedent is the insurer handing off part of a risk, and it stays fully liable to the policyholder no matter what happens next. The reinsurer is the carrier accepting that ceded risk — it underwrites portfolios and accounts rather than individual policyholders, and is paid to be there in the bad year, not just the average one.

Example. ABC Insurance (the cedent) sells a $5M policy and cedes $3M of it to XYZ Re (the reinsurer). If a $5M claim comes in, ABC pays the policyholder the full $5M, then collects $3M back from XYZ Re.

Not normally. There is no privity of contract between the policyholder and the reinsurer, because reinsurance is a separate contract between two carriers. The one common exception is a cut-through clause: a narrow, negotiated provision that lets claim payments move straight from reinsurer to policyholder if the cedent fails.

Example. A homeowner’s insurer goes bankrupt before paying a $200,000 claim. The homeowner cannot collect that money from the insurer’s reinsurer — unless the treaty happens to include a cut-through clause naming policyholders as beneficiaries.

A retention is the share of a risk an insurer keeps for its own account after ceding the rest. It matters because almost every other number in reinsurance is defined relative to it — it sets where a non-proportional layer starts to respond, and how much of a proportional treaty the cedent still carries itself.

Example. An insurer sets its retention at $1M per risk. On a $6M building, it keeps the first $1M of any loss itself and cedes the remaining $5M to reinsurers.

It is the doctrine that binds a reinsurer to a cedent’s good-faith claims decisions, so long as those decisions fall within the terms of the treaty. Without it, every claim payment would need to be relitigated between cedent and reinsurer, which defeats the point of transferring the risk in the first place.

Example. A cedent settles a liability claim for $400,000 rather than risk a bigger jury verdict at trial. As long as that call was reasonable and within the treaty terms, the reinsurer honours its share of the $400,000, even though it had no say in the decision.

A treaty is one contract that automatically covers a whole class of business for a period — both sides are bound to cede and accept without looking at each risk individually. Facultative reinsurance is the opposite: a one-off cession of a single risk, individually offered by the cedent and individually accepted or declined by the reinsurer.

Example. An insurer’s treaty automatically covers every commercial property policy it writes this year. For one unusually large $80M refinery outside its normal appetite, it instead negotiates a single facultative certificate just for that risk.

A tower is a full reinsurance programme built from cover stacked in layers above the cedent’s retention, with each layer priced, negotiated and placed as its own contract. A loss burns through it from the bottom up: the retention first, then each layer in turn, until the loss is exhausted or the tower runs out.

Example. A cedent retains the first $10M of any loss, then buys a $20M layer excess of $10M, and a $70M layer excess of $30M on top of that — three levels stacked to $100M of total protection.

A retrocession is a reinsurer buying its own reinsurance — protection for the book of business it has already assumed from cedents. It is the same mechanism as reinsurance, one link further up the chain.

Example. XYZ Re has assumed $50M of hurricane risk from several cedents. To protect its own balance sheet, XYZ Re buys a retrocession, ceding $20M of that assumed risk to another reinsurer.

A hybrid: under a facultative-obligatory treaty the cedent decides, risk by risk, whether to cede a given exposure, but once it does cede, the reinsurer is obligated to accept it. It gives the cedent an optionality a normal treaty does not have.

Example. A cedent looks at a new $15M risk and chooses to cede it under a fac-oblig treaty. The reinsurer has no choice in the matter — it is contractually obligated to accept.

Inuring reinsurance is the pecking order among a cedent’s own covers — which contract must apply and be netted down first, before another one is even asked to respond. Get the order wrong and a loss can be double-counted, or a layer can respond before it is meant to.

Example. A cedent has both a quota share and an excess-of-loss cover. The treaty wording says losses must be net of the quota share before the excess-of-loss layer is tested — that ordering is the inuring clause at work.

In a quota share, the reinsurer takes the same fixed percentage of every risk, every premium and every loss in the treaty — say 40% of everything, regardless of how large or small any one risk is.

Example. On a 50% quota share, the reinsurer takes 50% of every premium dollar and pays 50% of every claim dollar, on every policy in the treaty, big or small.

In a surplus share, the cedent keeps one fixed amount per risk (a “line”) and cedes only the surplus above it, so the cession percentage moves with the size of each risk — small risks may be kept entirely, large ones ceded heavily.

Example. A cedent’s line is $500,000. On a $2M risk it keeps $500,000 and cedes $1.5M (75%); on a $600,000 risk it keeps the full $500,000 and cedes only $100,000 (about 17%) — the split moves with the size of the risk.

A ceding commission is what the reinsurer pays back to the cedent for having acquired and serviced the business, covering the cedent’s acquisition costs and expenses. It only makes sense on proportional treaties, where the reinsurer is sharing in premium it did nothing to generate.

Example. A reinsurer takes 40% of a treaty’s premium and pays the cedent a 30% ceding commission on that share, to cover the cedent’s cost of acquiring and servicing the business.

Ceding commission is fixed, or on a sliding scale, and is paid regardless of results, to cover the cedent’s cost of writing the business. Profit commission is an extra payment made only if — and to the extent that — the treaty actually turns a profit, calculated after the underwriting year closes.

Example. The cedent receives its 30% ceding commission no matter what happens. If the treaty also finishes the year in profit, the cedent might receive an extra 10% profit commission on top — but only because there was a profit to share.

A bordereau is the periodic listing of every risk, premium and claim ceded under a treaty. Because a proportional treaty covers a whole class of business automatically, without the reinsurer underwriting each risk itself, the bordereau is the only record the reinsurer has of exactly what it is on risk for.

Example. Each month the cedent sends the reinsurer a spreadsheet (the bordereau) listing every policy ceded, its premium, and any claims — that list is how the reinsurer knows exactly what it is on risk for.

A sliding scale commission is a ceding commission that moves with the treaty’s loss ratio — rising when results are good and falling when they are bad, inside agreed floors and ceilings — so the cedent shares in both the upside and downside of underwriting performance.

Example. A sliding scale might pay a 35% commission if the loss ratio comes in at 40%, dropping to 25% if the loss ratio rises to 60% — worse results, lower commission, within agreed floors and ceilings.

In excess of loss, the reinsurer does not share a fixed percentage of every risk — it pays only the part of a loss that exceeds an agreed threshold, up to an agreed limit. Premium and losses are not shared proportionally; the reinsurer is paid to sit above that threshold and wait.

Example. On a $10M excess of $5M layer, a $3M loss gets nothing from the reinsurer (it never reaches the $5M threshold). A $12M loss gets $7M from the reinsurer — everything above $5M, up to the $10M limit.

The attachment point is where the reinsurer starts paying — everything below it is the cedent’s retention. The limit is the most the reinsurer will pay under that layer, however large the underlying loss grows; anything above attachment-plus-limit falls back on the cedent unless a higher layer picks it up.

Example. A layer described as “$20M excess of $10M” attaches at $10M and carries a $20M limit, so it responds to losses that fall between $10M and $30M.

Catastrophe XoL protects against the accumulation of many small claims from a single event — a hurricane touching thousands of policies — rather than one large claim on one policy. It responds to the combined loss from an occurrence, which is why the hours clause defining how long an event can run matters so much for it.

Example. A hurricane damages 5,000 homes an insurer covers, for a combined $60M loss. A catastrophe XoL layer responds to that $60M combined figure in one go, rather than treating each home as a separate claim.

An hours clause defines how long a catastrophe can run and still be treated as a single occurrence for a layer. Without it, a multi-day storm could arguably be sliced into several separate events, changing how many times a layer — and its reinstatements — get triggered.

Example. A treaty’s 72-hour clause lets a cedent bundle every loss from a hurricane that moved through over 60 hours into one single occurrence for the layer, instead of splitting it into several events.

A clash cover is excess-of-loss protection aimed specifically at the accumulation that happens when a single event or a single claimant triggers several policies at once — for example, one liability claimant hitting several lines the insurer wrote for the same defendant.

Example. One factory explosion injures workers covered under the insurer’s workers’ comp, general liability, and auto policies at the same time. A clash cover lets the insurer combine the loss across all three policies to test against a single reinsurance layer.

A stop loss responds to the total of all net losses over a period, often expressed as a loss ratio, rather than to any single event. It protects the account’s overall result for the year, where excess of loss protects against any one loss or occurrence being too large.

Example. A stop loss might trigger once the cedent’s loss ratio for the year exceeds 80%, however many small claims added up to get there — quite different from an XoL layer, which only cares about one single large loss.

A reinstatement puts a layer’s limit back to its original amount after it has been eroded by a loss, so the cover is available again for the rest of the period. It costs extra because the reinsurer is re-extending a limit it has already paid out once.

Example. A $10M layer is fully used up by one loss. One paid reinstatement restores the full $10M limit, so it is available again if a second loss happens later in the year.

Ultimate net loss is the contractual definition of the loss figure that actually gets tested against a layer’s attachment point — what counts (indemnity, defence costs, salvage) and what does not. Because it decides whether, and how much, a layer responds, the wording is negotiated line by line.

Example. A $5M jury verdict plus $300,000 of defence costs both count toward ultimate net loss, while $50,000 the cedent later recovers from a co-defendant is subtracted back out before the layer is tested.

Rate on line is premium expressed as a percentage of the limit purchased. A rate on line of 10% means the premium is 10% of the layer’s limit, which also tells you, roughly, how many clean years of premium it takes to repay one total loss to that layer.

Example. A $10M layer priced at $1M of premium has a 10% rate on line — meaning ten loss-free years of premium would equal one full $10M loss to that layer.

IBNR stands for Incurred But Not Reported — the reserve set aside for losses that have already happened but have not been reported yet. It matters because a book can look profitable on reported claims alone while a large IBNR liability quietly builds underneath it.

Example. An insurer has $2M of reported, unpaid claims on its books, but actuaries estimate another $500,000 of claims have already happened and simply have not been reported yet — that $500,000 is the IBNR reserve.

Loss ratio is incurred losses divided by earned premium, expressed as a percentage. It is the single most common yardstick in the industry — a quick read on whether the premium charged is covering the claims experience, before expenses are even considered.

Example. An account earns $10M of premium and incurs $6M of losses — a 60% loss ratio, meaning 60 cents of every premium dollar is going straight to claims.

A claims control clause gives the reinsurer the right to direct, or at minimum approve, the handling of a claim once it is large enough to threaten its layer. Reinsurers ask for one because, on a large claim, they may end up paying most of the loss without having had any say in how it was handled.

Example. Once a claim on a $5M excess layer looks likely to exceed $3M, the claims control clause lets the reinsurer step in and direct how it is defended or settled, since it stands to pay most of the eventual bill.

Experience rating prices a layer from the account’s own loss history, trended and developed forward. Exposure rating builds a price from first principles using a severity curve, independent of that account’s actual claims — useful when there is not enough loss history to trust, such as for a high excess layer.

Example. A cedent with ten years of clean loss history on a layer is likely to get experience-rated pricing based on that track record. A brand-new line of business with no claims history yet would instead be exposure-rated, using an industry severity curve.