The mental model
A primary carrier collects premium from policyholders, pays losses, and keeps what's left. To smooth volatility, protect against catastrophic losses, and free up capital for new business, the carrier transfers some of its risk to a reinsurer. The reinsurer is, mechanically, an insurance company whose customers are other insurance companies.
A reinsurance contract sits behind the scenes. The original policyholder never sees it, never signs it, has no contractual relationship with the reinsurer. The primary carrier remains fully obligated to the policyholder; reinsurance simply lets the primary recover from the reinsurer for losses ceded under the treaty. This separation is fundamental, and it is why reinsurance rarely shows up in the policy file an underwriter sees on a single risk. Reinsurance lives one layer up, in the relationship between carriers and their reinsurance partners, in the annual renewal of treaties, and in the modeled aggregations that determine how much capital the primary needs to hold.
Reinsurance is risk transfer between carriers, not between a carrier and its policyholder. The primary always pays the original loss to the insured. Reinsurance recovery flows after, between the primary and the reinsurer. This is why a "reinsurance dispute" never affects the policyholder directly, and why the primary's credit risk on its reinsurers (the security panel) is one of the things rating agencies care about most.
Treaty vs facultative
The first major split in reinsurance is whether the reinsurer is taking a portfolio of risks or a single risk.
How they coexist
A primary carrier typically has both. The treaty handles the bulk of the book on a portfolio basis. Facultative is used as a top-up: when a single account has a limit larger than the treaty per-risk capacity, the underwriter cedes the excess facultatively. The fac placement is essentially its own mini-broking exercise, with the cedent's underwriter contacting reinsurers for a one-off line on a one-off risk.
The economics
Treaty pricing reflects the portfolio's expected loss ratio plus a risk margin and the reinsurer's expense load. Facultative pricing reflects the specific risk, often at higher rates because the reinsurer is undertaking adverse selection: the cedent only sends fac on risks the treaty cannot or will not handle.
A heavy facultative market in a class often signals that primary capacity is constrained. When primary carriers are forced to fac more limit because their treaties cannot absorb it, fac rates tend to harden first, and treaty rates follow at the next renewal. Watching fac volumes is one of the leading indicators of cycle dynamics.
Proportional reinsurance
The second major split is whether the reinsurer takes a percentage of every risk (proportional) or pays only after losses exceed a threshold (non-proportional). Proportional reinsurance is the older form and remains the simplest to understand.
Quota Share (QS)
The cedent and the reinsurer split every premium and every loss in fixed percentages. A 60/40 quota share means the cedent keeps 60% of premium and 60% of losses; the reinsurer takes 40% of each. Pure pro rata. Used heavily in casualty, financial lines, and program business. Simple to administer. Provides the cedent with capital relief proportional to the cession.
Surplus Share (Surplus)
The cedent retains a defined "line" of every risk and cedes the rest to the reinsurer. If the cedent's retention is $1M per risk and a policy has a $5M limit, the cedent keeps $1M (20% of $5M) and cedes $4M (80%) to the reinsurer. Premium and losses are split in the same proportion as the limit split, which varies by risk. Smaller risks may be wholly retained; larger risks have a higher cession percentage. Common in commercial property.
Ceding commission
Under proportional treaties the reinsurer typically pays the cedent a "ceding commission" to compensate for acquisition cost, expenses, and the premium tax the cedent has already paid on the original premium. Ceding commissions can be flat (e.g., 30%) or sliding (varying with the actual loss ratio). Sliding scales align incentives: a low loss ratio earns a higher ceding commission, sharing the upside; a high loss ratio earns a lower commission, sharing the pain.
Profit commission
Many proportional treaties include a profit commission: a share of the underwriting profit returned to the cedent above a defined threshold. Mechanism for sharing favorable outcomes back to the originating carrier and aligning interests in selection quality.
Non-proportional reinsurance
Non-proportional reinsurance pays only after losses exceed an attachment point. The reinsurer is not on every dollar of every loss; it is on losses above a threshold. Excess of loss (XOL) is the dominant non-proportional form.
Per-risk excess of loss
The reinsurer responds when an individual loss exceeds the attachment. Example: a property treaty with $4M xs $1M (attaches at $1M, $4M of cover above) responds on any single property loss between $1M and $5M. Used to protect the cedent's net retention against severity on a per-risk basis.
Per-occurrence excess of loss
The reinsurer responds when total losses from a single occurrence exceed the attachment. The aggregation is by event, not by risk. Critical for catastrophe coverage, where one storm produces many policy losses that together exceed any single-policy threshold. Treated separately as cat XOL (next section).
Aggregate excess of loss / Stop loss
The reinsurer responds when the cedent's total losses across all risks during the period exceed an aggregate attachment, often expressed as a percentage of premium (e.g., losses above a 75% loss ratio). Caps the cedent's worst-case annual loss ratio. Common in classes with frequency-driven loss patterns (workers' compensation aggregates, crop, agriculture).
Reinstatements
An XOL layer typically has a defined number of reinstatements. After a loss exhausts the layer, a reinstatement premium pays for the layer to "refill" for subsequent losses during the same period. A typical cat XOL might have one or two paid reinstatements; some have unlimited free reinstatements (very rare and very expensive). Reinstatement provisions are central to pricing in any year with high loss frequency.
An XOL is described by attachment and limit: "$50M xs $50M" means $50M of cover attaching above $50M of loss. The full tower is described layer by layer. Each layer can be placed with different reinsurers. A typical large cedent's cat program might have layers running from $50M xs $50M up to $500M xs $1B, each layer subscribed by a different mix of reinsurers, all renewing on the same date.
Catastrophe XOL
Cat XOL deserves its own treatment because it is the most economically significant reinsurance segment and the one most affected by modeling, climate, and capital market dynamics.
How cat XOL works
The cedent buys a tower of per-occurrence excess layers protecting against catastrophic events. The attachment is set at the level of loss the cedent is willing to retain per event (often a 1-in-5 or 1-in-10 year event, depending on appetite). The exhaustion is set at a remote return period (1-in-100 to 1-in-250 year, depending on rating agency requirements and capital). Layers between attachment and exhaustion are placed across multiple reinsurers, with each layer often having its own pricing reflecting the modeled probability of loss to that specific layer.
The hours clause
Cat XOL contracts define what counts as a single "occurrence" using an hours clause: a defined time window (typically 72 hours for windstorm, 168 hours for earthquake) during which all losses from a named event are aggregated as one occurrence. Outside the window, separate occurrences. The hours clause is one of the most negotiated provisions in cat treaties because it determines whether one cat event hits one layer or splits across two.
Cat models
Pricing for cat XOL relies on third-party catastrophe models (Verisk/AIR, RMS, KCC, ARA being the major ones, with Moody's RMS and Verisk holding the largest market share). The cedent provides exposure data (locations, replacement cost, construction, occupancy, protection) and the model produces estimated annual losses (EL), occurrence exceedance probability (OEP), and aggregate exceedance probability (AEP) curves. Pricing is anchored to model output, with adjustments for non-modeled perils and for cedent-specific factors. Model output is also the principal lever for capital deployment by reinsurers.
The 2023 hard market and the 2024 evolution
The 2017-2022 period of elevated cat losses (Hurricane Harvey, Irma, Maria, Camp Fire, Hurricane Ian, the 2022 European drought, repeated severe convective storm losses in the US) led to the hardest cat reinsurance market in roughly two decades, peaking at the 1/1/2023 renewal. Attachment points moved up substantially, terms tightened, and the industry's appetite for low-attaching cat layers shrank. The 2024 renewals saw modest softening at the top of the tower as new capital arrived, but attachment levels remained well above the 2017 baseline.
When attachment points on cat XOL move up, primary carriers are required to retain more loss before reinsurance recovers. This forces primary rates higher, can drive carriers out of cat-exposed states (Florida and California being the recent examples), and reshapes appetite across the entire admitted property market. Reinsurance dynamics flow downstream into primary policyholder pricing, sometimes with a one-to-two year lag.
The cedent / reinsurer relationship
Reinsurance is a relationship business. Treaties are renegotiated annually but the reinsurer / cedent partnership often spans decades, with reinsurers maintaining continuous lines on a given cedent's program through soft and hard markets alike.
Utmost good faith
Reinsurance contracts are governed by an unusually strong duty of good faith (uberrimae fidei) flowing from the cedent to the reinsurer. The cedent is expected to disclose all material information about the underlying portfolio. Failure to disclose material facts can result in rescission. The doctrine arose because the reinsurer is not pricing each individual risk and depends on the cedent's representations about the portfolio.
Follow the fortunes / follow the settlements
Two related doctrines govern claims handling under treaties. "Follow the fortunes" obliges the reinsurer to follow the cedent's underwriting fortunes (good and bad) under the treaty. "Follow the settlements" obliges the reinsurer to follow the cedent's reasonable claim settlements without re-litigating coverage issues the cedent has already adjudicated. Both doctrines have nuances and exceptions, and both are central to reinsurance dispute resolution.
Reinsurance arbitration
Most reinsurance contracts include arbitration clauses. Disputes go to a panel of industry arbitrators rather than to courts, and confidentiality is typical. The result is that reinsurance dispute case law is thinner than most other commercial dispute case law, and many high-value disputes resolve outside the public record.
Security and counterparty risk
The cedent's claims-paying ability depends on the reinsurer's claims-paying ability. Cedents diversify across multiple reinsurers per layer to limit concentration. Rating agencies (AM Best, S&P, Fitch) maintain ratings on reinsurers, and most cedents have minimum rating thresholds for the reinsurers they will use. Collateral (LOCs, trust funds) is sometimes required from non-rated or lower-rated reinsurers, especially in regulated US-domiciled programs where statutory credit for reinsurance requires admitted or collateralized security.
Retrocessions and the chain
Reinsurers themselves buy reinsurance, called retrocession. A reinsurer that has accepted a layer of cat XOL from a cedent may turn around and cede part of that layer to another reinsurer (the retrocessionaire). This is how risk is spread across the global capital base.
Why retrocession exists
- Capital relief. Same logic as primary reinsurance, one layer up.
- Aggregation management. A reinsurer accepting cat XOL from many cedents builds up aggregate exposures to the same underlying perils. Retro is how the reinsurer manages those aggregates.
- Capacity matching. A reinsurer may have appetite for the relationship but not for the full size of the line. Retro lets it write the relationship and lay off the excess.
The retro market
Retrocession is a smaller, more concentrated market than primary reinsurance. A handful of major retrocessionaires (some of them dedicated retro syndicates at Lloyd's, some of them large global reinsurers, some of them ILS funds) underwrite the bulk of the global retro market. Hard retro markets translate directly into hard reinsurance markets a year or so later.
Spirals and the historical cautionary tale
If reinsurers retrocede to each other in a closed circle, losses can re-enter the same balance sheets multiple times before settling. The "LMX spiral" of the late 1980s and early 1990s (a series of retrocessional covers within the Lloyd's market that overlapped catastrophically when the same losses hit) bankrupted multiple syndicates and was a defining event in the modernization of Lloyd's. Modern systems and contractual safeguards make a repeat spiral harder, but the structural risk of mutual reinsurance among a small group of counterparties remains a reason to read retrocessional security panels carefully.
ILS and cat bonds
Insurance-Linked Securities (ILS) are reinsurance contracts repackaged as financial instruments and sold to capital markets investors. Cat bonds are the most visible form. Together with other ILS structures (industry loss warranties, sidecars, collateralized reinsurance funds), ILS represents roughly $100B+ of dedicated reinsurance capital, a meaningful share of total cat reinsurance capacity.
How a cat bond works
A primary or reinsurance cedent (the sponsor) sets up a special purpose vehicle (SPV) and issues cat bond notes to investors. The investors' principal is held in collateral and invested in low-risk securities (typically money market funds). If a defined trigger event occurs during the bond's term, principal is reduced and paid to the cedent as a reinsurance recovery. If the trigger does not occur, the principal is returned to investors at maturity. Investors earn a coupon (the reinsurance premium plus the underlying rate) for taking the risk.
Trigger types
- Indemnity. Pays based on the cedent's actual losses. Closest to traditional reinsurance. Requires loss audit and verification, longer dispute periods.
- Industry loss. Pays based on a defined industry-wide loss measure (PCS in the US, PERILS in Europe). Faster to pay, basis risk for the cedent (industry losses may not match cedent losses).
- Parametric. Pays based on a defined physical parameter of the event (wind speed at landfall, earthquake magnitude). Fastest to pay, highest basis risk.
- Modeled loss. Pays based on a model run of the event using fixed exposure data. Compromise between indemnity and parametric.
Why investors want cat bonds
Cat bond returns are largely uncorrelated with broader financial markets. A hurricane has no relationship to interest rates or equity returns. For institutional investors looking for diversification, this uncorrelated return profile is valuable. The sponsor base (primary carriers, reinsurers, government risk pools) benefits from access to a deep pool of capital that does not need to earn an insurance company's cost of equity.
Sidecars and collateralized reinsurance
Sidecars are special-purpose reinsurance vehicles created alongside an existing reinsurer to take a quota share of a defined book, funded by external investors. Collateralized reinsurance is a single-cedent placement where the reinsurance is fully collateralized in trust, often by an ILS fund. Both structures channel capital markets investment into traditional reinsurance contracts without the formal bond issuance machinery.
Reinsurance brokers and the renewal cycle
The reinsurance broking market is highly concentrated. Three firms (Aon Reinsurance Solutions, Guy Carpenter, Gallagher Re) account for the majority of intermediated treaty reinsurance globally. The broker's role is structural: building the cedent's reinsurance program, modeling exposures, developing the marketing strategy, identifying the reinsurer panel, negotiating terms, placing the program, and supporting throughout the year on claims, accounting, and actuarial questions.
The annual renewal cycle
Most treaty reinsurance renews on January 1 (1/1). Property cat in particular is heavily concentrated at 1/1, with secondary peaks at April 1 (Japan), July 1 (US, Australia, Latin America), and other dates depending on the cedent. The 1/1 renewal is the single most important annual event in the global reinsurance market. The cycle typically begins in September with broker presentations to reinsurer markets, builds through October with quotes coming in, runs through November and December with negotiation and finalization, and concludes at year-end with bound treaties effective January 1.
The broker's analytics function
Modern reinsurance brokers maintain large analytics teams that produce the catastrophe modeling, loss triangulation, exposure aggregation, and capital optimization analysis that underpins each cedent's reinsurance program. The broker is doing real underwriting work in support of the placement, not just intermediating. The broker's analytics output is one of the principal deliverables in the renewal cycle.
The Monte Carlo
The Rendez-Vous de Septembre in Monte Carlo is the traditional opening of the reinsurance renewal season. A few thousand executives, brokers, and underwriters gather in early September to set the tone for the upcoming 1/1 renewal. Substantive negotiations do not happen in Monte Carlo, but signals about appetite, pricing direction, and capacity availability are exchanged that shape the rest of the cycle. The Baden-Baden meetings in October play a similar role for the European market.
Where IDP earns its keep
Reinsurance is a smaller IDP market than primary insurance by document count, but the documents that exist tend to be high value, dense, and used by senior people on tight deadlines. The 1/1 renewal cycle compresses an enormous volume of complex analytical and contractual work into a few months, and the document workflows during that period have meaningful automation potential.
The reinsurance workflows where Indico shows up
- Submission pack ingestion. A typical treaty submission pack includes the cedent's underwriting narrative, multi-year loss triangulations, exposure summaries, prior year slip and wording, broker presentation, model output, and supplemental supporting exhibits. Extracting structured fields from this pack into the reinsurance underwriter's workspace is high-value automation during the renewal compression.
- Wording comparison. Year-over-year wording changes are a central focus during renewal. Highlighting and summarizing changes between the expiring wording and the proposed wording lets the underwriter focus on what is different rather than re-reading hundreds of pages.
- Loss triangle normalization. Different cedents present loss triangles in different formats (development quarterly vs annual, paid vs incurred, with or without IBNR). Normalizing into a consistent structure for the actuarial team is mechanical work that benefits from automation.
- Catastrophe model output extraction. Extracting EP curves, PML metrics, and exposure aggregates from cat model PDFs into structured form for the underwriter's portfolio aggregation models.
- Facultative submission processing. Fac is a one-off broker submission flow that maps directly onto the primary submission intake problem, but with reinsurance-specific data (underlying primary policy details, original limits, ceded share, ceded premium).
- Treaty accounting reconciliation. Cedent quarterly statements vs reinsurer ledger. Mostly structured but with attached exhibits that require document parsing.
- Bordereaux for proportional treaties. The same coverholder bordereaux problem from Chapter 10 applies here for proportional treaties where the reinsurer needs cession-level reporting.
For a reinsurance underwriter during 1/1 compression, the demo that wins is the submission pack ingestion agent paired with year-over-year wording comparison: take the broker's renewal pack, surface the changes from the expiring slip, pull the loss triangles into normalized form, and produce a one-page summary the underwriter can read in two minutes before joining the call. The compression of analyst hours into the period when every minute counts is the value, and the audit trail back to the source documents builds the trust the reinsurance underwriter needs to commit capacity on tight deadlines.