The rulebook that keeps the underwriter in the chain
Lloyd's new framework for broker-originated reinsurance structures reads like a compliance list. Read it as a design, and it is an operating model — a seating plan for who owns, funds, prices and evidences the risk.
What it is
On Friday 16 October, Lloyd's published a market bulletin setting out, for the first time, a full framework for broker-originated reinsurance structures, effective 1 January 2027. The context is the market's loudest argument of the autumn: the reported Aon-Blackstone vehicle codenamed Cortina, a structure that would put a fund manager's capital behind a broker's placements without an insurer in the chain to vet the risk or set the price.
The framework is neither a rejection nor an endorsement. It is a set of guardrails, and they are specific. A structure may take no more than 10% of any individual broker's treaty portfolio. The broker cannot own or capitalise the vehicle. The Lloyd's member aligned to the managing agent must provide at least 19.99% of the capital, and lead reinsurers at least 25% of the capacity, with the panel agreed by the managing agent. Once participation is set, a cedent must cede either all of its eligible portfolio with that broker to the structure, or none of it. The managing agent must be able to challenge the broker's pricing assumptions, with monthly reporting at bind level and quarterly data sharing with Lloyd's. A structure that has not reached 70% indexation by year three can be wound up. Efficiencies must flow back to protection buyers through a client dividend.
Read as a list, that is compliance. Read as a design, it is an operating model: who owns the vehicle, who funds it, who prices the risk, who challenges the pricing, what data proves the challenge, and what happens if the model fails. That is the interesting reading, and the one the market should spend the next few months on.
Why it matters now
Because the month-long argument about Cortina was never really about one vehicle. It was an argument about which node of the risk-taking chain each market participant owns. Brokers argued they should be able to put any form of capital to work for clients. Underwriters argued that a price set by nobody is not a price. Lloyd's has now answered in writing, and the answer is a seating plan: the broker originates, the member's capital carries skin in the game, the lead reinsurers price, the managing agent challenges and owns oversight, the cedent commits, and Lloyd's supervises.
The timing matters. The framework landed days before the 2027 capacity auction season went fully public, the moment every October when the market publishes its own evidence of who owns what. It landed in the same week two new syndicates received approval, Generali's Red Bridge and Credeq's 1998, both arriving with their underwriting decisions carried inside. And it landed a fortnight after JP Morgan became the first major analyst to warn, on the record, about facilities backed by private credit. The demand for these structures is not going away; Lloyd's own bulletin says it expects demand to increase. What changed on Friday is that the demand now has a specification attached.
The deeper reason it matters: this is the market codifying the thing it has never before had to write down, because it used to be embedded in everyone's behaviour. The chain, broker, underwriter, capital, evidence, always held together because the people in it knew each other and the firms owned the process end to end. Facilities and third-party capital pulled the chain apart into components. A rulebook is what you write when the shared understanding stops being shared.
What firms should do
Read the framework as an operating model, not a compliance list
Every guardrail maps to a node with an owner and an evidence requirement. The 10% cap is portfolio design. The 19.99% and 25% floors are ownership design. The challenge function is a capability specification. The monthly bind-level reporting is the data layer. The 70% indexation test is a performance gate. The client dividend is the value flow. Draw your own structure against that map and the gaps are immediately visible.
Brokers: treat the all-or-none rule as a strategy question
Once participation is set, a cedent cedes its whole eligible portfolio with that broker to the structure or none of it. That means the portfolio segmentation decision, which risks go in and which stay out, is now a one-time, board-level act, not a placing tactic. The brokers who will make these structures work are the ones who can show, line by line, which business belongs inside and why.
Managing agents: build the challenge function before the capital arrives
The framework requires the managing agent to test the broker's assumptions, with the pricing and reinsurance capability to do it and the bind-level data to prove it. That is not a compliance chore; it is a hiring and data-investment decision. A managing agent that cannot challenge a broker's pricing is not running the structure, it is hosting it, and Lloyd's has just made clear it can tell the difference.
Capital providers: understand where your money is allowed to sit
Under the framework, external capital arrives behind a member with 19.99% skin in the game, beside 25% of lead reinsurer capacity, inside a structure the managing agent governs with evidence. That is a very different proposition from passive index exposure, and it is deliberately so. The structure is designed to keep the underwriting decision attached to the money.
The opmodal perspective
This is what the Architecture Canvas was built for: making the chain visible. The framework is, in effect, a canvas with ownership labels. Map the risk-taking process end to end, broker, capital, member, lead reinsurers, managing agent, cedent, and mark on each node who owns it, what evidence they owe, and where the handoffs run. Current state and target state in identical structures, so the gap between the two is a plan rather than a hope. Any firm designing a facility, or any capital provider sizing one up, can use that map to see which requirements it already meets and which it cannot evidence.
The market has always understood that its core asset is the pricing decision, owned and evidenced. What Friday's bulletin does is make that understanding explicit for the new structures, in the same way the auction makes it explicit for the old ones: an evidence exam with public deadlines. The firms that can show their workings, at bind level, on day one, will move fastest through the new entrant process and negotiate from strength everywhere else. The firms that treat the framework as paperwork will find out that the supervision is the point.
A month ago the market asked who decides what a risk is worth. On Friday it got a written answer: the underwriter, named, capitalised with skin in the game, challenged monthly, and evidenced at bind level. The question for 2027 is no longer who decides. It is who can prove they did.
