Most P&C executives we talk to assume that Follow the Fortunes and Follow the Settlements are background doctrines, the kind of thing that ride along with a reinsurance treaty whether the wording mentions them or not. They believe that if the cedent acted in good faith, the reinsurer is bound to follow the cession. That belief has been comfortable for a long time and is now incorrect on the law in the Second Circuit.
In Utica Mutual Insurance Co. v. Clearwater Insurance Co., 906 F.3d 12 (2d Cir. 2018), the Second Circuit vacated the cedent's claim to Follow the Settlements protection, holding that Utica had not demonstrated the reinsurer was bound to indemnify per the underlying settlement.1 The doctrines are contractual rather than common-law, and a cedent that wants the protection has to point to the wording in the contract. The Eleventh Circuit reached a similar position in Public Risk Management of Florida v. Munich Reinsurance America, Inc., 38 F.4th 1298 (11th Cir. 2022), where the court declined to imply a follow-the-fortunes obligation absent express contractual language.2
For a regional or specialty carrier with AI anywhere in the claims pipeline, that holding is more consequential than the average board has processed. We think it deserves to be on the agenda this quarter.
What These Doctrines Actually Do
Follow the Fortunes (FTF) and Follow the Settlements (FTS) are two related contractual provisions that, when present in treaty wording, require the reinsurer to follow the cedent's good-faith claims-handling and settlement decisions even where the reinsurer would have decided differently. FTF is the broader of the two, covering the cedent's underwriting and claims fortunes generally. FTS is narrower, focused on settlement decisions on individual claims.
Their commercial purpose is straightforward. Treaty reinsurance only works if the cedent can cede a claim and reasonably expect the reinsurer to honor the cession without re-litigating every coverage question. Without that expectation, the cedent would face a second adversarial proceeding on every claim, this time against its own reinsurer, on top of whatever fight it just settled with the original policyholder. The express FTF/FTS clause is the contractual shorthand for "we are not going to do that to each other."
For most of the modern history of the US reinsurance market, those clauses have been so common that practitioners began to treat them as effectively automatic. They are not, and Utica v. Clearwater is the reminder.
What the Second Circuit Held
The dispute in Utica v. Clearwater arose from facultative reinsurance certificates Utica had ceded to Clearwater. Utica argued that the certificates should be read to include a Follow the Settlements obligation, even though the certificates themselves did not contain the language. The cedent's position rested on a long industry practice of treating the doctrine as background law.
The Second Circuit rejected the cedent-friendly framing. The court vacated and remanded on the cross-appeal because the cedent had not demonstrated entitlement to the Follow the Settlements protection it claimed.1 The legal authority a cedent might want to invoke against a reinsurer who second-guessed a cession would have to come from the document itself. There was no judge-made backstop available to fill the gap.
The Eleventh Circuit reached a comparable conclusion in 2022 in Public Risk Management of Florida v. Munich Reinsurance America, Inc., declining to imply Follow the Fortune obligations into a reinsurance contract whose express terms required proof of coverage.2 With two federal circuits now pointing the same direction on different fact patterns, the cedent's traditional fallback argument has narrowed considerably.
The takeaway is narrow and important. A cedent looking at its reinsurance contracts today should know with certainty whether each one contains express FTF and FTS language, because everything downstream of that question, including the treatment of AI-influenced cession decisions, depends on the answer.
Why This Sharpens the AI Question
The conversation about AI in insurance claims has so far focused on speed, accuracy, and customer experience. The reinsurance dimension has been mostly absent. Utica v. Clearwater is one of the reasons that absence matters.
Consider a regional carrier that has deployed AI across its claims operation. Some combination of intake chatbots, automated coverage triage, AI-assisted reserving, and ML-based fraud flagging now sits between a notice of loss and a settlement check. The actual settlement decision still nominally rests with a human adjuster, but the inputs the adjuster sees, the recommended reserve, and the path the file took to her desk were all shaped by an algorithm.
When that carrier ultimately cedes a developed claim to its reinsurer, the question of whether the cession reflected the good-faith judgment of an experienced adjuster or the lightly-supervised output of a model is one the reinsurer is now in a strong position to ask. Without express FTF/FTS language in the treaty, the question comes up with very little doctrinal constraint on the reinsurer. Even where the express language is present, the doctrine carries well-known voiding triggers, and the reinsurer can invoke any of them on the right factual record.
Those voiding triggers are the second piece most boards have not put on the agenda.
The Voiding Triggers Inside FTF/FTS
Even where Follow the Fortunes and Follow the Settlements are written into the treaty, the doctrine has standard exceptions that the reinsurer can invoke to avoid a cession. The major categories are well-established in reinsurance law:
- Bad faith on the part of the cedent in handling the underlying claim.
- Ex gratia payments, meaning settlements paid where no coverage actually existed under the underlying policy.
- Settlements that fall outside the underlying coverage as written.
- Gross negligence in claims-handling.
Each of these categories is a place where AI-influenced decisions create new arguments for a reinsurer who wants one. A coverage attorney representing a reinsurer in 2028 who needs to defeat FTS protection on a developed claim has more to work with when the cedent's claims process included an algorithm whose internal reasoning the cedent itself cannot fully reconstruct.
We are not predicting bad faith findings against carriers using AI. The doctrine of bad faith requires a high evidentiary threshold and reinsurance disputes generally settle before contested findings of fact are made. The relevant point operates one step earlier. The presence of AI in the claims pipeline gives the reinsurer's counsel a richer factual record to work from when negotiating which cessions get paid in full, which get paid net of arbitration costs, and which get refused outright. The settlement leverage moves toward the reinsurer in proportion to how much of the cedent's claims process the cedent cannot fully explain.
The same logic runs against ex gratia challenges. When a coverage decision was made or substantially shaped by a model, the question of whether the underlying policy actually covered the loss can become harder to defend cleanly. If the reinsurer's review concludes the model approved a payment the policy did not require, the cession is exposed under the ex gratia exception even where FTS is in the wording.
We covered the upstream version of this argument in the canonical post on why reinsurance is the existential AI risk for insurers, and the disclosure mechanics around it in the companion piece on the treaty clause your AI strategy lives or dies on. Utica v. Clearwater is the specific case that closes the door on the assumption that doctrine alone, without disclosure and without express wording, will protect the cession.
What a Defensible Posture Looks Like
To be precise about the precedent: no published case has yet rescinded a treaty or refused a cession on the ground that AI made the underlying claims decision. The legal architecture is in place, and Utica v. Clearwater is one of the load-bearing components, though the litigated AI case has not yet been written. The infrastructure exists ahead of the history.
Three things move a carrier into a defensible posture before the litigated case arrives.
The first is a treaty audit that maps, for each layer in the program, whether express FTF and FTS clauses are present and what the carve-outs say. Many CROs assume the answer is yes everywhere. The audit often finds it is no in places, particularly in older facultative covers and specialty layers placed through newer markets.
The second is a written record of human judgment on AI-influenced claims decisions, especially on larger files. The defensive value of an algorithm in the claims pipeline depends entirely on whether the cedent can show that a qualified person reviewed and adopted the algorithm's recommendation as her own judgment, with reasoning the cedent can reproduce three years later in front of a reinsurer's coverage counsel. Without that record, the cession is harder to defend on the standard FTS exceptions even where the treaty wording is friendly.
The third is the discipline of treating large or unusual cessions as if FTS were not in the contract. Where the dollar value or factual circumstances warrant, the cedent's claims and reinsurance functions should be able to articulate why the cession is sound on the merits, independent of doctrinal protection. Building that articulation into normal workflow on the cessions where it matters most reduces the carrier's exposure on the cessions where the reinsurer eventually pushes back.
The CEO Question for the Next Cycle
The conversation about AI in claims is still mostly about productivity. The reinsurance counterparty is having a different conversation, and the federal courts have already answered, in Utica v. Clearwater and the Eleventh Circuit's 2022 follow-up, a threshold doctrinal question that most boards have assumed would never need to be answered.
The question for the next board meeting is short.
What does our treaty actually say about Follow the Fortunes and Follow the Settlements, and is the wording sufficient to cover cessions on claims that an algorithm meaningfully shaped?
Answering it requires reading contracts the carrier is already a party to. The window to fix what the audit finds sits before the next renewal cycle, with no equivalent window after the next disputed cession.
Related
- The Reinsurance Treaty Is Where AI Risk Becomes Existential
- The One Treaty Clause Your AI Strategy Lives or Dies On
- The Lemonade "AI Jim" Tweet Is a Reinsurance Case Study, Not a PR Story
- Your Reinsurer Has Already Formed an Opinion About Your AI
Footnotes
-
"The Second Circuit also vacated and remanded on the cross-appeal as the plaintiff did not demonstrate its entitlement to a judgment that the defendant was bound to indemnify the plaintiff according to the plaintiff's settlement with the insured." — The Daily Record (NY): Second Circuit – Reinsurance: Utica Mutual Ins. Co. v. Clearwater Ins. Co. (October 16, 2018) ↩ ↩
-
"The court determined that the reinsurance agreement contained language expressly inconsistent with follow-the-fortune obligations… The Eleventh Circuit declined to imply such a clause absent explicit contractual language." Underlying decision: Public Risk Management of Florida v. Munich Reinsurance America, Inc., 38 F.4th 1298 (11th Cir. 2022). — Christian & Small LLP: Eleventh Circuit Reiterates Uphill Battle for Cedents Seeking to Imply Follow-the-Fortune Clauses (October 10, 2022) ↩ ↩
