Insurance / Specialty (Re)insurance · Deep dive

Convex Insurance

Stephen Catlin's 2019 do-over — a Bermuda + London specialty (re)insurer purpose-built for the post-XL Catlin hard market, now recapitalised at ~$7B by Onex and AIG after the IPO route closed.

emerging

The question that decides it: **Does Convex's 2019-vintage specialty book hold a sub-95% combined ratio through the 2024-2026 softening cycle** while Fidelis Insurance Holdings (NYSE: FIHL) and Berkshire Hathaway Specialty Insurance deepen their London Market and Bermuda attach, and hyperexponential/Ki-style algorithmic follow markets take the actuarial cold-start advantage away from every new specialty carrier — or does the greenfield disadvantage (no long-tail loss data, no reserve cushion built in prior soft markets) show up as reserve strengthening on the 2019-2021 vintage years the moment pricing gives up its 2023 peak?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Convex Insurance is the 2019 do-over Stephen Catlin built after selling Catlin Group to XL for £2.79B and watching the merged XL Catlin franchise get absorbed into AXA. Launched April 2019 in Bermuda and London with a ~$1.7B founding capital raise co-led by Onex Corporation and PSP Investments, topped up with a ~$1B follow-on in 2020, Convex bet that the specialty (re)insurance cycle was about to harden and that a clean-balance-sheet greenfield carrier — no legacy reserves, no runoff drag — could compound book value faster than the incumbent Lloyd’s syndicates and Bermuda peers still digesting the 2017-2019 catastrophe years. That thesis worked: gross written premium scaled from ~$1.1B in 2020 to roughly $5.9B in 2025, and the group’s combined ratio moved from 173% at the 2020 start (all expense, no earned premium) to ~87% by 2025. In February 2026, Onex and AIG jointly injected ~$7B of new capital in a recapitalisation that replaced the expected IPO route, took PSP out, and installed Onex CEO Bobby Le Blanc as chair after Catlin’s March 2026 step-down. The company matters because it is the single largest greenfield specialty balance sheet built in the London Market since Aspen and Allied World in the early 2000s, and because its next five years will settle whether the “start a Bermuda carrier at the top of the cycle” playbook still works once softening sets in and AI-native and MGA-first structures compete for the same accounts.

Founding story

The Convex story does not start in 2019 — it starts in 1984, when Stephen Catlin left his underwriter’s seat at Steel Burrill Jones and founded Catlin Group on the Lloyd’s platform. Over the next thirty years Catlin built the group into a top-five Lloyd’s specialty carrier, took it public on the London Stock Exchange, expanded into Bermuda and Zurich, and sold it to XL Group in 2015 for £2.79B — one of the largest London Market transactions of that decade. Catlin and long-time Chief Underwriting Officer Paul Brand stayed on at the merged XL Catlin, but the AXA acquisition of XL in 2018 accelerated their exit. By late 2018 both were unemployed, in their late-50s / mid-50s, and staring at what they believed was the beginning of the largest specialty-insurance hardening cycle since 9/11.

Onex Corporation — the Toronto-listed alternative-asset manager (TSX: ONEX) — and PSP Investments, one of Canada’s largest pension funds, agreed with them. In April 2019 the two co-anchored a ~$1.7B founding capital raise for a new Bermuda + London group called Convex, with Catlin as Executive Chairman and Brand as Group CEO. The Onex-PSP anchor was topped up by CDPQ, GIC, and other sovereign co-investors. The capital deployed into two regulated subsidiaries: Convex Insurance UK Limited (a UK PRA-regulated insurer writing London-market and European specialty) and Convex Re Limited (a Bermuda-domiciled reinsurer). AM Best assigned an initial A- (Excellent) financial-strength rating in 2019.

The founding team was not just Catlin and Brand. Convex pulled a large fraction of its senior underwriting bench directly from XL Catlin — a hire pattern that led to a widely-reported non-solicit dispute with AXA XL in the first eighteen months. Adrian Spieler joined as Group CFO from Novae. Chris Newman joined as Group COO. The pitch to underwriters was simple: no legacy claims book to manage, no bureaucracy, and Catlin himself was picking the classes.

How it works

Specialty (re)insurance is not the personal-lines P&C most people think of when they hear “insurance”. Convex sells cover to large commercial buyers — Fortune 500 boards, oil majors, aviation operators, marine fleets, financial institutions, energy traders — either directly through Lloyd’s syndicates and open-market broker slips, or as a reinsurer to other primary carriers. A typical account is placed by a global broker (Aon, Marsh, Willis Towers Watson, Guy Carpenter, Gallagher Re) into a subscription market: no single insurer takes the whole line, so Convex might be the lead underwriter on a $500M limit and take a 15% share, with the balance placed with a dozen other carriers. Each participation earns Convex a premium, generates a claims-reserve liability, and puts capital at risk against the modelled loss distribution.

Two things make specialty (re)insurance mechanically different from primary personal-lines: (1) the tail. A commercial casualty claim from 2020 might not be reported until 2027 and not settled until 2030-2032 — meaning Convex is still adding reserves to underwriting years long after the premium was booked. (2) The cycle. Specialty pricing moves in multi-year cycles; premium adequacy in 2022-2023 could be 30-40% above the 2019 baseline, meaning the same $1B of GWP represents a very different quality book depending on which vintage the premium was written in. Convex’s 2019-2021 book was written into a hardening cycle where pricing had not fully caught up to loss trend; its 2022-2024 book was written into the peak. The 2025-2026 book is being written into softening — and every specialty carrier CEO on an earnings call this cycle has said the same thing about discipline that every prior cycle’s specialty carrier CEO said before them.

The other mechanical piece is investment income. Convex writes premium, holds it as loss reserves for years, and invests those reserves. In a 5%+ short-rate environment (2023-2025) the investment yield became a material contributor to bottom-line ROE — perhaps 200-400 bps of the group’s combined-ratio equivalent. When rates fall, that contribution shrinks, and the underwriting book has to work harder to hit target ROE.

Product and business overview

Convex writes across most of the specialty product surface a Lloyd’s + Bermuda group would target:

• Property (direct & facultative) — commercial property risk for large industrial, energy and retail estates, plus fac reinsurance participation on primary carrier’s largest single-risk exposures.
• Marine and energy — hull, cargo, marine liability, offshore energy operators, energy trading exposures.
• Aviation — airline hull and liability, general aviation, war risk.
• Political violence, terrorism, war — one of the classes that hardened most aggressively after 2022.
• Cyber — a class Convex expanded into over 2021-2023 as pricing tripled off the 2019-2020 base.
• Professional lines — D&O, E&O, financial-institution E&O.
• Casualty — general liability, excess casualty, US-exposed umbrella.
• Financial institutions — bankers-blanket-bond, professional indemnity, crime.
• Reinsurance — treaty and facultative participation across the same product set.

Distribution is 100% wholesale — Convex does not sell direct. Business flows in through the London Market broker channel (Aon, WTW, Gallagher, Guy Carpenter, Lockton, McGill & Partners), plus a growing Miami and Dubai footprint for LatAm and MENA-facing risks. In April 2025 Convex launched Syndicate 1984 at Lloyd’s, giving the group additional stamp capacity and access to Lloyd’s international licensing network for the first time.

Business model and pricing

Convex earns premium (gross written premium, GWP), cedes some to retrocessionaires (net earned premium, NEP), pays claims (loss ratio), pays commissions and internal cost (expense ratio), and adds the two into a combined ratio. Below 100% is an underwriting profit; the industry benchmark for a well-run specialty carrier is 90-95%. Convex’s combined ratio moved from 173% in 2020 (the year the expense ran while premium was still ramping) to roughly 87% in 2025 per the disclosures accompanying the Feb 2026 recap.

Real pricing depends on the class. A cyber policy for a Fortune 500 buyer that cost $1M in 2019 cost $3-4M in 2023 and is settling back toward $2.5-3M in 2026. Marine hull rates hardened 30-50% between 2020 and 2023 and are giving back 10-15% per renewal. Political violence and war remained hard through 2025 because of Ukraine and Red Sea exposures, with 2026 the first softening year.

Traction over time

Figures reflect published disclosures, Reuters and Insurance Insider coverage of the Feb 2026 recap, and AM Best rating rationales; Convex is private and does not file GAAP financials, so precision year-over-year varies.

Market analysis

The global specialty (re)insurance market is roughly $250-300B of GWP (Aon Reinsurance Aggregate, 2025), of which Lloyd’s alone accounted for £52.1B in 2024 (Lloyd’s Corporation annual report, 2024). Bermuda accounts for another ~$130B of specialty and reinsurance GWP. The market has three structural forces moving it: (1) climate-driven cat volatility, which is expanding reinsurance limits demand faster than primary premium; (2) cyber and casualty severity, which have re-priced the primary market; and (3) capital cyclicality — every hardening attracts new capacity (Vantage 2020, Inigo 2020, Ark recap 2020, Fidelis 2015 relaunched), which mean-reverts pricing.

Competitive intel

Convex competes on every major London Market and Bermuda specialty slip against the incumbents named in the frontmatter. The two structurally hardest competitors are BHSI — because Berkshire’s AA balance sheet allows it to underprice specialty peers on any long-tail line — and Ki Insurance, because Ki’s algorithmic follow model runs at an expense ratio Convex cannot match. Fidelis MGU + Fidelis Insurance Holdings is the closest structural analogue to Convex on strategy, and its June 2023 IPO gives the public market a live benchmark for what Convex would trade at.

History and evolution

• 1984 — Stephen Catlin founds Catlin Group at Lloyd’s.
• 1997-2004 — Catlin IPOs on LSE, expands to Bermuda and Zurich.
• 2015 — XL Group acquires Catlin for £2.79B.
• 2018 — AXA acquires XL Group; Catlin and Brand exit.
• April 2019 — Convex launches with ~$1.7B from Onex + PSP; AM Best initial A-.
• 2020 — ~$1B follow-on raise; combined ratio 173% as expense runs ahead of earned premium.
• 2020-2021 — Non-solicit dispute with AXA XL over hires resolves; underwriting bench largely built.
• 2022 — First sub-100% combined-ratio year; Hurricane Ian tests the property book.
• April 2023 — AM Best upgrades flagship subsidiaries to A (Excellent).
• 2023-2024 — GWP compounds through the hard market; IPO rumours begin.
• April 2025 — Syndicate 1984 opens at Lloyd’s.
• February 2026 — Onex + AIG ~$7B recapitalisation; PSP exits; Bobby Le Blanc becomes non-executive chair.
• March 2026 — Stephen Catlin steps down as Executive Chairman; Paul Brand remains Group CEO.

What people say

The case for. Broker feedback in the trade press consistently rates Convex among the more decisive specialty markets for large complex risks — quotes in Insurance Insider and The Insurer name Convex among the “top 5 leaders” London brokers approach on E&O and D&F property. AM Best’s April 2023 upgrade rationale cited “very strong risk-adjusted capitalisation” and “disciplined underwriting”. Glassdoor reviews of Convex skew above the London Market average, with employees calling out compensation and the “grown-up culture” post-XL. The Feb 2026 Onex + AIG recap price implies underwriters and their advisers believe the 2019-2024 book is a real going concern rather than a portfolio to be run off.

The complaints. The first two years’ combined ratios (173% in 2020, ~120% in 2021) were bad enough that AM Best held the rating at A- for longer than the founding team wanted, and the trade press ran regular “when does Convex actually turn a profit” pieces through 2021. Greenfield specialty carriers carry a structural risk peers with long-tail claim histories do not: their 2019-2021 vintage reserves are set on modelled loss trends without the benefit of two decades of prior-year settlement patterns, and specialty market history is full of carriers that reserved 2019-2021 into the ground once actual settlements matured (see the 2000-vintage Bermuda class of 2001). The February 2026 recap replaced an IPO — which many industry observers read as a signal that public-market investors were unwilling to pay the multiple Onex wanted on the 2019-vintage book. Employee reviews on Glassdoor flag rapid growth strain and inconsistent line-of-business autonomy after the 2023 rating upgrade brought a more centralized operating model. There is also a leadership-continuity question: Catlin’s March 2026 exit removes the founder-underwriter voice from the board, replaced by Onex’s CEO — a governance shift that reads as more PE-owned insurance company and less founder-led franchise.

Outlook: the open question

The answer conditions. Convex works as an investment if two things are true across 2026-2029: (1) the 2019-2024 vintage-year reserves prove adequate — no cumulative reserve strengthening greater than ~5-7% of net reserves — and (2) the group holds a sub-95% combined ratio through the softening cycle, with the Feb 2026 AIG minority relationship translating into distribution advantage (US admitted paper access, LatAm and Asian broker relationships) that Convex could not build on its own. Convex fails as an investment if either fails: reserve strengthening on the 2019-2021 book would signal the greenfield actuarial cold-start was more expensive than the founding team modelled, and a combined-ratio drift above 95% into softening would put Convex closer to Aspen and Lancashire on the mediocre-middle-tier scoreboard than to Fidelis and BHSI on the durable-franchise one. The Feb 2026 recap kicked the IPO question down the road but did not answer whether a specialty (re)insurer built in 2019 can compound book value through a soft-market vintage.

How to attack it

The specific wedge to attack Convex is not “start another Bermuda specialty carrier” — that model has now been executed by Vantage, Inigo, Ark, Fidelis and Convex itself, and the marginal Bermuda + London capital is priced accordingly. The wedge is AI-native specialty MGA + fronted paper — a de novo underwriting team that writes on someone else’s admitted balance sheet (Accelerant, Skyward, Aspen, a segregated Bermuda cell) with an AI-native pricing platform (hyperexponential-backed or bespoke) doing what Convex’s underwriters do manually. Structurally the MGA takes 25-30% of the premium as ceding commission and profit share instead of Convex’s roughly 80-85% net earned share, but it also does not need $2-10B of equity — a $30-50M seed and a fronting deal is the whole capital gate.

Convex’s exploitable weaknesses:

• Greenfield reserve risk. The 2019-2021 book has no long-tail settlement history. Any AI-native MGA competing today has the advantage of choosing which vintage to enter and can price against historical development the same way BHSI does off Berkshire’s data. Source: AM Best rating rationales through 2023 have flagged the “limited operating history” for the group.
• Expense ratio. Convex reportedly runs an expense ratio in the mid-30s, which is credible for a full-service Lloyd’s + Bermuda group but 15-20 pts above Ki Insurance’s algorithmic follow model.
• Distribution dependence. ~100% wholesale — Convex depends entirely on Aon, WTW, Gallagher and Guy Carpenter routing accounts to its slip. Any MGA + fronting attacker can build the same broker relationships without the fixed-cost balance sheet Convex carries.
• Governance and founder-continuity. Stephen Catlin’s March 2026 exit removes the underwriter-signal that anchored the 2019 capital raise. A well-funded specialty MGA led by a departing Beazley or Hiscox line-head could replicate the underwriter-brand advantage Convex had at inception.
• AIG relationship risk. The Feb 2026 AIG minority likely comes with some form of preferential quota-share or reinsurance treaty — a dependence that creates counterparty-concentration risk if AIG’s own specialty appetite shifts.
• Cyber cold-start. Convex expanded cyber aggressively in 2021-2023, which is exactly the vintage that will be re-priced most heavily as cyber claims mature.

Adjacent-segment play

Convex’s core capabilities — a scale specialty balance sheet, disciplined underwriting bench, London Market and Bermuda regulatory perimeter, and Lloyd’s Syndicate 1984 — could be repackaged for at least five adjacent segments. First, US admitted paper — Convex writes excess & surplus business into the US through the Lloyd’s surplus-lines channel, but a targeted acquisition of a small US admitted carrier (an Amtrust runoff shell or an Ambac-adjacent stub) would let Convex sell to US buyers who require admitted paper for regulatory reasons. Second, MGA-in-a-box hosting — Convex could offer its Bermuda and Lloyd’s paper to third-party MGAs looking for a fronting counterparty, the same play Accelerant, Skyward and Vantage’s MGU arm have monetised, adding a fee-income stream that is not correlated to Convex’s own underwriting cycle. Third, LatAm regional — the Miami office already writes some LatAm-facing energy and political-violence risk; a scaled push into São Paulo or Mexico City specialty aligned with the nearshoring wave would open a distinct market where Bermuda paper is scarce. Fourth, retro / ILS pivot — Convex could seed a Bermuda ILS fund that raises third-party capital to write retrocession behind its own primary book, converting Convex’s underwriting franchise into a capital-light fee business (Fidelis’s MGU + Fidelis Insurance split, but staged rather than post-IPO). Fifth, run-off management — the same disciplined underwriting bench that built the primary book could be redeployed to run off legacy specialty portfolios in Europe or US, competing with Enstar and Randall & Quilter for the increasing volume of legacy specialty reserves European regulators want off-balance-sheet. Not attractive: SME retail — Convex has no channel and no advantage against Beazley, Hiscox, or Vouch on SME cyber and D&O.

Sources and further reading

• Convex Group Ltd. — Corporate site, 2026 (product surface, offices, leadership)
• Onex Corporation press release — Convex recapitalisation, February 2026 (recap structure)
• Insurance Insider — “Onex, AIG lead $7B Convex recap”, February 2026 (deal coverage)
• Reuters — “AIG takes minority stake in Convex Insurance”, February 2026 (AIG involvement)
• The Insurer — “Convex launches Syndicate 1984 at Lloyd’s”, April 2025 (Lloyd’s platform)
• AM Best rating rationale — Convex Insurance UK Limited, April 2023 (A upgrade)
• Lloyd’s Corporation Annual Report 2024, Lloyd’s, 2024 (£52.1B market size)
• Aon Reinsurance Aggregate 2025, Aon Reinsurance Solutions, 2025 (specialty market sizing)
• Financial Times — “Stephen Catlin returns with Convex”, April 2019 (founding)
• Fidelis Insurance Holdings — Form 10-K FY2024, March 2025 (comparable public specialty peer)
• Ki Insurance — corporate overview, 2025 (algorithmic follow market)