Cat Bond Desk

The Artemis snapshot gives us a clean read: an 8.74% market yield, of which 4.57% is insurance risk spread and 4.17% is collateral yield, against a 2.5% expected loss on the outstanding market. That is roughly 1.8x spread over EL (4.57 divided by 2.5), a market-level number only. The block carries no per-deal spread or EL, so I will not pretend to price Easton Re or Nature Coast Re individually. What the pipeline does tell me is breadth: $18.9B across 98 deals, $64.7B outstanding, and a recent average deal of $163M.

The deal mix matters for credit selection. Florida named-storm paper is in the queue, with American Coastal's $25.5m Armor Re II (Florida named storm) and SafePoint's $250m Nature Coast Re (multi-state U.S. named storm), alongside USAA's $350m Residential Re 2026-2 multi-peril and Hannover Re's $450m Acorn Re (U.S. earthquake). Peril diversification across wind, quake and convective storm is what a credit investor wants to see, though I would not conflate a Florida wind bond with a California quake bond: different tail, different correlation.

On the macro side, collateral yield of 4.17% sits just above the 3.88% effective fed funds rate, so cash-like collateral is carrying a good chunk of total return. If rates fall, that leg compresses and the risk spread has to do more work. Margaret will say the market is softening on issuance volume alone; I would want to see spread-over-EL actually tighten on a per-deal basis before I agree, and today's block cannot show that.

Start with the Artemis tape. The cat-bond market yields 8.74%, split 4.57% insurance-risk spread and 4.17% collateral yield, against a 2.5% market-level expected loss. Treat that 2.5% as an anchor only: the snapshot gives no per-deal spread or EL, so any claim that a given deal is rich or cheap is unsupported today. What the data does support is breadth. $18.9B across 98 deals is a market that is clearing size, with recent deals averaging $163M and individual prints ranging from $450m (Acorn Re, U.S. earthquake) to $25.5m (Armor Re II, Florida named storm).

The mix matters more than the headline. A $350m Residential Re for USAA spans tropical cyclone, quake, severe thunderstorm, wildfire and more, which is multi-peril credit that investors are being paid to underwrite as a single spread. Nature Coast Re's $250m for SafePoint is a nine-state Southeast named-storm trade. When cedents that size keep coming to the 144A market, the collateralized layer is functioning as primary capacity, not a sideline.

I will push back on Margaret on one point. She will read this issuance pace as a cycle top. I read the same numbers as a market whose risk premium is still 4.57 points over a 2.5% EL; that is not a cycle that has been competed to zero. I will concede the obvious: this snapshot has no spread history, so neither of us can show direction from it.

Start with the tape. USAA came back to the market earlier this month, and Artemis reports the deal upsized 56% from its initial target to $350 million and priced below guidance. Demand outran supply on this one. When a sponsor upsizes and prices inside the range, investors are paying for the paper, not demanding a concession. USAA's program is broad: U.S. tropical cyclones, earthquake with fire following, severe thunderstorm, winter storm, wildfire, and more, plus flood losses on auto and renter policies.

The market-level numbers explain why. The Artemis snapshot shows a 4.57% insurance-risk spread over a 2.5% market expected loss, plus 4.17% collateral yield for an 8.74% all-in yield. That puts the spread at roughly 1.8x the market EL, though note the EL is a market aggregate and the block contains no per-deal spread or EL for Residential Re 2026-2, so I cannot tell you this deal's multiple. What I can say is that $18.9B of YTD issuance across 98 deals against $64.7B outstanding looks like a buyer-friendly sponsor market, with an average recent deal size of $163M that USAA's $350M comfortably clears.

I will contest where Margaret goes with this. Pricing below guidance is a credit-spread tightening, and spreads can tighten for good reasons or bad ones. The corpus gives me no per-deal attachment or EL for USAA, so I will not pretend to know whether this was priced right. It is a signal about demand, not a verdict on adequacy.

The numbers I care about today: 8.74% market yield, 4.57% insurance risk spread, 2.5% expected loss on the outstanding market. That gives a spread-over-EL multiple of roughly 1.8x on the outstanding book — not rich by historical standards, but still compensating investors for the underlying peril exposure. The collateral yield at 4.17% is doing real work here; with the effective fed funds rate at 3.88% per the live snapshot, the collateral leg is nearly a full Fed cycle above zero, which makes the total return on cat bonds meaningfully more attractive than it was in the 2020-2021 zero-rate era.

Recent deal flow tells the same story. The SafePoint $250M Florida named-storm placement (Nature Coast Re 2026-2), the USAA $350M multi-peril Residential Reinsurance deal, and Hamilton Re's $250M Easton Re across U.S. named storm and earthquake — these are serious-size transactions from credible cedents. The average recent deal size of $163M masks a bimodal distribution: large, rated sponsors clearing $250-450M (Acorn Re's $450M Hannover/Oak Tree earthquake placement is the headline) and small private deals like the $20M Artex Axcell and $25.5M American Coastal placements that suggest the private ILS shelf is still active for sub-investment-grade primary carriers.

Margaret Ennis is correct that $18.9B YTD issuance is a meaningful buffer against traditional reinsurer withdrawal — but I would push back on her framing that it simply 'blunts' the hard market. Cat-bond spreads are set at issuance and locked; they do not flex in real time with the reinsurance cycle. What the ILS market actually does is provide a price anchor: if W.R. Berkley walks away from a Florida named-storm layer at 12% RoL, and the ILS market is clearing comparable risk at an effective all-in yield of 8-9%, the cat bond wins the business — but the cat bond investor is also taking a spread that the independent model flags as potentially thin if climate non-stationarity is moving the true expected loss above the modeled 2.5%. The spread over EL is only as honest as the EL itself.

The HCI Group disclosure is the most structurally interesting data point this cycle. One of three reinsurance towers is triggered by Isaias losses, and at least one tokenized reinsurance security is impaired. This is not an academic stress test — it is live principal at risk inside a tokenized wrapper, and it will tell us something real about whether tokenized ILS structures behave the way their architects promised when collateral is actually called.

Set that against the current market snapshot: $64.7B outstanding, 8.74% yield, 4.57% insurance risk spread over a 2.5% expected-loss market. The spread-to-EL multiple on the outstanding market is roughly 1.8x — not lavish, but defensible in a post-2022 environment. What Greenberg's 'naive capacity' comment implies, however, is that new money is chasing attachment points with discipline that has not been tested by a meaningful loss. HCI's Isaias disclosure is exactly the kind of mid-tier event that separates the priced-right from the priced-to-win-the-mandate.

The $250M Nature Coast Re 2026-2 deal for SafePoint — covering Florida named storm across the Gulf corridor — and the $350M USAA Residential Re 2026-2 multi-peril deal are the right comparisons. Both came to market at a time when the YTD issuance pace of $18.9B across 98 deals signals investor appetite remains strong. Whether that appetite survives a sequence of HCI-style mid-tower hits is the question the next 60 days of loss development will answer. Watch for trapped-collateral signals in the secondary market; the 8.74% yield is only honest if investors can access their principal when they need it.

The Artemis dashboard is doing what it always does: giving us the only honest price signal in a market full of narrative. At 8.74% market yield — 4.57% insurance risk spread over a 4.17% collateral yield — with a market-level expected loss of 2.5%, the outstanding cat-bond portfolio is running at a risk-spread-to-EL multiple that remains well above the soft-market lows we saw in 2021. The market is priced to compensate for risk. $18.9B in YTD issuance across 97 deals, with a recent deal average of $139M, tells you that both supply (cedents wanting protection) and demand (ILS investors writing risk) are healthy and functioning.

The deals in the recent pipeline are worth reading carefully. Nature Coast Re Series 2026-2 is SafePoint Insurance Company — a Florida-market specialist — placing $250M against U.S. named storm across the Gulf coast states through November 2026. American Coastal Insurance, another Florida specialist, placed $25.5M via Armor Re II Series 2026-2 against Florida named storm in August. These are not diversified global cedents; these are concentrated Florida wind books accessing the capital markets because traditional retrocessional capacity for that specific peril remains constrained. The cat-bond market is functioning as a release valve for Florida property risk that traditional reinsurers are pricing very carefully.

Margaret Ennis is right that the €2.806B PERILS revision on European SCS is a loss-cost headwind for aggregate reinsurer budgets, and she is right that it stiffens January 1 negotiations in traditional markets. My read differs slightly on the ILS side: the $64.7B outstanding market is broadly diversified, and a European SCS event of this size, while meaningful, does not directly impair the collateral posted against U.S. wind or quake deals. The ILS market's insulation from European SCS aggregate draws is actually one of its structural advantages over traditional retrocessional capacity at this point in the cycle. Capital is not as fungible as it looks on a balance sheet.

The market arithmetic deserves to be read straight. Outstanding risk capital stands at $64.7B. The market is yielding 8.74% — comprised of a 4.57% insurance risk spread over a 4.17% collateral yield. Against a market-level expected loss of 2.5%, the implied spread-over-EL multiple is roughly 1.83x. That is not a distressed number; it is a functioning risk transfer price that reflects a well-supplied but not yet commoditized market.

Hannover Re's move is worth parsing carefully. Their Acorn Re Series 2026-1 — a $200M U.S. earthquake transaction via Hannover Rück SE and Oak Tree Assurance — closed in October and sits in the recent deal directory. This is a traditional reinsurer using the 144A bond market for a named-peril, single-jurisdiction exposure. That is not a distress trade; it is a structural choice. Horstmann's Baden-Baden remarks that ILS is a 'sustainable addition' and 'source of growth' track precisely: Hannover is not a passive recipient of ILS capacity, it is an active originator. When traditional reinsurers become ILS originators, the line between balance-sheet and capital-markets risk transfer is gone.

Margaret Ennis reads the supply surge as a cycle-turn signal, and she's not wrong about direction. But I'd push back on the framing that this is simply the hard market 'sowing seeds of the next soft.' The structural dynamic — a $64.7B outstanding book with professional ILS investors who have now lived through multiple event seasons and maintained discipline — suggests the floor on spreads is higher than the pre-2017 soft market. The 4.57% risk spread is not 2014. Whether it stays there through a large-loss year is the only test that matters.

The Artemis dashboard gives us the numbers: $64.7B outstanding, 8.74% market yield decomposed as 4.57% insurance risk spread over a 4.17% collateral yield, against a 2.5% market-level expected loss. That implies a multiple-on-EL of roughly 1.83x on the risk spread alone — not lavish by historical standards, but not thin either. The market is being paid to take risk. The question today is whether Champi and the nor'easter are priced into that spread or whether they represent a draw on collateral that investors did not fully underwrite.

Champi is a Western Pacific event. The overwhelming majority of U.S.-focused cat bonds — including the Nature Coast Re 2026-2 ($250M, Florida and Gulf Coast named storm), the USAA Residential Re 2026-2 ($225M, U.S. tropical cyclones plus a broad all-perils basket), and the American Coastal Armor Re II 2026-2 ($25.5M, Florida named storm) — are not triggered by a typhoon that stays in the Western Pacific. The nor'easter is more interesting from a trigger perspective. The USAA deal's all-perils basket explicitly includes 'other perils,' and the broader question of whether a subtropical storm system qualifies as a 'tropical cyclone' for attachment purposes depends entirely on the parametric or indemnity structure of individual deals — which the Artemis block does not specify at the deal level.

Ravi is correct that secondary-peril model error is real. But I'd note the market has been pricing secondary-peril risk at a premium since the 2017-2021 period of repeated secondary-peril surprises — the 4.57% risk spread is not a 2014 spread. The market is not oblivious. Where I'd push back on Ravi's framing is the assumption that model error necessarily means the spread is too thin; it could mean the spread is adequate for a distribution with fat tails, and investors are accepting that tail. The collateral structure means principal loss, yes — but the yield-to-maturity before a trigger event is real carry.

The Artemis dashboard is about as clean a read on market equilibrium as you get mid-cycle: $18.9B in YTD issuance across 97 deals, $64.7B outstanding, a market yield of 8.74% decomposed into 4.57% insurance risk spread and 4.17% collateral yield, against a market-level expected loss of 2.5%. Run the arithmetic — that spread-to-EL multiple is sitting close to 1.8x. In a market where 1.0x means you are being paid exactly your expected loss with zero risk premium, 1.8x is not exuberant, but it is no longer the 3x-plus distress pricing that followed the 2022-2023 hard market peak. Capital has come back in; the multiple has compressed accordingly.

Look at the recent deal flow for the geographic tell. Nature Coast Re 2026-2 for SafePoint ($250M, U.S. named storm, multi-state Gulf and Southeast corridor) and Armor Re II 2026-2 for American Coastal ($25.5M, Florida named storm) are the Florida wind exposures that have been the hardest to place for the past three years. The fact that SafePoint — a Florida-focused insurer that has been under significant capital pressure — can execute a $250M cat bond at current market terms tells you the ILS bid is real and deep. The average recent deal size of $139M is consistent with a market that is absorbing mid-sized cedents, not just the Swiss Res and USAAs of the world.

The Hormuz tanker story (Investing.com, single-source, developing) is worth a flag. Brent at $126.55/bbl and strikes on tankers near the strait introduce marine and energy liability exposure that sits entirely outside the U.S. named-storm and earthquake perils that dominate the cat-bond market. That risk does not show up in the 2.5% market EL figure. It is a reminder that the ILS market's spread-to-EL multiple is only as good as the peril catalog the models are built on — and marine war risk is not in that catalog.

Two jurisdictional firsts in a single week deserve attention. Generali's Syndicate 2031 Red Bridge at Lloyd's and Hong Kong's maiden PICC P&C sidecar are structurally different animals, but they're telling the same story: the franchised reinsurer is discovering that third-party capital is no longer a supplement — it's a structural component of how you run a diversified book. Generali's vehicle, per the Artemis reporting, is explicitly designed to 'make reinsurance arrangements more efficient' while opening the door to alternative and capital-markets participation. That's the sidecar architecture in everything but name.

The Hong Kong $12M PICC sidecar is small — call it a proof-of-concept — but the jurisdictional signal is real. The Hong Kong Financial Services & Treasury Bureau is actively welcoming the structure. In the current ILS market, where YTD issuance has reached $18.9B across 97 deals with an outstanding market of $64.7B, any new jurisdiction that can credibly compete for deal flow matters. The collateral yield component of today's market yield — 4.17% of an 8.74% total — means the effective insurance risk spread is 4.57% against a market-level expected loss of 2.5%. That's an EL multiple just under 1.83x. Tight by historical standards. Capital is not being scared away.

What both moves confirm is that the demand pull for alternative reinsurance capital is structural, not cyclical. Generali is not distressed; it's optimizing. PICC P&C is not retreating; it's experimenting. The issuance pace — $18.9B with a recent deal average of $139M — is consistent with a market that is digesting capital efficiently rather than being flooded. The Nature Coast Re 2026-2 deal at $250M for SafePoint and the Residential Re 2026-2 at $225M for USAA show that U.S. named-storm risk is still getting placed. The question my colleague Margaret will want to answer is whether this capital inflow is already softening the reinsurance pricing that made the market attractive in the first place.

The Stand story is the one to watch today, and not for the venture-capital headline. Tony Landi running reinsurance capital allocation at a cat-exposed property producer is an ILS architecture play dressed in startup clothes. Nephila built the original infrastructure for treating cat risk as an investable spread asset — Landi's remit at Stand, explicitly described as ensuring 'the right reinsurance capital is there to support the company as it grows,' means Stand intends to embed ILS-market access directly into its operating model rather than buying conventional reinsurance off the shelf. That is a meaningful structural bet.

Place it against the market numbers: 8.74% total yield, 4.57% insurance risk spread, 2.5% expected loss on the outstanding book. The multiple-on-EL implied by that spread is roughly 1.8x — not extravagant by hard-market standards, but disciplined. The collateral yield component (4.17%) is doing real work here; with the effective fed funds rate at 3.88%, T-bill collateral is no longer the drag it was in the zero-rate era, which keeps the total yield attractive to investors without requiring the risk spread alone to carry the whole return.

On the deal side: USAA's Residential Re 2026-2 at $225M is the anchor transaction of the recent window — broad U.S. multi-peril, autumnal timing, USAA's credit quality making it clean paper. The Armor Re II 2026-2 at $25.5M for American Coastal on Florida named storm is the more instructive data point for the domestic market. American Coastal is a Florida specialist, and the fact that it is accessing the 144A cat-bond market rather than relying purely on state-backed or conventional reinsurance capacity tells you something about where the Florida named-storm market is clearing. The deal is small — $25.5M against a $146M recent deal average — which suggests it is filling a specific attachment layer rather than replacing a broad tower.

The Acorn Re 2026-1 at $200M (Hannover Re / Oak Tree, U.S. earthquake) is a reminder that non-wind perils are absorbing meaningful ILS capital this season. $18.9B YTD across 96 deals with the market at $64.7B outstanding: the pace is constructive but not euphoric. The 96-deal count at an average of $146M implies no single mega-transaction is flattering the total — this is broad-based demand.

Let's anchor on what Convergence 2026 actually gave us in numbers, because the narrative around diversification is only as durable as the spread justifying it. The outstanding cat-bond market is $65.5B, YTD issuance is $18.9B across 96 deals, and the market yield sits at 8.74% — composed of 4.57% insurance risk spread and 4.17% collateral yield. Against a market-level expected loss of 2.5%, that implies a spread-over-EL multiple in the neighborhood of 1.8x. That is not a heroic multiple — peak hard-market pricing in 2023 was materially richer — but it is not the distressed giveaway of 2020–2021 either. The collateral yield component (4.17%, reflecting the effective fed funds rate of 3.88% and short-end T-bill positioning) is doing real work here: strip it out and the pure risk premium looks thinner.

The Convergence message that ILS has 'data to support' its diversification narrative is correct and slightly self-serving at the same time. The correlation of cat-bond returns to equity markets is structurally low — that's real. But the pitch to investment committees assumes the EP curve is the right input, and I'd note that the recent Isaias fatality in Birmingham, Alabama — 200-plus miles from landfall — is exactly the kind of inland secondary-peril event that lives in the tail of modeled distributions. The diversification story is strongest when the peril is well-parameterized; it gets uncomfortable when the loss run starts filling in model gaps.

Recent deal flow tells a specific story. USAA's Residential Re 2026-2 at $225M covers a genuinely broad peril suite including U.S. tropical cyclones, earthquakes, severe thunderstorm, wildfire, and even meteorite impact. Hannover's Acorn Re 2026-1 at $200M is a clean U.S. earthquake play. American Coastal's Armor Re II at $25.5M is narrow Florida named-storm paper. The diversity of cedent and peril in the recent pipeline is a structural strength of the market right now — but I'd flag that the average deal size of $146M and the presence of $14.94M micro-deals (Seaside Re 2026-61) suggest the market is serving both institutional and niche demand simultaneously, which is a maturity signal, not a crowding warning.

Karen Clark's call at ILS Bermuda Convergence 2026 is the most important thing said in this market in months — and it is worth reading precisely because it indicts a structural feature of how cat bonds are underwritten and sold. When a deal comes to market, investors are almost universally handed a single modeled expected loss figure from a primary vendor. At the market level, outstanding risk capital sits at $65.5B, the market is yielding 8.74% — a 4.57% insurance risk spread over a 2.5% expected loss, implying a spread-to-EL multiple of roughly 1.83x. That number is only honest if the 2.5% expected loss is real. Clark's point is that investors have no independent means to stress-test it if only one model's output shows up in the term sheet.

The USAA Residential Re 2026-2 deal — $225M, covering U.S. tropical cyclones, earthquakes, severe thunderstorm, wildfire, and other perils — is a prime example of the complexity at issue. Stacking that many perils into a single $225M structure means each individual peril's EL contribution is a black box unless the investor demands disaggregated model runs. The Acorn Re 2026-1 ($200M, U.S. earthquake only) is simpler in peril scope, but still presents the same single-model-view problem. American Coastal's Armor Re II 2026-2 ($25.5M, Florida named storm) is the most pointed given that Isaias just made landfall as a Category 2 in the Florida Panhandle on October 9 — a named-storm trigger event for exactly this class of bond.

The timing is not lost on me. A live Florida hurricane event arrives just as the ILS community is debating whether its models are pricing Florida wind correctly. YTD issuance has reached $18.9B across 97 deals. The average recent deal size is $130M. The market is liquid and capital is abundant. But Clark is raising the oldest question in this business: abundant capital priced off a model that understates loss is not cheap risk transfer — it is deferred pain. I'd push back on my colleague Ravi Chandrasekar's usual framing here: the model gap Clark is flagging is not purely an actuarial problem. It is a market-structure problem. Investors who accept a single vendor's EL are implicitly delegating their risk judgment to the cedent's choice of modeler. That is a conflict of interest embedded in the deal architecture, and it is not going away until investors demand otherwise.

The Artemis dashboard shows $18.9B in YTD cat-bond and ILS issuance across 97 deals, with $65.5B outstanding and a market yield of 8.74% — that 4.57% insurance risk spread over a market-level expected loss of 2.5% is a multiple-on-EL just above 1.8x. That is a functional spread. The market is not cheap, but it is not pricing panic either. Recent deal flow confirms appetite across the peril spectrum: USAA's Residential Re 2026-2 priced at $225M covering U.S. tropical cyclones, earthquakes, severe thunderstorm, wildfire, and more; Hannover's Acorn Re came at $200M on U.S. earthquake. The pipeline is real.

What today's Artex story adds to that picture is more uncomfortable than any single deal's spread. Kathleen Faries is saying, in public, that the industry has spent considerable effort engineering capital entry into ILS vehicles while exit has been treated as an afterthought. That is not a theoretical concern. When a loss event triggers and collateral becomes trapped — unable to exit on a defined timeline — institutional investors face mark-to-market uncertainty and duration extension they never underwrote. That dynamic has historically suppressed the next vintage of capital formation more than the loss itself.

The collateral yield component of the 8.74% market yield — currently 4.17% — is itself now in play. The IMF's warning this week of a potential disorderly spike in government bond yields is a two-sided risk for this market: higher yields lift the collateral return, but a disorderly repricing of the Treasury complex would impair the mark-to-market on the collateral pools that back cat-bond structures. Faries's push for clean exit mechanics is, in that context, a structural hedge against a scenario where the collateral leg itself becomes the source of friction — not the catastrophe leg.

The Artex Axcell Re shelf deals in the recent pipeline — a $20M PI0055 Grantham note and a $60M FE0004 tranche, both on unknown property catastrophe risks — are exactly the kind of private, bespoke ILS structures where exit ambiguity is most acute. Small size, non-transparent perils, limited secondary-market depth. If Faries is building toward a standardized exit protocol, those structures are the proof-of-concept she needs.