Insurance Desk

Cat Bond Desk Soren Vaeth

The Artemis dashboard is telling you something clear: at 9.29% market yield against a 2.5% market-level expected loss, the outstanding cat-bond market is running an implied multiple-on-EL of approximately 3.7x. For context, that is a comfortable cushion — not 2017-level thinness, not the giddy excess of 2007, but a spread over EL that reflects a market which repriced hard after Ian and still has not given it all back. The 5.53% insurance risk spread is the honest number; the 3.76% collateral yield is what money-market rates are doing for you while you wait for a storm that may or may not arrive.

The deal mix this week is instructive. Matterhorn Re at $345M for Swiss Re covers U.S. and Canada named storm and earthquake — that is a core cedent keeping a familiar structure. 3264 Re at $200M for Hannover Re is the same perils, same geography. These are not exotic bespoke trades; they are blue-chip cedents doing routine capital management at mid-year. The Harbor Crest Re trade for Porch Group at $100M is the more interesting data point: a non-traditional cedent accessing ILS markets for a multi-peril U.S. book including wildfire and winter storm. That is the market deepening, not just the usual suspects.

The macro backdrop — VIX 14.63, HY OAS 2.71%, dollar index down 1.47 over 30 days — is exactly the environment that fattens ILS inflows. When risk appetite is on and collateral yields are north of 3.75%, institutional allocators are getting paid twice: once for the insurance risk, once for the Treasury strip sitting in the collateral account. The ICI data showing $6.5B net into taxable bonds this week even as domestic equity funds lost $18.1B is the capital rotation story in miniature. Some of that fixed-income allocation finds its way to ILS. YTD at $18.9B across 92 deals, average deal size $145M — the pace is brisk but not frothy.

The 123 Lights Re transaction deserves a separate line: $100M from the Los Angeles Department of Water & Power for California wildfire risk. A public utility accessing the cat-bond market directly is not nothing. It signals that LADWP has modeled its wildfire liability with enough confidence to structure a parametric or indemnity trigger, found investors willing to take that risk at a clearing spread, and elected cat bonds over traditional insurance procurement. After the 2025 Los Angeles fire season, that is a statement about both cedent sophistication and investor appetite.

At a market yield of 9.29% against a 2.5% expected loss — roughly 3.7x multiple-on-EL — the cat-bond market is priced generously relative to history, supported by tight credit conditions, with the LADWP wildfire deal signaling deepening cedent sophistication.

Bias flag — Treats the 3.7x multiple-on-EL as an honest price signal; underweights model error on wildfire and secondary perils that Chandrasekar flags — the EL denominator may be understated, compressing the true multiple.

The Cycle Margaret Ennis

Soren is right that the spread looks comfortable today, but I want to name what that comfort is resting on: an absence of events so far this season, not a structural improvement in peril. YTD issuance of $18.9B across 92 deals at an average of $145M per transaction tells you that cedents are eager sellers and investors are eager buyers. In a hard-to-soft transition, that is precisely the condition you would expect — cedents locking in favorable terms before spreads compress, investors chasing yield in a world where HY OAS is already at 2.71% and traditional fixed income is crowded.

The Hannover Re and Swiss Re transactions — both covering U.S./Canada named storm and earthquake — are mid-year renewals of programs that have been running for years. The fact that Hannover and Swiss Re are still in the market at these volumes tells you reinsurers are not yet net sellers of cat risk; they are still passing it through to the capital markets. The question I keep coming back to is when that stops. When ILS spreads compress enough that the economics of retrocession look better than cedent-to-capital-market direct issuance, you will see primary reinsurers step back from sponsoring new cat bonds and start buying them instead. We are not there yet, but at 3.7x multiple-on-EL in a benign loss environment, the clock is running.

Peak Atlantic hurricane season runs August through October. Every year in a benign first half, investors get lulled. The capital that came back after Ian's repricing — and the new capital that came in attracted by post-Ian spreads — is now sitting exposed through October. If a major named storm makes U.S. landfall in the next 60 days, the cycle narrative flips fast: trapped collateral, secondary-market spread widening, and suddenly the Jan-1 2027 renewal conversations get a lot more interesting. Hard markets sow the seeds of the next soft market, yes — but soft markets also sow the seeds of the next hard market, and the current softening pressure is real.

Brisk mid-year ILS issuance reflects cedent eagerness to lock in favorable spreads before further softening, but the market's comfort rests on a loss-free first half — peak hurricane season through October is the test that will determine whether the current cycle extends or snaps.

Bias flag — Mean-reversion framing assumes the current soft-market pressure will eventually snap back; may underweight the possibility that structural capital deepening in ILS is a regime shift, not just a cycle phase.

Modeled Loss Dr. Ravi Chandrasekar

Margaret's point about the timing of this comfort is exactly right, and I want to put a model frame around why the 2.5% market-level expected loss figure deserves scrutiny rather than acceptance. That number is an output of the models underlying each deal's risk analysis — RMS, AIR, KCC — aggregated across a $65.6B outstanding portfolio. It represents the probability-weighted average annual loss from the modeled peril set. What it does not capture, by construction, is the degree to which the historical event catalog used to calibrate those models is itself non-stationary.

The 123 Lights Re transaction is the clearest example. LADWP is structuring a California wildfire cat bond. The wildfire peril model is the youngest and least-calibrated of the major cat perils — the event catalog for large-scale interface fires in Southern California under current climate and development conditions spans roughly 15 years of useful data. The attachment probability assigned to that deal is a model output derived from that thin catalog, extended with climate adjustments that are themselves contested among modelers. I am not saying the deal is mispriced; I am saying the confidence interval around the expected loss is wider than the single point estimate suggests.

Similarly, the multi-peril Harbor Crest Re for Porch Group covers severe convective storm, winter storm, and fire-following-earthquake alongside named storm. Secondary perils — severe convective storm above all — have consistently produced actual losses above modeled expectations for the past decade. The 2021-2024 severe convective storm loss experience in the U.S. was structurally above model in ways that took the industry years to acknowledge. A multi-peril aggregate that bundles secondary perils into a single expected loss figure is aggregating a set of model hypotheses with very different error bars. The market yield of 9.29% may look generous at the single-point EL of 2.5%, but if the true expected loss on the secondary-peril component is materially higher than modeled, that multiple-on-EL shrinks faster than the spread compression does.

The 2.5% market-level expected loss is a model output, not a ground truth — California wildfire and secondary-peril exposures carry wider confidence intervals than the single-point figure implies, and the Harbor Crest multi-peril structure concentrates that uncertainty.

Bias flag — Focuses on model uncertainty in the EL numerator; underweights the social-inflation and litigation-driven loss development that affects primary carriers more than cat-bond collateral pools.

Carrier Books Theo Marchetti

From a carrier-fundamentals lens, the weekend's most actionable signal is not in the news corpus but in the SEC filing novelty data. The insurance sector shows 8-of-8 leaders filing on the latest 10-K cycle, with Item 1A Risk Factors averaging 30.3% novelty — moderate rewriting — but the distribution is skewed: PRU at 66.8% novelty (304 sentences added, 148 deleted) and TRV at 47.2% (246 added, 251 deleted) are doing meaningful risk-language overhauls, while CB at 16.6% is barely touching its disclosures. BRK-B's MD&A leads the sector at 73.5% novelty, which is notable given Berkshire's scale and the typical conservatism of its disclosures.

High novelty in Risk Factors is not by itself a red flag — it can reflect genuine strategic pivot, new line of business, or simply better legal drafting. But when TRV adds 246 and deletes 251 sentences in its risk language, that is a substantial rewrite of how it is characterizing its exposure to investors. TRV writes a large personal and commercial lines book with meaningful cat exposure; a near-wholesale replacement of its risk factor language in this filing cycle warrants a read-through to understand what old risks it stopped disclosing and what new ones it started. PRU's rewrite at 66.8% novelty is dominated by life and financial services risks rather than P&C, so the insurance-relevance there is different in kind.

On the macro backdrop: the 10-year/2-year spread at +51 basis points and effective fed funds at 3.63% are a modestly supportive environment for carrier investment portfolios — particularly for life carriers running long-duration asset-liability matches. The dollar index down 1.47 over 30 days is a marginal headwind for carriers with offshore reinsurance subsidiaries reporting in dollars. VIX at 14.63 and HY OAS at 2.71% suggest the equity market is not pricing in near-term tail risk for financial stocks, which tracks: the ICI data showing $6.5B into taxable bonds but $21.3B out of domestic equity is a rotation story, not a panic story. Carrier book values are not under threat from the macro today; they are under threat from the 60-day window that Dr. Chandrasekar and Margaret are both pointing at.

TRV's near-wholesale Risk Factor rewrite (47.2% novelty, 246 sentences added and 251 deleted) in its latest 10-K is the most actionable carrier disclosure signal this cycle — it warrants a detailed read-through to understand what exposures changed characterization.

Bias flag — Anchors on SEC filing novelty as a disclosure-quality signal; the novelty score indicates volume of language change, not direction — a high score could reflect improved transparency or increased concealment of risk.